0195162854

521

Upload: psgcu

Post on 14-Dec-2014

177 views

Category:

Documents


28 download

TRANSCRIPT

Page 1: 0195162854
Page 2: 0195162854

The Behavior ofthe Laboratory Rat

Page 3: 0195162854
Page 4: 0195162854

A Handbook with Tests

Edited by

IAN Q. WHISHAW

BRYAN KOLB

Department of Psychology and Neuroscience

Canadian Centre for Behavioural Neuroscience

OXFORDUNIVERSITY PRESS

2005

THE BEHAVIOR OFTHE LABORATORY RAT

Page 5: 0195162854

OXFORDUNIVERSITY PRESS

Oxford New York

Auckland Bangkok Buenos Aires Cape Town Chennai

Dar es Salaam Delhi Hong Kong Istanbul Karachi Kolkata

Kuala Lumpur Madrid Melbourne Mexico City Mumbai Nairobi

Sao Paulo Shanghai Taipei Tokyo Toronto

Copyright © 2005 by Oxford University Press, Inc.

Published by Oxford University Press, Inc.

198 Madison Avenue, New York, New York, 10016

www.oup.com

Oxford is a registered trademark of Oxford University Press

All rights reserved. No part of this publication may be reproduced,

stored in a retrieval system, or transmitted, in any form or by any means,

electronic, mechanical, photocopying, recording, or otherwise,

without the prior permission of Oxford University Press

Library of Congress Cataloging-in-Publication Data

Whishaw, Ian Q., 1939-

The behavior of the laboratory rat : a handbook with tests /

edited by Ian Q. Whishaw, Bryan Kolb.

p. cm. ISBN 0-19-516285-4

1. Rats—Behavior. 2. Rats as laboratory animals.

I. Kolb, Bryan, 1947- II. Tide.

QL737.R666W52 2004 616'.02733—dc22 2004041514

9 8 7 6 5 4 3 2 1

Printed in the United States of America

on acid-free paper

Page 6: 0195162854

To Samuel Anthony (Tony) Barnett (1915-2003)

Behavioral neuroscientists know S.A. Barnett best for hisbook The Rat: A Study in Behavior, first published in 1963. Hewas the author of over 150 papers (the last of which is thesecond chapter of this volume) and nine books. He was agraduate of Oxford University in Zoology. He advised theBritish Government on plague during WWII and was theChair of Zoology in Glasgow from 1971 to 1983. Tony wasan enthusiastic broadcaster, contributing for years to theBBC series Occam's Razor.

Page 7: 0195162854

This page intentionally left blank

Page 8: 0195162854

Preface

The principal function of the nervous systemis to produce behavior. Thus, the ultimategoal of most behavioral work with laboratoryanimals in neuroscience is to understand howmolecular events in the nervous system cometo produce behavior and, as a corollary, howchanges in molecular events produce differ-ences in behavior. Understanding these issuesoffers hope for understanding the nature ofthe human mind, which some may argue isthe fundamental question in neuroscience.But perhaps even more important is that un-derstanding brain-behavior relationships of-fers a way to find treatments for dysfunctionsof behavior, whether they are in the provinceof neurology or psychiatry. Advances in mo-lecular and cellular neuroscience have beendramatic over the past two decades, but mostof these advances have been independent ofan understanding of how they relate to be-havior. This is changing. Neuroscientists ori-ented toward molecular research are increas-ingly looking to the ultimate function of thephenomenon that they have been studying—behavior. For the majority of behavioral stud-ies, this means studying the behavior of thelaboratory rat.

This book has three objectives. Our firstobjective is to present an introduction of ratbehavior to neuroscience students. In choos-ing the rat as the subject species, we made theassumption that this species will remain, as ithas in the past, the primary subject used thelaboratory investigations of behavior. Oursecond objective is to describe the organiza-tion and complexity of rat behavior. The ma-jor theme emerging from many lines of re-search on rat behavior is that understandingthe rules of behavioral organization will be

central in understanding the structural basisof behavior. Our third objective is to update,as much as is possible, previous compendiumsof rat behavior. Behavioral neuroscience con-tinues to be a diverse field of research in whichthere remain many competing experimentalmethods and hypotheses, and we believe thatcollectively, the chapters of this volume re-flect that diversity.

As we have noted, advances in the fieldof neuroscience have kindled an interest in be-havior by researchers in many of its subdisci-plines. For many of these researchers, behav-ior may previously have seemed unrelated totheir studies and of little direct interest. Manyresearchers with primary training in diversefields such as genetics or biochemistry arenow looking at brain-behavior questions forthe first time and, like all fields, the literaturecan be bewildering to the novice. We there-fore asked the authors of the chapters of thisbook to imagine that students from anotherdiscipline were coming to them with ques-tions related to how they could incorporatebehavior into their research programs. For ex-ample, we asked them to imagine a studentcoming from such areas as medicine, chem-istry, or genetics who had no special trainingin behavior but now saw behavior as relevantand necessary to research questions in whichthey were interested. The challenge we pre-sented was, "Could they summarize their fieldof expertise so that the novice student wouldgain an understanding of the questions, meth-ods, and potential findings in that line of re-search?" We also asked them to make theirsummary brief, to emphasize methodology,and to minimize as much as possible literaturereviews. Our expectation is that a novice stu-

vii

Page 9: 0195162854

viii PREFACE

dent would read their chapter as a first intro-duction to the study of rat behavior. This in-troduction would then guide the student as heor her gained further expertise in the practi-cal application of that information and instudy of the larger body of literature.

In our view, the problem that we posedis not fictive. During the past few years, we re-ceived large numbers of telephone calls, e-mails, and queries regarding behavior frompeople with diverse backgrounds in neuro-science. We have also been asked to speak atmeetings about behavior to scientists who wecould not have imagined would have been in-terested in behavior or in the research that wedo. The tone of this interest is best representedby the comments of an acquaintance, a mo-lecular biologist, who stated: "I could not haveimagined that this psycho mumbo-jumbo wasgoing to be important, but now I see it as theonly game in town." We realize that in answerto such interest, it is simply not good enoughto say, "We could look at that for you," or"Perhaps you should collaborate with an ex-pert on that behavior." Rather, it is likely thatbehavioral methods will become a part ofmany lines of research, so the challenge for be-haviorists is to make their science accessible toother investigators. Indeed, we have encoun-tered the problem in reverse as we have addedmore molecular techniques to our behavioralstudies. Accordingly, we asked the authors ofthe chapters in this book to make their re-search accessible to new students of behavior.

Of course, this book is about rat behav-ior. The rat, Rattus norvegicus, was the firstspecies to be domesticated for the purpose ofscientific research. In the 100 years since therat was first introduced to the laboratory, ithas generated an incredibly large body of lit-erature. It has been the primary subject withwhich psychological theories have beentested; it has been the primary subject for thestudy of behavioral pharmacology; and it hasbeen the primary subject for the investigationof brain chemistry, anatomy, and physiology.One reason for the popularity of the rat is that

it is a behavioral generalist. Rats are found invirtually every ecological niche on earth, andthey have proved to be adaptable and suc-cessful in all of them. For the purposes of an-swering behavioral questions, we assert thatas a behavioral generalist, the rat will remainthe primary species for continued investiga-tions into the organization of brain and be-havior and the structural basis of behavior. Inbeing a generalist, the rat is very much likethe human, a species with which it is com-mensal. It is likely that the genes, neural struc-ture, and behavior of generalists have prop-erties that are similar. This is why the ratcontinues to be the primary model used tostudy a wide range of questions related to hu-man behavior and health.

With the development of genetic engi-neering, a line of inquiry that uses the mouseas the prime vertebrate model, we should ad-dress the question of whether the mousemight not have been a better species aboutwhich to compile a behavioral book. The lab-oratory mouse is the laboratory rat's closestdomestic relative, but we do not think that thetwo species are so similar that one can be sub-stituted for the other. This is especially so withrespect to behavior. Both species have beenused for behavioral research for approxi-mately 100 years, and each has apparentlyfound its laboratory niche. There is littledoubt that for many questions related to mo-tor functions, regulatory functions, and espe-cially cognitive functions, the rat has been thespecies of choice. We think that it will con-tinue to be so. In contrast, mice will likely con-tinue to be the subject of choice in geneticstudies in neuroscience, and presumably thestudy of many behavior genetic questions willretain the mouse as the primary laboratorysubject. However, the behavioral study of themouse in genetics is fundamentally differentfrom the primary questions addressed in be-havioral studies in the rat and must be the sub-ject of a separate volume.

Two excellent, and still relevant, bookshave been previously devoted to rat behavior.

Page 10: 0195162854

Preface ix

Norman L. Munn's 1950 "Handbook of Psy-chological Research on the Rat: An Introduc-tion to Animal Psychology" is directed tomany of the same questions as is the presentbook. It describes general activity, unlearnedbehavior, sensory processes, learning, socialbehavior, and rat models of neuropsychiatricdiseases. It also emphasizes methods of study.S. A. Barnett's 1963 book, "The Rat: A Studyin Behavior," covers much of the same groundbut with more emphasis on the ethogram, orprofile of behavior, of the wild rat. In whatway is the present book different from thesepredecessors? We think that the primary ad-vancement in understanding rat behavior isthe emergence of understanding how rat be-havior is organized. For example, rat groom-ing, play and aggression, exploration, cogni-tion, and other activities are organized withboth fixed and open syntax. The understand-ing of this organization provides new avenuesfor the investigation of genetic, neural, andhormonal regulation of behavior. This organ-ization has also led to the development ofcomputer-based behavioral analysis systemsthat aid in using behavior as an assay for otherscientific manipulations.

Although this book consists of 43 chapterson different aspects of rat behavior, and thus is

comprehensive, it is not exhaustive. Our majordifficulty in editing the book was in insisting thatauthors substantially shorten their chapters tomake the book manageable as a single volume.Indeed, we could have doubled the number ofchapters without covering every aspect of ratbehavior, but we believe that the selection ofchapters presented here provide more than ad-equate grist for an introduction to the study ofthe rat in behavioral brain research.

We express a special thanks to all of theauthors who generously contributed time towrite a chapter for this book. We also expressour thanks to Fiona Stevens of Oxford Uni-versity Press, who approached us and per-suaded us to compile this handbook and gaveus the liberty of selecting a structure of ourown choosing.

Lethbridge, Alberta, Canada. I. Q. W.B. K.

SUGGESTED READINGS

Barnett SA (1963) The rat: a study in behavior. Chicagoand London: The university of Chicago Press.

Munn NL (1950) Handbook of psychological researchon the rat: an introduction to animal psychology.Boston: The Riverside Press Cambridge.

Page 11: 0195162854

This page intentionally left blank

Page 12: 0195162854

Contents

Contributors, xiii

Part I Natural History

1. Evolution, 3Andrew N. Iwaniuk

2. Ecology, 15S. Anthony Barnett

3. Strains, 25Lauren Gerard Koch and Steven L Britton

4. Individual Differences, 37Guy Mittleman

Part II Sensory Systems

5. Vision, 49Glen T. Prusky and Robert M. Douglas

6. Somatosensation, 60Linda Hermer-Vazquez,Raymond Hermer-Vazquez, andJohn K. Chapin

7. Pain, 69Daniel Le Bars and Samuel W. Cadden

8. Vibrissae, 81Richard H. Dyck

9. Olfaction, 90Burton Slotnick, Heather Schellinck, andRichard Brown

10. Taste, 105Alan C. Spector

Part III Motor Systems

11. Posture, 121Sergio M. Pellis and Vivien C. Pellis

12. Orienting and Placing, 129Tim Schallert and Martin T. Woodlee

13. Grooming, 141J. Wayne Aldridge

14. Locomotion, 150Gillian Muir

15. Prehension, 162Ian Q. Whishaw

16. Locomotor and Exploratory Behavior, 171Ilan Golani, Yoav Benjamini, Anna Dvorkin,Dina Lipkind, and Neri Kafkafi

17. Circadian Rhythms, 183Michael C. Antle and Ralph E. Mistlberger

Part IV Regulatory Systems

18. Eating, 197Peter G. Clifton

19. Drinking, 207Neil E. Rowland

20. Foraging, 217Ian Q. Whishaw

21. Thermoregulation, 226Evelyn Satinoff

22. Stress, 236Jaap M. Koolhaas, Sietse F. de Boer, andBauke Buwalda

23. Immune System, 245Hymie Anisman and Alexander W. Kusnecov

Part V Development

24. Prenatal Behavior, 257Scott R. Robinson and Michele R. Brumley

25. Infancy, 266Jeffrey R. Alberts

26. Adolescence, 278Russell W. Brown

xi

Page 13: 0195162854

xii CONTENTS

27. Maternal Behavior, 287Stephanie L. Rees, Vedran Lovic, andAlison S. Fleming

28. Play and Fighting, 298Sergio M. Pellis and Vivien C. Pellis

29. Sex, 307William J. Jenkins and Jill B. Becker

30. Environment, 321Robbin L. Gibb

Part VI Defense and Social Behavior

31. Antipredator Defense, 335D. Caroline Blanchard and Robert J. Blanchard

32. Aggressive, Defensive, and SubmissiveBehavior, 344

Klaus A. Miczek and Sietse F. de Boer

33. Defensive Burying, 353Dallas Treit and John J. P. Pinel

34. Social Learning, 363Bennett G. Galef, Jr.

35. Vocalization, 371Greta Sokoloff and Mark S. Blumberg

Part VII Cognition

36. Object Recognition, 383Dave G. Mumby

37. Piloting, 392Etienne Save and Bruno Poucet

38. Dead Reckoning, 401Douglas G. Wallace and Ian Q. Whishaw

39. Fear, 410Matthew R. Tinsley and Michael S. Fanselow

40. Cognitive Processes, 422Robert J. Sutherland

41. Incentive Behavior, 436Bernard W. Balleine

Part VIII Models and Tests

42. Neurological Models, 449Bryan Kolb

43. Psychiatric Models, 462Henry Szechtman and David Eilam

44. Neuropsychological Tests, 475Gerlinde A. Metz, Bryan Kolb, andIan Q.Whishaw

Index, 499

Page 14: 0195162854

Contributors

JEFFREY R. ALBERTSDepartment of Psychology

Indiana University

Bloomington, Indiana

J. WAYNE ALDRIDGEDepartments of Neurology and Psychology

University of Michigan

Ann Arbor, Michigan

HYMIE ANISMANInstitute of NeurosciencesCarelton University

Ottawa, Ontario, Canada

MICHAEL C. ANTLEDepartment of Psychology

Columbia University

New York, New York

BERNARD W. BALLEINEDepartment of Psychology and the Brain Research

InstituteUniversity of California, Los Angeles

Los Angeles, California

S. ANTHONY BARNETT*Aranda, Australia

JILL B. BECKERDepartment of PsychologyReproductive Sciences Program and Neurosciences

ProgramUniversity of MichiganAnn Arbor, Michigan

YOAV BENJAMINIDepartment of Zoology

Tel Aviv University

Tel Aviv, Israel

D. CAROLINE BLANCHARDDepartment of NeurobiologyUniversity of HawaiiHonolulu, Hawaii

Deceased.

ROBERT J. BLANCHARDDepartment of NeurobiologyUniversity of Hawaii

Honolulu, Hawaii

MARK S. BLUMBERGDepartment of Psychology

Indiana University

Bloomington, Indiana

STEVE L. BRITTONFunctional Genomics Laboratory

Medical College of Ohio

Toledo, Ohio

RICHARD BROWNDepartment of Psychology

Dalhousie University

Halifax, Nova Scotia, Canada

RUSSELL W. BROWNDepartment of PsychologyEast Tennessee State University

Johnson City, Tennessee

MICHELE R. BRUMLEYDepartment of Psychology

University of Iowa

Iowa City, IA

BAUKE BUWALDADepartment of Animal PhysiologyUniversity of Groningen

Haren, The Netherlands

SAMUEL W. CADDENThe Dental School

University of Dundee

Dundee, Scotland

JOHN K. CHAPINDepartment of Physiology and PharmacologySUNY Downstate Medical Center

Brooklyn, New York

xiii

Page 15: 0195162854

xiv CONTRIBUTORS

PETER G. CLIFTONDepartment of Psychology

University of Sussex

Brighton, United Kingdom

SIETSE F. DE BOERDepartment of Animal PhysiologyUniversity of GroningenHaren, The Netherlands

ROBERT M. DOUGLASCentre for Macular Research

Department of Ophthalmology and Visual Sciences

University of British Columbia

Vancouver, British Columbia, Canada

ANNA DVORKINDepartment of ZoologyGeorge S. Wise Faculty of Life SciencesTel Aviv UniversityTel Aviv, Israel

RICHARD H. DYCKDepartment of Psychology

Department of Cell Biology and Anatomy

University of Calgary

Calgary, Alberta, Canada

DAVID ElLAMDepartment of ZoologyTel Aviv University

Tel Aviv, Israel

MICHAEL S. FANSELOWDepartment of Psychology

University of California, Los Angeles

Los Angeles, California

ALISON S. FLEMINGDepartment of PsychologyUniversity of Toronto at Missassauga

Missassauga, Ontario, Canada

BENNETT G. GALEF, JR.Department of Psychology

McMaster University

Hamilton, Ontario, Canada

ROBBIN L. GlBBCanadian Centre for Behavioural NeuroscienceDepartment of Psychology and Neuroscience

University of LethbridgeLethbridge, Alberta, Canada

ILAN GOLANIDepartment of Zoology

George S. Wise Faculty of Life SciencesTel Aviv University

Tel Aviv, Israel

LINDA HERMER-VAZQUEZDepartment of Physiology and PharmacologySUNY Downstate Medical CenterBrooklyn, New York

RAYMOND HERMER-VAZQUEZDepartment of Physiology and Pharmacology

SUNY Downstate Medical Center

Brooklyn, New York

ANDREW N. IWANIUKDepartment of PsychologyUniversity of AlbertaEdmonton, Alberta, Canada

WILLIAM J. JENKINSDepartment of Psychology

Reproductive Sciences Program and Neurosciences

Program

University of Michigan

Ann Arbor, Michigan

NERI KAFKAFIMaryland Psychiatry Research CenterUniversity of Maryland

College Park, Maryland

LAUREN GERARD KOCHFunctional Genomics Laboratory

Medical College of Ohio

Toledo, Ohio

BRYAN KOLBCanadian Centre for Behavioural NeuroscienceDepartment of Psychology and NeuroscienceUniversity of Lethbridge

Lethbridge, Alberta, Canada

JAAP M. KOOLHAASDepartment of Animal Physiology

University of Groningen

Haren, The Netherlands

ALEXANDER W. KUSNECOVDepartment of PsychologyBiopsychology and Behavioral Neuroscience ProgramRutgers, The State University of New Jersey

Piscataway, New Jersey

Page 16: 0195162854

Contributors XV

DANIEL LE BARSInstitut National de la Sante et de la Recherche

Medicale (INSERM)

Paris, France

DlNA LlPKINDDepartment of Zoology

George S. Wise Faculty of Life SciencesTel Aviv University

Tel Aviv, Israel

VEDRAN LovicDepartment of Psychology

University of Toronto at Missassauga

Missassauga, Ontario, Canada

GERLINDE A. METZCanadian Centre for Behavioral NeuroscienceDepartment of Psychology and Neuroscience

University of Lethbridge

Lethbridge, Alberta, Canada

KLAUS A. MICZEKDepartment of Psychology, Psychiatry, Pharmacology,

and Neuroscience

Tufts University

Medford, Massachusetts

RALPH E. MISTLBERGERDepartment of PsychologySimon Fraser University

Burnaby, British Columbia, Canada

GUY MlTTLEMAN

Psychology DepartmentUniversity of Memphis

Memphis, Tennessee

GILLIAN MUIRBiomedical SciencesWestern College of Veterinary Medicine

University of Saskatchewan

Saskatoon, Saskatchewan, Canada

DAVE G. MUMBYDepartment of Psychology

Concordia UniversityMontreal, Quebec, Canada

SERGIO M. PELLISCanadian Centre for Behavioral NeuroscienceDepartment of Psychology and Neuroscience

University of Lethbridge

Lethbridge, Alberta, Canada

VIVIEN C. PELLISCanadian Centre for Behavioral Neuroscience

Department of Psychology and Neuroscience

University of Lethbridge

Lethbridge, Alberta, Canada

JOHN J.P. PlNELDepartment of PsychologyUniversity of British Columbia

Vancouver, British Columbia, Canada

BRUNO POUCETLaboratoire de Neurobiology de la Cognition,

UMR 6155CNRS—Universite Aix-Marseille I

Marseille, France

GLEN T. PRUSKYCanadian Centre for Behavioural Neuroscience

Department of Psychology and Neuroscience

University of Lethbridge

Lethbridge, Alberta, Canada

STEPHANIE L. REESDepartment of PsychologyUniversity of Toronto at Missassauga

Missassauga, Ontario, Canada

SCOTT R. ROBINSONDepartment of Psychology

University of Iowa

Iowa City, Iowa

NEIL E. ROWLANDDepartment of Psychology

University of Florida

Gainesville, Florida

EVELYN SATINOFFDepartment of Psychology

University of DelawareNewark, Delaware

ETIENNE SAVELaboratoire de Neurobiology de la Cognition,

UMR 6155CNRS—Universite Aix-Marseille I

Marseille, France

TlM SCHALLERT

Department of Psychology

University of Texas at Austin

Austin, Texas

Page 17: 0195162854

xvi CONTRIBUTORS

HEATHER SCHELLINCKDepartment of Psychology

Dalhousie UniversityHalifax, Nova Scotia, Canada

BURTON SLOTNICKDepartment of Psychology

University of South FloridaTampa, Florida

GRETA SOKOLOFFDepartment of PsychologyIndiana UniversityBloomington, Indiana

ALAN C. SPECTORDepartment of PsychologyUniversity of FloridaGainesville, Florida

ROBERT J. SUTHERLANDCanadian Centre for Behavioural Neuroscience

University of LethbridgeLethbridge, Alberta, Canada

HENRY SZECHTMANDepartment of Psychiatry and Behavioural

NeurosciencesMcMaster UniversityHamilton, Ontario, Canada

MATTHEW R. TINSLEYDepartment of PsychologyUniversity of California, Los AngelesLos Angeles, California

DALLAS TREITDepartment of PsychologyUniversity of AlbertaEdmonton, Alberta, Canada

DOUGLAS G. WALLACECanadian Centre for Behavioural NeuroscienceDepartment of Psychology and NeuroscienceUniversity of LethbridgeLethbridge, Alberta, Canada

IAN Q. WHISHAWCanadian Centre for Behavioural NeuroscienceDepartment of Psychology and NeuroscienceUniversity of Lethbridge

Lethbridge, Alberta, Canada

MARTIN T. WOODLEEDepartment of PsychologyUniversity of Texas at AustinAustin, Texas

Page 18: 0195162854

Natural History I

Page 19: 0195162854

This page intentionally left blank

Page 20: 0195162854

Evolution

ANDREW N. IWANIUK1

ON THE ORIGIN OFRattus norvegicus

The Norway, or "laboratory," rat (Rattusnorvegicus) has been used in behavioral, neu-ral, physiological, and other forms of researchfor more than a century. The evolutionary his-tory of this species is often dismissed as unim-portant in psychological and biomedical re-search because the aim is not to understandevolutionary biology but rather to use the ratas a model system to investigate a specific as-pect of organismal biology. It is not our in-tention to critique these experiments becausethey are integral to our understanding of an-imal behavior, anatomy, molecular biology,and physiology. It is, however, important toacknowledge that the rat did not evolve in avacuum and that the morphological, physio-logical, and behavioral changes imposed by"domestication" are still a result of the evolu-tionary process.

This chapter addresses the evolution ofthe laboratory rat from the origins of rodentsin general to the speciation of the genus Rat-tus. This is not meant to be a complete reviewof all the taxonomy and phylogenetic historyof Rattus and higher-level taxonomic ranks,because discussions of this are provided else-where (Carleton and Musser, 1984; Luckettand Hartenberger, 1985; Musser and Carle-ton, 1993; Nowak, 1999). Instead, I provide asummary of the evolutionary events that ledto R. norvegicus. Because palaeontology, tax-onomy, and phylogenetics are intimately re-lated to one another, this chapter is organized

in terms of the taxonomy of R. norvegicus(Table 1-1). Evolutionary relationships andpalaeontological history are discussed withreference to other groups of the same taxo-nomic rank. For example, the order Rodentiais placed in the context of other mammalianorders. By summarizing the evolutionary his-tory of R. norvegicus, we aim to provide a ba-sic understanding of how the species hasevolved that may be instructive in interpret-ing the results of behavioral experimentationand/or comparative analyses.

ORDER RODENTIA

To understand the evolution of R. norvegicus,it is necessary to begin with the history of therodents in general and their relationship toother mammalian taxa. The order Rodentia isthe most abundant of all of the mammalianorders, numbering close to 2000 species. Ro-dents are found on every continent, exceptAntarctica, and account for almost half of allplacental mammals. They are readily distin-guishable from other mammals by an array ofmorphological features (Luckett and Harten-berger, 1993, 1985), the most prominent ofwhich is their distinctive dental morphology.Rodent incisors are large, unrooted, and per-sistently growing teeth with enamel only onthe upper surface that maintains a beveled cut-ting edge. The surface morphology of the mo-lars is also distinctive, and the jaw structureexhibits adaptations that allow considerablemovement during grinding (Hand, 1984).

3

Page 21: 0195162854

NATURAL HISTORY

Table 1-1. Taxonomy of the Genus Rattus

Class MammaliaOrder Rodentia

Superfamily MuroideaFamily Muridae

Genus RattusSpecies adustus

argentiventer*bontanuscollettienganusexulans*foramineusgiluwensishoffinannijobiensiskorinchilosea*lutreolusmarmosurusmollicomulusmordaxnativitatis^norvegicus*osgoodipelurusranjiniaesanilasimalurensissteinitanezumitimorensistunneyi

villosissimus

annandaleibaiuensisburruselaphinuseverettifeliceusfusdpeshainaldihoogerwerfikoopmanileucopuslugensmackaritfmindorensismontanusmorotaiensisnitidus*novaeguineaepalmarumpraetorrattus*sikkimensissordidusstoicustawitawiensistiomanicus*turkestanicusxanthurus

Note: The taxonomy is taken from Guy and Musser (1993).*Commensal species.fSpecies that recently became extinct.

Despite some broad similarities in mor-phology, rodents are a morphologically andbehaviorally diverse order. They span a rangeof locomotor behaviors that include gliding,climbing, swimming, underground digging,hopping, and running. Not only do rodentsexhibit this range of locomotor behaviors; inmost instances, they have evolved indepen-dently many times. For example, subterraneanlocomotion has evolved at least three times(Muroidea, Geomyoidea, Bathyergoidea). Sim-ilarly, there is a broad range of social systemsfrom uniparental monogamy/polygamy to

complex, multimale/multifemale societies havealso evolved independently many times.

This behavioral diversity belies the fact thatthe rodents form a monophyletic group. Thatis, all rodents share a common ancestry that isnot shared with nonrodent species. Althoughthe issue of rodent monophyly was questionedby molecular studies of the guinea pig (Caviaporcdlus} (Graur et al., 1991; D'Erchia et al.,1996), more recent studies agree that rodentsare monophyletic (Adkins et al., 2001; Madsenet al., 2001; Murphy et al., 2001a,b; Huchon etal., 2002). There is, however, some debate re-garding the position of Rodentia relative toother mammalian orders in phylogenetic trees.

Traditionally, the order Lagomorpha(hares, rabbits, and pikas) is considered to bemost closely related to rodents (i.e., a sister-group) based on their morphological similar-ity (Shoshani and McKenna, 1998) (Fig. 1-lA).The rodents and lagomorphs together com-prise a clade termed Glires. Morphologicalsimilarities also link the elephant shrews (or-der Macroscelidea) as the sister-group to theGlires (Fig. 1-lA). Molecular studies havedemonstrated markedly different relation-ships between many traditional mammalianorders, but they all agree that rodents andlagomorphs should be placed together (Hu-chon et al., 2001; Madsen et al., 2001; Murphyet al., 2001a,b) (Fig. 1-1B). The broadly basedstudies also agree in placing the Glires assister-group to a clade containing primates(Madsen et al., 2001; Murphy et al., 2001a,b).

In terms of dating the origin of the Ro-dentia, the superorder Glires likely originatedbetween 64 and 104 million years ago (mya)(Archibald et al., 2001; Murphy et al., 200la).The diversification of rodents is therefore esti-mated at 65 mya at the earliest on the basis ofboth palaeontological (Alroy, 1999; Archibaldet al., 2001) and molecular (Bromham et al.,1999; Foote et al., 1999; Eizirik et al., 2001)data. Thus, the predecessors of rodents mayhave coexisted with dinosaurs, but true ro-dents did not evolve until after the Cretaceous-Tertiary Period boundary (i.e., <65 mya),

4

Page 22: 0195162854

Monotremata

Marsupialia

Xenartha

Afrosoricida

Eulipotyphla

Lagomorpha

RodentiaMacroscelidea

Scandentia

Primates

Dermoptera

Chiroptera

Pholidota

Carnivora

Tubulidentata

Cetartiodactyla

Perissodactyla

Hyracoidea

Sirenia

Proboscidea

Monotremata

Marsupialia

Proboscidea

Hyracoidea

Sirenia

Macroscelidea

Afrosoricida

Tubulidentata

Xenartha

Primates

Scandentia

Dermoptera

Lagomorpha

RodentiaEulipotyphla

Chiroptera

Pholidota

Carnivora

Perissodactyla

Cetartiodactyla

Figure 1-1. Two alternative phylogenetic hypotheses of the interordinal relationships of all mammals. Con-trary to "traditional" taxonomies, the Insectivora is broken up into two orders: Afrosoricida and Eulipoty-phla. Also, the Cetacea and Artiodactyla are merged into the singular order Cetartiodactyla. The first tree(A) is based on morphological characters and places rodents as a sister-group to the lagomorphs (Shoshaniand McKenna, 1998). The second tree (B) differs in many of the interordinal relationships but retains thesister-group relationship between Rodentia and Lagomorpha (Murphy et al., ZOOla). Although the mono-tremes (Monotremata) were not sampled in the analysis presented by Murphy et al. (2001a), they are includedhere as the basal group to provide consistency between the two trees.

A

B

5

Page 23: 0195162854

NATURAL HISTORY

contrary to some estimates (Hedges et al.,1996; Kumar and Hedges, 1998).

SUPERFAMILY MUROIDEA/FAMILY MURIDAE

The superfamily Muroidea is the most diverserodent taxon, with more than 1300 species(Musser and Carleton, 1993). Within this su-perfamily, there is consensus that only onefamily is represented by extant species: theMuridae (Musser and Carleton, 1993). Thisfamily includes the "true" rats (Rattus) as wellas numerous genera of mice (e.g., spiny mice,deer mice), voles, lemmings, gerbils, and ham-sters. The diversity in life history and mor-phology within murids is almost as great as thatof rodents as a whole, with a wide span of lo-comotor modes, social systems, and ecology.Murids are generally distinguished from otherrodent families by their molar morphology(Carleton and Musser, 1984), but they also pos-sess a "primitive" middle ear structure (Lavo-cat and Parent, 1985) and several unique de-velopmental features (Luckett, 1985).

The kangaroo-rats (superfamily Dipo-doidea) are generally considered to be thesister-group to the Muridae (Fig. l-2b). Thisis based on a number of shared morphologi-cal features (Luckett and Hartenberger, 1985)as well as molecular evidence (Nedbal etal., 1996; Adkins et al., 2001). More recently,Huchon et al. (2002) demonstrated a possiblerelationship between the Geomyoidea (pocketgophers) and the Muridae, using the mostcomprehensive sampling of rodent species inmolecular phylogenetics thus far (Fig. l-2a).Although this relationship was strongly sup-ported in their analyses, several alternativearrangements could not be discounted. Fur-ther investigation of this arrangement is there-fore warranted as this could result in severaltaxonomic changes at the superfamily level ofrodents.

The diversification of the rodent super-families and families is estimated to have oc-

curred in late Palaeocene to early Eocene(==55 mya) (Hartenberger, 1998). A basal lin-eage leading to the Muridae branched off atthis point, with "modern" murids apparent inthe middle Eocene (36.5 to 49 mya). The old-est murid discovered thus far, the hamster-likeCricetodon, was found in Mongolia and hasbeen dated as late Eocene (34.2 to 36.5 mya)(Li and Ting, 1983, in Carleton and Musser,1984). Fossil taxa more closely resemblingRattus do not appear to occur for another 20million years (see later).

SUBFAMILY Murinae

Within the Muridae, the subfamily Murinae isthe most speciose lineage. There are morethan 500 recognized species within the Muri-nae that span more than 120 genera (Musserand Carleton, 1993; Nowak, 1999). The sub-family Murinae encompasses a broad range ofspecies of diverse ecological and morphologi-cal forms that include several specialized gen-era that resemble species from other sub-families, families, and orders. For example,Hydromys resemble aquatic shrews; Echio-thrix, elephant shrews; Komodomys, gerbils;Crateromys, squirrels; and Nesokia, gophers.Other genera are more "typically" murine,such as Rattus and the laboratory or house,mouse (Mus musculus). Other murid rodents,such as hamsters, gerbils, and voles, are di-vided into several other subfamilies (Fig. 1-3).

In terms of the relationships between themurines and other murid subfamilies, there isgeneral agreement that Gerbillinae is thesister-group to Murinae on the basis of bothmorphological (Flynn et al., 1985) and molec-ular (Dubois et al., 1999; Michaux et al., 2001)evidence (Fig. 1-3). The position of severalother subfamilies has, until recently, remainedcontroversial (Chevret et al., 1993; Dubois etal., 1999; Michaux et al., 2001).

There is some debate over what repre-sents the first murine (Flynn et al., 1985). Someauthors claim that it is the middle Miocene (15

6

Page 24: 0195162854

Chapter 1. Evolution

Figure 1-2. These two phylogenetictrees presented illustrate two alter-native topologies of the interrela-tionships between rodent families.(A) The maximum likelihood treederived from reconstructions ofthree nucleotide sequences (Huchonet al., 2002). Here, the Muridae(shown in bold) forms a clade withthe pocket gophers (Geomyoidea) asa sister-group. (B) In contrast, thissecond tree represents the one ofseveral topologies depicted in Ad-kins et al. (2001). Although differ-ences were present in the interfa-milial relationships between sometrees, all of them agree with placingthe jerboas and kangaroo rats(Dipodoidea) as sister-group to theMuridae (shown in bold). The otherfamilies and superfamilies includesquirrels (Sciuridae), mountainbeaver (Apladontia rufa) (Aplodonti-dae), dormice (Gliroidea), scaly-tailedflying squirrels (Anomaluridae),Springhare (Pedetes capensis) (Pedeti-dae), beavers (Castoridae), gundis(Ctenodactylidae), cane rats (Thry-onomyoidea), African mole rats(Bathyergoidea), Old World porcu-pines (Hystricoidea), New Worldporcupines (Erethizontoidea), chin-chillas (Chinchillidae), degus (Octo-dontidae), and guinea pigs and cavies(Cavoidea).

Sciuridae

Aplodontidae

Gliroidea

MuroideaGeomyoidea

Anomaluridae

Pedetidae

Dipodoidea

Castoridae

Ctenodactylidae

Hystricoidea

Thryonomyoidea

Bathyergoidea

Octodontoidea

Erethizontoidea

Chinchilloidea

Cavioidea

Sciuridae

Aplodontidae

Gliroidea

MuroideaDipodoidea

Pedetidae

Castoridae

Geomyoidea

Ctenodactylidae

Hystricoidea

Thryonomyoidea

Bathyergoidea

Erethizontoidea

Octodontoidea

Chinchilloidea

Cavioidea

to 16 mya) Antemus from Thailand (Jacobs,1977; Hand, 1984; Jaeger et al., 1986). The in-clusion of Antemus in the Murinae has, how-ever, been questioned because of morpho-logical similarities with other subfamilies.Therefore, others consider Progonomys to bethe earliest murine (Flynn et al., 1985) at amore recent age of 11 to 12 mya. Due to thefragmentary material recovered thus far, noresolution has been posited for the affinity ofAntemus, but an origin of 15 to 16 mya agreeswith molecular estimates (see later).

The point at which the murines divergedfrom the other subfamilies is difficult to pin-

point, as they appear to have first evolved incentral Asia, where mammalian palaeontol-ogy is still in its early stages. Based on theavailable evidence, a divergence time of be-tween 16 and 23.8 mya (i.e., late Oligocene toearly Miocene) seems likely (Hartenberger,1998; Tong and Jaeger, 1993). This is sup-ported by molecular estimates of 17.9 to 20.8mya (Michaux et al. 2001). In contrast, thehamsters (Cricetinae) are known from middleEocene deposits (Hartenberger, 1998) datingbetween 36.5 and 49 mya. Thus, the murinesunderwent an explosive radiation relativelylate in murid rodent evolution.

7

A

B

Page 25: 0195162854

- Nanospalax ehrenbergi

- Rhizomys pruinosus

- Tachyoryctes sp.

- Macrotarsomys ingens

- Nesomys rufus

- Steatomys sp.

- Dendromus mystacialis

- Saccostomus campestris

- Cricetomys gambianus

- Mystromys albicaudatus

- Calomyscus mystax

- Dicrostonyx torquatus

- Clethrionomys glareolus

- Neotoma fuscipes

- Peromyscus maniculatus

- Myospalax sp.

- Phodopus roborovskii

- Mesocricetus auratus

- Cricetulus migratorius

- Tatera gambiana

- Gerbillus henleyi

- Lophuromys sikapusi

- Deomys ferrugineus

- Uranomys ruddi

- Acomys cahirinus

- Rattus norvegicus

- Micromys mintttus

- Otomys angoniensis

- Mus musculus

NATURAL HISTORY

Spalacinae

Rhizomyinae

Nesomyinae

Dendromurinae

Cricetomyinae

Mystromyinae

Calomyscinae

Arvicolinae

Sigmodontinae

Cricetinae

Gerbillinae

Acomyinae

Murinae +Otomyinae

Figure 1-3. This phylogenetic tree depicts the interrelationships between subfamilies within the Muridaebased on a combined analysis of LCAT and vWF genes (Michaux et al., 2001). Note that the spiny mice(Acomyinae) and gerbils (Gerbillinae) form a sister-group relationship with a clade composed of the Murine(in bold) and the vlei rats (Otomyinae). The other subfamilies include hamsters (Cricetinae), voles and lem-mings (Arvicolinae), blind mole rats (Spalacinae), bamboo rats (Rhizomyinae), New World murids (e.g., deermice, wood rats, muskrat) (Sigmodontinae), mouse-like hamsters (Calomyscinae), pouched mice and rats(Cricetomyinae), and several diverse African clades (Nesomyinae, Dendromurinae, and Mystromyinae).

GENUS Rattus

The status of the genus Rattus has undergonenumerous changes since its first use by Fischer

genera that were once considered to belongto the genus Rattus have been further subdi-vided into other genera, such as Praomys,Mastomys, and Apomys (Nowak, 1999). There

in 1803 (Musser and Carleton, 1993). Several is, in fact, still some debate regarding how to

8

Page 26: 0195162854

Chapter 1. Evolution

define the genus itself (Carleton and Musser,1984; Musser and Holden, 1991; Musser andCarleton, 1993). Despite these problems, Rat-tus can generally be distinguished from othermurine genera by long body fur, sparselyhaired tails with overlapping scales, and stoutskulls with relatively large auditory bullae andprominent coronoid processes (Watts andAplin, 1981). These characters are by no meansdefinitive, however, as they are also present inother genera. A precise description of charac-ters that typify the genus is currently wanting.

Coinciding with the uncertainty of theboundaries of the genus Rattus are uncertain-ties of the relationships between murine gen-era. A number of studies have depicted phy-logenetic relationships within the Murinae(Robinson et al, 1997; Martin et al., 2000;Suzuki et al., 2000; Michaux et al., 2001), butWatts and Baverstock (1995) provided themost comprehensive number of species. Intheir analysis, Watts and Baverstock (1995)found that four biogeographical clades couldbe recognized: southeast Asian, Australasian,New Guinean, and African (Fig. 1-4). Basedon their analyses, they suggested that the firstmurines arose 20 mya, leading to the basalmembers of the lineage. This is supported byfossil evidence that estimated the divergencetime between Rattus and Mus at approxi-mately 12 mya (Jaeger et al., 1986). At 8 mya,the African, southeast Asian, and Australasianclades underwent a rapid speciation event,resulting in "bushy" phylogenetic trees andthe aforementioned problems of delimitinggeneric boundaries. The age at which Rattusdiverged from other members of the Asianclade is uncertain, but based on Watts andBaverstock's (1995) phylogeny, it would haveoccurred within the past 8 million years (seeFig. 1-4). This estimate agrees with the earli-est Rattus fossil being recorded from the latePliocene (<3 mya) of China (Xue, 1981, inJaeger et al., 1986). Such a recent origin of Rat-tus suggests that their occurrence in Australa-sia occurred well after other murines first col-onized the area 4 to 6 mya (Hand, 1984).

Figure 1-4. This phylogenetic tree depicts intergeneric re-lationships within the Murinae based on microcomplementfixation of albumin (Watts and Baverstock, 1995). Both Rat-tus and Mus are shown in bold, and the date of divergencebetween these two genera is approximately 12 million years(Jaeger et al., 1986). Note that although Acomys, Lophuromys,and Uranomys were included in the murine phylogeny ofWatts and Baverstock (1995), more recent studies haveshown that they form a monophyletic clade separate fromthe Murinae (Dubois et al., 1999; Michaux et al., 2001) (seeFig. 1-3).

9

MaxomysNiviventerTokudaiaLeopoldomysBunomysKomodomysSundamysStenomysRattusBullimusTaeromysNesokiaBandicotaParuromysPapagomysBerylmysParahydromysHydromysLeggadinaPseudomysNotomysConilurusMesembrionomysLeporillusZyzomysMayermysNeohydromysPseudohydromysLeptomysCrossomysXeromysUromysSolomysMelomysLorentzimysPogonomysChiruromysAnisomysCoccymysHyomysMacruromysMallomysApodemusMusMyomysMastomysPraomysOtomysHylomyscusStochomysHybomysGrammomysThallomysRhabdomysPelomysLemniscomysAethomysArvicanthisDasymysMillardiaMicromysVandeleuraPhloeomysAcomysUranomysLophuromys

Page 27: 0195162854

10 NATURAL HISTORY

Rattus norvegicus

Following this pattern of most speciose line-ages, the species R. norvegicus is a member ofthe most speciose genus of murid rodentswith up to 50 species being recognized (Table1-1) (Musser and Carleton, 1993; Nowak,1999). As discussed earlier, the status of thisgenus is uncertain, but in general they have asimilar appearance to R. norvegicus. In Aus-tralia, R. norvegicus differs from native Rattusprimarily by their longer tails and larger skulls,but also there are differences in tail and footpad coloration, a tapered snout, and coarse fur(Watts and Aplin, 1981).

Another feature that can aid in distin-guishing R. norvegicus from other Rattusspecies is that it occurs in urban and other dis-turbed environments. This is not, however, aunique feature of R. norvegicus as severalRattus species are commensal. That is, theyare found primarily close to human habitationand prefer this habitat to others. Some of thesespecies are well known pests in a variety of lo-cales, such as the Polynesian Rat or Kiore (R.exulans), Black Rat (R. rattus}, and R. norvegi-cus. The other species (see Table 1-1) are lesswell known by Western researchers but areno less destructive (e.g., Leung et al., 1999).

Some of the remaining Rattus species canalso be found in close proximity to humansbut generally occur in more remote habitats.For example, the Australian Bush Rat (R.juscipes) is often found in suburban areas(Watts and Aplin, 1981; Menkhorst, 1995;Strahan, 1998) and the Long-haired Rat(R. villossissimus) can overwhelm towns andfarms when in plague proportions (Watts andAplin, 1981; Strahan, 1998). These are not con-sidered to be commensal species, however, asthey do not depend on human-altered envi-ronments. In fact, most Rattus species preferundisturbed areas in environments that in-clude arid plains, rain forests, coastal heathand subalpine regions.

Structurally, there are few differences inbehavior between R. norvegicus and other

Rattus species (Begg and Nelson, 1977; Bar-nett et al., 1982; Beeman 2002). In general,R. norvegicus tend to be more aggressive intheir social interactions than other Rattusspecies that have been examined (Barnett etal., 1982), but this is true only for "wild"norvegicus as laboratory strains tend be moredocile (see Chapters 2 and 3). Behaviorally,all Rattus are best described as generalistspecies that are flexible to environmentalperturbations. It is perhaps this flexibility thathas enabled their success as introducedspecies. It should be noted, however, thatthis behavioral flexibility is not unlimited andthat some species are susceptible to extremeenvironmental changes. For example, twospecies of Rattus native to Christmas Island(macleari and nativitatis) have become extinctsince European settlement (Nowak, 1999). Inaddition, two species are currently listed asendangered (baluensis and enganus) and 13more as vulnerable (International Union forConservation of Nature and Natural Re-sources, 2002). Thus, even the seemingly lim-itless behavioral adaptability of Rattus speciesappears to have its limits.

In terms of their interspecific relation-ships, relatively little is known. Musser andCarleton (1993) described five groupings ofRattus species: (1) norvegicus (1 species), (2) aR. rattus group composed of various Asianspecies (21 species), (3) a native Australiangroup (6 species), (4) a native New Guineagroup (8 species), and (5) a Sulawesi/Phillip-ines Islands group (5 species). The remaining10 species that Musser and Carleton (1993) in-clude in Rattus are described as possessing un-resolved phylogenetic affinities. Of particularinterest is that norvegicus is sufficiently dis-tinctive to warrant its own monotypic group.The distinctiveness of norvegicus is further sup-ported by several molecular studies (Chan,1977; Chan et al., 1979; Baverstocket al., 1986;Verneau et al., 1997; Suzuki et al., 2000;Dubois et al., 2002; but see Pasteur et al., 1982)(Fig. 1-5). How R. norvegicus is related to otherRattus species and divergence dates between

Page 28: 0195162854

Chapter 1. Evolution 11

-R. tunneyi

• R. villosissimus

. R. colletti

• R. leucopus

• R. fuscipes

. R. norvegicus

• R. rattus

. R. tiomanicus

. R. exulans

• R. argentiventer

• R. norvegicus

• Berylmys bowersii

• R. annandalei

Figure 1-5. These two phylogenetic trees depict the rela-tionships between Rattus species. (A) This tree is derivedfrom Baverstock et al. (1986) and includes the endemic Aus-tralian species and R. norvegicus (shown in bold). (B) Thissecond tree is derived from Chan et al. (1979). Berylmysbowersii is included because Rattus annandalei was shown tobe more closely related to it than other Rattus species.

the different species has remained largelyunexplored.

R. norvegicus itself is presumed to haveoriginated in Asia, but its long associationwith humans makes a more precise locationdifficult to determine. Two suggested pointsof origin are the steppes north of the CaspianSea (Matthews, in press) and northern China(Musser and Carleton, 1993). From this gen-eral area, it spread throughout Europe in themid-1700s. By the late 1800s, norvegicus, andR. rattus, had spread across much of NorthAmerica as well as successfully invading in-numerable ocean ports around the world.

WHAT DOES IT ALL MEAN?

Given what is known about the evolutionaryhistory R. norvegicus, how can this informationbe used to understand the biology and be-havior of R. norvegicus^ Cross-species analysesof anatomy, behavior, and ecology can yieldprofound insights into the ultimate mecha-nisms underlying species phenorypes as wellas their possible functions. For example, thecomparative approach has been instrumentalin understanding the evolution of behavior(Martins, 1996; Lee, 1999) and the nervoussystem (Butler and Hodos, 1996).

From a comparative perspective, infor-mation on evolutionary history can be usedto examine the evolution of the behavior ofR. norvegicus. One example is provided by acomparative analysis of social play behaviorin muroid rodents (Pellis and Iwaniuk, 1999).R. norvegicus possesses a highly complex formof social play (see Chapter 28) compared withother rodents. There are a number of reasonsthat complex play has evolved in R. norvegi-cus, such as evolutionary history, social sys-tem, ecology, and others. Using a phylogenyof 13 murids, Pellis and Iwaniuk (1999) ana-lyzed several features of murid rodent socialplay and assessed the relative importance ofphylogeny and social system on the evolutionof these features. Contrary to many other be-haviors that are closely linked with phyloge-netic history, the complexity of social play didnot conform to any observed pattern acrossthe phylogeny. Further analyses indicated thatthe degree of sociality was positively corre-lated with the complexity of social play. Thus,the more social a murid, the more complex isits play. Last, using play traits to constructtheir own "phylogeny," Pellis and Iwaniuk(1999) demonstrated that the play of R.norvegicus most closely resembled that of theGolden Hamster (Mesocricetus auratus) ratherthan other murines (see Fig. 1-3). This meansnot only that play is variable within muroidsbut also that convergent forms of play can andhave evolved in divergent rodent lineages.

A

B

Page 29: 0195162854

12 NATURAL HISTORY

The implications that this has for theories con-cerning the evolution of play in murids andother mammals are far-reaching and are stillbeing investigated (e.g., Iwaniuk et al., 2002;Pellis and Iwaniuk, 2004).

Without a phylogeny to use with phylo-genetically based statistical methods, thestudy of Pellis and Iwaniuk (1999) would havebeen superficial. In fact, further interspecificcomparisons of the behavior of murids andmurines could yield similar insight into the be-havioral evolution of R. norvegicus. Compara-ble data are not, however, available for manyother rodents. This curtails our ability to de-termine whether observations made on thebehavior and biology of R. norvegicus are rep-resentative of rodents or, for that matter,other mammals. For example, are many of thebehaviors of R. norvegicus typical of all Rattusspecies or specific to R. norvegicust Thus, ourunderstanding of R. norvegicus behavior wouldbe greatly enlightened by further studies onother murines and Rattus species.

CONCLUSION

The evolution of the laboratory rat is charac-terized by a series of explosive radiations thatoccurred within relatively narrow time framesat the level of order, family, subfamily, andgenus. As a result of these evolutionary radi-ations, the phylogenetic relationships be-tween many taxa remain unresolved. Simi-larly, the fossil record is patchy and the firstappearance of many taxa is debatable due tomultiple instances of convergent evolution ofmorphology throughout the evolutionary his-tory of rodents. The recent use of more com-prehensive species sampling and multiplegenes is painting a clearer picture of rodentevolution, but there are many unansweredquestions. In particular, there are many un-certainties regarding the origin and relation-ships between and within the genus Rattusthat could aid in delineating the boundaries of

this speciose genus. Furthermore, future re-search into the radiation of the genus Rattuscould yield insight into the general features ofradiative evolution as well as the potential forcommensalism to aid in dispersal and specia-tion. From the perspective of understandingthe behavior of R. norvegicus, such researchwill yield insight into how and why somebehaviors and not others have evolved inR. norvegicus.

ACKNOWLEDGMENTS

I would like to thank the editors, Ian Whishaw and Bryan Kolb,for inviting my contribution to this volume and for their edi-torial assistance and Karen Dean for reviewing earlier drafts ofthis chapter. Financial support was provided by a Monash Uni-versity Postgraduate Publications Award to the author.

REFERENCES

Adkins RM, Gelke EL, Rowe D, Honeycutt RL (2001)Molecular phylogeny and divergence time estimatesfor major rodent groups: evidence from multiplegenes. Molecular Biology and Evolution 18:777-791.

AlroyJ (1999) The fossil record of North American mam-mals: evidence for a Paleocene evolutionary radia-tion. Systematic Biology 48:107-118.

Archibald JD, Averianov AO, Ekdale EG (2001) LateCretaceous relatives of rabbits, rodents, and otherextant eutherian mammals. Nature 414:62-65.

Barnett SA, Fox IA, Hocking WE (1982) Some social pos-tures of five species of Rattus. Australian Journal ofZoology 30:581-601.

Baverstock PR, Adams M, Watts CHS (1986) Biochem-ical differentiation among karyotypic forms of Aus-tralian Rattus. Genetica 71:11-22.

Begg RJ and Nelson JE (1977) The agonistic behaviourof Rattus villosissimus. Australian Journal of Zoology25:291-327.

Beeman C (2002) An analysis of skilled forelimb move-ments in the Bush Rat (Rattus fuscipes). Unpublishedhonours thesis, Monash University, Clayton.

Bronham L, Phillips MJ, Penny D (1999) Growing upwith dinosaurs: molecular dates and the mam-malian radiation. Trends in Ecology and Evolution14:113-118.

Butler AB and Hodos W (1996) Comparative vertebrateneuroanatomy: evolution and adaptation. NewYork: Wiley-Liss.

Page 30: 0195162854

Chapter 1. Evolution 13

Carleton MD (1984) Introduction to rodents. In: Orderand families of recent mammals of the world (An-derson S and Jones JK Jr, eds.), pp. 255-265. NewYork: John Wiley & Sons.

Carleton MD and Musser GG (1984) Muroid rodents.In: Order and families of recent mammals of theworld (Anderson S and Jones JK Jr, eds.), pp.289-380. New York: John Wiley & Sons.

Chan KL (1977) Enzyme polymorphism in Malayan ratsof the subgenus Rattus. Biochemical Systematicsand Ecology 5:161-168.

Chan KL, Dhaliwal SS, Yong HS (1979) Protein varia-tion and systematics of three subgenera of Malayanrats (Rodentia: Muridae, genus Rattus Fischer).Comparative Biochemistry and Physiology B 64:329-337.

Chevret P, Denys C, Jaeger J-J, MichauxJ, Catzeflis FM(1993) Molecular evidence that the spiny mouse(Acomys) is more closely related to gerbils (Gerbilli-nae) than to true mice (Murinae). Proceedings ofthe National Academy of Sciences U.S.A. 90:3433-3436.

D'Erchia AM, Gissi C, Pesole G, Saccone C, Arnason U(1996) The guinea-pig is not a rodent. Nature 381:597-600.

Dubois J-YF, Catzeflis FM, Beintema JJ (1999) The phy-logenetic position of "Acomyinae" (Rodentia, Mam-malia) as sister group of Murinae + Gerbillinaeclade: evidence from the nuclear ribonuclease gene.Molecular Phylogenetics and Evolution 13:181-192.

Dubois J-YF, Jekel PA, Mulder PPMFA, Bussink AP,Catzeflis FM, Carsana A, Beintema JJ (2002) Pan-creatic-type ribonuclease 1 gene duplications in ratspecies. Journal of Molecular Evolution 55:522-533.

Eizirik E, Murphy WJ, O'Brien SJ (2001) Molecular dat-ing and biogeography of the early placental mam-mal radiation. Journal of Heredity 92:212-219.

Felsenstein J (1985) Phylogenies and the comparativemethod. The American Naturalist 126:1-25.

Flynn LJ, Jacbos LL, Lindsay EH (1985) Problems inmuroid phylogeny: relationship to other rodentsand origin of major groups. In: Evolutionary rela-tionships among rodents: a multidisciplinary analy-sis (Luckett WP and Hartenberger J-L, eds.), pp.589-618. New York: Plenum Press.

Foote M, Hunter JP, Janis CM, SepkoskiJJJr (1999) Evo-lutionary and preservational constraints on originsof biologic groups: divergence times of eutherianmammals. Science 283:1310-1314.

Graur D, Hide WA, Li W-H (1991) Is the guinea-pig arodent? Nature 351:649-652.

Hand S (1984) Australia's oldest rodents: mastermariners from Malaysia. In: Vertebrate zoogeogra-phy and evolution in Australasia (Archer M and

Clayton G, eds.), pp. 905-912. Sydney, NSW: Hes-perian Press.

Hartenberger J-L (1998) Description de la radiation desRodentia (Mammalia) du Paleocene superieur auMiocene; incidences phylogenetiques. ComptesRendus de I'Academie Sciences, Paris, Sciences dela Terre et des Planetes 326:439-444.

Hartenberger J-L (1996) Les debuts de la radiation adap-tative des Rodentia (Mammalia). Comptes Rendusde rAcademie Sciences, Paris, Sciences de la Terreet des Planetes 323:631-637.

Harvey PH and Pagel MD (1991) The comparativemethod in evolutionary biology. Oxford: OxfordUniversity Press.

Hedges SB, Parker PH, Sibley CG, Kumar S (1996) Con-tinental breakup and the ordinal classification ofbirds and mammals. Nature 381:226-229.

Huchon D, Catzeflis FM, Douzery EJP (2000) Varianceof molecular datings, evolution of rodents and thephylogenetic affinities between Ctenodactlyidaeand Hystricognathi. Proceedings of the Royal Soci-ety of London, Series B 267:393-402.

Huchon D, Madsen O, Sibbald MJJB, Ament K, Stan-hope MJ, Catzeflis F, de Jong WW, Douzery EJP(2002) Rodent phylogeny and a timescale for theevolution of Glires: evidence from an extensivetaxon sampling using three nuclear genes. Molecu-lar Biology and Evolution 19:1053-1065.

International Union for Conservation of Nature andNatural Resources (2002) 2002 IUCN red list ofthreatened species. Available at: http://www.iucn.org. Acessedjune 5, 2003.

IwaniukAN, Nelson JE, Pellis SM (2001) Do big-brainedanimals play more? comparative analyses of playand relative brain size in mammals. Journal of Com-parative Psychology 115:29-41.

Jacobs LL (1977) A new genus of murid rodent from theMiocene of Pakistan and comments on the origin ofthe Muridae. Paleobios 25:1-11.

Jaeger J-J, Tong H, Buffetaut E (1986) The age of Mus-Rattus divergence: paleontological data comparedwith the molecular dock. Comptes Rendus deI'Academie Science, Paris, Sciences de la Terre etdes Planetes 302:917-922.

Kumar S and Hedges SB (1998) A molecular timescalefor vertebrate evolution. Nature 392:917-920.

Lavocat R and Parent J-P (1985) Phylogenetic analysisof middle ear features in fossil and living rodents.In: Evolutionary relationships among rodents: amultidisciplinary analysis (Luckett WP and Harten-berger J-L, eds.), pp. 333-354. New York: PlenumPress.

Lee PC (ed.) (1999) Comparative primate socioecology.Cambridge: Cambridge University Press.

Page 31: 0195162854

14 NATURAL HISTORY

Leung LKP, Singleton GR, Sudarmaji R (1999) Ecologi-cally-based population management for the rice-fieldrat in Indonesia. In: Ecologically-based managementof rodent pests (Singleton GR, Hinds LA, Leirs H,Zhang Z, eds.), pp. 305-318. Canberra: AustralianCentre for International Agricultural Research.

Luckett WP and Hartenberger J-L (1993) Monophyly orpolyphyly of the order Rodentia: possible conflictbetween morphological and molecular interpreta-tions. Journal of Mammalian Evolution 1:127-147.

Luckett WP and Hartenberger J-L (1985) Evolutionaryrelationships among rodents: comments and con-clusions. In: Evolutionary relationships among ro-dents: a multidisciplinary analysis (Luckett WP andHartenberger J-L, eds.), pp. 685-712. New York:Plenum Press.

Madsen P, Scally M, Douady CJ, Kao DJ, DeBry RW,Adkins R, Amrine H, Stanhope MJ, de Jong WW,Springer MS (2001) Parallel adaptive radiations intwo major clades of placental mammals. Nature409:610-614.

Martin Y, Gerlach G, Schlotterer C, Meyer A (2000) Mo-lecular phylogeny of European muroid rodentsbased on complete cytochrome b sequences. Mo-lecular Phylogenetics and Evolution 16:37-47.

Martins EP (ed.) (1996) Phylogenies and the compara-tive method in animal behavior. Oxford: OxfordUniversity Press.

Mathews F (2004) Rattus norvegicns. Mammalian Species,in press.

Menkhorst PW (1995) Mammals of Victoria: distribu-tion, ecology and conservation. Melbourne: OxfordUniversity Press.

Michaux J, Reyes A, Catzeflis F (2001) Evolutionary his-tory of the most spetiose mammals: molecular phy-logeny of muroid rodents. Molecular Biology andEvolution 18:2017-2031.

Murphy WJ, Eizirik E, Johnson WE, Zhang YP, RyderOA, O'Brien SJ (2001a) Molecular phylogenetics andthe origins of placental mammals. Nature 409:614-618.

Murphy WJ, Eizirik E, O'Brien SJ, Madsen O, Scally M,Douady CJ, Teeling E, Ryder OA, Stanhop MJ, deJong WW, Springer MS (200Ib) Resolution of theearly placental mammal radiation using Bayesianphylogenetics. Science 294:2348-2351.

Musser GG and Carleton MD (1993) Family Muridae.In: Mammal species of the world: a taxonomic andgeographic reference, 2nd edition (Wilson DE andReader DM, eds.), pp. 501-756. Washington, DC:Smithsonian Institution Press.

Musser GG and Holden ME (1991) Sulawesi rodents(Muridae, Murinae): morphological and geographi-cal boundaries of species in the Rattus hoffmannigroup and a new species from Pulau Peleng, MalayArchipelago. Bulletin of the American Museum ofNatural History 206:322-413.

Novacek MJ (1992) Fossils, topologies, missing data, andthe higher level phylogeny of eutherian mammals.Systematic Biology 41:58-73.

Nowak RM (1999) Walker's mammals of the world, 6thedition. Baltimore: Johns Hopkins University Press.

Pasteur N, Worms J, Tohari M, Iskandar D (1982) Ge-netic differentiation in Indonesian and French ratsof the subgenus Rattus. Biochemical Systematicsand Ecology 10:191-196.

Pellis SM and Iwaniuk AN (2004) Evolving a playfulbrain: a levels of control approach. InternationalJournal of Comparative Psychology, in press.

Pellis SM and Iwaniuk AN (1999) The roles of phylogenyand sociality in the evolution of social play inmuroid rodents. Animal Behaviour 58:361-373.

Robinson M, Catzeflis F, Briolay J, Mouchiroud D (1997)Molecular phylogeny of rodents, with special empha-sis on murids: evidence from the nuclear gene LCAT.Molecular Phylogenetics and Evolution 8:423-434.

Shoshani J and McKenna MC (1998) Higher taxonomicrelationships among extant mammals based onmorphology, with selected comparisons of resultsfrom molecular data. Molecular Phylogenetics andEvolution 9:572-584.

Strahan R (1998) Mammals of Australia. Sydney, NSW:Reed Books/New Holland Press.

Suzkui H, Tsuchiya K, Takezaki N (2000) A molecularphylogenetic framework for the Ryukyu endemicrodents Tokudaia osimensis and Diplothrix legata.Molecular Phylogenetics and Evolution 15:15-24.

Tong H and Jaeger J-J (1993) Muroid rodents from theMiddle Miocene Fort Ternan locality (Kenya) andtheir contribution to the phylogeny of muroids.Paleontographica Abteilung A 229:51-73.

Verneau O, Catzeflis F, Furano AV (1997) Determina-tion of the evolutionary relationships in Rattus sensolato (Rodentia: Muridae) using LI (LINE-1) ampli-fication events. Journal of Molecular Evolution45:424-436.

Watts CHS and Aplin HJ (1981) Rodents of Australia.Sydney, NSW: Angus & Robertson Press.

Watts CHS and Baverstock PR (1995) Evolution in theMurinae (Rodentia) assessed by microcomplementfixation of albumin. Australian Journal of Zoology43:105-118.

Page 32: 0195162854

Ecology

S. ANTHONY BARNETT2

Among mammals, the miscalled "Norwayrat" (Rattus norvegicus), known also as the"brown rat" although it is often gray butsometimes black, is rivaled as a pest only bythe house mouse (Mus domesticus vel muscu-lus) and the "ship" or "black" rat (Rattus rat-tus), which as a rule is not black. R. norvegicusalso displays peculiar features that make it ofspecial interest to both ethologists and exper-imental psychologists. This chapter providesan overview of the ecology of the wild R.norvegicus.

ENVIRONMENTS: RATSAS COMMENSALS

The hordes of rats that are present in somehuman communities are due to our growingand storing of great concentrations of food,construction of shelter in buildings and drainsand at the edges of fields, and killing of thecarnivorous species that prey on small mam-mals. Moreover, we often kill rats only whenthey are numerous. The resulting slaughtermay be impressive, but the likely result is asurviving population that can breed at a highrate.

In agricultural land and gardens and onthe banks of canals and streams (in which Nor-way rats readily swim), they make extensiveburrows that form irregular systems ofbranching and conjoining passages in whichthey breed.

Norways climb well, although not as wellas R. rattus. Until the 18th century in north-

ern Europe, R. rattus was without competi-tors, but when Norways invaded from theeast, they largely replaced the black rat. Build-ings in cities often housed R. rattus in the at-tics and Norway rats on the lower floors. Nor-ways are also numerous in large sewers; inthose of at least one American city, they aresaid to be the main food of a population offeral alligators. However, urban environ-ments will harbor few rats if they have well-maintained buildings, a modern sewer sys-tem, and a high standard of hygiene.

Rarely, Norways live independent of peo-ple, such as on seashores.

Seeking shelter is a feature of rat behav-ior. When food can be carried, it is eaten in anest or under cover. It may also be hoardedfor consumption later.

In burrows and elsewhere, rats buildnests of straw or of fragments of textile ma-terial or paper dexterously shredded andarranged. In addition to nesting material,items sometimes carried to the nest includefragments of wood, stones, and cakes of soap.Nest building is enhanced by cold.

Females rear their litters in nests thatthey defend against intruders. The move-ments of a female retrieving strayed young re-semble those of a rat hoarding objects.

EXPLORATION, NEOPHILIA,AND NEOPHOBIA

In stable conditions, members of establishedcolonies move, usually at night, on beaten

15

Page 33: 0195162854

16 NATURAL HISTORY

tracks between nests and sources of food, wa-ter, and bedding. They commonly run withtheir vibrissae and fur in contact with a verti-cal surface. Rubbing of the fur leaves a smearfrom skin secretions; the odor of the smear isthought to attract other rats.

Norways are also highly exploratory andrange widely from settled pathways. Move-ment in the living space (home range) can beanalyzed under three categories. (1) Move-ments may be a search for food, water, nest-ing material, or shelter; movements thencease when the objective is reached. (2) Smallmammals such as Norway rats regularly pa-trol the whole of their home range; places thathave not been recently visited are likely to beapproached first. (3) Exploratory movements,especially of new places, may continue whenall basic needs seem to be satisfied. The sec-ond and third types of movement are exam-ples of neophilia, or the tendency to approachwhat is unfamiliar.

The three aspects of movement can bedistinguished experimentally. Figure 2-1shows a "plus maze" in which a wild Norwaymay live for many days. Suppose a rat, alreadyfamiliar with the maze, is confined in the nestbox, without food, for some hours, and thearms of the maze are then opened. The ratsoon moves out and, during the next fewhours, spends much time eating and drinking;each meal, however, is followed by a brief pa-trol of all parts of the environment, includingthose that contain no food or other reward.

If different palatable foods are offered inthe arms, the rat makes a meal of one; then,as it patrols the maze, it also samples the otherfoods. Sampling is an aspect of neophilia. It isappropriate for an omnivore, and, as we seelater, it is important for food selection.

A rat may also be allowed to become fa-miliar with a maze with access to only threearms. If the fourth arm is then opened and sogives access to a new place, it is quickly ex-plored. Such neophilic behavior is typical ofsmall mammals.

Wild Norways are, however, atypical intheir response to strange objects. A commonexperience for farmers and warehousekeepersis to lay traps or bait on runways and to findthat the Norways then give up their usualnightly visits, as if aware of danger. This neo-phobia is a response to a discrepancy. A novelobject in a familiar place is avoided. It is theprincipal explanation of the popular reputa-tion of rats for intelligence. If, in a plus maze,a strange object is put in an arm that has al-ready been explored, farther entry into thearm is delayed. Similar avoidance has beensystematically observed among rats in freepopulations.

A decline in food consumption may beused as an index of the avoidance of a new ob-ject. However, the automatic recoil, as shownby wild Norways, is an indiscriminate avoid-ance of any new object in a familiar place, ed-ible or inedible, dangerous or harmless, usu-ally at a distance. It is not a response to astrange odor or taste.

In their response to new objects, the do-mestic varieties differ greatly from the wildtype. Kept in small cages and offered food ina strange container, domestic rats soon ap-proach the container and continue to eat asbefore. Wild Norways, in identical conditions,may remain at the back of the cage and stopeating for days.

Other species of Muridae that are de-pendent on human communities are neopho-bic, but the independent species of the genus,Rattus, so far studied, is not. Hence neopho-bia is probably a result of natural selectionamong rats that have lived with our ancestorsat least since a settled agriculture began.

The neophobic response is, however, notirrevocably fixed ("innate" or "instinctive").Wild Norways infesting a landfill, in whicheverything is, in effect, likely to be a new ob-ject, are hardly neophobic. In their unstableenvironment, they habituate to incessantchange and so develop a novel response tonovelty.

Page 34: 0195162854

Chapter 2. Ecology 17

Figure 2-1. A residential maze that allows experiments that last many days. An animal lives in the centralbox. Food, water, nesting material, or other objects can be offered at the ends of the arms. Visits to the armsand duration of stay are recorded and analyzed with a computer.

FOOD

MEALS AND ENERGY INTAKE

Like other rodents, Norways are equipped tocope with hard foods such as seeds and nuts;they can even gnaw through lead pipes. Typ-ically, a small, hard object, such as a wheatgrain, is held in the forefeet while it is eaten.Individuality, however, is shown in feedingpatterns; one rat, eating flour, may bury itsnose in the food, while another sits, dexter-ously scooping the food into its mouth withone paw; yet another may use both forepaws.

In a stable environment, wild rats eat inthe darkness, a few grams at a time, at regu-lar intervals. In the short term, adult Norwaysadjust these meals to maintain a steadyweight, but they grow slowly throughoutmost of their adult lives, up to about 700grams in the most favorable conditions.

The mealtimes of rats can be altered bytraining. If food is made available at only one

time in each 24-hour period, rats regularly as-semble at that time. In a colony of wild Nor-ways, individual mealtimes may also be influ-enced by dominance relationships: the oldestor heaviest rats have priority at food sources.

FOOD SELECTION

Wild Norways with access to a variety offoods may seem to be indiscriminate. Theymay prey on small birds or mammals; someeat snails; and those that live on the banks ofstreams may eat fish. Wild Norways that feedlargely on wheat grains quickly accept newlyavailable alternative foods, such as cabbageleaves or raw meat.

The selection of foods is influenced by fla-vor, odor or texture. Rats favor sweet mix-tures, whether they contain sugar or saccha-rin. Wild rats also prefer finely dividedwheatmeal to whole grains. Some edible oils,such as arachis, added to grains increase ac-ceptance, but butyric acid or aniseed oil is a

Page 35: 0195162854

18 NATURAL HISTORY

deterrent. (Some domestic Norways readilyaccept both.)

The most notable influences on feedingare related to the internal effects of foods. Aprominent feature of selection by Norway ratsis their ability to acquire aversions. The pre-ceding section describes the avoidance ofnovelty (neophobia), which is typical of thespecies. Another, very different kind of with-drawal, shared with other mammals, arisesfrom adverse, individual experience.

A pile of unfamiliar food on a rats' run-way is at first avoided: it is a "new object" (theneophobic response). The delay is greater ifthe food is in an unfamiliar container. How-ever, eventually the food is sampled. Afterthis, the food is again avoided, as if "fear" and"curiosity" (or hunger) are opposed; the for-mer gradually gives place to unhesitating con-sumption.

If, however, the food contains a poison,the small amount first taken may cause illnessbut not be lethal. For a time, the animal thenstops eating. When it resumes eating, as a ruleit refuses the toxic mixture—it has become"bait shy." It may even reject constituents ofthe mixture with a distinct taste, such as sugar.The survival value of neophobia, combinedwith acquired aversions to poisoned foods,has been observed in many field experiments.These two features make infestations of Nor-ways (and of R. rattus) difficult to manage.

For experimental psychologists, suchaversions have two unexpected features. (1) Inconventional experiments on learning, habitsare usually acquired slowly and require manytrials, but aversions arise from a single expe-rience. (2) In laboratory experiments, the in-terval between a stimulus and the animal's re-sponse is usually brief, but the intervalbetween ingesting a poison and the develop-ment of illness may be several hours. It is anexample of learning after a long delay. Evi-dently, when illness is involved, an immedi-ate impact is not needed for learning. This hasbeen confirmed in many experiments on lab-oratory rats.

Acquiring an aversion is the obverse ofthe ability to choose favorable foods (dietaryself-selection), which also has been extensivelyanalyzed in experiments on domestic Nor-ways. The importance of sampling foods is ev-ident when rats are offered a cafeteria-type ar-ray of food with several mixtures of differentnutritional values. (This resembles conditionsin freedom. The chow used in laboratories isa complete diet but is quite unnatural.) Nu-tritionally deficient animals tend to choose un-familiar foods and so make finding the bestone more likely.

SOCIAL INTERACTIONS

All accounts of the social lives of animals useexpressions derived from human social action.Examples are status system (or dominance hier-archy), dominance and subordinacy, courtship,and territory. Such uses are inevitable but in-cur the danger of anthropomorphism, that is,describing the animals as if they were human.The account that follows tries to avoid thishazard.

SOCIAL INFLUENCES ON FEEDING

Nineteenth-century writings contain a per-sistent and popular story of rat warning oth-ers about the dangers of poison bait. Mostsuch tales are preposterous. Nonetheless, it isreasonable to ask whether a rat's feeding be-havior is affected by encounters with otherrats. In this aspect of social behavior, domes-tic rats resemble wild Norways. Seemingly,the domestic varieties retain remarkable abil-ities that are presumably crucial for the sur-vival of wild rats in freedom.

Wild Norways are "central place for-agers." When they return home after feeding,they smell of the food that they ate. Experi-ments have tested the effects of such odors.One question was: Would another rat, givena choice, be influenced against the food if thefirst rat were ill? No such effect has been

Page 36: 0195162854

Chapter 2. Ecology 19

found. Food aversions, it seems, are not ac-quired by detecting illness in others.

Young rats can, however, be drawn awayfrom toxic foods by the behavior of adults.When young rats become independent, theytend to go where other rats are already feed-ing. Such local enhancement may lead wildNorways, in freedom, to eat nourishing ratherthan harmful food.

In contrast to aversions, positive prefer-ences are socially influenced. The choice offood of a wild or a laboratory Norway rat canbe changed by the postprandial odor of an-other. Hence, both wild and laboratory Nor-ways can transmit information about food totheir neighbors.

SOCIAL RELATIONSHIPS

In social interactions other than feeding, wildNorways differ greatly from the domesticvarieties. All domestic Norway rats—white,hooded, even brown or black—are, in ordi-nary laboratory conditions, peaceful. Theyrepresent a typical result of domestication, be-cause they can be caged and moved aroundwith little regard to their companions. (In spe-cial conditions, individuals of some strainscan, however, be induced to interact quiteviolently.)

The absence of structured social interac-tions among laboratory rats once led to astrongly held belief that rats have little sociallife. (Until they saw the contrast on film, someexperimenters refused to accept that the so-cial interactions of domestic Norways differgreatly from those of the wild type.)

Male wild Norways, trapped fromcrowds, are sometimes found to be scarred,evidently due to bites by other rats. They havetherefore been suspected of continuous strife.Yet, if several adult males are put together ina large cage or enclosure, with plenty of food,water, and nest sites, they play, grow, andlook sleek. The postures they adopt includethose described later as accompanying conflictbut are harmless. In such conditions, no indi-

vidual has the opportunity to establish aterritory.

When, however, wild male Norways meetas strangers, one of which is resident in its liv-ing space, the outcome is different. Initially, oneis likely to crawl under the other (Fig. 2-2). Per-haps this distinctive performance deters attack.Another usually peaceful act is grooming(strictly, allogrooming); while huddling, one ratnibbles at the fur of another. These interactionsmay be put under the heading of "social seda-tion," in contrast to "social stress."

An enigmatic act, seen during meetingsbetween males, has been called the threat pos-ture (TP) (Fig. 2-3). "Threat," however, im-plies an intention to punish or hurt, but noconvincing means exists of deciding what a ratintends. In the TP, the back is arched, the legsare fully extended, the hairs are erect, and thehead is usually turned toward the opponent.The posture often precedes an attack andsometimes also occurs after it, or two rats maybe in the TP at the same time.

The most striking interaction is attack(Fig. 2-4), in which one male leaps at anotherwith rapid adductions of the forelimbs (visible

Figure 2-2. During encounters, one rat often crawls underanother. (From Barnett, The rat: A study in behavior.Drawn by Gabriel Donald from a photograph.)

Page 37: 0195162854

20 NATURAL HISTORY

Figure 2-3. The "threat posture," often seen during en-counters by male Norways, is common to all of the speciesof Rattus studied. (From Barnett, The rat: a study in be-havior. Drawn by Gabriel Donald from a photograph.)

only on sped-up cinefilm). Sometimes it is ac-companied by a brief bite. Rarely, the attackerbites and holds on.

Attacks are not relentless but are often in-terrupted: the agonists turn to "boxing" (Fig.2-5) or to other positions, especially the TP.During these intervals, if the agonists haveseparated, an attacked rat may approach theattacker.

Interactions do not depend only on visionand contact. Rats regularly sniffother rats, andobjects are often also marked with urine. Wildrats of all species possess an array of glandsthat secrete pheromones, and an attackingmale urinates and defecates as it approachesan intruder. The odors that are important dur-ing a clash are not established.

Figure 2-4. Attack by a male Norway. (From Barnett, Therat: a study in behavior. Drawn by Gabriel Donald from aphotograph.)

Figure 2-5. This posture, although it occurs during clashesand is called boxing, is nonviolent. (From Barnett, The rat:a study in behavior. Drawn by Gabriel Donald from aphotograph.)

Scent marking by mammals was at firstassumed to be a defense of a territory—an in-stance of a common presumption that socialbehavior is predominantly combative. Later,many scent marks have been shown to be at-tractive. Some observers, however, believethat wild Norways attack other males only ifthey possess a strange odor.

Social interactions are accompanied bysounds. A hostile encounter between Nor-ways may begin with percussion—tooth chat-tering by the attacker. During encounters,while they perform the postures describedearlier, male Norways also utter pure whis-tles, harsh screams, and intermediates be-tween them. During attack and boxing, bothagonists scream and whistle, but when one ap-proaches or "threatens" another, only the an-imal approached sounds off.

Clashes among wild Norways can bequantitatively analyzed by introducing astrange male into the living space of another.For maximum effect, the resident should havefemale companionship. Males might thereforebe supposed to fight for females. However,this is not so: in small colonies, a female Nor-way in estrus is followed by several males,which take turns. Females not in estrus are ig-nored; no pairs are formed.

Page 38: 0195162854

Chapter 2. Ecology 21

In experiments, a female need not bepresent during the actual encounter. Somehave been staged with a single resident male,while the female was shut away. The visitorthen typically makes the first approach, butlively action is nearly always begun by the res-ident: the newcomer fends off the attacker orruns away. The encounters therefore hardlyrate as fights, for they are one-sided.

The encounters fall in the category of ter-ritorial behavior. A territory is a region, occu-pied by an individual, a family, or a largergroup, from which other members of thespecies are excluded. (Some ethologists re-serve the term for a defended region.) It re-quires learning, by each individual, about theenvironment and companions or neighbors.Attacks by a resident on intruders are exam-ples of territorial defense. Another example,already mentioned, is defense of a nest by afemale with a litter.

SOCIAL STATUS AND ENIGMATIC DEATH

When adult males are introduced to groupsof wild Norways of both sexes, no groupaction occurs. Three kinds of adult maleemerge. "Alpha" males are always large andmove about freely and initiate attacks on in-truders; small rats do not overcome muchlarger ones. Others, the "beta" males, adaptthemselves to an inferior role; they keep awayfrom the alphas but feed well and gain weight.They attack intruders only if the alphas are firstremoved; they may then adopt the status of analpha. Both alphas and betas also appear amongAustralian longhaired rats (R. villosissimus).

Last are the "omega" males. Although at-tacks are intermittent and brief, after a day ortwo under attack some rats decline in weight,move slowly, and eat hesitantly with a droop-ing posture and bedraggled appearance; even-tually, they die. Similar debility can resultfrom encounters between black and long-haired rats. In human terms, the omegas seemdejected or seriously depressed. This condi-tion has been fancifully likened to "voodoo

death," in which a person dies after beingcursed by a witch doctor.

Deaths during clashes have, less implau-sibly, been attributed to killing by other rats.However, in several hundred closely observedencounters, many quite vigorous, betweenwild rats of five species, no rat was killed byanother rat. An unwounded omega rat waslikely to collapse and die, sometimes quicklybut more often after hours or days of in-creasing debility. Unexplained deaths havealso been reported in free populations and inconditions in which some rats could reachfood only by running the gauntlet of othersnesting near the food.

This baffling phenomenon has beencalled "death of unknown origin," or DUO,by analogy with the physician's "pyrexia ofunknown origin," or PUO.

"SOCIAL STRESS"

The search for the causes of DUO has in-volved experiments on several species of Rat-tus and, in particular, on tree shrews (Tupaiabelangeri), which also display the three kindsof individuals named here alpha, beta, andomega.

Findings from early studies of DUO sug-gested that both attacked and attacking ratswere adapting to stressful conditions. Duringconflict, the adrenal glands enlarge. The ad-renal response helps to prepare an animal forexertion. Correspondingly, the blood sugarlevel of attacked wild Norways is high. Col-lapse is not due to hypoglycaemia.

Observations on wild rodent populations,however, led to conjectures on the involvementof infection, especially of the kidneys. And, incontrolled conditions, glomerulonephritis wasfound in rats that were socially stressed. If, then,a bacterial infection contributes to DUO, itshould be possible to prevent it; in experiments,the antibiotic neoterramycin prevented deathamong socially stressed rats.

Attention was therefore directed to re-sistance to pathogens. The flaring of infections

Page 39: 0195162854

22 NATURAL HISTORY

during social stress suggests immune sup-pression. During the response to stressors,glucocorticoids secreted by the adrenal cortexinteract with the immune system. In severelyadverse conditions, these hormones can di-minish the immune response.

Hence an immunological account of DUOseems possible. One hypothesis concerns cy-tokines, soluble proteins secreted by lympho-cytes and other cells during infection. Cy-tokines have an anomalous "side effect"; theycan induce loss of appetite and lethargy, and somay contribute to the collapse of omegas.

If so, they provide an example of a re-sponse that is usually adaptive but can be fa-tal. Here, therefore, is an anomaly. In biologyit is usual to ask of a trait: How does it con-tribute to survival? Yet to ask this of DUOsounds ludicrous. Can death have survivalvalue? The question, however, implies an as-sumption that is often made but is incorrect:that every feature of an organism is a directresult of natural selection, and therefore aids(or has aided) survival or breeding. Manytraits are, however, only indirect conse-quences of natural selection: they go inciden-tally with advantageous features and do notthemselves contribute to survival. Darwincalled this phenomenon correlated variation.

DUO therefore has five major peculiari-ties. (1) Death may occur without woundingor other obvious cause. (2) Most of the phys-iological changes found during the collapse ofomegas resist the effects of stressors andshould therefore prevent death. (3) The pos-tures of lethal encounters resemble those seenalso during harmless interactions ("play"). (4)Betas evidently possess special features, notyet identified, that enable them to adapt to at-tack. (J) DUO has no obvious explanation interms of survival value.

ETHOLOGICAL QUESTIONS

The social interactions of wild rats alsoraise questions concerning social signals andaggression.

The "Social Signal"Like genetics and experimental psychology,social ethology began with simple concepts.Standard signals, said to be "innate," wereidentified for each species studied and called"releasers"; each represented a distinct state ofthe signaler and reliably evoked a similarly"fixed action pattern" or standard responsefrom another.

The social signals of rats (like those ofmany other animals) do not correspond to thisdigital concept. They form a fluctuating com-plex that evokes a similar diversity of re-sponses. Some ethologists call such encoun-ters "negotiations"; indeed, like many politicaland financial transactions, they are difficult tointerpret.

"Aggression" and "Drive"Another difficulty arises because the responseof a resident rat to an intruder is often calledaggression.

Calling defense "aggressive" is an in-stance of the widespread custom of puttingdiverse activities, such as hunting and birdsong, under this one heading. In ordinaryspeech, "aggression" signifies ungovernedviolence or unprovoked assault intended tocause injury. However, defense of an oc-cupied area is not unprovoked, nor is itungoverned.

Further, animals are sometimes held tohave an aggressive drive that compels themto attack members of their own species. Thedrive may said to build up internally if it is notexpressed. But territorial dense occurs only inclearly defined circumstances. An adult malewild rat deprived of encounters with strangemales does not turn on members of its owngroup as substitutes.

Each of the many activities called ag-gressive requires separate analysis. In suchanalysis, exemplified by the preceding ac-count, imagined drives do not help; the useof the blanket term "aggression" obscuresthe special features of the behavior understudy.

Page 40: 0195162854

Chapter 2. Ecology

BREEDING AND POPULATIONS

The elaborate social interactions of wild Nor-ways allow them to exist in large and denselycrowded populations. The conditions that en-courage the growth of the populations weredescribed earlier, but they do not tell us whatlimits population growth. Possible checks onnumbers include, shortage of food or water,lack of shelter, predators, pathogens, and neg-ative social interactions.

If a population is isolated and in constantconditions, increase may follow the S-shaped"logistic" curve. Increase is slow at first, be-comes rapid, and then slows and stops. Aftermany rats have been killed by poisoning,some populations of Norways seem to followthis ideal curve. Those reduced by only about50% breed at a high rate and quickly restoretheir numbers; others, brought to about 10%,recover slowly. (These findings have implica-tions for pest management.)

The decline at the top of the curve sug-gests an adverse influence, of which the effectbecomes greater in proportion to populationdensity. If such density-related factors couldbe revealed, we could learn much about howpopulations are regulated. Sometimes, in-deed, altering a single feature, such as foodsupply or shelter, can have a distinct effect onnumbers. Often, however, no single factorcan be identified as crucial. Nor do popula-tions remain constant when they havereached an apparent maximum. The logisticcurve is notoriously a model to which few nat-ural populations conform.

Crowding, measured by numbers to anarea or volume, is influenced by territorialbehavior. When, however, the food supplyis lavish, population densities are reachedmuch above those in other environments.Territorial spacing seems then to be, upto a point, in abeyance. Here, evidently, isan instance of density-related factors inter-acting.

Populations are continually influencedby the conditions we offer them. Reduced

23

cover and shortage of nesting material makeconstructing nests and rearing young moredifficult. They also alter behavior. Prominentfeatures of groups of Norways are contact,crowding, and fecundity; hence, a completepicture should include the balance betweensocial sedation and social stress. However, thefavorable effects of crowding, or at least ofcontact, have been little acknowledged.

To understand what regulates popula-tions, we need to know not only about therats' food, shelter, and diseases but also abouthow these interact and how they influencetheir social lives.

DIVERSITY

The Norway rat, even the domestic Norway,is sometimes called "the rat." In fact, however,about 300 species of the genus Rattus havebeen described. Of these, much is knownabout the polymorphic black rat. This species,which has been very successful as a com-mensal of Homo sapiens, can be tamed, yet ithas never been domesticated. (It even some-times throws up an occasional albino mutant.)

Another notable species is the rice rat(Rattus argentiventer), which has been de-scribed as the most important rodent pest insoutheast Asia. It not only eats large amountsof rice but also cuts down the rice plants atthe base, after the fields have been drained.Yet its biology remains not fully known.

Not all "rats" are members of the genusRattus. Indian mole rats (Bandicota bengaknsis)have a large distribution in southeast Asia.Dense populations live in the godowns of Cal-cutta and other Indian cities where, early inthe twentieth century, they replaced R. rattus;they are also a conspicuous pest in rice fields.

We have no explanation of the geogra-phy of these species, nor any full account oftheir interactions. They do, however, remindus that the Norway, despite its prominence, isonly one species among many, all of whichpresent unsolved problems.

Page 41: 0195162854

24 NATURAL HISTORY

FURTHER READING In: Social learning in animals (CM Heyes and BG Galef,eds.) pp. 49-64. San Diego: Academic Press.

Barnett SA (1975) The rat: a study in behavior, 2nd edi- Hoist D v (1998) The concept of stress and its relevancetion. Chicago: University of Chicago Press. for animal behavior. Advances in the Study of Be-

Barnett SA (2001) The story of rats. Sydney: Allen & havior 27:1-131.Unwin. Singleton GR (ed.) (1999) Ecologically-based manage-

Galef BG (1996) Social enhancement of food preferences. ment of rodent pests. Canberra: ACIAR.

Page 42: 0195162854

Strains

LAUREN GERARD KOCH AND STEVEN L. BRITTON

An understanding of the development of in-bred strains relates directly to the use of ex-perimental design for probing at cause and ef-fect at all levels of biological organization.

THEORETICAL BASIS OFINBRED STRAINS

HARDY-WEINBERG PRINCIPLE

The rudimentary ideas of the Hardy-Weinbergprinciple are a good starting point for under-standing the development and use of inbredanimals. The Hardy-Weinberg principle is ahypothetical representation of genotypic fre-quencies that will occur in randomly matingpopulations. This principle has use because itsimplifies to a multinomial expression that de-scribes the genetic content of a population(Falconer and Mackay, 1996a).

For an entire population, there are nu-merous alleles (alternate forms of the samegene) for most genetic loci. For example, theA locus could have five variants in a popula-tion designated as AI, A2, A3, A4, and A5. Eachindividual within a population, however, hasonly two alleles at each locus and these canbe either identical (homo2ygotic, e.g., AiAi orA2A2, etc.) or different (heterozygotic, e.g.,AiA2 or A^A4}.

To understand the Hardy-Weinbergprinciple, consider a large population thatcontains only two alleles (Ar and A2) for a spe-cific gene. The frequency of allele Al is theproportion of all A alleles in the population

that are of the Al type. This includes individ-uals who are homozygous (A^) and half ofthe alleles in individuals who are heterozy-gous (AiA2). The frequency of allele A2 is theproportion of all A alleles in the populationthat are of the A2 type. Let p equal the fre-quency of all A! alleles and q equal the fre-quency of all A2 alleles such that in fractionalrepresentation, p + q = 1.

Because each gamete contains only oneof the A variants, each A locus has the proba-bility of being represented in progeny pro-portional to its allelic frequency representedin the parental population if mating is ran-dom. Thus, as alleles recombine from theparental gene pool via fertilization, there is pprobability of an individual acquiring an Aj al-lele and p X p = p2 probability of acquiringan A±AI combination. Likewise, there isq = 1 — p chance of acquiring an A2 allele andq2 probability of both variants being of the A2

type (A2A2 genotype). Extending this logic,there is pq chance that the first allele is A! andthe second allele is A2. The chance of the firstallele being A2 and the second allele being Al

is qp. Therefore, the combined chance of ob-taining a heterozygote (AjA2 or A2A:) for theA locus is (p X q) + (q X p) = 2pq.

These ideas form the basis of the Hardy-Weinberg principle that describes a theoreti-cal relationship for gene frequencies and geno-type frequencies in a population. By thisprinciple, picking an individual randomlyfrom a population is tantamount to pickingtwo genes at random from the entire genepool of a population. As from above, the prob-

25

3

Page 43: 0195162854

26 NATURAL HISTORY

ability of obtaining A-^Ai is pz, the probabilityof AiA2 is 2pq, and the probability of A2AZ isq2. The sum of all the fractional probabilitiesaccounts for all of the A: and A2 variants asgiven by the Hardy-Weinberg principle in itsusual form:

p2 + 2pq + q2 = 1 (Eq. 1)

Not obvious from first inspection, theHardy-Weinberg equation models only forpopulations in which gene pool frequenciesdo not change. The Hardy-Weinberg princi-ple operates with the assumption that the fol-lowing five conditions are met:

1. The population contains an infinite number ofindividuals.

2. Genotypes do not influence mate choice.3. No mutation or natural selection occurs.4. No migration in or out of the population oc-

curs (closed population).5. An equal evolutionary fitness exists among the

individuals.

A population that conforms to these con-ditions is declared to be in Hardy-Weinbergequilibrium. This principle thus provides a the-oretical standard for analysis of conditions inwhich changes do occur, such as with in-breeding. Inbreeding is the mating between in-dividuals related closely enough to produce anonrandom distribution of genotypes. Theconsequence of inbreeding is to increase thefrequency of homozygous genotypes (A^and A2A2) and thus decrease the frequency ofheterozygous genotypes (A iA2). Stated an-other way, inbreeding represents deviationfrom Hardy-Weinberg proportions (p2 +2pq + q2) by reduction of heterozygous geno-types, with no attendant loss of individual al-lelic frequencies.

DEFINING AN INBRED STRAIN

In diploid organisms, sister-brother (full-sib)mating represents the closest form of in-breeding. At the most, two parents can have

four different allelic variants and thus producefour different genotypes among the progeny.The probability that the sibs have the samegenotype is 1:4. The magnitude of inbreedingcan be estimated from the relative loss of het-erozygosity across generations and is termedthe coefficient of inbreeding (F). The measure ofthe coefficient of inbreeding ranges from 0 to1 and reflects deviation from Hardy-Weinbergproportions (Fig. 3-1). Consider a baselinepopulation that was initially in Hardy-Wein-berg equilibrium such that the frequency ofheterozygous genotypes was 2pq. The coeffi-cient of inbreeding (F) after t generations ofinbreeding is estimated as

F = (2pq)i - (2pq)c/(2pq)i (Eq. 2)

where (2pq)i is the initial frequency of het-erozygotes at baseline, and (2pq)c is the cur-rent frequency of heterozygotes after t gener-ations of inbreeding.

When F = 0, the genotype frequenciesare not different from Hardy-Weinberg pro-portions. When F = 1, inbreeding is com-

Figure 3-1. Idealized graphic depicting loss of heterozy-gosity with 20 generations of sister-brother inbreeding. Ini-tial conditions presume two kinds of alleles (Aj and A2)available at locus A. Twenty-five percent are present in theAjAj homozygotic form; 50%, as heterozygotes (A^); and25%, in the A2A2 homozygotic form. The coefficient of in-breeding (F) increases and the remaining heterozygosity(H) decreases with each generation.

Page 44: 0195162854

Chapter 3. Strains

plete, no heterozygous genotypes are present,and the population contains only homozy-gous genotypes in p (A^) and q (A2A2)frequencies.

F can also be expressed as 1 — H, whereH is the frequency of heterozygotes. There-fore, H = 1 in a population that is in Hardy-Weinberg equilibrium. The reduction inheterozygosity with inbreeding via full-sibmating can be calculated from this recursionrelation (Hedrick, 2000):

27

Exponential Decrease In Heterozygosity with Inbreedin

where Ht = the heterozygosity at the tthgeneration.

So, consider the start of inbreeding witha closed base population that is in equilibrium(H0 =1). From the above recurrence relation,the heterozygosity would decrease with eachsubsequent generation (Ht) in a sequenceH0 =1, Hi = 0.75, H2 = 0.625, H3 = 0.5,H4 = 0.406, H5 = 0.328, or given by the ratios2:2, 3:4, 5:8, 8:16, 13:32, 21:64, etc, approach-ing zero as a limit. Conveniently, but perhapsnot apparent from initial inspection, the de-nominators double and the numerators forthe declines in Ht follow a Fibonacci sequence(Atela et al., 2002) where each subsequentnumber is the sum of the previous two num-bers (Crow, 1986). From this sequence, the re-maining Ht at any generation can be calcu-lated from solution of the above sequencethat, as shown in Figure 3-2, follows an ex-ponential relationship: Ht = 0.944e~°-2117*.

Likewise, the coefficient of inbreeding inthe tth generation of full-sib mating can alsobe expressed as a recursion relationship(Hedrick, 2000):

An inbred strain is defined as one that hasbeen sister-brother mated for at least 20 gen-erations as agreed on by The International

Figure 3-2. On average, decreases in heterozygosity witheach generation of sister-brother mating follow an expo-nential decay. The numerator of this decay is representedby a Fibonacci sequence. In addition to breeding schemes,many natural phenomena, such at spiral growth patterns inplants and seashells, follow this pattern (Atela et al., 2002).

Committee on Standardization Nomencla-ture for Mice (Silver, 1995). As shown (seeFigs. 3-1 and 3-2), 20 generations of inbreed-ing produce a population of animals that haveon average 1.4% remaining heterozygosityand are thus about 98.6% homozygotic (Haitiand Clark, 1988).

VARIATION

Traits can be divided into either mendelian orquantitative. Mendelian traits are those forwhich a genetic difference at a single locus issufficient to cause a difference in phenotypicexpression of a given character. Quantitativetraits do not manifest as discrete phenotypesin populations but distribute with continuousvariation. Continuous variation is the result ofthe variable presence and expression of manygenes (i.e., polygenic) within a population asthey interact with the environment. Mosttraits of physiological, morphological, clinical,or behavioral interest are of the quantitativetype, and their inherent complexity enhancesthe importance of well-defined inbred strains.

Two statements about the componentsof variation summarize the usefulness of in-bred strains: (1) environmental factors ac-count for within-strain variation, and (2)genetic factors account for between-strainvariation. Thus, a common starting point is to

Page 45: 0195162854

28 NATURAL HISTORY

assess both the within- and between-strainvariations in a panel of inbred strains for a traitof interest.

HERITABILITY

The major assumption of quantitative genet-ics is that variation from the mean value of atrait is caused by the additive influence ofgenetics plus environment (Falconer andMackay, 1996b) and can be expressed as

VP = VG + VE (Eq. 5)

This equation implies that variance inphenotype (VP) can be partitioned into geno-typic variance (VG) and environmental vari-ance (VE). The measure most widely used asa descriptor of the contribution of genetic fac-tors is heritability, and two types can be con-sidered. Broad sense heritability (H2) is thesum of all genetic factors that influence thephenotype at the population level and is givenby the expression H2 = VG/VP. Broad senseheritability is also called the degree of genetic de-termination. In theory, the phenotypic variancecan be estimated if one of the two components(G or E) is eliminated. Experimentally, geneticvariance can be more closely controlled froma highly inbred line with identical genotypes.The major assumption with this approach isthat the environmental variance is similar be-tween inbred strains and that the genotypesrespond similarly to the environment. This isproblematic because of gene-environment in-teractions. Indeed, some strains have beenfound to be more variable and sensitive toenvironmental differences for a given trait(Crabbe et al, 1999).

The more useful and commonly referredto narrow sense heritability (h2) or simply her-itability is expressed as h2 = VA/VP, where VA

is defined as the additive genetic variance. Ad-ditive genetic variance is that portion of geneticvariance that causes offspring to resemble theirparents and is considered the variance associ-

ated with the average effects of substituting oneallele for another. Estimates of heritabilityrange from 0.0 to 1.0. If h2 = 0, there are nogenetic contributions to phenotypic variance.If h2 = 1, all phenotypic variation can be ac-counted for by genetic factors. (Note that thesymbol h2 is itself heritability, and not thesquare of the term in the arithmetic sense, ause chosen by Sewall Wright [1921].)

Two approaches can be used to estimateh2 for a given trait. The first uses informationon the resemblance between relatives and iseasier conceptually but more difficult experi-mentally. In a widely heterogenic population(outbred), it is more likely for relatives to pos-sess the same allelic variants for genes. That is,the offspring from parents high for a trait wouldalso be expected to be high for a trait. In con-trast, offspring from parents that demonstratelow for a trait would more likely be low for atrait. These ideas form the basis for using theregression of mean offspring values on themean value of the parents (mid-parent value)as an estimate of h2. If the trait is inherited ad-ditively with complete fidelity, such that thevalues of offspring are highly similar to the par-ents, then the slope of the regression line (h2)equals 1. In contrast, if no additive similarityexists between parent and offspring, h2 = 0.

As a second approach, h2 can also be es-timated for measures of a trait from a panelof inbred strains and is based on two as-sumptions related to the properties of inbreds.First, individuals within each inbred strain aresimilar genetically. Trait variation within astrain is thus attributed to the environmentalvariance such that an estimate of VE can beobtained. Second, variation between strainsderives from genetic differences and can beused to estimate VA. The critical factor is thateach inbred strain represents almost exclu-sively homozygous genotypes.

As indicated earlier, narrow sense heri-tability (h2) is estimated from the ratio of ad-ditive genetic variance (V^) to phenotypicvariance (VP) (Falconer and Mackay, 1996b),

Page 46: 0195162854

Chapter 3. Strains 29

with the phenotypic variance being the sumof the VA and environmental variance (V£):

VA and VE can be estimated by partition-ing Equation 6 into the between-group vari-ance component (ZB2) and the within-groupvariance component (ZW2) from analysis ofvariance using these expressions:

where MSB is the mean square betweenstrains, MSW is the mean square withinstrains, and n is the number of animals in eachstrain. ZB2 equals 2VA, and VE is approximatedby ZW2. Combining Equations 6, 7, and 8yields an estimate of h2 (Hegmann and Possi-dente, 1981):

The relationship between strain differ-ences and VA as they relate to the effect of in-breeding on character variance are presentedby Crow and Kimura (1970) and summarizedby Hegmann and Possidente (1981). In brief,VA within an inbred line [VA(i)J increases rel-ative to the VA in a randomly mating popula-tion [VA(r)] as the coefficient of inbreeding (F)increases, as given by

VA(i) = VA(r) (1 + F) (Eq. 10)

As a result, if F approaches 1, the estimate ofvariance among inbred strains should be twicethe VA for a trait in a randomly bred popula-tion from which the inbred strains were de-

rived. This explains why the estimate of VA inEquation 9 is divided by 2.

If the number of rats is not equal for allstrains, a weighted average value for n can becalculated as suggested by Sokal and Rohlf(1981):

(Eq. 11)

where a is the number of strains, and n^ is thenumber of rats in each strain.

DEVELOPMENT OF MODELS

INBRED STRAINS DERIVEDFROM SELECTED LINES

A very broad idea emerges from considerationof the above fundamentals on genetic vari-ance and heritability. Based on R. A. Fisher's1930 theorem of natural selection, traits pe-ripherally associated with evolutionary fit-ness, such as morphology, behavior, and com-plex physiology, demonstrate more additivegenetic variance (i.e., h2} than do traits essen-tial to fitness because of less pressure from nat-ural selection. This generalization is consis-tent with the demonstration of success inartificial selection for traits such as blood pres-sure (Knudsen et al., 1970; Yamori et al., 1972)and aerobic capacity (Koch and Britton, 2001)in rats.

Many of the currently available inbred ratstrains were not developed from lines first se-lectively breeding for a trait but were simplyinbred to increase genetic uniformity. As aconsequence, evaluation of a panel of inbredstrains may yield only minimal between-strainvariance for a desired trait. In this case, two-way artificial selection can be used to createlow and high lines widely different for a trait

Page 47: 0195162854

30 NATURAL HISTORY

from which inbred strains can be subsequentlyproduced.

AVAILABILITY OF GENETICALLYHETEROGENEOUS RATS

The development of the widely heteroge-neous outbred stock of N:NIH rats from theAnimal Resource Center of the National In-stitutes of Health (NIH) is a major resourcefor the creation of artificially selected lines.These rats are available from the NIH and aresomewhat ideal as a founder population. Thisgenetically segregating stock of rats originatedfrom the intentional crossbreeding of eightdisparate inbred strains of independent origin(AxC 9935 Irish [ACI], Brown Norway [BN],Buffalo [BUF], Fischer 344 [F344], Marshall520 [M520], Maudsley reactive [MR], Wistar-Kyoto [WKY], and Inbred Wistar [WN]) byHansen and Spuhler in 1979 [1984]. (See alsoRat Genome Database, http://www.rgd.mcw.edu/strains) The outbred stock is managedsuch that the gene frequency remains stabi-lized, the inbreeding is minimized within theclosed colony, and the genetic variability re-mains substantial. Each rat from the N:NIHstock is a genetic admixture of the eightfounder inbred strains and in theory can be auseful source for genetic analysis of behavioraltraits (Mott et al., 2000).

GENERAL APPROACH TO SELECTION

Selective breeding begins by measuring thetrait of interest in a large founder populationthat has wide genetic heterogeneity. The con-trasting lines for the trait (low and high) arestarted by breeding rats that demonstrate theextreme values of the founder population.Then, at each subsequent generation, prog-eny are phenotyped and selected as the "best"for the trait and bred to create the next gen-eration. This process is repeated until thechange in the population mean produced byselection (selection response) plateaus, whichtypically indicates exhaustion of additive ge-

netic variance for the trait (Fig. 3-6). The de-gree of heterozygosity in a selected popula-tion can be increased above that random bredby making the contributions from each fam-ily more equal; this can be accomplished bytaking the "best" female and male from eachmating and using them as parents in the nextgeneration (within-family selection). Within-family selection coupled with a systematic ro-tational breeding design keeps mating be-tween relatives at the minimum to maintainthe rate of inbreeding (AF) per generation justless than 1% if at least 13 families are used inboth the low and high selected lines. AF =1 /(4N), where N = number of individual par-ents in each line at each generation (Falconerand Mackay, 1996a).

SELECTION ON THEGENETIC COMPONENT

In general, the more information that is avail-able per each individual about the trait, themore accurate is the selection. For example, iffive repeat measures of running capacity wereavailable, it would seem logical to select on theaverage of these five estimates of running ca-pacity (Nicholas, 1987). Despite this, we took adifferent approach when we started large-scaleselection for aerobic treadmill running capacityin 1998 (Koch and Britton, 2001). For each rat,the single best value of five trials was deemedas the measure most closely associated with thegenetic component of running capacity. Thisidea of estimating the genetic component fromthe best trial rather than the average of all tri-als has two origins. (1) The environment canhave an infinite negative influence on capacity(i.e., a detrimental environment can take the ca-pacity to zero). Factors such as subtle differ-ences in daily housing or handling conditionscould cause a genetically superior rat to per-form below its maximal ability for a given trial.(2) However, the environment can have only afinite positive influence on a trait. That is, a fa-vorable environment cannot cause a rat to ex-ceed values above its genetically determined

Page 48: 0195162854

Chapter 3. Strains 31

upper limit of capacity. Thus, we reasoned therat's best trial of five comes closest to the ge-netically determined upper limit of its capacity.Although we used only the best day for selec-tion, it should be noted that testing across 5 daysmatches the average estrus cycle for female ratsand thus eliminates this as a variable.

EXPERIMENTAL USE OF INBREDS

Inbred strains of animals have been a core sub-strate in the analysis of complex phenotypesand remain as one of the basic tools forprogress in functional genomics. The majorvalue of inbred strains emanates from theclose genetic uniformity that facilitates phe-notyping, genotyping, and the opportunity formultiple investigators to repeatedly evaluatethe same genotype.

Our laboratory group has scanned panelsof inbreds for measures of physical capacitythat include aerobic endurance treadmill run-ning capacity, sensorimotor capacity, andstrength (Barbato et al., 1998; Biesiadecki etal, 1998; Koch and Britton, 2003).

Figure 3-3 A shows the degree of variationin sensorimotor ability among 11 inbredstrains of rats (Biesiadecki et al., 1998). Senso-rimotor capacity was estimated on the basis ofthree separate tests. Rats from each strain weretested to determine how long each could re-main on (1) a rotating cylinder as the velocityof rotation increased every 5 seconds (one-direction rotation test), (2) a rotating cylinderthat reversed direction every 5 seconds andincreased velocity every 10 seconds (two-direction rotation test), and (3) a platform thatwas tilted 2° every 5 seconds from 22° to 47°(tilt test). The distribution among the strainsfor each test was continuous and normally dis-tributed. On all three tests, the Black hoodedPVG strain was consistently the highest rank-ing strain (for both males and females),whereas the Copenhagen (COP) and MilanNormotensive Strain (MNS) strains were con-sistently the lowest ranking strains. The large

Figure 3-3. (A) Variation in sensorimotor capacity basedon three tests of motor ability (one-way rotation, tilt test,and two-way rotation) measured as timed performance(seconds) in a panel of 11 inbred strains. The PVG straindemonstrated the greatest ability, whereas rats of the MNSand COP strains represented strains with the lowest per-formance on all three tests. (B) For aerobic running capac-ity (distance run to exhaustion in meters), the DA and COPdisplayed the greatest difference. (Adapted from Biesiadeckiet al. (1998) and Barbato et al. (1998.)

differential in sensorimotor performance be-tween PVG and MNS or COP strains suggeststhat these strains can be used as contrastingmodels of sensorimotor capacity.

The same panel of 11 inbred strains wastested for aerobic treadmill endurance runningcapacity to exhaustion as estimated from dura-tion of the run, distance run, and vertical workperformed to take into account variation inbody weight (Barbato et al., 1998). The COPrats were the lowest performers and the DArats were the highest performers on all esti-mates of running. The wide divide in per-formance between COP and DA strains of ratsrepresents identification of genetic substrate forexploration of aerobic endurance capacity andrelated traits such as economy of running, oxy-gen transport pathways, and heart function.

Page 49: 0195162854

32 NATURAL HISTORY

Figure 3-4. Regression of mean off-spring on mid-parent value for one-way rotation test for sensorimotorcapacity. Each point represents themean value of one set of parents(along x-axis) and mean value oftheir offspring (along y-axis). Axesintersect at the mean value for allparents and all offspring and aremarked at 10 second intervals. Theslope of the line estimates the heri-tability (h2) and has a value of 0.44.(Data from Koch and Britton [2003].)

Figure 3-4 displays an estimate of heri-tability for generalized sensorimotor capacity(in N:NIH stock) as estimated from time on arotorod. The mean of the offspring from eachof 19 families was regressed on the mid-par-ent value of each family and yielded a slopeof 0.44. In the panel of 11 inbred rats that weretested, h2 for sensorimotor capacity as esti-mated by comparing the between- and within-strain variances averaged 0.39 in females and0.48 in males (Koch and Britton, 2003). Theseresults demonstrate a wide phenotypic varia-tion and a heritable component to sensori-motor capacity sufficient for success in the de-velopment of contrasting rat genetic modelsby divergent artificial selection.

COSEGREGATION ANALYSIS

Contrasting inbred strains can be characterizedat organ, tissue, cell, and molecular levels oforganization to identify physiological, behav-ioral and morphological differences that maybe responsible for variations in the trait; theseare termed "likely determinant phenotypes"(Jacob and Kwitek, 2002). The identification ofthe genes responsible for a given phenotypeusing divergent inbred strains is based on twowidely held principles of biology: (1) genescause traits, and not vice versa, and (2) genesthat cause a given trait will remain associated

with that trait and other genes will segregaterandomly relative to the trait.

Understanding the use of inbreds incosegregation studies is one path of geneticanalysis that has theoretical and heuristicvalue. Like individuals, two inbred strains candemonstrate wide variation for a given trait.The central idea is to follow the association ofgenes or downstream gene products (such asmessenger RNA, protein, or subordinatephysiological and biochemical traits) with val-ues of the phenotype in a segregating popu-lation. A segregating population is one inwhich alleles recombine randomly to yieldnew genotypes in mating crosses. This ap-proach works because loci and pathwayscausative of a given trait will remain associ-ated with that trait and other genes and nonas-sociated traits will segregate randomly. Themost informative segregating population iscreated from two sequential crosses of inbredstrains. The first cross is between two differ-ent inbred parental strains (Pi and P2) that dif-fer significantly for a trait. The assumption isthat the two strains have contrasting allelicvariants that dictate the phenotypic differ-ence. These original contrasting strains are of-ten referred to as the "low strain" and "highstrain" to indicate the directional difference intrait measure. The PI X P2 cross yields an Fx

(first filial) population composed of close to

Page 50: 0195162854

Chapter 3. Strains 33

identical heterozygotes. A subsequent F: XFI cross yields the desired segregating F2 (sec-ond filial) population in which allelic variantsrecombine randomly, producing a variation ofgenotypes and thus a variation in the distri-bution of phenotypic trait values (Fig. 3-5).

Following a single locus that distributesvia discrete mendelian methods allows one tounderstand the usefulness of cosegregationwhen extended to analysis of a polygenic trait.Consider two allelic variants (Ai and A2) thatproduce sufficient variation that the homozy-gotic (A^! and A2A2) and the heterozygotic(A:A2) genotypes can be distinguished by phe-notype. Assign the low strain (P:) as the ho-mozygotic genotype A^ and the high strain(P2) as the homozygotic genotype A2A2.Crossing Pl X P2 produces only AjA2 het-erozygotes in the F! population. These crossescan be followed using Punnett squares:

Gametes from Low Strain (Pj)

A, A,

Gametes

from High

Strain (P2)

(F! Population)

An intercross between F! heterozygotesproduces a segregating population that yieldsall possible genotypes in the ratio of 1:2:1 (1AjA1( 2 AiA2, 1 A2A2):

Gametes from F! Male

A! A2

Gametes

from F!Female

(F2 Population)

Thus, for a single locus with a large effect,segregation of the phenotype can be followedin an F2 population. That is, phenotypic ex-pression of the homozygotes (A^ and A2A2)separates to express the low and high valueswhile the heterozygotes demonstrate interme-diate values. Note the simple genotype associ-ated with creation of an F2 population origi-nated from two inbred strains. For each locus,only three allelic combinations are possible andcan be followed for association between geno-type and phenotype as shown in Figure 3-5.Genetic markers are identifiable physical loca-tions on a chromosome (loci) whose inheri-tance can be followed similarly as described for

Figure 3-5. The biological determi-nants of the differences in trait be-tween two inbred strains (Pt and P2)can be evaluated via cosegregationanalysis. Two sequential crosses pro-duce an F2 population in which alle-les recombine randomly and can beused to determine which genes anddownstream products cosegregatewith the distribution of the pheno-type. (Adapted from Britton andKoch [2001].)

A!

A2

A,A2

A,A2 A2A2

A,A2

2

A2 AjA2 AiA2

A,A2 A,A2

Page 51: 0195162854

34 NATURAL HISTORY

Mendelian traits and form the basis for ge-nomic scans (see Jacob and Kwitek, 2002).

THE BENEFITS OF INBREDS DERIVEDFROM SELECTED LINES

Four properties make inbred lines originallycreated from two-way artificial selection use-ful (Fig. 3-6). First, as indicated earlier, mod-els can be created for traits for which there ex-ists only minimal variance between thealready available inbred strains. Second, be-cause the selection process often carries thephenotypic means of the low and high linesbeyond the range of the extremes of thefounder population (Falconer and Mackay,1996a), the signal measurements of the traitcan be made to differ substantially betweenthe lines. Third, by maintaining a high levelof heterozygosity at each generation during

Figure 3-6. An idealized selective breeding paradigm. Se-lective breeding begins by measuring a trait in a large groupof widely heterogenous animals from which selections ofchoice breeders are made to concentrate alleles expressingextreme values of the phenotype. At each generation, prog-eny are phenotyped, selected as the "best," and bred to cre-ate the next generation. This process is repeated until thechange in population mean (selection response) hasplateaued or a desired difference is attained between the se-lected lines.

selection, the main complement of contrast-ing alleles causative of trait difference will beconcentrated in both the divergent lines andthe subsequently developed inbred strains.Fourth, selection across many generations in-terprets into inadvertent selection for insensi-tivity to subtle changes in environment. Thislack of influence by environmental effects inselected strains can be of great benefit to thereproducibility of the phenotype. Inbredstrains that differ markedly for a trait but didnot originate from selected lines often demon-strate wide variation in phenotype in responseto similar controlled experimental conditions(Crabbe et al., 1999).

THE FUNDAMENTAL CHALLENGEFOR ALL MODELS

Although we focused here on genetic models,the development and use of any kind of ani-mal model are associated with subtle prob-lems, especially for models of complex dis-eases. First, physical and chemical maneuversapplied to mimic disease conditions such as lig-ation of cerebral arteries (stroke), intracerebralinjection of 6-hydroxydopamine (Parkinson'sdisease), or administration of streptozocin (di-abetes mellitus) more accurately reflect re-sponse to injury rather than emulating the pro-gression of disease. Second, although it seemsthat disease models derived from selectionwould be a highly informative alternative, di-rect selection for a disease is also problematic.Selection would be based on currently knownmeasurable traits of the disease and not the fullarray of underlying mechanisms. The issue isthat traits are not mechanisms, such that se-lection based on measures thought to charac-terize a disease will leave components eitherunrepresented or weighted inappropriately.Third, the problem in selecting on a diseasetrait is amplified because it appears that dis-eases emerge not as discrete events but as bi-ological complexes with coordinately regu-lated gene clusters, such as the pathogeniccascades represented by Parkinson's disease

Page 52: 0195162854

Chapter 3. Strains 35

(Tieu et al., 2003) and metabolic syndrome X(Lopez-Candales, 2001).

To address this challenge, we currentlyhypothesize that clinically relevant modelscan emerge from two-way artificial selectionfor high-order, complex, physiological traitssuch as sensorimotor function. We predictthat selection for the extremes of fundamen-tal trait variation will create two populations:(1) a low line that displays low physiologicalfunction and disease-related traits and (2) ahigh line that displays high physiological func-tion, no disease-like traits, and resistance tothe development of disease.

This hypothesis derives from our highlyspeculative ideas about the association be-tween our evolutionary history and oxygen(Britton, 2003). The underlying proposal is that2 billion years of evolution in an oxygen envi-ronment has determined that oxygen metabo-lism occupies a central feature of our biology(DesMarias, 2000). It appears that evolution fol-lowed the increased free energy transfer af-forded by the widened redox potential whenoxygen is the final electron acceptor in oxida-tion reactions (Baldwin and Krebs, 1981).Obligatory for the use of oxygen in energytransfer pathways was the simultaneous co-evolution of enzymes that detoxify the reactiveoxygen species that are byproducts of oxidationreactions. Thus, the pathways that mediateboth oxidation reactions and oxygen detoxifi-cation reactions constitute a large part of ourbiology (Young and Woodside, 2001; Myers etal., 2002). Our extension is that essentially alldisease will resolve at the molecular level intoproblems associated with oxygen utilization.For example, it has been shown that an in-creased expression of genes involved in oxida-tive phosphorylation contributes to human di-abetes (Mootha et al., 2003) and that increasedproduction of reactive oxygen species is relatedto the cellular dysfunction and biochemical al-terations present in Parkinson's disease (Tieuet al., 2003). Development and study of well-defined animal models will be useful in resolv-ing mechanisms of cellular injury and death.

ACKNOWLEDGMENTS

This work was supported by grants from the U.S. Public HealthService, National Institutes of Health (Heart, Lung and BloodInstitute grant HL-64270 and the National Center for ResearchResources grant RR-17718) to S.L.B. and L.G.K.

REFERENCES

Atela P, Gole C, Hotton S (2002) A dynamical systemfor plant pattern formation: a rigorous analysis.Journal of Nonlinear Science 12:641-676.

Baldwin JE and Krebs H (1981) The evolution of meta-bolic cycles. Nature 291:381-382.

Barbato JC, Koch LG, Darvish A, Cicila GT, Metting PJ,Britton SL (1998) Spectrum of aerobic endurancerunning performance in eleven inbred strains ofrats. Journal of Applied Physiology 85:530-536.

Biesiadecki BJ, Brand PH, Koch LG, Metting PJ, BrittonSL (1998) Phenotypic variation in sensorimotor per-formance among eleven inbred rat strains. Ameri-can Journal of Physiology 276:R1383-R1389.

Britton SL (2003) Is there an answer? IUBMB Life55:429^130.

Britton SL and Koch LG (2001) Animal genetic modelsfor complex traits of physical capacity. Exercise andSport Sciences Reviews 29:7-14.

Crabbe JC, Wahlsten D, Dudek BD (1999) Genetics ofmouse behavior: interactions with laboratory envi-ronment. Science 284:1670-1672.

Crow JF (1986) Basic concepts in population, quantita-tive and evolutionary genetics. New York: W.H.Freeman & Company.

Crow JF and Kimura M (1970) An introduction to pop-ulation genetics theory. New York: Harper andRow.

DesMarias DJ (2000) When did photosynthesis emergeon earth? Science 289:1703-1705.

Falconer DS and Mackay TFC (1996a) Introduction toquantitative genetics, 4th edition. Essex, England:Addison Wesley Longman, Ltd.

Falconer DS and Mackay TFC (1996b) Heritability. In:Introduction to quantitative genetics, pp. 160-183.Essex, England: Addison Wesley Longman Limited.

Fisher RA (1930) The genetical theory of natural selec-tion. Oxford, England: Clarendon Press.

Hansen C and Spuhler K (1984) Development of the Na-tional Institutes of Health genetically heteroge-neous rat stock. Alcoholism, Clinical and Experi-mental Research 8:477-479.

Hard DL and Clark AG (1988) Principles of populationgenetics, 2nd edition. Sunderland, MA: Sinauer As-sociates, Inc.

Hedrick PW (2000) Quantitative traits and evolution. In:

Page 53: 0195162854

36 NATURAL HISTORY

Genetics of populations, pp. 445-500. Sudbury, MA:Jones and Bartlett Publishers.

HegmannJP and Possidente B (1981) Estimating geneticcorrelations from inbred strains. Behavior Genetics11:103-114.

Jacob HJ and Kwitek AE (2002) Rat genetics: attachingphysiology and pharmacology to the genome. Re-view. Nature Review Genetics 3:33-42.

Knudsen K, Dahl LK, Thompson K, Iwai J, Leith G(1970) Effects of chronic salt ingestion: inheritanceof hypertension in the rat. Journal of ExperimentalMedicine 132:976-1000.

Koch LG and Britton SL (2001) Artificial selection for in-trinsic aerobic endurance running capacity in rats.Physiological Genomics 5:45-52.

Koch LG and Britton SL (2003) Genetic component ofsensorimotor capacity in rats. Physiological Ge-nomics 13:241-247.

Lopez-Candales A (2001) Metabolic syndrome X: a com-prehensive review of the pathophysiology and rec-ommended therapy (review). Journal of Medicine32:283-300.

Mootha VK, Lindgren CM, Eriksson KF, SubramanianA, Sihag S, Lehar J, Puigserver P, Carlsson E, Rid-derstrale M, Laurila E, Houstis N, Daly MJ, Patter-son N, MesirovJP, Golub TR, Tamayo P, Spiegel-man B, Lander ES, Hirschhorn JN, Altshuler D,Groop LC (2003) PGC-1 alpha-responsive genes in-volved in oxidative phosphorylation are coordi-

nately downregulated in human diabetes. NatureGenetics 34:267-273.

Mott R, Talbot CJ, Turn MG, Collins AC, Flint J (2000)A method for fine mapping quantitative trait loci inoutbred animal stocks. Proceedings of the NationalAcademy of Sciences of the United States of Amer-ica 97:12649-12654.

Myers J, Prakash M, Froelicher V, Do D, Partington S,Atwood JE (2002) Exercise capacity and mortalityamong men referred for exercise testing. New En-gland Journal of Medicine 346:793-801.

Nicholas FW (1987) Veterinary genetics. New York: Ox-ford University Press.

Silver LM (1995) Mouse genetics: concepts and applica-tions. New York: Oxford University.

Sokal RR and Rohlf FJ (1981) Biometry: the principles andpractice of statistics in biological research, 2nd edi-tion. San Francisco: W.H. Freeman and Company.

Tieu K, Ischiropoulos H, Przedborski S (2003) Nitric ox-ide and reactive oxygen species in Parkinson's dis-ease. IUBMB Life 55:329-335.

Wright S (1921) Systems of mating. Genetics 6:111-178.Yamori Y, Ooshima A, Okamoto K (1972) Genetic fac-

tors involved in spontaneous hypertension in rats:an analysis of F2 segregation generation. JapaneseCirculation Journal 36:561-568.

Young IS and Woodside JV (2001) Antioxidants in healthand disease. Journal of Clinical Pathology 54:176-186.

Page 54: 0195162854

Individual Differences

GUY MITTLEMAN4

Intrinsic to any discussion of human behavioris the notion that people display vast individ-ual differences in behavior in response to thesame situation. Knowledge of individual dif-ferences has fueled a variety of disciplines, in-cluding personality and social psychology, andindividual differences in the behavior of hu-man beings have long been recognized as atool in the investigation of the neuropsy-chophysiological bases of a variety of diseases.The purpose of this chapter is to suggest thatrats, like humans, despite going through a ge-netic bottleneck when domesticated, displayprofound individual differences that have rel-evance for understanding a variety of normaland pathological conditions.

Vast individual differences in the behav-ioral response to drugs of abuse as well as sub-sequent patterns of self-administration havebeen observed in humans. As stated byO'Brien et al. (1986), "Some addicts go formonths or years using heroin or cocaine onlyon weekends before becoming a daily (ad-dicted) user. Others report that they had suchan intense positive response that they becameaddicted with the first dose. Similar variationin initial response may also be observed in an-imals." These individual differences in the re-inforcing effects of addictive drugs (de Wit etal., 1986) are considered by many clinicians tobe a primary factor in the vulnerability to ad-diction shown by some individuals (O'Breinet al., 1986). Although much research has beendevoted to determining the substrates of ad-diction (for review, see Koob and Bloom,1988), there have been relatively few investi-

gations into the mechanisms that underlie in-dividual differences.

GENESIS OF THE CONCEPT OFINDIVIDUAL DIFFERENCES IN RATS

Elliot Valenstein was the first to suggest thatrats displayed consistent individual differ-ences. He observed that in response to elec-trical stimulation of the lateral hypothalamus(ESLH), rats displayed a variety of differentbehaviors, including eating, drinking, gnaw-ing, hoarding, grooming, aggression, retrievalof young, and male copulatory behavior (fora review, see Valenstein, 1975). Some animalsconsistently responded to the electrical stim-ulation by eating or drinking, whereas othersdisplayed only an increase in locomotor ac-tivity. The explanation that these differencesin the behavioral response to ESLH could beattributed to corresponding differences inelectrode placement within the lateral hypo-thalamus was eliminated. Even with indistin-guishable electrode placements and experi-mental procedures, different rats showed verydifferent responses to brain stimulation (Coxand Valenstein, 1969; Valenstein et al., 1970).Moreover, rats with electrodes implanted atdifferent hypothalamic sites showed a strongtendency to display the same behavior fromboth electrodes (Valenstein et al., 1970). Fur-thermore, Wise (1971) tested rats for evokedeating and drinking using moveable hypo-thalamic electrodes. Animals that displayed anevoked consummatory response both ate and

37

Page 55: 0195162854

38 NATURAL HISTORY

drank, and continued to do so as electrodeswere advanced as much as 1.5 mm in adorsal-to-ventral direction through the hypo-thalamus. Within the limits of this positivearea, movement of the electrode had little ef-fect on current threshold for evoking eatingand drinking. Animals that displayed no re-sponse failed to eat or drink at any stimulatedsite. Wise concluded that the variability be-tween animals reflected individual response ten-dencies rather than differences in the site ofstimulation. Bachus and Valenstein (1979) de-stroyed hypothalamic cells and fibers sur-rounding stimulating electrodes that evokeddrinking in Long-Evans rats. Although thelarger lesions extended up to 3.0 mm from thecenter of the electrode and higher current lev-els were required to excite more distal neu-ronal elements, all animals continued to drinkwhen stimulated. Thus, stable characteristicsof animals, and not the precise neuroanatom-ical locus of the electrode, accounted for theresponse to hypothalamic stimulation. It wasalso demonstrated that there were strain dif-ferences (Long-Evans versus Sprague-Dawley)in the likelihood of evoking consummatorybehavior with ESLH, suggesting a genetic link(Mittleman and Valenstein, 1981). Valenstein(1969) concluded that this evidence providedjustification for postulating that individual an-imals had a "prepotent" and intrinsic tendencyto respond in a characteristic manner toESLH.

INDIVIDUAL DIFFERENCESAND RELATED PSYCHOLOGICAL

AND NEUROPHYSIOLOGICALCHARACTERISTICS

Because eating and drinking elicited by ESLHare forms of nonregulatory ingestive behav-ior, it seemed logical to compare behaviorselicited by ESLH with those occurring in an-other situation that elicited nonhomeostaticconsumption. One of the best-known meth-ods for eliciting nonregulatory ingestion is the

schedule-induced polydipsia paradigm (Falk,1961). When rodents, for example, are foodbut not water deprived and given intermittentpresentations of small amounts of food, manydevelop excessive fluid consumption, calledschedule-induced polydipsia (SIP). Such ex-cessive drinking has been reported in manyanimal species and occurs under a variety ofschedules of reinforcement (for reviews, seeWallace and Singer, 1976; Roper, 1981;Wetherington, 1982). It has been suggestedthat SIP is an example of the larger categoryof adjunctive behaviors that occur in situa-tions where strongly motivated appetitive orconsummatory behaviors are interrupted orthwarted and probably shares characteristicsin common with displacement behaviors thatoccur in more natural settings (Tinbergen,1952; Falk, 1966, 1969, 1971).

Mittleman and Valenstein (1984) im-planted 42 adult, male Long-Evans rats withhypothalamic electrodes and screened themfor eating and drinking elicited by ESLH.Twenty-four animals initially ate or drank(ESLH-positive) in response to stimulation,whereas 18 (ESLH-negative) did not. Whentested in SIP, ESLH-positive animals rapidlyacquired drinking, which increased from 2 mlon day 1 to more than 11 ml on day 10,whereas ESLH-negative rats hardly drank atall on day 1, and consumption increased toonly 4 ml by day 10. This study shows that in-dividual differences in response to ESLH arepredictive of differences in SIP. Thus, the be-havioral differences observed in ESLH are notunique to this paradigm but are representa-tive of a more global characteristic. Pre-dictability extended to cognitive ability and"emotionality." That is, SIP-positive animalsmore rapidly learned an active avoidance re-sponse and showed less freezing when con-fronted with an aggressive resident rat in theresident-intruder paradigm (Dantzer et al.,1988).

The individual differences observed inthese paradigms also have been associatedwith biological differences in the response

Page 56: 0195162854

Chapter 4. Individual Differences 39

properties of mesolimbic dopamine systems(e.g., Antelman and Szechtman, 1975; Rob-bins and Koob, 1980; Wallace et al., 1983;Fibiger and Phillips, 1986). Individual differ-ences in the behavioral response to amphet-amine were first investigated as an indirectmeans of assessing dopaminergic activity.Long-Evans rats that readily engage in SIPconsistently show an enhanced response to asingle injection of D-amphetamine (Mittlemanand Valenstein, 1985). In addition, in responseto repeated injections of this drug, SlP-posi-tive rats (Fig. 4-1, left, solid line), in compar-ison to negative animals (dashed line), showedmore rapid behavioral sensitization of overallstereotypy (Mittleman et al., 1986). Becausedopaminergic activity is the primary substrateof stereotyped behavior (Creese and Iversen,1975), these results suggested differences indopaminergic neural systems in SIP-positiveand -negative rats. Neurochemical evidencefor such a difference was confirmed using anindirect measure of dopamine turnover (seeFig. 4-1, right). Foot shock was used to in-crease dopamine utilization in SIP-positive(solid bar) and -negative (open bar) Long-

Evans rats. Clear differences between thegroups emerged. Positive rats showed signifi-cantly greater increases in dopamine turnoverin the striatum and nucleus accumbens thandid SIP-negative (open bar) animals, as as-sessed by ratios of 3,4-Dihydroxyphenylaceticacid to dopamine and high-performance liq-uid chromatography-with electrochemical de-tection (Mittleman et al., 1986). That Long-Evans rats have significantly different individualresponse profiles to single and repeated injec-tions of amphetamine as well as associated re-gional differences in dopamine metabolism wasconfirmed by Segal and Kuczenski (1987).

In agreement with these lines of research,more recent investigations have confirmedthe general concept that individual differencesin the unconditioned response to psychomo-tor stimulants are related to the responseproperties of dopaminergic systems. Sabeti etal. (2003) reported that, in outbred Sprague-Dawley rats, individual differences in the lo-comotor response to cocaine are directly re-lated to corresponding differences in the rateof dopamine clearance in the nucleus accum-bens. The mechanism underlying this indi-

Figure 4-1. (Left) The development of amphetamine-induced stereotypy in SIP-positive (solid line) and-negative rats (dashed line). Both groups received an injection of 3.0 mg/kg D-amphetamine sulfate once every3 days for 27 days (a total of nine injections). Overall stereotypy was rated on an 8 point scale. The averagerating of the 2 hour test session is shown. (Right) SIP-positive (solid bar) and -negative (open bar) rats received20 minutes of intermittent footshock (1.3 mA for 0.5 second every 15 seconds). Dopamine and dihydroxy-phenylacetic acid (DOPAC) levels in the nucleus accumbens and striatum were determined, and theDOPAC/dopamine ratio was used as an indirect measure of dopamine utilization. Results are expressed asthe percentage increase over an unshocked control group composed of both positive and negative animals.(Modified from Mittleman et al. [1986].)

Test Day Brain Region

Page 57: 0195162854

40 NATURAL HISTORY

vidual difference was differential inhibition ofnucleus accumbens dopamine transporters.

Because individual differences in ESLH,SIP, and the behavioral responsiveness to psy-chostimulants can be linked to correspondingneurochemical differences in forebrain dopa-mine, it was predicted that this relationshipwould extend to individual differences in ad-dictive liability. This possibility was confirmedby Piazza and colleagues using an animal modelof intravenous drug self-administration (Weeks,1962; Schuster and Thompson, 1969). Piazza etal. (1993) tested male Sprague-Dawley rats forspontaneous locomotion in a novel environ-ment. Animals could be divided into "high" and"low" responders based on the amount of spon-taneous locomotion exhibited when tested in acircular corridor for 2 hours (Piazza et al.,1989). Animals were then implanted withintracardiac catheters and allowed to self-administer a low dose of D-amphetamine sul-fate (10 /xg/20 /zl/inj). Behavioral differencesin self-administration were highly correlatedwith differences in activity levels (Piazza et al.,

1989). Figure 4-2 (left) shows the significantdifferences in the number of drug requests (asindicated by nose pokes into the "active" hole)made by rats that showed high and low lev-els of locomotion in response to a novel en-vironment. Also shown is the number of re-sponses made into the "inactive" hole by thesetwo groups. Responding in the inactive holewas quite similar, suggesting that these twogroups differed specifically in the number ofdrug requests, which minimized the role ofnonspecific activational effects of amphet-amine in the observed difference in drugrequests.

The rats were then gradually food de-prived to 85% of their baseline weight andtested for SIP using standard procedures. Asindicated (see Fig. 4—2, right), animals thatwere predisposed to self-administer ampheta-mine (SA+) acquired SIP significantly more rap-idly than animals that did not self-administer(SA—). This figure also shows a subgroup ofSA— rats that were tested for an additional 5days in the self-administration paradigm, dur-

Active Inactive

Hole Test Days

Figure 4-2. (Left) Self-administration testing was conducted in test chambers with a hole (2 cm above thefloor) in each of the two short walls. Each hole was monitored by an infrared photocell beam such that nosepokes in one of the holes (denned as "Active") turned on the infusion pump for 2 seconds, injecting 10 /tgof D-amphetamine sulfate into the animal's venous system. Nose pokes in the other hole (defined as "Inac-tive") had no effect. During each 30 minute session, the number of nose pokes in both the active and the in-active holes was recorded for animals that were predisposed to self-administer amphetamine (SA+) and thosethat did not (SA—). (Right) Rats were given 10 daily 30 minute SIP tests in standard operant cages equippedwith a drinking tube connected to a water-filled graduated cylinder. After each test, the total amount of wa-ter consumed was recorded. Filled squares designate SA+ animals (those that acquired SA); open trianglesindicate D-amphetamine-negative rats. Also shown (open diamonds) is the performance of SA (imposed) ratsthat were yoked to SA+ animals so that they received the same amount and timing of amphetamine infu-sions. (Modified from Piazza et al. [1993].)

Page 58: 0195162854

Chapter 4. Individual Differences 41

ing which drug requests from a randomly se-lected SA+ rat were "played back" in the or-der that the rat received the amount and tim-ing of amphetamine administrations. TheseSA— imposed animals were then tested in theSIP paradigm to determine if the amount ofprior amphetamine exposure was a factor thatinfluenced responsiveness in SIP. As shown(see Fig. 4-2, right), these animals did not dif-fer from the SA— animals, suggesting thatprior drug exposure was not an important fac-tor influencing the development of individualdifferences in SIP.

Taken together, a number of possibleconclusions are suggested by these results.First, they suggest a relationship between in-dividual differences in the behavioral responseto activating or "arousing" conditions such asSIP or ESLH and profound individual differ-ences in forebrain dopamine systems asdemonstrated by behavioral, psychopharma-cological and neurochemical means. Rats thatreadily engaged in excessive drinking duringSIP showed significantly greater behavioral re-sponsiveness to single or repeated injectionsof D-amphetamine. These drug-induced dif-ferences in behavior were related to signifi-cantly greater stress-induced neurochemicalresponsiveness of the mesolimbic dopaminesystem. Second, these results further indicatedthat the consistent behavioral differences ob-served in SIP are related to individual differ-ences in the predisposition to self-administerthe psychoactive compound amphetamine.This was demonstrated by the high level ofconcordance between amphetamine SA andSIP. This result is particularly important be-cause it provides further evidence that theseindividual differences in responsiveness arepart of a more global characteristic. From theresults cited earlier, it appears that this"global" characteristic may include cognitive,emotional, pharmacological, and neurochem-ical components.

These results support the notion that con-sistent individual differences in behavior can beused as a means for investigating the neuro-

physiological and psychological underpinningsof behavior, but they do not, by themselves,indicate that the investigation of individualdifferences in rats has relevance toward un-derstanding the individual differences in vul-nerability to addiction observed in humans.Nevertheless, a national survey by the Sub-stance Abuse and Mental Health Services Ad-ministration (SAMHSA) (2003, p. 55) indicatesthat in 2002, approximately 9.4% of the popu-lation of the United States was classified as sub-stance abusers or substance dependent, basedon the criteria specified in the Diagnostic andStatistical Manual of Mental Disorders, 4th edition(American Psychiatric Association, 1994).

INDIVIDUAL DIFFERENCES INRATS HAVE FACE VALIDITY WITHRESPECT TO HUMAN INDIVIDUAL

DIFFERENCES AND CANNOT BEEXPLAINED BY PROXIMAL

EXPERIENTIAL FACTORS

In an effort to determine if the observed indi-vidual differences in self-administration we hadobserved in rats were similar to the vast indi-vidual differences seen in humans, we investi-gated the acquisition of amphetamine self-administration in naive rats. The goals of thisexperiment were to (1) document individualdifferences in the acquisition of drug taking, (2)simultaneously determine any dose prefer-ences during the acquisition process, and (3) in-vestigate any experiential factors that might ex-plain the development of individual differencesin amphetamine self-administration.

Male Long-Evans rats (Harlan; n = 207;age range, 50 to 125 days) were maintainedon food and water ad libitum throughout test-ing. They were initially tested for locomotionin a novel environment using the methods ofPiazza et al. (1993). The animals were thenimplanted with a catheter and tested in ex-perimental chambers in which one curved wallcontained a row of five 2.5 cm round holes (acentral hole and two on each side) illuminated

Page 59: 0195162854

42 NATURAL HISTORY

by a yellow LED. A nose poke response intoeach hole was detected by an infrared photo-beam. A house light located near the ceiling ofthe experimental chamber signaled the avail-ability of drug. Drug self-administration contin-gencies were programmed such that an animalreceived a different drug dose depending onthe hole where a nose poke was made. Eachinjection was followed by a 30 second time-out period, during which the house light andthe lights within each hole were extinguishedand any additional nose poke responses wererecorded but had no programmed conse-quences. Drug concentration and pump de-livery rate (2.0 //,1/s) were kept constantthroughout an experimental session, whereasunit dose of drug per injection was controlledby varying the duration of pump action (i.e.,volume of injected solution). In this experi-ment animals could sample five doses (0.0,0.018, 0.032, 0.056, and 0.10 mg/kg per injec-tion) by making a nose poke response into thefive different stimulus "holes."

Rats had 23 hours of daily access to thedrug for a total of 10 days. The animals' firstexperimental session was composed of a sam-pling component and a choice component,whereas subsequent experimental sessionswere composed of the choice componentonly. During the sampling component, ratshad to request each drug dose before accessto all doses was permitted. Provided that thefive doses were sampled before the end of theexperimental session, the choice componentbegan. During this period, the programmedexperimental contingencies were identical tothose described for the sampling componentexcept that all five drug doses were continu-ously available after an injection-producedtime-out. The dose associated with each holewas counterbalanced using a Latin square toreduce any response bias. During the first 5days of testing, animals responded on a fixedratio 1 (FR 1) schedule. This was changed toan FR 2 schedule for the final 5 test days todetermine if the animals were willing to ex-pend more effort to receive the drug.

A total of 146 animals completed the ex-periment. Figure 4-3 (top) describes the rangeof individual differences in amphetamine self-administration. Average daily drug taking inindividual rats ranged from a mean of 0.08 to28.2 mg/kg. Because of this vast range of re-sponding, we arbitrarily divided the groupinto thirds using the standard deviation. Lowresponders were defined as those animals self-administering the mean amount or less(range, 0.81 to 2.02 mg/kg per day; n = 116,or 80% of the population). Middle respondersranged from the mean to 1 SD above themean (range, 2.03 to 5.90 mg/kg per day;n = 15, or 10% of the population), and highresponders were those animals that self-

Figure 4-3. (Top) The sample distribution of the 146 ratstested for 10 days of D-amphetamine self-administration.Average daily drug infused ranged from 0.81 to more than28 mg/kg per day. (Bottom) Self-administration behavior ofthe low, middle, and high groups (see text for definition).Moderate to high levels of amphetamine self-administrationwere associated with a "binge and crash" pattern. Theswitch to an FR 2 schedule reduced responding in the highand middle groups.

Page 60: 0195162854

Chapter 4. Individual Differences 43

administered amounts greater than 1 SD abovethe mean (range, 5.91 to 28.2 mg/kg per day;n = 15, or 10% of the population). Analyses in-dicated that the groups differed significantlyfrom one another and that responding in thehigh and middle groups was significantly re-duced by the switch to an FR 2 schedule. Asshown in Figure 4—3 (bottom), both of thesegroups showed a characteristic "binge andcrash" pattern of responding for drug.

Significant dose preferences that weregroup dependent appeared on test day 1. Asindicated in Figure 4—4, rats in the high groupshowed the clearest preference. These ani-mals preferred the middle dose (0.032 mg/kgper injection) over the 0.0 control dose. Lowresponders took significantly more of the mid-dle (0.032 mg/kg per injection) and high (0.10mg/kg per injection) doses in comparison tothe control dosage. Surprisingly, rats in themiddle group failed to develop a dose prefer-ence. When considered over the 10 test days(Fig. 4-5), dose preferences remained consis-tent only in the high-responder group. Theseanimals self-administered the middle dosepreferentially throughout testing. Perhaps be-cause of the lower levels of self-administration

TEST DAY 1 (FR 1) PREFERENCES

Figure 4-4. Dose preferences of the low, middle, and highgroups when first given the opportunity to self-administeramphetamine on test day 1. Both the low and high groupsrapidly developed a significant preference for the 0.032(mg/kg per injection) dose. In contrast, the middle groupdid not show an initial dose preference. **p < .01 differ-ence from the 0.0 mg/kg per injection dose.

DRUG DOSE (mg/kg/inj)

Figure 4-5. Dose preferences of the low, middle, and highgroups over the 10 days of testing. Although the low groupshowed an initial preference, it was not maintainedthroughout testing. The middle group never exhibited aconsistent preference over the 10 days; only the high groupmaintained a consistent dose preference throughout test-ing. **p < .01 difference from the 0.0 mg/kg per injectiondose. SED, standard error of the difference in means.

and corresponding reduced experience withthe drug, animals in the middle and lowgroups failed to show a significant dose pref-erence when considered over the 10-day test-ing interval.

We also wanted to determine if a varietyof factors were related to the vast individual dif-ferences in amphetamine self-administration.Because age is a significant factor in humanabuse (SAMHSA, 2003), it was included as afactor. It also seemed reasonable that differ-ences in self-administration might be a simpleoutgrowth of differences in activity. Thus,more active animals might be more likely tonose poke for drug. Activity in a novel circu-

DOSE (mg/kg/inj)

DOSE PREFERENCES ACROSS ACQUISITION (10 days)

Page 61: 0195162854

44 NATURAL HISTORY

lar corridor and activity in the test chamber(day 1) before the initial infusion were in-cluded for these reasons. The first drug doseselected was included as a representative ex-periential variable that might influence subse-quent abuse. Because there are experiential orenvironmental influences on the likelihood ofdrug taking (Everitt et al., 2001) we reasonedthat self-administration might be influencedby the first dose selected. For example, itseemed possible that the likelihood of self-administration would be reduced if the firstdose selected had no effects (0.0 mg/kg perinjection) or was potentially aversive to anaive animal (0.10 mg/kg per injection). Fi-nally, to determine if characteristics of initialdrug taking predicted individual differences inself-administration, the latency to the firstdrug request and total amount of drug infusedon day 1 were also included.

Table 4-1 presents the results of theseanalyses. Age was significantly negatively cor-related with individual differences in amphet-amine self-administration, in that, much like

Table 4-1. Summary of the Relationship between

Predictor and Performance Variables

Predictor Variables

Age (days)

Activity countsin circularcorridor

Activity beforefirst drugrequest (day 1)

First drug doseselected(day 1)

Latency to firstdrug request(day 1)

Total druginfused(day 1)

Average Total Drug

Infused per Day*

-.2106.0107

146.1132.1739

146-.0382

.6672129

.0318

.7034146-.1634

.0487146

.5602

.0000146

*Values given as Pearson correlations,No. of observations.

two-tailed P values, and

humans, larger amounts of drug taking wereassociated with younger ages. Differences inself-administration were not a byproduct ofgeneral activity. Neither activity in the circu-lar corridors nor activity in the test chamberswas related to the amount of self-administration.Initial experience in self-administration alsofailed to significantly influence subsequentself-administration, as the first dose selectedwas unrelated to the amount of drug infused.Not surprisingly, initial behavior related toself-administration was predictive of individ-ual differences in drug taking. The latency tocommence self-administration was negativelycorrelated with drug taking in that rats withshorter latencies to self-administer took moreamphetamine. Self-administration on day 1also significantly predicted drug taking overthe 10-day test interval.

Considered together, these results indi-cate that, similar to human individual differ-ences in vulnerability to addiction, about 10%of the rats showed high levels of abuse. Thesedifferences could be categorized along two di-mensions: the amount of drug requested andin terms of dose preferences. These individualdifferences in amphetamine self-administrationwere related to age of the animal and initialself-administration but not to differences in ac-tivity or initial contact with the drug. Twoconclusions are suggested. First, rats appear tobe good models for studying human vulnera-bility to addiction. They show a wide rangeof individual differences in drug taking and,with repeated drug contact, a similar per-centage go on to show substantial abuse. Sec-ond, it seems most parsimonious to attributethese observed individual differences to somesort of innate biological predisposition thatvaries with age.

CONCLUSIONS

Variability in the response to drugs, lesions,or various other experimental manipulationsare frequently observed in animals, but they

Page 62: 0195162854

Chapter 4. Individual Differences 45

are typically acknowledged as simply repre-senting experimental error along with the nat-ural range of variation occurring between in-dividuals of the same species. Although thistraditional attitude does have merit, results ofthe investigations presented in this chaptersuggest that much within-experiment vari-ability consists of individual differences. As il-lustrated in this chapter, exploiting these in-dividual differences has a potentially largepayoff in terms of modeling the behavioraland physiological differences observed be-tween humans as well as providing a meansof exploring the factors that control the ex-pression of such differences. In addition, be-cause behavioral and neurochemical differ-ences are preexisting, it is unnecessary toinduce differences by using artificial meanssuch as lesions, pharmacological agents, orradical changes in the environment. Thus, us-ing an individual differences approach, the be-havioral and neurochemical sequelae of thesedifferences may be investigated in intact or-ganisms. It should be noted that individual dif-ferences are being identified in many differenttasks. A current search on PubMed using thekeywords "rat," "individual differences," and"behavior" returned more than 300 articles.

ACKNOWLEDGMENTS

The author wishes to thank the following for their assistancewith some of the experiments reported in this chapter: includ-ing Carrie L. Van Brunt, Rachel Chase, Mary Houts, Pat LeDue, Paul Rushing, Peter Pierre, and Paul Skjoldagger. The self-administration experiment was supported by NIDA grant1R29DA07517.

Dr. William Marks, a cognitive psychologist and long-time supporter of the Neuroscience Program at the Universityof Memphis, died unexpectedly while this chapter was beingwritten. This work is dedicated to his memory.

REFERENCES

American Psychiatric Association (1994) Diagnostic andstatistical manual of mental disorders, 4th edition.Washington, DC: American Psychiatric Association.

Antelman SM and Szechtman H (1975) Tail pinch in-

duces eating in sated rats which appears to dependon nigrostriatal dopamine. Science 189:731-733.

Bachus SE and Valenstein ES (1979) Individual behav-ioral responses to hypothalamic stimulation persistdespite destruction of tissue surrounding the elec-trode tip. Physiology and Behavior 23:421-426.

Cox VC and Valenstein ES (1969) Distribution of hypo-thalamic sites yielding stimulus-bound behavior.Brain Behavior and Evolution 2:359-376.

Creese I and Iversen SD (1975) The pharmacological andanatomical substrates of the amphetamine responsein the rat. Brain Research 83:419-436.

Dantzer R, Terlouw C, Tazi A, Koolhaas JM, Bohus B,Koob GF, Le Moal M (1988) The propensity forschedule-induced polydipsia is related to differencesin conditioned avoidance behaviour and in defensereactions in a defeat test. Physiology and Behavior43:269-273.

De Wit H, Uhlenthuth EH, Johanson CE (1986) Indi-vidual differences in the reinforcing and subjectiveeffects of amphetamine and diazepam. Drug and Al-cohol Dependence 16:314-3 60.

Everitt BJ, Diskinson A, Robbins TW (2001) The neu-ropsychological basis of addictive behavior. BrainResearch Brain Research Reviews 36:129-138.

FalkJL (1961) Production of polydipsia in normal ratsby an intermittent food schedule. Science 133:195-196.

FalkJL (1966) The motivational properties of schedule-induced polydipsia. Journal of the ExperimentalAnalysis of Behavior 9:19-25.

FalkJL (1969) Conditions producing psychogenic poly-dipsia in animals. Annals of the New York Academyof Science 157:569-593.

FalkJL (1971) The nature and determinants of adjunc-tive behavior. Physiology and Behavior 6:577-588.

Fibiger, HC, Phillips AG, (1986) Reward, motivation andcognition: psychobiology of meso-telencephalicdopamine systems. In: Handbook of physiology:higher neural functions (Bloom FE and Plum F,eds.), pp. 647-675. Washington, DC: AmericanPhysiological Society.

Koob GF and Bloom FE (1988) Cellular and molecularbasis of drug dependence. Science 242:715-723.

Mittleman G and Valenstein ES (1981) Strain differencesin eating and drinking evoked by electrical stimula-tion of the hypothalamus. Physiology and Behavior26:371-378.

Mittleman G and Valenstein ES (1984) Ingestive be-havior evoked by hypothalamic stimulation andschedule-induced polydipsia are related. Science224:415-417.

Mittleman G, Castaneda E, Robinson TE, Valenstein ES(1986) The propensity for non-regulatory ingestivebehavior is related to differences in dopamine sys-

Page 63: 0195162854

46 NATURAL HISTORY

terns: behavioral and biochemical evidence. Behav-ioral Neuroscience 100:213-220.

Mittleman G and Valenstein ES (1985) Individual dif-ferences in non-regulatory ingestive behavior andcatecholamine systems. Brain Research 348:112-117.

O'Brien CP, Ehrman, RN, Terns JN (1986) Classical con-ditioning in human opioid dependence. In: Behav-ioral analysis of drug dependence (Goldeberg SRand Stolerman IP, eds.), pp. 329-338. London: Aca-demic Press.

Piazza. PV, Mittleman G, Deminiere JM, Le Moal M,Simon H (1993) Relationship between schedule-induced polydipsia and amphetamine self-adminis-tration: individual differences and the role of expe-rience. Behavioural Brain Research 55:185-194.

Piazza PV, Demeniere JM, Le Moal M, Simon H (1989)Factors that predict individual vulnerability to am-phetamine self-administration. Science 245:1511-1513.

Robbins TW and Koob GF (1980) Selective disruptionof displacement behavior by lesions of the mesolim-bic dopamine system. Nature (London) 285:409-412.

Roper TJ (1981) What is meant by the term "schedule-induced" and how general is schedule induction?Animal Learning and Behavior 9:433-440.

Sabeti J, Gerhardt GA, Zahniser NR (2003) Individualdifferences in cocaine-induced locomotor sensitiza-tion in low and high cocaine locomotor-respondingrats are associated with differential inhibition ofdopamine clearance in nucleus accumbens. Journalof Pharmacology and Experimental Therapeutics305:180-190.

Schuster CR and Thompson T (1969) Self-administra-tion and behavioral dependence on drugs. AnnualReview of Pharmacology 9:483-502.

Segal DS and Kuczenski R (1987) Individual differencesin responsiveness to single and repeated ampheta-

mine administration: behavioral characteristics andneurochemical correlates. Journal of Pharmacologyand Experimental Therapeutics 242:917-926.

Substance Abuse and Mental Health Services Adminis-tration (2003) Results from the 2002 National Sur-vey on Drug Use: National Findings. Rockville, MD:Office of Applied Studies, NHSDA Series H-22,DHHS publication No. SMA 03-3836.

Tinbergen N (1952) "Derived" activities: their causation,biological significance, origin, and emancipation dur-ing evolution. Quarterly Review of Biology 27:1-32.

Valenstein ES (1969) Behavior elicited by hypothalamicstimulation: a prepotency hypothesis. Brain Behav-ior and Evolution 2:295-316.

Valenstein ES (1975) Brain stimulation and behaviorcontrol. In: Nebraska Symposium on Motivation(Cole JK, Sonderegger TB, eds.), pp. 251-292. Lin-coln: University of Nebraska Press.

Valenstein ES, Cox VC, Kakolewski JW (1970) A reex-amination of the role of the hypothalamus in moti-vation. Psychological Review 77:16-31.

Wallace M and Singer G (1976) Schedule induced be-havior: a review of its generality, determinants andpharmacological data. Pharmacology Biochemistryand Behavior 5:483-490.

Wallace M, Singer G, Finlay J, Gibson S (1983) The ef-fects of 6-OHDA lesions of the nucleus accumbensseptum on schedule-induced drinking, wheel run-ning and corticosterone levels in the rat. Pharma-cology Biochemistry and Behavior 18:129-136.

Weeks JR (1962) Experimental morphine addiction:method for automatic intravenous injections in un-restrained rats. Science 138:143-144.

Wetherington CL (1982) Is adjunctive behavior a thirdclass of behavior? Neuroscience and BiobehavioralReviews 6:329-350.

Wise RA (1971) Individual differences in the effects ofhypothalamic stimulation: the role of stimulationlocus. Physiology and Behavior 6:569-572.

Page 64: 0195162854

Sensory Systems II

Page 65: 0195162854

This page intentionally left blank

Page 66: 0195162854

Vision

GLEN T. PRUSKY AND ROBERT M. DOUGLAS5

In some ways the rat is an odd choice as a lab-oratory animal. For one, the rat is nocturnal,whereas most laboratories operate on a diur-nal schedule with testing done in relativelybright light. Second, the rat evolved in acrowded ground environment of forest un-derbrush, wetlands, grassy fields, and under-ground burrows. Such places are as rich in ol-factory, auditory, and tactile information asthey are in visual information. Laboratory con-ditions are the opposite: open, uncrowded, andclean. Testing boxes and rooms thus providean unnatural visual environment and a mini-mum of other sensory information. The con-sequences for vision and visually guided be-havior are not obvious, and the situation hasnot been helped by the limited experimentaldata about rat vision. Despite an early start byLashley and others, the psychophysical analy-sis of rat vision has lagged as vision re-searchers focused their efforts on larger mam-mals like cats and primates. This relative stateof ignorance has lead to two widespread butcommon misconceptions that we wish to cor-rect in this chapter. First, many vision scien-tists believe the rat is almost blind and thus isa poor choice as an experimental model of vi-sion. In fact, the rat has a typical mammalianvisual system that functions quite well, andrats can learn demanding visual tasks rivalingthose used in primates. Second, those inter-ested in animal cognition often assume thatthe rat sees what a human sees, or that all ratssee equally. Differences between stimuli maybe obvious to the experimenter, but they maynot be so for a rat, let alone for different strains

of rats. Furthermore, even if visual informa-tion is available in an experimental setting andwould be used by a primate in a task likereaching, the visual information may not nec-essarily be used (e.g., Whishaw and Tolmie,1989). In an effort to place rat vision in a sen-sible experimental setting, we discuss in thischapter some of the methodologies used tomeasure rat vision, the known visual capabil-ities of rats, and the implications of rat visionfor those researchers who use rats in vision-based experiments.

EXPERIMENTAL METHODS FORMEASURING RAT VISION

Lashley's jumping stand may have been thefirst method used to quantify rat vision (Lash-ley, 1930; Seymoure and Juraska, 1997), and itis still used to a limited extent. Y-mazes (Sey-moure and Juraska, 1997), conditioned aver-sion (Dean, 1978), and operant tasks (Keller etal., 2000; Jacobs et al, 2001) have all been usedwith some success, but in general, these meth-ods require a considerable amount of time totrain and test rats, which probably accountsfor their limited popularity. Some experi-menters have also used a modification of theMorris water task, in which rats learn to swimto a platform that is raised above the water'ssurface (Morris et al., 1982). However, view-ing distances are hard to control in this sit-uation, making quantitative measurementsnearly impossible, and the task confuses visualdetection with visual acuity (this issue is dealt

49

Page 67: 0195162854

50 SENSORY SYSTEMS

with in more detail later in the chapter). Onthe other hand, swimming to an exit platformand using distant visual cues seem to comenaturally to the rat and prompted us to de-velop the visual -water task (Prusky et al, 2000)to psychophysically measure visual discrimi-nation thresholds. As shown in Figure 5-1, a

Figure 5-1. Visual water task. (A) Top view. Apparatusconsists of a trapezoidal tank containing water with a mid-line barrier creating a Y-maze. Two monitors face into thearms of the maze and display either a + (reinforced) or a —(nonreinforced) stimulus. A platform is always submergedbelow the + stimulus regardless of its left/right position.Rats are released into the pool at the narrow end and thenswim to the divider. They inspect each picture from thisvantage and then choose to swim toward one of them. Ifthey select the + stimulus, they are rewarded quickly withescape from the water and the trial is scored as correct. Ifthey choose the — stimulus, they are obligated to swim un-til they find the escape platform on the opposite side, andthe trial is scored as incorrect. The path of a rat executinga correct response is illustrated. (B) Front view of appara-tus and visual stimuli configured for typical testing of vi-sual acuity; gray Vs grating.

trapezoidal-shaped tank is made into a Y-mazewith a central divider and with computermonitors placed behind a glass wall at the endof each arm. A platform is submerged belowa positive stimulus displayed on one of thetwo monitors. Once rats are taught that theimages on the screens are clues to the loca-tion of the platform, the animals stop at theend of the divider and inspect both screens be-fore making their choices. At no time have wehad to explicitly reinforce this behavior; ratsspontaneously stop and compare the screensbefore taking a chance on going the wrongway. The water aids in dispersing odor trailsand focuses the rat's attention on the com-puter monitors without generating a greatdeal of stress. On land, rats have a rich arrayof sensory inputs to consider, but when in wa-ter, rats seem to know that vision is the bestmodality to use and that the visual cues willbe some distance away. Besides the task ex-ploiting ecologically relevant behavior, theuse of computer-generated stimuli is also ad-vantageous. The computer monitors allowstimuli to be presented that are impossible toproduce with printed cards; we have usedstimuli that consist of moving patterns andvarying contrasts with a wide range of values.In addition, computer control permits the au-tomatic interleaving of stimuli and animals,which greatly increases throughput in the lab-oratory. Perhaps the major advantage of thevisual water task is that the training and test-ing of visual thresholds can be completedmuch faster than with the other methods citedabove.

THE RAT EYE

The small eye of the rat is reasonably efficientat gathering light, but it has relatively poor op-tics. The rat retina appears homogeneous andlacks a fovea or area centralis. The retina con-tains both rods and cones, with the propor-tion of cones being about 1% (LaVail, 1976).The cone system is often overlooked in rats;

Page 68: 0195162854

Chapter 5. Vision 51

however, almost all behavioral testing re-ported in rats has taken place in photopic (lab-oratory lighting) conditions in which the rodsare not functioning. Rats are typical mam-malian dichromats with short-wavelengthcones comprising about 10% of the total conepopulation (Szel and Rohlich, 1992). Thereis a twist, however: the short-wavelengthcone is most sensitive in the ultraviolet, withthe peak at 359 nm (Fig. 5-2A). The mid-wavelength cone has peak sensitivity at 510 nm,and there is little sensitivity beyond 650 nm.The evolutionary significance of the ultravio-let sensitivity is not known, but in a series ofelegant experiments, Jacobs et al. (2001) haveshown that rats can make color discrimina-tions using their two cone systems.

SPATIAL VISION

Acuity is the most common measure of visionand is the smallest spatial pattern that can beresolved. Distinguishing between two smalldots and one large one, for example, dependson having adequate acuity. Most pigmentedrats have acuities of 1.0 to 1.1 cycles per de-gree (c/d) when measured with vertical oroblique gratings (Dean, 1978; Burch and Ja-cobs, 1979; Keller et al., 2000; Prusky et al.,

2000). With horizontally oriented gratings,their acuity is 1.4 c/d (Bowden et al, 2002).This is much lower than normal human vi-sual acuity (20:20) of 30 c/d. However, hu-mans have a highly specialized fovea, and therat acuity is not greatly dissimilar from that ofhuman peripheral retina. In addition, much ofthe world that interests a rat will be viewedat close distances and they likely see these ob-jects in considerable detail (see Figs. 5-3 and5-4 for examples).

Humans can detect single elementssmaller than the acuity limit if they have suf-ficiently high contrast. For example, we cansee stars even though they subtend anglessmaller than 1 minute of arc, which corre-sponds to 30 c/d. The brightness is averagedover the larger area and the element is seenas blurred, covering the wider area with alower contrast. Humans also have good con-trast sensitivity (>100, corresponding to <1%difference between light and dark), so we candetect the presence of very small single ele-ments. Although rats have an inverted U-shaped contrast sensitivity curve characteris-tic of all vertebrates, the peak sensitivity islower than that of humans (Keller et al., 2000).As shown in Figure 5-2B, rats have a peak sen-sitivity of about 25, meaning they can detecta 4% difference in a pattern of 0.1 to 0.2 c/d.

Wavelength (nm) Spatial Frequency (cpd)

Figure 5-2. (A) Relative sensitivities of the two cone pigments in the rat. The short-wavelength pigment (S)has a sensitivity that peaks in the ultraviolet range. This is quite different from the human short-wavelengthpigment (leftmost arrow). The more abundant medium wavelength (M) pigment has a sensitivity more simi-lar to that of human M and L wavelength pigments. (Redrawn from Jacobs et al. [2001]). (B) Contrast sensi-tivity for Long-Evans rats. Contrast sensitivity is the reciprocal of the threshold contrast (right side) at eachspatial frequency. Peak sensitivity is about 0.2 c/d, and the acuity cutoff is just above 1.0 c/d.

Page 69: 0195162854

52

Page 70: 0195162854

Chapter 5. Vision 53

Figure 5-3. It is difficult to judge whether a rat can see a vi-sual cue in a behavioral task because the visibility of the cuedepends on its size, its distance from the rat, and the acuityof the rat. This series of 10 images of the authors (G.P. onleft; R.D. on right; width of image is 45 cm) can be used intwo ways. First the images model what animals with differ-ent acuities would see at a distance of 1.0 m. The acuities (inc/d) are shown in the top right corner of each image. Theimages have been filtered to emulate different acuities by re-moving all spatial frequencies above the frequency listed. Fre-quencies near the acuity limit are modeled as an attenuation.Second, the series of images can also be used to gauge whata normal wild or pigmented rat would see at different dis-tances (viewing distances are shown at the right of each im-age). For animals with other acuities, the figure can be usedin the same way by simply multiplying the distances by theacuities. For example, an albino rat with an acuity of 0.5 c/dwill see at 50 cm what a normal rat sees at 1.0 m. Conversely,a Fisher-Norway rat with an acuity of 1.5 c/d can be seen at1.5 m and see the same as a normal rat at 1.0 m.

Although rats have relatively low-resolu-tion retinas, subsequent neural processing ap-pears to make excellent use of the informa-tion provided by the retina. For example, wehave found that rats can discriminate gratingsthat differ in orientation by less than 3° (Bow-den et al., 2002). We have also obtained evi-dence for rat hyperacuity in a Vernier acuitytask. Hyperacuity in orientation and Vernierdiscriminations is thought to reflect the spe-cialized processing of visual cortical circuitry.We also have evidence that rats can detectmotion coherence in a field of moving dots; acapability that in primates is thought to be dueprimarily to extrastriate processing (Neve etal., 2002). The coherence thresholds for rats

Albino Wild/Pigmented Fisher-Norway Human

Figure 5^t. The visual environment for rats in the laboratory is often radically different from what theywould experience in nature. The top picture is a 360 degree panorama of an outdoor setting where rats mightlive. The picture has been filtered as in Figure 5-3 to model that spatial frequencies that a wild rat can see.Note that despite its low acuity, when one considers the large visual field, there is a considerable amount ofvisual information available to the animal. The middle panel is a 360 degree panorama of a Morris watermaze testing room at the Canadian Centre for Behavioural Neuroscience, filtered as above and taken fromthe center of the pool. Although the large, high-contrast features of the cues that have been placed on thewalls are visible to the rat, many fine details are not. For example, there is a clock on the wall in the middleof the panel that is visible as an oval, but the interior details are blurred. The degree of blurring of the dockfor different strains of rats (Albino, wild/pigmented, Fisher-Norway) and humans (30 c/d; 20/20 Snellen acu-ity) is presented in the bottom row of pictures. The degree of visual detail available to any rat is much lessthan what a human experimenter could see, but there is also significant variability between rat strains. Alsonote the reduced contrast of the heavy black rim for the albino rat, which will make even large objects likethe clock less useful as navigation cues for these animals.

Page 71: 0195162854

54 SENSORY SYSTEMS

in this task are higher than those of humans,possibly reflecting the much greater corticalprocessing possible in the human brain, but itis still remarkable that rats can even do thetask and detect motion coherence when asfew as 25% of the dots in a display move inone direction. Besides having this "global"motion perception system, rats also have amore localized system for detecting the mo-tion of individual dots. The optimal dot dis-placement of 2° for the rat is close to that of1.0° for humans (Braddick, 1980).

In summary, the rat visual system is spe-cialized for low-light and low-resolution vi-sion. However, it has a typical mammaliancontrast sensitivity function, dichromaticcolor vision, and a visual cortical system ca-pable of precise spatial and temporal analysis.In addition, various visual thresholds can bepsychophysically quantified in the visual wa-ter task as efficiently, or more efficiently, thanthose thresholds can be generated in primates.

IMPLICATIONS OF RAT VISIONFOR INVESTIGATORS

The laboratory rat has a reputation amongmany vision scientists as an animal with poorvision. As a consequence, the rat has not beenwidely adopted as an animal model for inves-tigating the structure and function of themammalian visual system. In some ways, thisreputation is deserved; the acuity and contrastsensitivity of the laboratory rat are lower thanthat of other popular models, such as cats andferrets, and the rat does not have trichromaticcolor vision, a fovea, or the elaborate func-tional divisions of the lateral geniculate nu-cleus, striate, and extrastriate cortex that domany primates, including humans. Thesefacts alone, however, do not fully explain whyrats as models of mammalian vision have notbeen widely embraced, because the rat modelhas many other merits that would be of greatvalue for studying the nature of mammalianvision. Instead, the lack of simple behavioral

techniques to psychophysically quantify rat vi-sion has probably curtailed studies of visionmore than anything else. The recent devel-opment of the visual water task to readilymeasure rat spatial vision, motion sensitivity,orientation sensitivity, and so on will un-doubtedly increase the popularity of rats asfundamental models of mammalian vision.

In contrast to the relative minor use ofrats in specific vision-based studies, rats arecommonly used as subjects in visuobehavioraltasks to investigate the neural substrates ofmammalian cognitive function. Althoughmost would agree that reduced visual com-petence in a rat should negatively influencethe measurement of its cognitive function inthese tasks, little attention has been paid tothis issue. Again, this is likely due to the his-torical lack of simple and fast behavioral tasksto measure vision in control experiments. In-breeding and transgenic manipulations in ratshave created many strains with desirable traitsfor experimental brain research. Among theseare strains with mutations that affect the vi-sual system specifically and, as such, providevaluable models of visual system diseases oroffer experimental advantages for studyingthe structure and function of the visual sys-tem. Some mutations, such as albinism, how-ever, can negatively affect the visual system,whereas unique genetic combinations in somestrains may augment normal visual function.

These and other characteristics of the lab-oratory rat visual system have major implica-tions for experimental studies of the mecha-nisms of mammalian vision and for using ratsas models for experimental brain research ingeneral. Here we provide three examples ofsuch implications from our work.

EFFECTS ON VISION OF RATDOMESTICATION AND INBREEDING

To assess the effects of rat domestication andinbreeding on visual function, we measuredthe visual acuity of six different strains (inbred,outbred, and albino) of laboratory rats and

Page 72: 0195162854

Chapter 5. Vision 55

compared their acuity with that of wild rats.We found no significant effect of inbreedingor outbreeding on acuity; however, we didfind that all albino strains had acuity measure-ments about half those of wild rats.

Albinism produces a number of struc-tural abnormalities in the visual system, in-cluding neuroretinal abnormalities resultingfrom nonpigmented retinal pigment epithe-lium (Jeffery, 1998), abnormal decussation ofretinal ganglion cell axons at the optic chiasm(Lund et al., 1974), and abnormal interhemi-spheric connections of the visual cortex (Abeland Olavarria, 1996). The most likely expla-nations for reduced visual acuity in albino ratsare that excessive light scattering within theretina (Abadi et al., 1990) make the albino eyea rather poor image-forming device and thatlight-induced retinal degeneration (Birch,1977) results in poor spatial sampling. It is alsopossible that deficient central visual process-ing plays a role in the acuity deficits of albinorats. The absence of melanin or the melanin-related agent (Rice et al., 1999) responsible foranomalous axonal decussation at the opticchiasm in albinos (Jeffrey, 1997) may also pro-duce errors of interhemispheric connectivityof the visual cortex (Abel and Olavarria, 1996)and result in anomalous visual cortical pro-cessing. Given the numerous visual deficits inalbino rat strains, it is almost certainly wise toavoid using them in behavioral experimentsthat depend on vision.

Surprisingly, our study also found thatone pigmented strain, Fisher Norway, had asignificantly higher acuity than that of wildrats, as well as other pigmented and albinostrains tested: their grating threshold was ap-proximately 50% higher than that of other pig-mented strains and 150% higher than that ofthe albino strains. The most likely explanationfor this finding is that a genetic differenceaccounts for their enhanced acuity. Fisher-Norway rats are an F! cross between an in-bred Fisher 344 female and an inbred BrownNorway male. Although we did not measurethe acuity of the Brown Norway strain in this

study, we did measure the acuity of Fisher 344animals and found their acuity to be about 0.5c/d. It is possible that the Brown Norwaystrain possesses acuity higher than other pig-mented strains in our study, and that theFisher-Norway animals owe their high acuityto genes present in the Brown Norwaygenome. The superior visual acuity of theFisher-Norway strain also raises the possibil-ity that the acuity of 1.0 c/d we measured forwild rats is lower than that of native Rattusnorvegicus or that there is substantial hetero-geneity in the visual acuity of wild rats.

It is also possible that Fisher 344 andBrown Norway strains carry alleles that aredeleterious for high-resolution vision withinthe strains, but the unique combination ofgenes in Fisher 344/Brown Norway het-erozygotes results in alleles that are beneficialto visual acuity. This may produce animalswith enlarged eyes, smaller receptive fields inthe visual cortex, or other structural changesthat could lead to higher-resolution vision. Asa consequence, the Fisher-Norway strain maybe an attractive model for studying the mech-anisms of visual perception in nocturnalrodents.

Figure 5-3 graphically illustrates how theperception of a normal pigmented rat variesas a function of distance. The figure can alsobe used to compare the relative visual per-ception of rat strains with different acuities.

A recent study has shed some light on theimplications of variation in rat visual acuityfor performance in the Morris water task(Marker and Whishaw, 2002). When com-bined with the results of our study, the dataindicate that visual acuity alone does not pre-dict performance on place or matching-to-place versions of the Morris water task(Marker and Whishaw, 2002). For example,despite having superior (Fisher-Norway) orequal visual acuity (Dark Agouti) relative tothe Long-Evans strain, both of these strainswere relatively impaired. Even among albinostrains whose visual acuity did not differ inour study, Marker and Whishaw (2002) re-

Page 73: 0195162854

56 SENSORY SYSTEMS

ported that there are significant strain differ-ences in performance. These data do not ruleout the possibility that visual acuity can in-fluence rat behavioral performance in vision-dependent tasks, because all of the albinostrains with poor visual acuity (Sprague-Dawley, Wistar, Fisher 344) in Harker andWhishaw's (2002) study were impaired rela-tive to the Long-Evans (pigmented) strain.Therefore, it is likely that rat strains vary in anumber of brain functions, including visualfunction, and this variation can contribute todifferences in performance on complex be-havioral tasks.

SPATIAL LEARNING

The initial experiments that used the Morriswater task (Morris et al., 1982) to measure placelearning and memory anticipated that inter-pretations of the data could be confounded byabnormal visual, motor, or motivational func-tion. One configuration of the task that was de-veloped to control for these noncognitive fac-tors, including reduced visual function, was thecued-platform task. The rationale of this task wasthat if animals could not learn to swim directlyto a visible platform in the pool, they probablyhad noncognitive, possibly visual, impairmentsand should be excluded from further study.Conversely, it was reasoned that animals thatcould swim directly to the cued platform wererelatively free of these deficits, and therefore,their performance in a place configuration ofthe water maze most likely reflected their cog-nitive ability to learn and remember the plat-form location in space.

Recently, however, we performed a studythat challenges the ability of the cued-platformtask to detect animals with visual deficits suffi-cient to affect place learning (Prusky et al.,2000c). First, we binocularly deprived rats dur-ing the critical period for visual plasticity, whichreduced their adult visual acuity by about 30%.We then measured the place learning of theseanimals in the Morris water task and found thatthey had a significant impairment in learning

the task. These same animals, however, werenot impaired in their ability to locate a cued plat-form. These data indicate that place learningdeficits should be expected in animals with 30%to 100% reductions in visual acuity (blind ani-mals are known to be impaired in place learn-ing). However, screening animals with 30% re-duced acuity is not possible using a typical cuedplatform version of the Morris water task.

The major limitation of the cued plat-form task for identifying animals with visualdeficits is that the task does not accuratelymeasure the visual function required for ac-curate place learning. For example, any imagecan be analyzed as if it is formed by a set ofsine waves of different spatial frequencies, andthe receptive field organization of mammalianretinal and cortical cells appears to measurethe magnitude and location of the differentspatial frequency components. The ability toidentify cues and locate a place in visual spaceis limited by sensitivity to the highest spatialfrequencies. In a typical cued platform task,the single platform cue appears dark against awhite pool wall. The single object can be de-composed into many frequencies, any ofwhich can be used to locate the platform. Thatis, in the absence of distractors or multiple vi-sual cues that must be discriminated, detec-tion of a single large, high-contrast cue maybe possible with vision consisting of only thelowest spatial frequency detectors. Moreover,performance in the cued platform task is notan accurate measure of visual acuity becausethe cue is usually too large and there is littlecontrol of the viewing distance. Even from thefarthest viewing position in a 1.5 m pool, a 10cm cue will subtend about 5° of visual angle,corresponding to a grating acuity of 0.1 c/d,an order of magnitude lower than the 1.0 c/dthat normal rats can see.

Considering the relatively low visual acu-ity of normal rats («1.0 c/d), it is also possi-ble that apparent cognitive impairments in theMorris water task could be the result of in-sufficient visual cueing in the test room. Theoverall size, contrast, illumination, and stabil-

Page 74: 0195162854

Chapter 5. Vision

ity of the cues should be considered as well astheir relationship to the animal as it swimsabout the pool; stable distal cues should sub-tend more than 1 ° of visual angle after con-sidering the viewing positions from within thepool. As a guide to experimenters in selectingvisual cues, Figure 5-3 shows what a normalrat can see of a well-known object, humanheads, at a variety of distances. Figure 5-4models what rats see in a typical laboratorysetting for the Morris water maze.

The visual competence of rats is perti-nent to the interpretation of data from watermaze studies or from other visual based be-havioral tasks where visual function may becompromised. For example, a study by Lind-ner et al. (1997) reported that blind rats per-formed better than atropine-treated animalsin place learning, but the groups could not bedifferentiated in a cued platform task. Otherstudies have correlated water maze perfor-mance with the loss of photoreceptors in agingSprague-Dawley rats (e.g., Osteen et al., 1995).It is possible that visual deficits contributed todeficient place learning in these studies, how-ever, the only way this uncertainty can beresolved is if the visual function of atropine-treated animals or animals with retinal degen-eration is measured independently.

VISUAL PLASTICITY

A critical period for visual plasticity early inlife in which the nature of visual experienceshapes the structure and function of the de-veloping visual system was first delineatedby Wiesel and Hubel (1970). Although thisdemonstration in cats, and later in monkeys,provided a model for how abnormal visualexperience in developing humans can lead toamblyopia, the lack of simple psychophysicalmeasures of cat and monkey vision meant thatmost studies of visual cortical plasticity thatfollowed used electrophysiological measuresof visual function, such as cortical ocular dom-inance. This is surprising because the best clin-ical measure of human amblyopia is visual

57

perception, and the relationship between oc-ular dominance and vision is correlational, notcausal (Murphy and Mitchell, 1987). Cats andmonkeys also have a number of limitationsthat are absent in the rat. For example, in therat there is easy access to the whole of visualcortex, the cortex is flat, there are large stereo-taxically defined monocular and binocular re-gions of primary visual cortex, and an abun-dance of biochemical and molecular tools areavailable to investigate cellular function in therat. In addition, the rat shows a developmen-tal ocular dominance plasticity that is funda-mentally the same as that in other mammals(Fifkova, 1968; Fagiolini et al., 1994). We usedthe visual water task to demonstrate that im-poverishment of the visual environment dur-ing the rat's critical period leads to permanentamblyopia (Prusky et al., 2000b). Ocular dom-inance experiments in cats and primates havevirtually taken for granted that interocularcompetition is the mechanism of develop-mental visual plasticity. Our behavioral re-sults showing large deficits from binoculardeprivation suggest that this simple story isincomplete and that there are additionalprocesses that contribute to the developmentof amblyopia. The rat may also be one of thebest species in which to study these factors: ithas the advantages listed earlier; in addition,there are many transgenic and inbred strainsavailable, rats are easily housed and handled,rats have precisely tuned visual cortical re-ceptive fields (Girman et al., 1999), and wenow have the ability to quickly measure its vi-sual function. In fact, the rat may be the bestchoice to intergrate the results of many di-verse methodologies to study the mechanismsof how visual experience is translated into ma-ture visual function.

CONCLUSIONS

The development of the visual water task hasenabled researchers to rapidly and accuratelyquantify the vision of rats. Rats have a typically

Page 75: 0195162854

58 SENSORY SYSTEMS

mammalian visual system specialized for low-resolution vision, and many of the visual capa-bilities of rats are surprisingly good. Therefore,the rat visual system is a very good model forstudying the mechanisms of mammalian vi-sion. In addition, it is important to consider theunique visual capabilities of rats and strain dif-ferences in rat vision when designing behav-ioral experiments that depend on vision.

REFERENCES

Abadi R, Dickinson CM, Pascal E, Papas E (1990) Reti-nal image quality in albinos. A review. OphthalmicPaediatrics and Genetics 11:171-176.

Abel PL and Olavarria JF (1996) The callosal pattern instriate cortex is more patchy in monocularly enu-cleated albino than pigmented rats. NeuroscienceLetters 204:169-172.

Bowden WF, Douglas RM, Prusky GT (2002) Horizon-tal bias in rat visual acuity. Program No. 260.18. Ab-stract Viewer/Itinerary Planner. Washington, DC:Society for Neuroscience. Online.

Braddick OJ (1980) Low-level and high-level processesin apparent motion. Philosophical Transactions ofthe Royal Society of London. Series B Biological Sci-ences 290:137-151.

Birch D and Jacobs GH (1977) Effects of constant illu-mination on vision in the albino rat. Physiology andBehavior 19:255-259.

Birch D and Jacobs GH (1979) Spatial contrast sensitiv-ity in albino and pigmented rats. Vision Research19:933-937.

Dean P (1978) Visual acuity in hooded rats: effects of su-perior collicular or posterior neocortical lesions.Brain Research 156:17-31.

Fagiolini M, Pizzorusso T, Berardi N, Domenici L, Maf-fei L (1994) Functional postnatal development of therat primary visual cortex and the role of visual ex-perience: dark rearing and monocular deprivation.Vision Research 34:709-720.

Fifkova E (1968) Changes in the visual cortex of rats af-ter unilateral deprivation. Nature 220:379-381.

Girman SV, Sauve Y, Lund RD (1999) Receptive fieldproperties of single neurons in rat. primary visualcortex. Journal of Neurophysiology 82:301-311.

Harker KT and Whishaw IQ (2002) Impaired spatial per-formance in rats with retrosplenial lesions: impor-tance of the spatial problem and the rat strain inidentifying lesion effects in a swimming pool. Jour-nal of Neuroscience 22:1155-1164.

Hughes A (1977) The refractive state of the rat eye. Vi-sion Research 17:927-939.

Jacobs GH, FenwickJA, Williams GA (2001) Cone-basedvision of rats for ultraviolet and visible lights. Jour-nal of Experimental Biology 204(Pt 14):2439-2446.

Jeffery G (1997) The albino retina: an abnormality thatprovides insight into normal retinal development.Trends in Neurosciences 20:165-169.

Jeffery G (1998) The retinal pigment epithelium as a de-velopmental regulator of the neural retina. Eye12:499-503.

Keller J, Strasburger H, Cerutti DT, Sabel BA (2000) As-sessing spatial vision-automated measurement ofthe contrast-sensitivity function in the hooded rat.Journal of Neuroscience Methods 97:103-110.

Lashley KS (1930) The mechanism of vision: I. A methodfor rapid analysis of pattern vision in the rat. Jour-nal of General Psychology 37:453-460.

LaVail MM (1976) Survival of some photoreceptors inalbino rats following long-term exposure to contin-uous light. Investigative Ophthalmology and VisualScience 15:64-70.

Lindner MD, Plone MA, Schallert T, Emerich DF (1997)Blind rats are not profoundly impaired in the refer-ence memory Morris water maze and cannot beclearly discriminated from rats with cognitivedeficits in the cued platform task. Cognitive BrainResearch 5:329-333.

Lund RD, LundJS, Wise RP (1974) The organization ofthe retinal projections to the dorsal lateral genicu-late nucleus in pigmented and albino rats. Journalof Comparative Neurology 58:383^03.

Morris RG, Garrud P, Rawlins JN, O'Keefe J (1982) Placenavigation impaired in rats with hippocampal le-sions. Nature 297:681-683.

Murphy KM and Mitchell DE (1987) Reduced visual acu-ity in both eyes of monocularly deprived kittens fol-lowing a short or long period of reverse occlusion.Journal of Neurosciences 7:1526-1536.

Neve AR, Prusky GT, Douglas RM (2002) Perception ofmotion coherence in rats. Program No. 353.17. Ab-stract Viewer/Itinerary Planner. Washington, DC:Society for Neuroscience, 2002. Online.

O'Steen WK, Spencer RL, Bare DJ, McEwen BS (1995)Analysis of severe photoreceptor loss and Morriswater maze performance in aged rats. BehavioralBrain Research 68:151-158.

Prusky GT, West PWR, Douglas RM (2000a) Behavioralassessment of visual acuity in mice and rats. VisionResearch 40:2201-2209.

Prusky GT, West PWR, Douglas RM (2000b) Experi-ence-dependent plasticity of visual acuity in rats.European Journal of Neuroscience 12:3781-3786.

Prusky GT, West PWR, Douglas RM (2000c) Reducedvisual acuity impairs place but not cued learning in

Page 76: 0195162854

Chapter 5. Vision 59

the Morris water task. Behavioral Brain Research116:135-140.

Rice DS, Goldowitz D, Williams RW, Hamre K, JohnsoPT, Tan SS, Reese BE (1999) Extrinsic modulationof retinal ganglion cell projections: analysis of thealbino mutation in pigmented mosaic mice. Devel-opmental Biology 21:41-56.

Seymoure P and Juraska JM (1997) Vernier and gratingacuity in adult hooded rats: the influence of sex. Be-havioral Neuroscience 111:792-800.

Szel A and Rohlich P (1992) Two cone types of rat retinadetected by antivisual pigment antibodies. Experi-mental Eye Research 55:47-52.

Whishaw IQ and Tomie JA (1989) Olfaction directsskilled forelimb reaching in the rat. Behavioral BrainResearch 32:11-21.

Wiesel TN and Hubel DH (1970) The period of sus-ceptibility to the physiological effects of unilateraleye closure in kittens. Journal of Physiology (Lon-don) 206:419-436.

Page 77: 0195162854

Somatosensation

LINDA HERMER-VAZQUEZ,RAYMOND HERMER-VAZQUEZ,AND JOHN K. CHAPIN

6

In many ways, the rat somatosensory system,at both the sensory periphery and throughoutthe central nervous system (CNS), is homol-ogous to that found in other mammals, in-cluding primates. Conserved elements of thissystem include multiple types of cutaneous,proprioceptive, nociceptive, and thermal re-ceptors; somatosensory afferents from thespinal cord including the dorsal column path-ways, the spinothalamic and spinoreticulartracts, and the spinocervical tract; and ascen-sion through the dorsal column nuclei, mul-tiple thalamic nuclei including the ventralposterior lateral nucleus (VPL) (for the non-vibrissal body surface) and the ventral poste-rior medial nucleus (VPM) (for vibrissal in-formation) and at least two neocorticalsomatosensory areas (SI and SII; Krubitzer,1995; Paxinos, 1995). Likewise, systems forsensorimotor integration and learning, such asthe basal ganglia and cerebellum, have beenconserved across rats and other mammals, ashave the main motor output pathways (thecorticospinal, rubrospinal, vestibulospinal,and reticulospinal tracts). Additionally, thereis extensive homology across rats and othermammals in terms of cytoarchitecture, cyto-chemistry, and the roles played by differentneurotransmitters and neuromodulators (Pax-inos, 1995; Aboitiz, 2001). These facts under-score the general point that in many ways ratsare a good mammalian exemplar.

Nevertheless, rats are adapted to a dark,cluttered environment of underground bur-rows and densely wooded terrain, usuallyclose to water, making the sensory world ofthe rat quite different from that of most pri-mates, especially humans. The rat has its ownfoveal somatosensory system: its whiskers,perioral areas, and forepaws, as well as itskeen olfactory system. As the rat movesthrough a new environment, it constantlywhisks and sniff objects directly in front of itand uses the somatosensory receptors on thetip of its nose, in many cases, to determinewhere to next place its forepaw (L. Hermer-Vazquez and R. Hermer-Vazquez, unpub-lished data). Then it gingerly places itsforepaw on the object's surface where thenose had just been, as it continues to "map"the new object using somatosensory and ol-factory information. Thus, the rat relies on in-formation that is integrated from snout andforepaw sensory receptors.

Correspondingly, the somatosensorycortical areas devoted to representation of thenose, whiskers, mouth, forepaw skin, andlong, whisker-like hairs on the underside ofthe rat's wrists ("sinus hairs," as explainedlater) are greatly enlarged relative to the sen-sory representations of other body parts, andthe receptive fields for these foveal body re-gions are much more sharply defined (Fig.6-la andb) (Chapin and Lin, 1984). Moreover,

60

Page 78: 0195162854

Chapter 6. Somatosensation 61

Figure 6-1. (A) Map of the cutaneous representation of the rat's body in Si cortex. Note the specific cor-tical regions devoted to cutaneous maps of the forepaw (fp) and digits (d2-d5) and thumb (t), wrist sinushair (w), and mystacial vibrissaw (A-E, 1-8), which together comprise the rat's somatosensory fovea. T,trunk; hi, hindlimb; HP, hindpaw; dhp, dorsal hindpaw; dl-d5, hindpaw digits from 1 to 5; hm, hindlimbmuscle; vfl, ventral forelimb; dfl, dorsal forelimb; w, wrist sinus hairs; dfp, dorsal forepaw; d2-d5, forepawdigits from 2 to 5; t, thumb; uz, zone that was unresponsive during mapping; A-E, 1-8, rows (from dorsalto ventral) and numbers (from caudal to rostral) of facial whiskers; RV, rostral small vibrissae, N, nose;FBP, frontobuccal pads; UL, upper lip; LL, lower lip; LJ, lower jaw. (B) Receptive field centers for isolatedsingle units in SI for the palmar surface versus the dorsal hand. Note the finer representation of the skinon the palm in SI.

the rat has two primary cortical representa-tions of the forepaw, and one of them, nowreferred to as "caudal Ml," contains overlap-ping somatosensory inputs to layer 4 and mo-tor outputs descending from layer 5. Thereare also separate cortical representations forthe rodent-specific somatosensory peripheralorgans referred to earlier. The mystacial vi-brissae are represented as separate granularaggregates (in the cortex, "barrel fields") incaudolateral SI. Cells in this region are ex-tremely responsive to whisker manipulationbut much less so to the manipulation of theunderlying skin and fur adjacent to thewhiskers. In contrast, the more rostral corti-cal representation of the perioral surface haslarge receptive fields extending into the pe-ripheral cortical whisker zone, which aremuch more responsive to stimulation of theskin and fur. Finally, the rat has another dis-tinctive somatosensory apparatus: the sinushairs on its ventral wrists. These long, tinyhairs are represented cortically by their own

granular aggregates, similar to the granularaggregates that comprise the barrel fields forthe whiskers. Also as with the whiskers, theskin and fur adjacent to the wrist sinus hairsare represented in a more classic manner by thelarger caudal and rostral forepaw-forelimb re-ceptive fields (Chapin and Lin, 1984).

In this chapter, we describe five princi-ples for how sensorimotor behaviors arelearned and performed by rats, based onnew findings from neuroscience. We focusmainly on nonvibrissal somatosensory pro-cessing but discuss examples from thewhisker-tactile system or from other sensorymodalities when they illustrate a point well.Taking the necessary biological constraintsinto account, studies of rat somatosensorybehavior are pioneering for behavioral neu-roscience. Rats also make excellent subjectsfor experiments in that they eagerly performmovements repetitively and stereotypically,minimizing variation in how each trial isexecuted.

Page 79: 0195162854

62 SENSORY SYSTEMS

PRINCIPLE I: ANALYZEDSOMATOSENSORY FEEDBACK

INFORMATION IS CONSTANTLYINFLUENCING THE ASCENDING

SOMATOSENSORY DATASTREAM IN RATS

Both "bottom-up" stimulus feature-based in-formation and relatively more processed "top-down" information interact to produce re-sponses of cells in intermediate levels of thesomatosensory system such as the so-matosensory thalamus. This has been repeat-edly demonstrated in studies of the so-matosensory receptive fields of thalamic andcortical neurons, which have made clear thatboth ascending sensory information from theperiphery and descending corticofugal projec-tions determine somatosensory receptive fieldstructure. For example, Shin and Chapin(1990) found that motor cortical stimulationdecreased the response time in thalamic VPLneurons to mechanical stimulation of theforepaw. Extending that line of reasoning,Krupa et al. (1999) determined the short-latency and long-latency receptive fields ofthalamic VPM neurons, and then tested themagain under (I) muscimol inactivation of SIcortex or (2) muscimol SI inactivation com-bined with lidocaine inactivation of ascendingsomatosensory input. Under each type of re-versible chemical blockade, immediate recep-tive field reorganization occurred. Moreover,the findings suggested that GABAergic corti-cofugal feedback appeared to suppress short-latency responses caused mainly by ascendinginfluences in many thalamic neurons, whereascorticofugal glutamatergic excitatory influ-ences appeared to be necessary for many tha-lamic cells to exhibit long-latency responses.Results in rats using other sensorimotor pro-tocols as well support the general notion thatreceptive fields are created by multiple as-cending and descending influences and areheld under a dynamic tension that allows im-mediate reorganization after any change inthese inputs. Thus, top-down, cognitive in-

formation as well as bottom-up datastreams—which, crucially, are not always pre-dicted by the top-down processes—can betaken into account in the rat's analysis of itssomatosensory world.

PRINCIPLE II: RATS ARECONSTANTLY EVALUATING

INFORMATION ACROSS MULTIPLETIMESCALES TO MORE ACCURATELY

PREDICT WHAT WILL HAPPENIN THEIR WORLD

Consistent with the preceding data on de-scending influences on ascending sensory in-flow, it is widely agreed that rats are constantlyreevaluating the past at different timescales andcombining the resulting knowledge with theircurrent perceptions to predict the future at dif-ferent timescales (Llinas, 2001). That is, they usetheir intellect as well as their small size and noc-turnal nature to elude predators. The fact thatthese evaluative processes are occurring onmultiple timescales complicates one's interpre-tation of rat behavior.

Many older models of rat information pro-cessing are based on the simplifying assump-tion that tasks are learned and maintained as asimple function of the number of training tri-als up through reaching an asymptotic per-formance level and degraded as a simple func-tion of lack of practice or interference by newmemories (Baddeley, 1992). This view sug-gested that time flows forward linearly, in a reg-ular manner, throughout task acquisition andperformance; that is, the rat is using its experi-ence on the current trial to shape its perfor-mance on the next trial, until some maximumlevel is reached and maintained. For example,rats in a recent study of ours (Hermer-Vazquezet al., in press) were trained daily on the skilledreach-to-grasp-food task (Whishaw and Pellis,1990) described earlier. On day 1 of training,the sample of rats grasped the target correctlyon 27% of trials and improved daily and lin-early in their performance until day 6, when

Page 80: 0195162854

Chapter 6. Somatosensation 63

their success rate began to asymptote at ap-proximately 68% correct. Consistent with thisview of task learning, it was shown that the ratcaudal digit-wrist motor cortex, which containsan overlapping somatosensory representationof the forepaw (Chapin and Lin, 1984), expandsand grows new synapses during the learning ofthis task, whereas no similar changes were de-tected in the (more strictly motor) rostral Ml(Kleim et al., 2002). Moreover, the degree oflong-term potentiation in rat Ml synapses hasbeen found to correlate with reaching skill ac-quisition (Rioult-Pedotti et al., 2000).

Nevertheless, other data suggest that theinformation processing relevant to task learn-ing and performance flows in a more complexmanner, as rats attempt to better understandtheir past to better predict the future. Theseprocesses occur at multiple timescales through-out different levels of the nervous system. Forexample, at the somatosensory periphery,slowly adapting versus rapidly adapting cuta-neous receptors in the rat glabrous skin (Paxi-nos, 1995) allow simultaneous perception of dif-ferent aspects of the tactile world (Johnson,2001). At the same time, even at very low lev-els of the somatosensory-motor interface, mul-tiple feedback processes are occurring. For ex-ample, the spinal stretch reflex illustrates howproprioceptive feedback is constantly influenc-ing muscle tension. Modification of neural pro-cessing by "evaluative" feedback is seen atlonger timescales as well; for example, corti-cothalamic feedback projections modify sub-cortical sensory processing, as described in thesection on Principle I (e.g., Krupa et al., 1999).And at still longer time-scales, such as hours,considerable data now suggest that after ratsspend a day whisking, sniffing, and manipulat-ing objects in a novel environment, the ratsreevaluate their sensorimotor performance ontask from the prior day during slow-wave sleep,developing better internal models for thosetasks as evidenced by the fact that performanceoften improves after a complete sleep cycle withno additional overt practice (Lee and Wilson,2000; Poe et al., 2000). All of these processes,

occurring at distinctive timescales, help the ratuse its past to more accurately predict its future.

PRINCIPLE III: INFORMATIONFROM MULTIPLE SPATIAL SCALESIS PROCESSED SIMULTANEOUSLY

IN THE RAT

Three or four decades ago, many researchersadvocated the view that mammalian sensorysystems represent their sensory surfaces topo-graphically and with high resolution and thatthe high-resolution, body-centered informa-tion gained at the sensory surface was gradu-ally transformed at higher levels of the ner-vous system by cells with progressively largerreceptive fields into object-centered represen-tations. It is now widely recognized that in-formation is processed at both fine and broadspatial scales simultaneously. For instance, atthe somatosensory surface, type 1 cutaneousreceptors process mechanical somatosensorydata with high spatial resolution, and type IIreceptors have larger, less well-defined recep-tive fields (Vallbo and Johansson, 1984). Theprinciple of simultaneous processing of sen-sory data at multiple spatial scales holds athigher levels of the nervous system as well.

For instance, in many cases, different mi-crocircuits in the CNS—defined by their celltypes, neurochemistry, and connectivity—process information at their own, distinctivespatial scales. It is now known that in all thal-amic nuclei, including the sensory relay nuclei,a matrix of calbindin-immunoreactive cellsprojects diffusely throughout the cortex, re-gardless of sensory topography or even sensorydomain (Jones, 2001). In contrast, the preciselyprojecting, topographically ordered cells moreclassically associated with thalamic relay effer-ents stain positively for parvalbumin. Thus, forexample, the relay nucleus for body so-matosensory afferents, VPL, contains bothparvalbumin-positive, precisely projecting, so-matotopically organized afferents to layer 4 ofcortical area SI, with correspondingly small re-

Page 81: 0195162854

64 SENSORY SYSTEMS

ceptive fields, and much more widely project-ing, calbindin-staining cells whose axons targetthe superficial cortical layers in multiple sen-sory modalities. These diffusely projecting cellsplay a critical role in coordinating activityacross brain regions, particularly in view of ev-idence that many layer IV corticothalamic feed-back neurons target these transcortically pro-jecting matrix cells (Jones, 2002).

Likewise, at the primary somatosensorycortical level, the forepaw and perioral areas arerepresented by cells with small as well as cellswith large receptive fields (Chapin and Lin,1984), just as cells in the septa between preciselymapped cortical barrel fields have much largerreceptive fields (Brecht and Sakmann, 2002). Ratsomatosensory cortical and thalamic processingtherefore shares spatial-related features withothers subcortical entities such as the basal gan-glia, which contains topographically more or-dered regions that process information withhigh spatial resolution (striosomes), as wellmore diffusely organized region (matrix cells)(Brown et al., 2002), and the cerebellum, wheredistant regions of the body surface are adjacentlylocated in the cerebellar "mosaic" (Bower andWoolston, 1983). Thus, either rat behavioraltasks need to be extremely well controlled ifthey are to be based primarily on the hypothe-sis of high-resolution, topographic processing(e.g., by localizing perception and motion to asmall and isolated portion of the body) or theexperimenter's hypotheses need to account forprocessing occurring at multiple spatial scales.

PRINCIPLE IV: RAT SENSORYAND MOTOR PROCESSING

ARE CONSTANTLY INFLUENCINGONE ANOTHER

Both older and newer neuroanatomical andneurophysiological data indicate that all levelsof the neuraxis, including the spinal cord, "as-cending" sensory information, and "descend-ing" motor information, influence each other.For example, it has been known for many years

that at the level of spinal reflexes, propriocep-tive sensory information about the stretch stateof a muscle feeds back to regulate muscle ten-sion in rats and other mammals (Kandel et al.,2000). Newer data demonstrate that sensori-motor interplay occurs at higher levels of therat neural system as well. For instance, outputsfrom the rat's basal ganglia, once thought to bea motor-learning structure, project to layer 1 ofits somatosensory cortex and other primaryneocortical areas (McFarland and Haber, 2002),likely influencing the processing of all corticalcells with dendrites extending into layer 1. Datasuch as these indicate that sensory and motorareas are continuously interacting via multiple,parallel looping structures. Likewise, all "sen-sory" thalamic "relay" nuclei, including the VPL(processing somatosensory information), themedial geniculate nucleus (MGN) (processingauditory information), and the lateral geniculatenucleus (LGN) (processing visual data), receiveaxon collaterals of layer 5 efferents from Ml andthe premotor cortex (Guillery and Sherman,2002). A fact that is not widely appreciated isthat these axon collaterals constitute a largerportion of inputs to the thalamic relay nucleithan do the ascending sensory fibers! Thus, de-scending motor commands appear able tomodulate or even drive thalamic sensory pro-cessing. These facts are among many morepieces of anatomical and physiological data sug-gesting that sensory and motor processing areconstantly modulating each other.

The constant interaction of sensory andmotor datastreams is well illustrated by rat ol-factory as well as somatosensory behaviors.For example, the strength of an odor modu-lates sniffing intensity, even on the first sniffof the odor (Johnson et al., 2003). Further-more, this modulation of sniffing intensity oc-curs so rapidly that it is thought that corticalprocessing cannot be involved; the modula-tion must take place at brain stem or spinallevels (Johnson et al., 2003). In our studies ofrats performing an olfactory-driven, reach-to-grasp food task (Whishaw and Pellis, 1990),the presence or absence of a food-related odor

Page 82: 0195162854

Chapter 6. Somatosensation 65

mRN cell, FINAL SNIFF

Figure 6-2. Perievent histograms for representative single cells in the rat caudal primary motor cortex (Ml)and magnocellular red nucleus (mRN), centered around the final sniff of the food pellet before lifting the pawto initiate reaching. In each graph, the center horizontal line depicts the cell's firing rate, and the lines aboveand below it show 2 SDs from the mean (i.e., statistical significance of rate modulation). It can be seen thatat the final sniff moment, each cell's firing rate is significantly depressed. The binwidth for these perieventhistograms is 25 milliseconds. The percentages below each graph show the proportion of single units recordedin each area that displayed such significant modulation on final sniff.

determines whether the rat will lift its pawand guides the spatial accuracy of the reach(Hermer-Vazquez et al., unpublished data).Indeed, rats appear to sniff the target just be-fore lifting the paw to gain the initial spatialcoordinates for the impending reach and thentake several more "update sniffs" during theovert arm-movement phases of the task.These behavioral observations have a corre-sponding neurophysiology. For instance,while recording single units from the digit-wrist area of caudal Ml and the magnocellu-lar red nucleus as rats perform this task, wehave found that many neurons in both areasthat code the overt arm-movement phases arestrongly modulated by olfactory information

taken in during the final sniff of the foodtarget just before lifting the paw (Hermer-Vazquez et al., in press) (Fig. 6-2).

Somatosensory processing also appears toconstantly guide the rats' reaching maneuvers.We have found that many rat Ml units are par-ticularly active during phases of the reachingtask in which somatosensory information is be-ing evaluated, such as lifting the paw off theground, brushing the paw against the shelf onwhich the food target rests on the way to thetarget, and contacting the food pellet itself(Hermer-Vazquez et al., in press) (Fig. 6-3). Thecells whose firing rates increase as the rat's pawhits the shelf are likely responding to move-ment of the sinus hairs of the ventral wrist,

Ml cell, LIFTsig011a

Time (sec)

19% (SEM 8.4%)

Ml cell, ONSHELFsig012c

31% (SEM 13.4%)

Ml cell, CONTACT

sigOOSa

Time (sec)

26% (SEM 8.4%)

Figure 6-3. Perievent histograms for single cells recorded from the rat caudal Ml centered around each ofthree reaching-task events in which the cutaneous inputs to the paw change and are presumably evaluatedby Ml cells: lifting of the paw off the ground (left), the paw making contact with the shelf (middle), and thepaw making contact with the food pellet (right). Other elements of the graphs, such as the binwidth, are thesame as in Figure 6-2.

Ml cell, FINAL SNIFF

Page 83: 0195162854

66 SENSORY SYSTEMS

Ml cell, NOSE TAPsig002b

mRN cell, NOSE TAPsig020b

Figure 6-4. Perievent histogramsfor single cells recorded from therat caudal Ml and mRN, with upper-body motor fields, that also re-spond strongly to cutaneous inflowfrom the perioral region. Other el-ements of the graphs, such as thebinwidth, are the same as for Fig-ures 6-2 and 6-3.

Time (sec)

and the large percentage of cortical "on-shelf'cells suggests that this phase of the task is amajor calibration point for the final approachto the target.

The role of somatosensory processing inmotor cells' activity is especially clear once thetask is well learned. When reaching tasks arefirst being learned, the latencies of spike rateincreases in Si and Ml are relatively consis-tent with the "data in, cognitive transforma-tions performed, data out" task model, in thatSi units peak in their firing several tens of mil-liseconds before the Ml cells' peaking. In onestudy, the inadequacy of that model was strik-ingly illustrated. As animals became more pro-ficient at the task, roughly one third of allrecorded SI units developed much longer la-tencies, consistent with "motor" processing,and roughly one-third of all recorded Ml cellsdeveloped early, "somatosensory" latencies (J.Chapin, unpublished data). Consistent withthis fact, we have also found that many neu-rons in both the rat's caudal Ml and its mag-nocellular red nucleus respond as strongly, ifnot more strongly, to sensory input from thebody regions corresponding to their motorfields. For example, Figure 6-4 shows repre-sentative cells from Ml and the magnocellu-lar red nucleus (mRN) with an upper-bodymotor field responding to mechanical taps onthe ratis nose. The sharp increase in firing rateshown by the perievent histograms at the mo-ment of the tap, as with the other data pre-sented earlier, supports the view that task-related sensory and motor information dy-

namically and continuously interact even instructures such as the rat's red nucleus thatwere classically thought to be "motor."

PRINCIPLE V: RAT BEHAVIORSAPPEAR TO BE ORGANIZED INTOSURVIVAL-RELATED REPERTOIRES

THAT CAN BE ADAPTED TONOVEL CIRCUMSTANCES

Increasing evidence suggests that cortical,subcortical, and spinal motor circuits are or-ganized according to synergistic activity ofgroups of muscles, or perhaps even accordingto the whole, complex movements they pro-duce, rather than being organized somato-topically or musculotopically (Graziano et al,2002). As a minimum, it is clear that individ-ual cells can code for synergies of musclesrather than the movements of single musclesor joints (d'Avella et al., 2003). For example,we have found that during skilled reaching,red nucleus cells appear to code for combinedlimb movements and postural shifts (Hermer-Vazquez et al., in press). Moreover, new find-ings with primates as well as rats suggest thatwhole, survival-related movements, such asmoving the hand toward the mouth and open-ing the jaw, defensively blocking objects fromhitting the face, or manipulating objects in afrontal "workspace," are coded by motor-re-lated circuits from the spinal level (Strick,2002) to higher motor cortical regions(Graziano et al., 2002)

Time (sec)

Page 84: 0195162854

Chapter 6. Somatosensation 67

Rat behavioral research has produced re-sults consistent with this view of how move-ment control is organized in the brain. Forinstance, using the reach-to-grasp food taskdescribed earlier, Metz and Whishaw (1996)found that rats did not adjust their grip sizefor food pellets of varying diameters. On thebasis of these findings, they argued that man-ual movements in rats are organized in termsof stereotypic movements. Our observationsof how rats learn the standard reach-to-grasptask are also consistent with this view. For in-stance, on the first few training trials rats of-ten reach for the food pellet, and then, evenif they succeed in grasping it, they do not re-tract the pellet toward their mouth, open theirmouth, and place the pellet inside. Rather, af-ter barely starting to retract their paw, theydrop the pellet and let their arm go limp tem-porarily. Then, after several more trials theygradually pull the pellet closer and closer totheir mouth. Similarly, on many initial trialsrats fail to contact and grasp the target, anddo not even extend their paw forward a suffi-cient distance, but still they open their mouthand retract their pelletless hand (L. Hermer-Vazquez and R. Hermer-Vazquez, unpub-lished data). Thus, a crucial part of learning toperform the whole maneuver appears to belearning to conjoin two stereotypic move-ments: extending the paw outward and grasp-ing the object, and then pulling it toward andinside the mouth. We have evidence that inthe rat Ml, the neural control for skilled reach-ing and for the "reachlike" phase of locomo-tion, in which the forepaw is lifted off theground, projected forward, and placed backdown, is similar: In both cases, Ml cells pref-erentially encode the lift and paw-downphases of the movements (Hermer-Vazquezet al., in press). These findings suggest that theinterjoint timing required for one type ofmovement can be gradually shifted over thecourse of learning so that it produces a differ-ent, although related, movement.

The view that rats learn sophisticated mo-tor behaviors by shaping a preestablished reper-

toire to new circumstances has dramatic im-plications for one's task analysis. When the ratis learning a new and difficult motor task, forinstance, instead of learning a completely noveland lengthy series of joint torques and angles,the animal may start with a subset of its hard-wired movements and then combine and sub-tly adapt them to the current circumstances.

FINAL COMMENTS

Using recent findings from neuroscience, wehave begun to sketch a new view of the psy-chological information processing that occursas rats leam and execute a new somatosensory-motor task. This new view includes the factsthat the rat's mind is continuously processingaspects of the task at multiple temporal andspatial scales, that sensory and motor pro-cessing are highly fused, and that whole,stereotypic movements appear to be repre-sented in the rat's brain and mind. We believethat if researchers in the areas of rat percep-tion, cognition, and behavior base their taskdesigns and task analyses on these principles,it will facilitate developing a more accuratepsychological understanding of the rat's per-formance. For example, if a researcher hy-pothesizes that rats will learn a new task byadapting a set of partly known preexisting mo-tor behaviors, it will greatly simplify his or herunderstanding of how to kinematically codethe evolving motor sequence. However, re-searchers still must validate their task analy-ses by testing their hypotheses about how theanimals perform their tasks, because rats dohave distinctive somatosensory and otheradaptations whose deployment may not al-ways be obvious to the experimenter.

REFERENCES

Aboitiz F (2001) The origin of isocortical development.Trends in Neurosciences 24:202-203.

Baddeley A (1992) Human memory: theory and prac-tice. Boston: Allyn and Bacon.

Page 85: 0195162854

68 SENSORY SYSTEMS

Bower JM and Woolston DC (1983) Congruence of spa-tial organization of tactile projections to granule celland Purkinje cell layers of cerebellar hemispheres ofthe albino rat: vertical organization of cerebellarcortex. Journal of Neurophysiology 49:745-766.

Brecht M and Sakmann B (2002) Dynamic representa-tion of whisker deflection by synaptic potentials inspiny stellate and pyramidal cells in the barrels andsepta of layer 4 rat somatosensory cortex. Journalof Physiology 543(pt 1):49-70.

Brown LL, Feldman SM, Smith DM, Cavanaugh JR,Ackermann RF, and Graybiel AM (2002) Differen-tial metabolic activity in the striosome and matrixcompartments of the rat striatum during natural be-haviors. Journal of Neuroscience 22:305-314.

Chapin JK and Lin C-S (1984) Mapping the body repre-sentation in the SI cortex of anesthetized and awakerats. Journal of Comparative Neurology 229:199-213.

D'Avella A, Saltiel P, and Bizzi E (2003) Combinationsof muscle synergies in the construction of a naturalmotor behavior. Nature Neuroscience 6:300-308.

Graziano MS, Taylor CS, Moore T, and Cooke DF(2002) The cortical control of movement revisited.Neuron 36:349-362.

Guillery RW and Sherman SM (2002) Thalamic relayfunctions and their role in corticocortical commu-nication: generalizations from the visual system.Neuron 33:163-175.

Hermer-Vazquez L, Hermer-Vazquez R, Moxon KA,Kuo K-S, Viau V, Zhan Y, and Chapin JK (2003) Dis-tinct temporal activity patterns in the rat Ml andmagnocellular red nucleus during skilled versus un-skilled movement. Behavioural Brain Research, inpress.

Hermer-Vazquez L, Hermer-Vazquez R, and Chapin JK(2003) Olfactomotor coupling prior to skilled,olfactory-driven reaching. PNAS, under review.

Iwaniuk AN and Whishaw IQ (2000) On the origin ofskilled forelimb movements. Trends in Neuroscience23:372-376.

Johnson BN, Mainland JD, and Sobel N (2003) Rapid ol-factory processing implicates subcortical control ofan olfactomotor system. Journal of Neurophysiol-ogy 90:1084-1094.

Johnson KO (2001) The roles and functions of cutaneousmechanoreceptors. Current Opinion in Neurobiol-ogy 11:455-461.

Jones EG (2001) The thalamic matrix and thalamocorti-cal synchrony. Trends in Neuroscience 24:595-601.

Kandel ER, Schwartz JH, and Jessell TM (eds.) (2000)

Principles of neural science, 4th edition. New York:McGraw-Hill, Health Professions Division.

Kleim JA, Barbay S, Cooper NR, Hogg TM, Reidel CN,Remple MS, Nudo RJ (2002) Motor learning-dependent synaptogenesis is localized to function-ally reorganized motor cortex. Neurobiology ofLearning and Memory 77:63-77.

Krubitzer L (1995) The organization of neocortex inmammals: are species differences really so different?Trends in Neuroscience 18:408-417.

Krupa DJ, Ghazanfar AA, Nicolelis MA (1999) Immedi-ate thalamic sensory plasticity depends on corti-cothalamic feedback. Proceedings of the NationalAcademy of Sciences 96:8200-8205.

Lee AK and Wilson MA (2002) Memory of sequentialexperience in the hippocampus during slow wavesleep. Neuron 36:1183-1194.

Llinas RR (2001) I of the vortex: from neurons to self.Cambridge, MA: MIT Press.

McFarland NR and Haber SN (2002) Thalamic relay nu-clei of the basal ganglia form both reciprocal andnonreciprocal cortical connections, linking multiplefrontal cortical areas. Journal of Neuroscience 22:8117-8132.

Metz GA and Whishaw IQ (1996) Skilled reaching an ac-tion pattern: stability in rat (Rattus norvegicus)grasping movements as a function of changing foodpellet size. Behavioural Brain Research 116:111-122.

Paxinos G (ed.) (1995) The rat nervous system. SanDiego: Academic Press.

Poe GR, Nitz DA, McNaughton BL, Barnes CA (2000)Experience-dependent phase-reversal of hippocam-pal neuron firing during REM sleep. Brain Research855:176-180.

Rioult-Pedotti MS, Friedman D, Donoghue JP (2000)Learning-induced LTP in neocortex. Science 290:533-536.

Shin HC and Chapin JK (1990) Mapping the effects of mo-tor cortex stimulation on somatosensory relay neu-rons in the rat thalamus: direct responses and afferentmodulation. Brain Research Bulletin 24:257-265.

Strick L (2002) Stimulating research on motor cortex.Nature Neuroscience 5:714-715.

Vallbo AB and Johansson RS (1984) Properties of cuta-neous mechanoreceptors in the human hand relatedto touch sensation. Human Neurobiology 3:3-14.

Whishaw IQ and Pellis S (1990) The structure of skilledforelimb reaching in the rat: a proximally drivenmovement with a single distal rotatory component.Behavioural Brain Research 41:49-59.

Page 86: 0195162854

Pain

DANIEL LE BARS AND SAMUEL W. CADDEN7

Sherrington (1906) introduced the concept of"nociception" (from the Latin nocere, "toharm"). Within that concept, "nociceptive"stimuli are those that threaten the integrity ofthe body and/or directly activate a collectionof discrete sensory organs or nerves known asnociceptors. These stimuli set off a varied butlimited repertoire of somatic and vegetativereflex and behavioral responses that are asso-ciated with the perception of pain. Sherring-ton also described pain as "the psychical ad-junct of a ... protective reflex," underliningthat pain triggers reactions and induceslearned avoidance behaviours that may de-crease whatever is causing it. As a result, painmay limit the (potentially) damaging conse-quences, and it is the behaviors associatedwith this end point that define many "pain re-sponses" in rats.

The complexity of pain is underlined bythe definition of pain adopted by the Interna-tional Association for the Study of Pain (IASP):"an unpleasant sensory and emotional experi-ence associated with actual or potential tissuedamage or described in terms of such dam-age." The painful experience is more than asensory experience that discriminates the in-tensity, location, and duration of a stimulus;it is also characterized by an emotional aver-sive state, which pushes one to action (moti-vation). This emotion is a fundamental and in-extricable part of the painful experience andnot a reaction to the sensory aspect. There-fore, the pain gets our attention, interfereswith activity, and mobilizes strategies for de-fense. Zimmermann (1986) reinterpreted the

IASP definition of pain so that it could be ap-plied to animals: "an aversive sensory experi-ence caused by actual or potential injury thatelicits progressive motor and vegetative reac-tions, results in learned avoidance behaviour,and may modify species specific behaviour, in-cluding social behaviour."

The purpose of models in the rat is tomodel human pain. Thus, there is a need forthese models to reflect the different categoriesof human pain, such as physiological pain (no-ciceptive pain), inflammatory pain, and neu-ropathic pain (neuropathic: pertaining to dis-ease of the nervous system). The first twotypes usually occur after injury and are oftenassociated with each other. During an in-flammatory episode, the threshold for pain islowered so that (1) innocuous physical con-tact may become painful (allodynia) and/or(2) a nociceptive stimulus is perceived as moreintense pain than usual (hyperalgesia). Al-though these pains can exceed the duration ofthe stimulus, they usually disappear after therelated injury has healed.

By contrast, neuropathic pain resultsfrom an injury or a pathological transforma-tion of the somaesthesic system that evolvesan abnormal and unsuitable mode of func-tioning. In addition to the usual symptoms ofinflammatory pain, there are continuous orparoxysmal "spontaneous" pains (e.g., sensa-tions of "electrical" discharges), pains stem-ming from insensible regions (e.g., the para-doxical "anesthesia dolorosa"), paraesthesia(tingling, pricking, dullness), dysaesthesia(very unpleasant, but not painful, sensations),

69

Page 87: 0195162854

70 SENSORY SYSTEMS

and sometimes sympathetic disturbances.There is also the very paradoxical situationusually known as phantom pain that occursafter deafferentation (e.g., avulsion of thebrachial plexus or amputation of a limb) andthus in the absence of a nociceptive stimulusor nociceptors. Taken together, these symp-toms are described by patients as "strange" orsometimes even as not being pain but "worsethan pain." It is unclear whether these all havecorollaries in the rat.

SENSATION AND REACTION

It is fairly obvious that in the context of ratmodels of pain, psychological pain is difficultto monitor. The absence of verbal communi-cation in rats is undoubtedly an obstacle to theevaluation of pain. Nevertheless, most pain re-sponses displayed by humans can be objec-tively measured behaviorally, and very simi-lar behavioral responses can be measured inrats in response to similar stimuli. There arecircumstances when there can be little doubtthat a rat is feeling pain, such as when it is re-sponding to stimuli with vocal responses. Onthe other hand, it is far more difficult to cer-tify that at a given moment a rat feels no painbecause it is presenting no typical physicalsigns or overt behaviors. This is particularlyso given that we know that immobility orprostration is sometimes the only response ac-companying pain. The question of pain in ratscan be approached only with anthropomor-phic references, although differences probablydo exist by comparison with humans, notablywith respect to certain cerebral structures(Bateson, 1991).

By contrast with the polymorphic natureof the pain that is described as a sensation inhumans, that in rats can be estimated only byexamining their reactions. This is essentiallythe same difficulty as is faced by the pediatri-cian, the geriatrician, or the psychiatrist whendealing with patients incapable of expressingthemselves verbally. In those cases, too, the

semeiology is not unequivocal. It has to betaken in context and placed in an inventory,as its meaning will differ depending on the de-gree of maturation (or degradation) of thenervous system.

The study of behavioral reactions pro-vides the only indicator of the perceived, dis-agreeable sensation resulting from a stimulusthat would be algogenic (pain producing) ifexperienced by a human. But it must never beforgotten that these responses are often notspecific: for example, escape can result fromany disagreeable stimulus whether it is noci-ceptive or not. In addition, it should be re-membered that the existence of a reaction isnot necessarily evidence of a concomitant sen-sation (Hardy et al, 1952).

The observed reactions in the rat cover awide spectrum ranging from the most ele-mentary reflexes to far more integrated be-haviors (e.g., escape, avoidance). In almostevery case, it is a motor response that is mon-itored. By contrast, vegetative responses areconsidered only occasionally.

MODELS OF CHRONIC PAIN

There are two types of chronic pain modelsin the rat—models of rats with induced arthri-tis and models of rats with lesions of the cen-tral or peripheral nervous systems.

The injection of complete Freund's adju-vant in the rat brings about a severe generalmalaise associated with ankylosing spondyli-tis (Butler, 1989). In other models, the arthri-tis is limited to one joint.

Rat models of neuropathies are based ona total or partial deafferentation of an area ofthe body. The cutting of several neighboringdorsal roots triggers, within 2 to 3 weeks, atype of self-mutilation "autotomy " (Dong,1989; Kauppila, 1998). As with anesthesia do-lorosa in humans, this behavior is thought tobe elicited by the pain evoked by deafferenta-tion. However, such an interpretation is dis-puted by some because the introduction of a

Page 88: 0195162854

Chapter 7. Pain 71

female into the cage of an operated rat is suf-ficient to eliminate the autotomy behavior.

There are three types of rat models ofneuropathic pain triggered by partial deaf-ferentation (Seltzer, 1995; Bennett, 2001). Thechronic constriction injury model is producedby a loose ligature around the rat sciatic nerve,resulting in a loss of myelinated fibers but—for the main part—the survival of C-fibers(Bennett and Xie, 1988). The partial nervetransection model is produced by partial sec-tion of the rat sciatic nerve (Seltzer et al.,1990). The spinal nerve transection model isproduced by the section of the L5-6 spinal dor-sal roots, with the hindpaw remaining par-tially innervated through the L4 root (Kim andChung, 1992). In addition, there are somemodels that were designed to replicate knownanatomoclinical entities in humans (e.g., thediabetic neuropathy produced by streptozocinand neuropathies elicited by antineoplasticdrugs such as Taxol (paclitaxel), vincristine, orcisplatin). All of these models are character-ized by evidence of allodynia and hyperalge-sia when stimuli are applied to the affectedzone. With the notable exception of the ob-servation of autotomy, purely behavioral ap-proaches to spontaneous pain in the rat arerare. The same tests are used to evaluatestimulus-evoked pain in normal rats and inrats with chronic pain. The main difference re-sults from the fact that the tests are applied toa body and a nervous system with a differenthistory; however, the triggered behavior andthe other measured variables are the same.

reflexes" by Sherrington (1906). They include(I) basic motor responses (withdrawal, jump-ing, contractures, etc.), (2) neurovegetative re-actions, generally in the context of Selye's"alarm reaction," with an increase in sympa-thetic tone (tachycardia, arterial hypertension,hyperpnea, mydriasis, etc.), and (3) vocaliza-tion. The latter categories include conditionedmotor responses that result from a period oflearning and sometimes can be very rapid. Be-havioral reactions (escape, distrust of objectsresponsible for painful experiences, avoid-ance, aggression, etc.) or modifications of be-havior (social, food, sexual, sleep, etc.) are of-ten observed. It must be noted, however, thateven if active motor reactions are frequent,passive responses are seen just as often in an-imals, such as immobility, which allows theanimal to preserve a painless posture. Fur-thermore, motor atonia is a general responseto illness, regardless of whether the conditionis painful.

INPUT AND OUTPUT: THESTIMULUS AND THE RESPONSE

Behavioral tests in the rat need to be appro-priate. The stimuli have to be quantifiable, re-producible, and noninvasive (Beecher, 1957;Lineberry, 1981). All nociceptive stimuli canbe defined by a number of different parame-ters: (1) the physical nature of the stimulus, (2)the site of application of the stimulus, and (3)the history of this site of stimulation.

BEHAVIORAL REACTIONS

There are two main types of reaction to a no-ciceptive stimulus: (1) those organized by cen-ters that are relatively "low" within the hier-archy of the central nervous system and (2)more complex ones organized by higher cen-ters in the central nervous system.

The former can be elicited in decerebrateanimals and were termed "pseudo-affective

PHYSICAL NATURE OF THE STIMULUS

Regardless of whether a stimulus is electrical,thermal, mechanical, or chemical, it is essen-tial that three of its parameters are controlled:the intensity, the duration, and the surfacearea to which it is applied. These three pa-rameters determine the "global quantity ofnociceptive information" that will be gener-ated and carried toward the central nervoussystem by the peripheral nervous system.

Page 89: 0195162854

72 SENSORY SYSTEMS

SITE OF APPLICATION OF THE STIMULUS

Clinical pain can originate from somatic, vis-ceral, articular, or musculotendinous tissues.In nociceptive tests, stimuli are usually appliedto cutaneous and, to a lesser extent, visceralstructures. We know that some areas of skincan have a specific, particular function. For ex-ample, the rat tail, a structure used in manynociceptive tests, is an essential organ for ther-moregulation and balance (see Chapters 12and 21).

PREVIOUS HISTORY OFTHE STIMULATED SITE

Tests for acute pain involve healthy tissuesand, occasionally, acutely inflamed tissues (ofa few days' standing at most). Tests forchronic pain relate to rheumatic or neuro-pathic conditions that last for longer periodsof time (from weeks up to months).

Because the application of the stimulusmust not produce lesions, one often defines alimit for how long the animal should be ex-posed to the stimulus (the "cut-off time").This limit is absolutely necessary when the in-tensity of the stimulus is increasing. Further-more, the repeated application of a stimuluscan sensitize peripheral receptors and/or pro-duce central sensitization.

it is not always easy to confirm that it is be-ing achieved. For example, the appearance ofa flexion reflex does not inevitably mean thatthe stimulus is nociceptive or that it is a no-ciceptive flexion reflex. Indeed, flexion re-flexes are not triggered exclusively by noci-ceptive stimuli (Schomburg, 1997). It must bepossible in the behavioral model to differen-tiate responses to nociceptive stimuli from re-sponses to non-nociceptive stimuli. In otherwords, the quantified response must be ex-clusively or preferentially triggered by noci-ceptive stimuli ("output specificity"). In thisrespect, one must remember that some in-nate and acquired behaviors can be triggeredby aversive stimuli that are not nocicep-tive/painful.

SENSITIVITY

It must be possible to quantify the responseand for this variable to be correlated withstimulus intensity within a reasonable range(from the pain threshold to the pain tolerancethreshold). In other words, the quantified re-sponse must be appropriate for a given typeof stimulus and monotonically related to itsintensity. The model must be sensitive to ma-nipulations, notably pharmacological, thatwould reduce the nociceptive behavior in aspecific fashion.

REQUIREMENTS FOR BEHAVIORALMODELS OF NOCICEPTION

Ideally, a behavioral model for nociception inthe rat possesses the following characteristics(Lineberry, 1981; Vierck and Cooper, 1984;Ramabadran and Bansinath, 1986; Hammond,1989; Watkins, 1989; Tj01sen and Hole, 1997;Le Bars et al., 2001; Berge, 2002).

SPECIFICITY

The stimulus must be nociceptive ("inputspecificity"). Although this is common sense,

VALIDITY

The model must allow the differentiation ofnonspecific behavioral changes (e.g., in motil-ity, attention, etc.) from those triggered by thenociceptive stimulus itself. In other words, theresponse being monitored must not be con-taminated by simultaneous perturbations re-lated to other functions, notably if these havebeen introduced by a pharmacological agent.The test validity (i.e., the degree to which thetest actually measures what it purports tomeasure) is undoubtedly one of the most dif-ficult issues to determine (Berge, 2002; Hans-son, 2003).

Page 90: 0195162854

Chapter 7. Pain

RELIABILITY

Consistency of scores must be obtained whenanimals are retested with an identical test orequivalent form of the test. In this context, therepeated application of the stimulus must notproduce lesions.

REPRODUCIBILITY

Results obtained with a test must be repro-ducible not only within the same laboratorybut also between different laboratories.

Before describing tests that endeavor tomeet these requirements, it is worth notingthat they can be divided into two categories:those measuring a threshold and those mea-suring supraliminal responses. However, bothcategories permit the study of only one pointon the stimulus-response curve, be it thethreshold or an arbitrary point farther up thecurve. As a result, they allow only a rough ap-preciation of the gain of the process (Tj01senand Hole, 1997).

TESTS BASED ON THEMEASUREMENT OF REACTIONTIME FOR ESCAPE BEHAVIOR

Escape tests are based mainly on the applica-tion of thermal stimuli to the skin. Heat con-stitutes a relatively selective stimulus for no-ciceptors, and radiant heat has the advantageof not producing a concomitant tactile stimu-lus. Nevertheless, heating is progressive andresults in thermoreceptors being activated be-fore nociceptors are recruited. Just as there isthis sequence of activation of thermorecep-tors and then nociceptors, there is a sequenceof a hot sensation followed by a painful one.As a result, the possibility that the same stim-ulus is successively a conditioning and a con-ditioned stimulus cannot be ruled out. In ad-dition, one has to address the question of themeaning that can be ascribed to measure-ments of reaction time when a stimulus is

73

gradually increasing in intensity (Le Bars et al.,2001).

TAIL-FLICK TEST

There are two variants of the tail-flick test.One consists of applying radiant heat to asmall surface of the tail. The other involvessubmersing the tail in water at a predeter-mined temperature. Both provoke a vigorouswithdrawal movement of the tail (d'Amourand Smith, 1941). It is the reaction time of thismovement that is recorded. The "tail-flick" isa spinal reflex as witnessed by the persistenceof at least its shorter-latency form after sec-tion or cold block of upper parts of the spinalcord.

PAW WITHDRAWAL TEST

This test is comparable to the tail-flick test butoffers the advantages that (1) it does not in-volve the preeminent organ of thermoregula-tion in rats (the tail) and (2) it can be appliedto freely moving animals (Hargreaves et al.,1988). However, the latter is also a disadvan-tage in that the position of the leg becomes afactor of variability, because the backgroundlevel of activity in the flexor muscles varieswith the position of the animal.

HOT PLATE TEST

This test consists of introducing a rat into anopen-ended cylindrical space with a floor con-sisting of a heated metallic plate (Woolfe andMacDonald, 1944). Two behavioral compo-nents can be measured by their reaction times:paw licking and jumping. Both are consideredto be supraspinally integrated responses.However, licking the forepaw is a typicalbehavior for heat dissipation (Roberts andMooney, 1974). Although such behavior is rel-atively stereotyped in the mouse, it is morecomplex in the rat, which sniffs, licks itsforepaws, licks its hindpaws, straightens up,stamps its feet, starts and stops washing itself,

Page 91: 0195162854

74

Figure 7-1. Learning phenomena in the hot plate test. Inthe experiments, measurements were made of the delay be-tween after the animal had been put on the hot plate andbefore the animal began to lick the paws. (A) Sandklihler etal. (1996) repeated the test daily in Sprague-Dawley rats. (B)Lai et al. (1982) repeated the test each week on a Wistarstrain of rat. It can be seen that in both cases, four or fivetests were sufficient to almost halve the reaction time.(Modified from Sandkiihler et al. [1996] and Lai et al.[1982].)

etc. (Espejo and Mir, 1993). Furthermore, thistest is very susceptible to learning, whichmakes it delicate to interpret (Fig. 7-1).

TEST BASED ON THEMEASUREMENT OF THRESHOLD

FOR ESCAPE BEHAVIOR

Threshold tests are based on the applicationof mechanical or electrical stimuli to the skinor an internal organ.

APPLICATION OF INCREASING PRESSURE

An increasing pressure is applied to a smallarea on the hindpaw and interrupted whenthe threshold is reached (Green et al., 1951).This produces successive responses of reflexwithdrawal of the paw, a more complexmovement whereby the rat tries to release itstrapped limb, a struggle, and, finally, a vocalreaction. Although the first of these is un-doubtedly a proper spinal reflex, the latter twoclearly involve supraspinal structures. Withthe aim of improving the sensitivity of thistest, Randall and Selitto (1957) proposed com-paring thresholds observed on a healthy pawand an inflamed paw.

SENSORY SYSTEMS

APPLICATION OF CALIBRATED PRESSURE

Skin pressure tests involve applying a fiber ofa given diameter to the skin (Handwerker andBrune, 1987). Pressure is applied until the fiberbends. The use of a range of fiber diameters(variously called Semmes-Weinstein fibers orvon Frey hairs) makes it possible to determinethe threshold for evoking a response in the an-imal (e.g., a flexion reflex). This test is a prizedtool in models of neuropathic pain (e.g., Kimand Chung, 1992). A technical difficulty withthis approach relates to the sensitivity of suchfibers to humidity (Moller et al., 1998).

DETENTION OF HOLLOW ORGANS

Distention tests include colorectal distentionby means of an inflatable balloon that pro-duces avoidance behavior, reflex activities inthe abdominal muscles, and quantifiable veg-etative responses (increased arterial pressure,tachycardia) (Ness and Gebhart; 1988). A de-velopment of this model involves first in-flaming the colon by the administration ofchemical agents. Distention of the vagina oruterus has also been used in female rats(Berkley et al., 1995).

APPLICATION OF ELECTRICAL STIMULI

The application of electrical stimuli has the ad-vantages of being quantifiable, reproducible,and noninvasive and of producing synchro-nized afferent signals. However, intense elec-trical stimuli nondifferentially excite all pe-ripheral fibers, including large-diameter fibersthat are not directly implicated in nociceptionas well as fine A$ and C fibers, which medi-ate thermoreceptive as well as nociceptiveinformation.

Electrical stimuli of gradually increasingintensities can be delivered in trains throughsubcutaneous electrodes in the tail of the rat(Carroll and Lim, 1960; Levine et al., 1984).The gradually increasing currents produce, insuccession, reflex movement of the tail, vo-

Page 92: 0195162854

Chapter 7. Pain 75

calization at the time of stimulation, and, fi-nally, vocalization, which continues beyondthe period of stimulation (a vocalization af-terdischarge). These responses are organizedon a hierarchical basis; they depend on the dif-ferent levels of integration of the nociceptivesignal in the central nervous system: the spinalcord, the brain stem, and the thalamus/rhinencephalon. The last of these can reflectaffective and motivational aspects of pain be-havior (Borszcz, 1995a).

Electrical stimuli can also be applied inthe form of single, short-duration pulses.These elicit, in succession, twitching, escapebehavior, vocalization, and biting of the elec-trodes. Again, these responses are hierarchi-cally organized, with the last one being themost coordinated. They also depend on dif-ferent levels of integration of the nociceptivesignal (Charpentier, 1968).

In an attempt to overcome the drawbackthat intense electrical stimuli excite non-nociceptive as well as nociceptive nerves (seeearlier), some investigators have stimulatedtissues in which they believe all the afferentnerve fibers are nociceptive. Most commonly,the dental pulp has been used for this purpose.Contrary to commonly held belief, however,it is not certain that all of the afferent fibers inthe dental pulp are nociceptive, although theproportion of them that are may be greater thanthat in other tissues (Le Bars et al., 2001). In ad-dition, the anatomical arrangement of the den-tal tissues in the rat is such that it is difficult toapply electrical stimuli that excite pulpal nerveswithout exciting (non-nociceptive) nerves in thesurrounding tissues (e.g., Hayashi, 1980; Jiffry,1981). Some workers have shown that the ex-clusive activation of rat pulpal nerves is pos-sible provided adequate care is taken (e.g., Ra-jaona et al., 1986, Myslinski and Matthews,1987), but it seems that this has not alwaysbeen done. Two types of response have beenmonitored in such models: the dysynaptic jawopening reflex (e.g., Vassel et al., 1986), whichis organized within the trigeminal structuresof the brain stem in a fashion similar to spinal

reflexes elsewhere in the body (e.g., Sumino,1971), and the appearance of more complexreactions such as scratching, head move-ments, and vocalization (e.g., Rajaona et al.,1986), which involve coordination at highercenters.

TESTS BASED ON THEOBSERVATION OF BEHAVIOR

The main types of these tests involve using anintradermal or intraperitoneal injection of anirritant, algogenic, chemical agent as the no-ciceptive stimulus. In this section, we also con-sider the description of complex vocal pat-terns elicited by electrical stimuli.

INTRADERMAL INJECTIONS OFIRRITANT AGENTS

The most commonly used substance for in-tradermal injections is formalin (the formalintest). When it is injected into the dorsal sur-face of the rat forepaw, formalin provokes a"painful" behavior that can be assessed on afour-level scale related to the posture of theinjected paw (see Fig. 7-2): 0 indicates normalposture; 1, paw remaining on the ground butnot supporting the animal; 2, paw clearly be-ing raised; and 3, paw being licked, nibbled,or shaken (Dubuisson and Dennis, 1977). Aninitial phase can be observed about 3 minutesafter the injection; then, after a quiescent pe-riod, there is a second phase between minutes20 and 30. The first phase results essentiallyfrom the direct stimulation of nociceptors,whereas the second phase involves a periodof sensitization during which inflammatoryphenomena occur.

INTRAPERITONEAL INJECTIONS OFIRRITANT AGENTS

The intraperitoneal administration of agentsthat irritate serous membranes provokes a verystereotyped behavior, which is characterized by

Page 93: 0195162854

76 SENSORY SYSTEMS

Figure 7-2. Behavior triggered by an intradermal injectionof formalin into the right forepaw (see text). (Modified fromDubuisson and Dennis [1977].)

abdominal contractions, movements of thebody as a whole (particularly of the hindpaws),twisting of dorsoabdominal muscles, reductionin motor activity, and motor incoordination(writhing behavior). These "behaviors" are allconsidered to be reflexes (Hammond, 1989).

STIMULATION OF HOLLOW ORGANS

Tests that involve the injection of algogenicsubstances directly into hollow organs areused as models for visceral pain. Administra-tion of formalin into the rat colon can producea complex, biphasic type of "pain" behaviorinvolving an initial phase of body stretching

and contraction of either the flanks or thewhole body and a second phase that predom-inantly involves abdominal licking and nib-bling (Miampamba et al., 1994). Similarly, anumber of models have been developed forbladder or uterine pain whereby reflexesand/or more complex behaviors have beenobserved after the administration of irritantsinto the organ (e.g., McMahon and Abel, 1987;Pandita et al., 1997; Wesselmann et al., 1998).

Giamberardino et al. (1995) studied thebehavior produced by the surgical introduc-tion of dental cement, to mimic a calculus,into the ureter. This produced something akinto episodes of writhing behavior over a 4-dayperiod. A concomitant hyperalgesia in the ab-dominal muscles provided evidence of vis-ceromuscular convergence. To the best of ourknowledge, this is the only animal model of"referred" pain.

VOCALIZATION ELICITED BYELECTRICAL SHOCKS

Complex vocal patterns can be produced byelectrical stimuli of short duration (Jourdan etal., 1995). Three types of emissions have beenidentified (Fig. 7-3).

1. Two distinct "peeps," the energies of which aredistributed across a wide range of audible fre-quencies without a defined structure, occur.The first peep results from activation of rela-tively rapidly conducting AS fibers, and the sec-ond peep results from activation of slowly con-ducting C fibers.

2. "Chatters" are characterized by formants com-posed of a fundamental frequency and its har-monics; these constitute a very elaborate re-sponse, the physical characteristics of which aresimilar to human words.

3. Ultrasonic emissions, inaudible to humans andmade up of a fundamental frequency, withoutharmonics, between 20 and 35 kHz, occur withmild modulations.

The characteristics of the first two peepsemitted by the rat are reminiscent of the phe-nomenon of "double pain" observed in humansafter a brief, sharp nociceptive stimulus. The

Page 94: 0195162854

Chapter 7. Pain 77

Figure 7-3. Example of a vocal response elicited by a single 20 mA, 2 millisecond square electrical pulse ap-plied to the base of the tail of a rat. (Top) Recording during the 6 seconds after the stimulus. This responseconsisted of eight components separated by silences: two peeps, two chatters, and four ultrasonic emissions,as defined by the corresponding spectrograms shown in the insets below. These spectrograms revealed thatthe first two peeps corresponded to strong emissions within a wide spectrum of frequencies. By contrast, thetwo subsequent chatters were pure, slightly modulated sounds with a fundamental frequency and corre-sponding harmonics. The last four components were pure ultrasound with a single, 21.7 kHz, frequency.(Modified from Jourdan et al. [1995].)

physiological meanings of the other compo-nents of the response are more difficult to un-derstand. In line with Pavlovian conditioning,the "chatters" can be triggered by a light signal(Borszcz, 1995b). The ultrasonic emissions mayreflect the affective state and the degree of anx-iety of the animal, because they can also berecorded in other experimental situations thatgenerate fear or stress (Sales and Pye, 1974;Haney and Miczek, 1994). In addition, they aresensitive to anxiolytic drugs (Tonoue et al.,1987; Cuomo et al., 1992).

FURTHER CONSIDERATIONSAND CONCLUSION

Behavioral tests of nociception are not en-tirely satisfactory (Le Bars et al., 2001). Their

first weakness lies in the difficulty of control-ling the stimuli used to trigger a nociceptivereaction. More important, even when thephysical parameters of the external stimuli arewell controlled, that does not necessarily re-sult in an equally well-controlled effectivestimulus. Indeed, the stimulus that affectivelyactivates the peripheral nociceptors is also in-fluenced by the physiological state of thetargeted tissue. This state is determined bythe vegetative system, mainly through ther-moregulation, systemic arterial blood pres-sure, and local vasomotor tone. For example,variation of local temperature may confoundmany experimental protocols, not only thoseusing heat as a stimulus (Tj01sen and Hole,1997). The second great weakness of thesemodels lies in the nature of the dependentvariable, which is usually the determination

Page 95: 0195162854

78 SENSORY SYSTEMS

of the threshold for a reaction. Most of themodels do not allow the study of stimulus-response relationships although these are re-ally an essential element of sensory physiol-ogy. Furthermore, it is often not the thresh-old itself that is measured but rather aresponse time to a stimulus of increasing in-tensity. Because when the skin is subjected toa constant source of radiation, the tempera-ture increases with the square root of time,such an approach is highly questionable whenradiant heat is used.

These considerations invite prudence inthe interpretation of results obtained usingtests of nociception in the rat. It is worthstressing that these considerations also relateto rat models of chronic pain insofar as thetests applied in these models are the sameones, or almost the same ones, as those de-scribed for acute pain. Faced with the sim-plistic and reductionist approaches available,one is inclined to conclude that valid behav-ioral approaches of pain should be promotedin animals, notably in the rat.

REFERENCES

Bateson P (1991) Assessment of pain in animals. AnimalBehavior 42:827-839.

Beecher HK(1957) The measurement of pain: prototypefor the quantitative study of subjective responses.Pharmacological Reviews 9:59-209.

Bennet GJ (2001) Animal models of pain. In: Methods inpain research (Kruger L, ed.), pp. 67-91. Boca Ra-ton: CRC Press.

Bennett GJ and Xie YK (1988) A peripheral mononeu-ropathy in rat that produces disorders of pain sen-sation like those seen in man. Pain 33:87-107.

Berge OG (2002) Reliability and validity in animal painmodeling. In: Pain and brain (Tj01sen A and BergeOG, eds.), pp. 77-89. Bergen: University of Bergen.

Berkley KJ, Wood E, Scofield SL, Little M (1995) Be-havioral responses to uterine or vaginal distensionin the rat. Pain 61:121-131.

Borszcz GS (1995 a) Increases in vocalisation and motorreflex thresholds are influenced by the site of mor-phine microinjection: comparisons following ad-ministration into the periaqueductal gray, ventral

medulla, and spinal subarachnoid space. BehavioralNeuroscience 109:502-522.

Borszcz GS (1995b) Pavlovian conditional vocalizationsof the rat: a model system for analyzing the fear ofpain. Behavioral Neuroscience 109:648-662.

Butler SH (1989) Animal models and the assessment ofchronic pain: critique of the arthritic rat model. In:Issues in pain measurement, advances in pain re-search and therapy, vol 12 (Chapman CR and LoeserJD, eds.), pp. 473-479. New York: Raven Press.

Carroll MN and Lim RK (1960) Observations of theneuropharmacology of morphine-like analgesia.Archives Internationales de Pharmacodynamie et deTherapie 125:383-403.

Charpentier J (1968) Methods for evaluating analgesicsin laboratory animals. In: Pain: proceedings of theInternational Symposium on Pain organised by theLaboratory of Psychophysiology (Soulairac A, CahnJ, Charpentier J, eds.), pp. 171-200. New York: Aca-demic Press.

Cuomo V, Cagiano R, De Salvia MA, Mazzoccoli M,Persichella M, Renna G (1992) Ultrasonic vocalisa-tion as an indicator of emotional state during activeavoidance learning in rats. Life Sciences 50:1049-1055.

D'Amour FE and Smith DL (1941) A method for deter-mining loss of pain sensation. Journal of Pharma-cology and Experimental Therapeutics 72:74-79.

Dong WK (1989) Is autotomy a valid measure of chronicpain? In: Issues in pain measurement, advances inpain research and therapy, vol 12 (Chapman CR andLoeser JD, eds.), pp. 463-472. New York: RavenPress.

Dubuisson D and Dennis SG (1977) The formalin test:a quantitative study of the analgesic effects of mor-phine, meperidine and brain stem stimulation in ratsand cats. Pain 4:161-174.

Espejo EF and Mir D (1993) Structure of the rat's be-haviour in the hot plate test. Behavioural Brain Re-search 56:171-176.

Green AF, Young PA, Godfrey El (1951) A comparisonof heat and pressure analgesiometric methods inrats. British Journal of Pharmacology 6:572-585.

Hansson P (2003) Difficulties in stratifying neuropathicpain by mechanisms. European Journal of Pain7:353-357.

Hammond DL (1989) Inference of pain and its modula-tion from simple behaviors. In: Issues in pain mea-surement, advances in pain research and therapy,vol 12 (Chapman CR and Loeser JD, eds.), pp.69-91. New York: Raven Press.

Handwerker HO and Brune K (1987) DeutschsprachigeKlassiker der Schmerzforshung [Classical Germancontributions to pain research]. Hassfurt: Tagblatt-Druckerei KG.

Page 96: 0195162854

Chapter 7. Pain 79

Haney M and Miczek KA (1994) Ultrasounds emitted byfemale rats during agonistic interactions: effects ofmorphine and naltrexone. Psychopharmacology(Berlin) 114:441-448.

Hardy JD, Wolff HG, Goodell H (1952) Pain sensationand reaction. Baltimore: Williams & Wilkins.

Hargreaves K, Dubner R, Brown F, Flores C, Joris J(1988) A new and sensitive method for measuringthermal nociception in cutaneous hyperalgesia. Pain32:77-88.

Hayashi H (1980) A problem in electrical stimulation ofincisor tooth pulp in rats. Exp Neurology 67:438-441.

Jiffiy MT (1981) Afferent innervation of the rat incisorpulp. Experimental Neurology 73:209-218.

Jourdan D, Ardid D, Chapuy E, Eschalier A, Le Bars D(1995) Audible and ultrasonic vocalisation elicitedby single electrical nociceptive stimuli to the tail inthe rat. Pain 63:237-249.

Kauppila T (1998) Correlation between autotomy-behavior and current theories of neuropathic pain.Neuroscience and Biobehavioral Reviews 23:111-129.

Kim SH and Chung JM (1992) An experimental modelfor peripheral neuropathy produced by segmentalspinal nerve ligation in the rat. Pain 50:355-363.

Kontinen VK and Meert TF (2003) Predictive validity ofneuropathic pain models in pharmacological stud-ies with a behavioral outcome in the rat: a sytem-atic review. In: Proceedings of the 10th World Con-gress on Pain, Progress in Pain Research andManagement, vol 24 (Dostrovsky JO and Carr DB,eds.), pp. 489-498. Seattle: IASP Press.

Lai YY and Chan SHH (1982) Shortened pain responsetime following repeated algesiometric test in rats.Physiology and Behavior 28:1111-1113.

Le Bars D, Gozariu M, Cadden SW (2001) Animalmodels of nociception. Pharmacological Reviews53:597-652.

Levine JD, Feldmesser M, Tecott L, Gordon NC,, Izdeb-ski K (1984) Pain induced vocalization in the rat andits modification by pharmacological agents. BrainResearch 296:121-127.

Lineberry CG (1981) Laboratory animals in pain re-search. In: Methods in animal experimentation, vol6, pp. 237-311. New York: Academic Press.

McMahon SB and Abel C (1987) A model for the studyof visceral pain states: chronic inflammation of thechronic decerebrate rat urinary bladder by irritantchemicals. Pain 28:109-127.

Miampamba M, Chery-Croze S, Gorry F, Berger F,Chayvialle JA (1994) Inflammation of the colonicwall induced by formalin as a model of acute vis-ceral pain. Pain 57:327-334.

Moller KA, Johansson B, Berge OG (1998) Assessing me-

chanical allodynia in the rat paw with a new elec-tronic algometer. Journal of Neuroscience Methods84:41-47.

Myslinski N and Matthews B (1987) Intrapulpal nervestimulation in the rat. Journal of NeuroscienceMethods 22:73-78.

Ness TJ and Gebhart GF (1988) Colorectal distension asa noxious visceral stimulus. Physiologic and phar-macologic characterization of pseudoaffective re-flexes in the rat. Brain Research 450:153-169.

Pandita RK, Persson K, Andersson KE (1997) Capsaicin-induced bladder overactivity and nociceptive be-haviour in conscious rats: Involvement of spinalnitric oxide. Journal of the Autonomic NervousSystem 67:184-191.

Rajaona J, Dallel R, Woda A (1986) Is electrical stimu-lation of the rat incisor an appropriate experimen-tal nociceptive stimulation? Experimental Neurol-ogy 93:291-299.

Ramabadran K and Bansinath M (1986) A critical analy-sis of the experimental evaluation of nociceptive re-actions in animals. Pharmaceutical Research 3:263-270.

Randall LO and Selitto JJ (1957) A method for mea-surement of analgesic activity on inflamed tissue.Archives Internationales de Pharmacodynamie et deTherapie 111:409-419.

Roberts WW and Mooney RD (1974) Brain areas con-trolling thermoregulatory grooming, prone exten-sion, locomotion, and tail vasodilation in rats. Jour-nal of Comparative Physiology and Psychology 86:470-480.

Sales GD and Pye D (1974) Ultrasonic communicationby animals. London: Chapman and Hall.

Sandkuhler J, Treier AC, Liu XG, Ohnimus M (1996)The massive expression of c-fos protein in spinaldorsal horn neurons is not followed by long-termchanges in spinal nociception. Neuroscience 73:657-666.

Schomburg ED (1997) Restrictions on the interpretationof spinal reflex modulation in pain and analgesia re-search. Pain Forum 6:101-109.

Seltzer Z (1995) The relevance of animal neuropathymodels for chronic pain in humans. Seminars inNeurosciences 7:211-219.

Seltzer Z, Dubner R, Shir Y (1990) A novel behavioralmodel of neuropathic pain disorders produced inrats by partial sciatic nerve injury. Pain 43:205-218.

Sherrington CS (1906) The integrative action of thenervous system. New York: C. Scribner's Sons.

Sumino R (1971) Central pathways involved in the jawopening reflex in the cat. In: Oral-facial sensory andmotor mechanisms (Dubner R and Kawamura Y,eds.). New York: Appleton-Century-Croft.

Tj01sen A and Hole K (1997) Animal models of analge-

Page 97: 0195162854

80 SENSORY SYSTEMS

sia. In: Pharmacology of pain (Dickenson AH andBessonJM, eds.), pp. 1-20. Berlin: Springer-Verlag.

Tonoue T, Iwasawa H, Naito H (1987) Diazepam andendorphine independently inhibit ultrasonic distresscalls in rats. European Journal of Pharmacology142:133-136.

Vassel A, Pajot J, Aigouy L, Rajaona J, Woda A. (1986)Effects, in the rat, of various stressing procedureson the jaw-opening reflex induced by tooth-pulpstimulation. Archives of Oral Biology 31:159-163.

Vierck CJ and Cooper BY (1984) Guideline for assessingpain reactions and pain modulation in laboratoryanimal subjects. In: Advances in pain research andtherapy, vol 6 (Kruger L and Liebeskind JC, eds.),pp. 305-322. New York: Raven Press.

Watkins LR (1989) Algesiometry in laboratory and man:

current concepts and future directions. In: Issues inpain measurement, advances in pain research andtherapy, vol 12 (Chapman CR and Loeser JD, eds.),pp. 249-265. New York: Raven Press.

Wesselmann U, Czakanski PP, Affaitati G, Giamber-ardino MA (1998) Uterine inflammation as a nox-ious visceral stimulus: behavioral characterizationin the rat. Neuroscience Letters 246:73-76.

Woolfe G and MacDonald Al (1944) The evaluation ofthe analgesic action of pethidine hydrochloride (De-merol). Journal of Pharmacology and ExperimentalTherapeutics 80:300-307.

Zimmermann M (1986) Behavioural investigations of painin animals. In: Assessing pain in farm animals (DuncanIJH and Molony Y, eds.), pp. 16-29. Brussels: Office forOfficial Publications of the European Communities.

Page 98: 0195162854

Vibrissae

RICHARD H. DYCK 8

The brain provides all animals with the meansto detect, interpret, and act on specific classesof sensory stimuli. Rats, like other animals, areequipped with sophisticated arrays of sensoryreceptors that are especially well suited to al-low them to negotiate their particular envi-ronments. Rats have evolved in complex anddiverse environments that are rich in olfac-tory, gustatory, auditory, visual, and tactualinformation. However, because rats are noc-turnal, by necessity they must be especially re-liant on nonvisual stimuli. One of the mostconspicuous of sensory receptor arrays, whichare readily apparent when one observes thebehaving rat, are the long, specialized hairsfound on the face and nose commonly calledtactile hairs, sinus hairs, whiskers, or vibrissae.

Rats, like all mammals (except for humansand a few nonprimate species), possess vibris-sae that are essential for their survival. Likecommon pelagic fur or hair, vibrissae are rigidcolumns of dead epidermal cells that are deeplyimbedded in epidermal hair follicles. What dis-tinguishes vibrissae from pelagic hairs are theirdistinct structure, length, sensory innervation,motor control, and, most important, their func-tion. Although pelagic hair is distributed widelyacross the skin, vibrissae are predominant onthe eyebrows, cheeks, lips, and chin with smallgroups sometimes appearing in other areas likethe abdomen or flexor surface of the wrist(Pocock, 1914; Sokolov and Kulikov, 1987). Al-though each of these groups of tactile hairs isimportant in guiding various aspects of the rat'sbehavior, the remainder of this chapter is con-cerned with the most widely studied, those lo-

cated on the rat's upper jaw, the so-calledmystacial (moustache) vibrissae.

Despite the apparent ecological andanatomical salience of tactual informationprovided by the vibrissae, there is a surprisingpaucity of studies that assess their role in guid-ing the rat's behavior (see Gustafson andFelbain-Keramidas, 1977, for the most recentreview). This is particularly true comparedwith the volumes of published studies thatdeal with the anatomy and physiology of therodent vibrissal somatosensory system. Thischapter provides a review of the anatomicalcharacteristics of the rat's vibrissa sensory ap-paratus and its growth and dynamic func-tional characteristics. This information is es-sential because knowledge of the use of thevibrissa sensory apparatus and the constraintsunder which it operates allows an under-standing of its role in the generation of be-havior in laboratory and natural settings.

So large and particular distribution of an exquisitely sensible

nerve, it is reasonable to suppose, must be for the purpose of

some sensible function.

—Broughton, 1823

STRUCTURE AND GROWTH

The rat provides a very useful model to un-derstand the mechanisms underlying the de-velopment, structure, and function of themammalian somatosensory system. In thevibrissal system, the exquisitely structuredtopographical organization of the periphery is

81

Page 99: 0195162854

82 SENSORY SYSTEMS

maintained in the one-to-one functional rep-resentation of the vibrissae in the brain stembarrelettes, ventrobasal thalamic barreloids,and barrels within layer 4 of the somatosen-sory vibrissa cortex (Jones and Diamond,1995). The usefulness of this system is exem-plified by the ability to unambiguously iden-tify these functional compartments using sim-ple anatomical stains (for example, see Fig.8-3). In addition, because the vibrissae arecontinuously replaced and spontaneously re-grow when trimmed or plucked, this modelsystem allows one to easily manipulate thesensory input at the level of the peripheral re-ceptor and to perform subsequent analyses us-ing the same peripheral array. Moreover, thissystem is amenable to within-animal designswhereby experimental and control manupu-lations occur in the same animal. This sectionis intended to briefly provide the reader witha basic understanding of several of these pa-rameters that should prove useful in the de-sign of experimental methods for behavioralstudies using the rat vibrissa system.

VIBRISSA FOLLICLE

Vibrissae are distinguished from other kindsof hair by connective tissue capsules that formaround the follicles. The capsule of each vi-brissa is composed of spongy cavernous tissuein the lower half and an open ring sinus in theupper portion, with both cavities filled withblood (Vincent, 1913). In addition to provid-ing metabolic requirements to the folliculartissues and nerves, the turgid blood sinus isbelieved to be essential for amplifying vibra-tory information from the vibrissae that acti-vates the sensory nerve endings in the follicle.

VIBRISSAL ORGANIZATION

The mystacial vibrissae are arrayed in a bilat-erally symmetrical, stereotypic manner on therat's cheek and upper lip (Fig. 8-1). The cau-dal array, consisting of about 35 vibrissae, aredistributed on the mystacial pad in five well-

defined caudorostrally oriented rows that arereferred to, from dorsal to ventral, as row Ato E (Fig. 8-2). The vibrissae are also arrangedin dorsoventrally oriented arcs numbered 1 to7 from caudal to rostral. The four longest, cau-dalmost vibrissae are not aligned within rowsand are therefore referred to as straddlers (ato 8; see Fig. 8-2). Collectively, because oftheir relative length, this group has also beenreferred to as the macrovibrissae. Vibrissaewithin the same arc are approximately thesame length, and they decrease in length, suc-cessively, caudorostrally. The straddlers are45 to 60 mm in length, with vibrissae in arcs1 through 4 projecting 40 to 44, 33 to 35, 23to 25, and 11 to 16 mm, respectively (Ibrahimand Wright, 1975).

The more rostral, furry buccal pad consistsof 40 to 70 vibrissae arranged in five rows thatline the rat's upper lip (see Figs. 8-1 and 8-2).Because of their shorter length (<7 mm), thisgroup of vibrissae has also been referred to asmicrovibrissae (Brecht et al., 1997).

SENSORY INNERVATION

The mystacial vibrissae are innervated by theinfraorbital branch of the maxillary division ofthe sensory portion of the trigeminal nerve

Figure 8-1. More than 70 specialized tactile hairs projectfrom each side of the rat's upper jaw. The caudal-most ar-ray consists of approximately 35 macrovibrissae, whereasthe rostral-most array lining the rat's upper lip consists of40 to 70 microvibrissae.

Page 100: 0195162854

Chapter 8. Vibrissae 83

Figure 8-2. The organization of the rat's mystacial vibris-sae on the skin of the rat's upper jaw. A-E correspond tothe five rows of larger vibrissae located on the mystacialpad. The vibrissae are also organized in the dorsal/ventralplane into seven arcs (1-7). The longest, caudal-most vi-brissae lie between rows and are referred to as "straddlers"(a-5). The furry buccal pad (FBP) is densely populated withnormal pelagic fur as well as vibrissae; however, only vi-brissae follicles contain blood sinuses, which are made vis-ible in this figure by staining with xylene. NV, nasal vibris-sae; NS, nostril; R, rostral; V, ventral. (Figure freelyavailable online from Barrels Web: http://www.neuro-bio.pitt.edu/barrels/pics.htm; courtesy of S. Haidarliu andE. Ahissar.)

(nerve V). The nerve is divided into fascicles,each of which innervates a row of vibrissae(Dorfl, 1985), indicating that vibrissae arefunctionally more closely affiliated withinrows than they are between rows (Simons,1983, 1985). Each vibrissa is exclusively inner-vated by as many as 200 large, myelinated sen-sory nerve endings that originate from cell bod-ies in the trigeminal ganglion (Vincent, 1913;Zucker and Welker, 1969), thereby differenti-ating these sensitive tactile structures fromnormal hair, which is sparsely innervated.

GROWTH AND REPLACEMENT

It is common in behavioral and neural scienceto establish the role of a particular structureor system by assessing the functional conse-quences of its removal or inactivation. A dis-

tinct advantage of the vibrissal sensory systemis that the vibrissae are continuously replacedthroughout the animals life and they effec-tively regenerate after they are plucked ortrimmed (Oliver, 1966), thereby providing thelaboratory researcher with a significant degreeof control over experimental parameters.

Vibrissae are present when rats are bornand reach their adult length in the secondpostnatal month (Ibrahim and Wright, 1975).The short, rostral vibrissae grow at a rate ofabout 0.5 mm/day, while the longest vibris-sae grow at a rate of 1.5 to 2 mm/day. Re-gardless of their position on the face, all vib-rissae require approximately 4 weeks to reachtheir maximum length, after which a new vib-rissa appears from the same follicle approxi-mately 1 week later. When it achieves a lengthof one-half to three-fourths of its final length,the old vibrissa falls out and its function istaken over by the new whisker (Ibrahim andWright, 1975). Trimming the old vibrissa at

Figure 8-3. Functional compartments (barrels) in layer 4of the somatosensory cortex show a one-to-one correspon-dence to individual vibrissae on the contralateral face.These compartments are readily identifiable using simplehistological methods, in this case staining for synaptic zinc(see Land and Akhtar, 1999 for methods), which exempli-fies the utility of the vibrissal sensory system in empiricalstudies assessing the relationship between brain structureand brain function. PMBSF, the posteromedial barrel sub-field, corresponds to the cortical representation of themystacial macrovibrissae. ALBSF, the anterolateral barrelsubfield, Corresponds to the cortical representation of themicrovibrissae. H, hindlimb; F, forelimb; L, lower lip.

Page 101: 0195162854

84 SENSORY SYSTEMS

or above the level of the skin does not affectthe rate of its growth or that of the newwhisker. However, when a whisker is pluckedfrom the follicle, a new whisker begins togrow immediately thereafter (Oliver, 1966).

In total, the life cycle of an individual vi-brissa is approximately 2 months.

MOTOR ASPECTS OFVIBRISSAE FUNCTION

In all mammals, tactile sensitivity is en-hanced by the relative movement of an ob-ject across the receptor periphery. By the sec-ond postnatal week of life (Welker, 1964),the rat's mystacial vibrissae are active duringexploratory and discriminative behaviors,with individual whiskers serving as elementsin a receptive array, scanning across objectsurfaces.

Three distinct behavioral states havebeen described in relation to whisking behav-ior. The first is quiet behavior, where whiskermovement is absent while the animals arestanding or sitting still (Fanselow andNicolelis, 1999). The second, referred to aswhisker-twitching behavior, also occurs duringimmobility but involves small-amplituderhythmic movements at a rate of 7 to 12 Hz(Semba and Komisaruk, 1984). Finally, duringactive exploration, the whiskers are rhythmi-cally swept forward (protraction) and back-ward (retraction) (Welker, 1964) at a frequencyof 5 to 9 Hz (Carvell and Simons, 1990; Bergand Kleinfeld, 2003), with these movementsfinely coordinated with sniffing and discretehead movements (Welker, 1964). A whiskingcycle lasts around 120 milliseconds, with two-thirds of it involving whisker protraction. Al-though the potential exists for individual cau-dal vibrissae to move independently of oneanother (Sachdev et al., 2002), for the mostpart, all mystacial vibrissae are observed tomove as a single unit, with bilateral symme-try (Vincent, 1912). Protractions are producedby contractions of muscles that form a sling

around the base of each whisker follicle. Re-traction is a passive process, largely mediatedby the viscoelastic properties of facial tissues(Dorfl, 1982; Carvell et al., 1991), but it can beactively assisted by muscles that move the un-derlying mystacial pad (Berg and Kleinfeld,2003).

Whisking movements provide a meansfor actively sensing the proximate environ-ment directly in front of the rat. Because mostsensory receptors respond preferentially tochanges in the sensed signal, active movementof the sensory organs enhances these changes,even when the environment is stationary.Whisking also facilitates hyperacuity, a processthat enables this sensory system to achieve ahigher effective resolution than is allowed bythe peripheral receptor size or spacing.

The muscles that control individual vi-brissae and mystacial pad movements are in-nervated by the buccal branch of the facialmotor nerve (nerve VII) (Dorfl, 1985). Thisnerve can be readily transected without af-fecting the sensory innervation of the vibris-sae (Semba and Egger, 1986). This paradigmhas been proved useful for assessment of therole of whisking in the tactile function of vi-brissae (Krupa et al., 2001) and in the devel-opment of the vibrissa sensory system(Nicolelis et al., 1996). The results of thesestudies indicate that the ability of rats to dis-criminate objects at high resolution, by usingonly their vibrissae, is dependent on intactwhisking behavior. Furthermore, these stud-ies have shown that the development of nor-mal tactile perception is dependent on intactwhisking behavior during early stages of de-velopment. As such, whisking performs thecritically important function of creating mo-tion between the vibrissa hair shaft and ob-jects that it contacts. Rats whisk as they probetheir immediate environment for the pres-ence of objects, obstacles, or food. This care-fully regulated motor action is, therefore,linked intimately to the acquisition and pro-cessing of ecologically important tactile sen-sory information.

Page 102: 0195162854

Chapter 8. Vibrissae 85

METHODS USED TO ANALYZETHE KINEMATICS OF WHISKING

The most effective means of observing andmeasuring the dynamic activity of vibrissae dur-ing exploratory whisking behaviors is by usinghigh-speed recording techniques. Early, low-res-olution analyses of vibrissae movements weredescribed by recording the rat with motion pic-ture film as it was engaged in exploratory be-havior (Welker, 1964). Videographic analysisreplaced film as photographic technologiesadvanced, however, both the spatial and tem-poral resolutions of videography are low, andthe quantitative analysis of these data is ex-tremely time consuming (Carvell and Simons,1990). On-line tracking of individual whiskermovements using optoelectronic and piezo-electric methods, which have very high spatial(26 /mi) and temporal (1 millisecond) resolu-tion, have provided significant advances in thekinematic analysis of vibrissae movements(Bermejo et al., 1998; Bermejo and Zeigler, 2000;Bermejo et al., 2002). However, behavioralanalysis is limited by the fact that the head ofthe rat must be fixed to isolate movements ofthe vibrissae alone. The continued developmentof high-speed digital video cameras now permitsmonitoring of whisker displacement in freelymoving animals along two dimensions (fromthe side and overhead) at a rate of as high as1000 frames per second using a shutter speed of200 microseconds (Hartmann et al., 2003). Thislevel of resolution allows the analysis of whiskermovements in behaving animals at frequenciesfar above typical whisking ranges, including res-onant frequencies, which are believed to be re-quired for texture discrimination (Hartmann etal., 2003; Neimark et al., 2003).

SENSORY ASPECTS OFVIBRISSAE FUNCTION

The important role of vibrissal sensation inguiding rat behavior was first systematicallyinvestigated in rats by Vincent (1912), who as-

sessed their ability to learn and navigate amaze after discrete sensory manipulations.The strategy that she and most of her succes-sors have used was to establish the normalfunction of vibrissae by assessing changes inbehavior that followed vibrissal removal.Moreover, it was readily apparent to her thatrats do not use the vibrissal sense in exclusionof other sensory systems. It is, therefore, nec-essary to exclude the potential contribution ofthe other senses in guiding behavior to assessthe particular role of vibrissae. By systemati-cally removing the vibrissae, the eyes, and theolfactory bulbs individually or in combina-tion, Vincent was able to conclude that thevibrissae are "delicate tactile organs, whichfunction in equilibrium, locomotion, and thediscrimination of surfaces in distinct ways ..."(1913, p. 69). Most of her conclusions havebeen confirmed and reaffirmed over the yearsby the few stalwart researchers who have alsotaken up this area of study. Many of the meth-ods have been refined or enhanced by tech-nological advances, and new details regardingthe role of vibrissae have emerged, but theconclusions are essentially the same—the ratuses its vibrissae to determine the position,size, texture, and shape (identity) of objects thatit encounters in its environment. The remain-der of this chapter details the most relevant ofthese studies that have been applied in an at-tempt to define the role of sensory processingby the vibrissal system in these functions.

OBJECT POSITION

The deflection of vibrissae in the sensory ar-ray, individually or in combination, encodesthe spatial location or position of objects, ver-tically and horizontally, in vibrissal space. Ahead-orientating movement seen in responseto active palpation of the vibrissae with a finewooden dowel is a simple and effective wayof assessing general sensitivity of the vibrissalsystem. Another simple task, the forelimbplacing task, takes advantage of the instinctualresponse that a rat makes, when suspended by

Page 103: 0195162854

86 SENSORY SYSTEMS

the torso with the limbs hanging freely, whenthe vibrissae touch a surface. Animals with anintact and functioning vibrissal sensory sys-tem reach with the forelimb ipsilateral to thestimulated vibrissae and place it on the sur-face of the object. Animals with lesions of thewhisker to cortex sensory pathway show animpaired response or none at all.

DISTANCE DETECTION

Rats can determine their distance from an ob-ject relying solely on information provided bythe whiskers. This ability is exemplified by asimple task that requires a rat to traverse twoelevated platforms that are separated by avariable sized gap to obtain a food reward(Hutson and Masterton, 1986). In the gap-crossing task, it is necessary for the animals tobe blinded, blindfolded, or tested in the dark,to restrict the salient cues to that solely pro-vided by vibrissae sensation. The rats aretrained to cross a gap that is widened at 1-cmintervals until they can no longer step acrossit. At this point the animals must stretch their

Figure 8-4. Rats use their vibrissae to scan the proximateenvkonment for objects and to determine their position,size, texture, and shape. Here, a rat whisks its vibrissaeacross a textured stimulus to make the appropriate dis-crimination in a forced choice test and then jumps to thecorrect platform to receive a food reward.

bodies over the gap and extend their vibrissaeto make contact with the reward platform andthen to use this information to accuratelyguide their leap. A "feel-before-jumping" be-havior must be reinforced in the first few daysof training, by widening the runway gap to 60cm. The rats cannot jump this distance, and ifthey try, they fall more than 32 cm to thebenchtop. Trained rats with an intact vibris-sae system can reliably span gaps of 16 cm.This distance is reduced to that spannable byits nose (—13 cm) when all of the vibrissae areremoved (Hutson and Masterton, 1986), orthe anatomical pathway from the periphery tothe cortical representation of the vibrissae isdamaged (Hutson and Masterton, 1986; Jen-kinson and Glickstein, 2000). Removal of sub-sets of the vibrissae reduces the spannablegap distance to that sensed by the remainingvibrissae.

DETERMINATION OF OBJECT SIZE

A novel and ingenious behavioral task that hasbeen used to determine the rat's ability to dis-criminate the size of an aperture using onlytheir vibrissae was described by Krupa et al.(2001). The apparatus consists of a large re-ward area connected to a small discriminationchamber by a small passage. Animals aretrained to discriminate the size of an aperturein the discrimination chamber, using onlytheir large facial vibrissae, and to indicate theirchoice (wide or narrow) by poking their noseinto either one of two spatially distinct sen-sors in the reward chamber. The animals areput on a water-restriction schedule so thatcorrect responses can be reinforced by a wa-ter reward. The difference between wide andnarrow apertures is systematically reduced tosensitively assess the rat's ability to undertakefine-grain distance detection. Rats were foundto be able to discriminate between small dif-ferences in aperture width (3 mm) after 30training sessions. By trimming the largemystacial vibrissae or inactivating the barrelcortex, the authors were able to establish that

Page 104: 0195162854

Chapter 8. Vibrissae 87

the ability to discriminate distances was de-pendent on an intact sensory pathway fromthe periphery to the cortex. However, aboli-tion of active whisking by transection of thefacial nerve had no deleterious effect, indicat-ing that movement of the body through spaceprovides sufficient mechanical activation ofthe appropriate vibrissae to make aperturesize discriminations.

TEXTURE DISCRIMINATION

We explore the texture of an object by mov-ing our fingers across it. Rats scan their envi-ronment and objects within it by moving theirvibrissae. The presence of an object or ridges,grooves, and irregularities across its surfaceare transduced by deflections of the vibrissaeshaft into signals from the sensory receptorsin the follicle.

Vibrissa-based texture discrimination inrats has been studied using several means, butthe simplest and easiest task was developed al-most 100 years ago (Richardson, 1909), withvariations on the design still effectively usedtoday (Hutson and Masterton, 1986; Guic-Robles et al., 1989; Carvell and Simons, 1990;Prigg et al., 2002). In their basic form, tests oftexture discrimination force the rat to make achoice between two stimuli varying only inthe degree of roughness. The apparatus con-sists of a start platform and two choice plat-forms separated by an adjustable gap. Duringeach trial, rats are required to stretch acrossthe gap from the start platform to palpate thesurface of a stimulus that is attached to thefront of the choice platforms (as in Fig. 8-4).The length of the gap is adjusted so that thestimulus can be palpated only by the mysta-cial vibrissae and the animals are forced tomake a choice by jumping to the correct plat-form. Rats are fitted with blindfolds (Carvelland Simons, 1990), binocularly occluded(Guic-Robles et al., 1989), or tested in the dark(with infrared light) to restrict the sensoriumto only vibrissal sensation. If they make a cor-rect choice, the food-restricted animals are

availed with a food reward that is concealedbehind a door on the correct platform. Ratstrained in this task have been shown to reli-ably detect differences in surface textures assmall as 30 /urn spaced at 90 fjim intervals, ata level comparable to primates using their fin-gertips (Carvell and Simons, 1990). Whentheir whiskers are trimmed, accuracy falls tochance levels with even the coarsest discrim-inanda (Guic-Robles et al., 1989).

OBJECT RECOGNITION

Lashley (1950) first suggested that vibrissal sen-sation might endow the rat with the ability ofform recognition. Although this hypothesishas often been restated, direct empirical con-firmation of this contention was absent untilvery recently. In 1997, Brecht et al. describedan object discrimination task that requiredsighted or blind rats to discriminate betweenan array of cookies of varying sizes and geo-metric forms using their vibrissae. Cookieswere either sweetened or embittered with caf-feine, which is odorless, to train the desired dis-crimination. Testing of sighted animals wasperformed under infrared illumination or in to-tal darkness. The target cookie (small, sweettriangle) was presented in random positionsamong 15 distractors in a 4 X 4 array. By ob-serving whisker movements and behavioralchanges after selective whisker removal, theauthors were able to doubly dissociate two dis-tinct vibrissal systems. They determined thatthe longer, laterally oriented macrovibrissaewere critically involved in spatial tasks, such asdistance detection, but were not necessary forobject recognition. On the other hand, the mi-crovibrissae were found to be essential for ob-ject recognition but not for spatial tasks.

VIBR1SSA SYSTEM PLASTICITY

The behavioral tasks that are described in thischapter provide the means with which to de-termine the functional capabilities of the rat's

Page 105: 0195162854

88 SENSORY SYSTEMS

vibrissa sensory system. Behavioral studies ap-plied to the vibrissal system are vastly out-numbered by those detailing its anatomy andphysiology. The amalgamation of these areasand their application to complex behavioralprocesses such as perceptual learning andmemory (Harris et al., 1999; Harris and Dia-mond, 2000), or the role of vibrissal sensationin activity- and experience-dependent modifi-cations of central neuronal networks, duringdevelopment and in adulthood (see Fox, 2002,for a review), is a large and growing field andholds promise for keeping behavioral scien-tists busy for decades to come.

CONCLUSIONS

Rats can use their mystacial vibrissae to per-form a variety of tactile discriminations andbehaviors. Using behavioral tasks, the vibrissalarray has been demonstrated to function as anactive, skin-like receptive surface used for fine-grained texture discriminations. In addition,the vibrissae can be likened to a retina-like sen-sor that uses the large, peripheral macrovibris-sae to detect coarse components of the sensoryfield, such as object location and distance, whilethe microvibrissae are analogous to the retinalfovea, involved in fine-grained resolution re-quired for the accurate discrimination of ob-jects. Because of the one-to-one correspondenceof a vibrissa with its cognate cortical barrel (Fig.8-3), the rat vibrissal system provides an excel-lent system for studies designed to understandthe relationship between sensory function andits central nervous system representation.

BIBLIOGRAPHY

Berg RW and Kleinfeld D (2003) Rhythmic whisking byrat: retraction as well as protraction of the vibrissaeis under active muscular control. Journal of Neuro-physiology 89:104-117.

Bermejo R and Zeigler HP (2000) "Real-time" monitor-ing of vibrissa contacts during rodent whisking. So-matosensory and Motor Research 17:373-377.

Bermejo R, Houben D, Zeigler HP (1998) Optoelec-tronic monitoring of individual whisker movementsin rats. Journal of Neuroscience Methods 83:89-96.

Bermejo R, Vyas A, Zeigler HP (2002) Topography ofrodent whisking—I. Two-dimensional monitoringof whisker movements. Somatosensory and MotorResearch 19:341-346.

Brecht M, Preilowski B, Merzenich MM (1997) Func-tional architecture of the mystacial vibrissae. Be-havioral Brain Research 84:81-97.

Broughton SD (1823) On the use of whiskers in felineand other animals. London Medical and PhysicalJournal 49:397-398.

Carvell GE and Simons DJ (1990) Biometric analyses ofvibrissal tactile discrimination in the rat. Journal ofNeuroscience 10:2638-2648.

Carvell GE, Simons DJ, Lichtenstein SH, Bryant P (1991)Electromyographic activity of mystacial pad mus-culature during whisking behavior in the rat. So-matosensory and Motor Research 8:159-164.

Dorfl J (1982) The musculature of the mystacial vibris-sae of the white mouse. Journal of Anatomy 135:147-154.

Dorfl J (1985) The innervation of the mystacial regionof the white mouse. Journal of Anatomy 142:173-184.

Fanselow EE and Nicolelis MA (1999) Behavioral mod-ulation of tactile responses in the rat somatosensorysystem. Journal of Neuroscience 19:7603-7616.

Fox K (2002) Anatomical pathways and molecular mech-anisms for plasticity in the barrel cortex. Neuro-science 111:799-814.

Guic-Robles E, Valdivieso C, Guajardo G (1989) Rats canlearn a roughness discrimination using only theirvibrissal system. Behavioral Brain Research 31:285-289.

GustafsonJW and Felbain-Keramidas SL (1977) Behav-ioral and neural approaches to the function of themystacial vibrissae. Psychological Bulletin 84:477-488.

Harris JA and Diamond ME (2000) Ipsilateral and con-tralateral transfer of tactile learning. Neuroreport11:263-266.

Harris JA, Petersen RS, Diamond ME (1999) Distribu-tion of tactile learning and its neural basis. Pro-ceedings of the National Academy of Science, USA96:7587-7591.

Hartmann MJ, Johnson NJ, Blythe Towel R, Assad C(2003) Mechanical characteristics of rat vibrissae:resonant frequencies and damping in isolatedwhiskers and in the awake behaving animal. Jour-nal of Neuroscience 23:6510-6519.

Hutson KA and Masterton RB (1986) The sensory con-tribution of a single vibrissa's cortical barrel. Jour-nal of Neurophysiology 56:1196-1223.

Page 106: 0195162854

Chapter 8. Vibrissae 89

Ibrahim L and Wright EA (1975) The growth of rats andmice vibrissae under normal and some abnormalconditions. Journal of Embryology and Experimen-tal Morphology 33:831-844.

Jenkinson EW and Glickstein M (2000) Whiskers, bar-rels, and cortical efferent pathways in gap crossingby rats. Journal of Neurophysiology 84:1781-1789.

Jones EG and Diamond IT (eds.) (1995) The barrel cor-tex of rodents. New York: Plenum Press.

Krupa DJ, Matell MS, Brisben AJ, Oliveira LM, NicolelisMA (2001) Behavioral properties of the trigeminalsomatosensory system in rats performing whisker-dependent tactile discriminations. Journal of Neu-roscience 21:5752-5763.

Land PW and Akhtar ND (1999) Experience-dependentalteration of synaptic zinc in rat somatosensory bar-rel cortex. Somatosensory and Motor Research16:139-150.

Lashley K (1950) Personal communication. In: Hand-book of psychological research on the rat (MunnNL, ed.). New York: Houghton-Mifflin.

Neimark MA, Andermann ML, Hopfield JJ, Moore CI(2003) Vibrissa resonance as a transductino mecha-nism for tactile encoding. Journal of Neuroscience23:6499-6509.

Nicolelis MA, De Oliveira LM, Lin RC, ChapinJK(1996)Active tactile exploration influences the functionalmaturation of the somatosensory system. Journal ofNeurophysiology 75:2192-2196.

Oliver RF (1966) Histological studies of whisker regen-eration in the hooded rat. Journal of Embryologyand Experimental Morphology 16:231-244.

Pocock RJ (1914) On the facial vibrissae in the mam-malia. Proceedings of the Zoological Society ofLondon 889-912.

Prigg T, Goldreich D, Carvell GE, Simons DJ (2002) Tex-ture discrimination and unit recordings in the ratwhisker/barrel system. Physiology and Behavior77:671-675.

Richardson F (1909) A study of sensory control in therat. Psychology Reviews Monograph Supplement12:1-124.

Sachdev RN, Sato T, Ebner FF (2002) Divergent move-ment of adjacent whiskers. Journal of Neurophysi-ology 87:1440-1448.

Semba K and Komisaruk BR (1984) Neural substrates oftwo different rhythmical vibrissal movements in therat. Neuroscience 12:761-774.

Semba K and Egger MD (1986) The facial "motor" nerveof the rat: control of vibrissal movement and ex-amination of motor and sensory components. Jour-nal of Comparative Neurology 247:144-158.

Simons DJ (1983) Multi-whisker stimulation and its ef-fects on vibrissa units in rat SmI barrel cortex. BrainResearch 276:178-182.

Simons DJ (1985) Temporal and spatial integration inthe rat SI vibrissa cortex. Journal of Neurophysiol-ogy 54:615-635.

Sokolov VE and Kulikov VF (1987) The structure andfunction of the vibrissal apparaus in some rodents.Mammalia 51:125-138.

Vincent SB (1912) The functions of the vibrissae in the be-havior of the white rat. Behavior Monographs 1:7-81.

Vincent SB (1913) The tactile hair of the white rat. Jour-nal of Comparative Neurology 23:1-36.

Welker WI (1964) Analysis of sniffing of the albino rat.Behaviour 12:223-244.

Zucker E and Welker WI (1969) Coding of somatic sen-sory input by vibrissae neurons in the rat's trigem-inal ganglion. Brain Research 12:138-156.

Page 107: 0195162854

Olfaction

BURTON SLOTNICK, HEATHER SCHELLINCK,AND RICHARD BROWN

9

RATS ARE MACROSMATIC

Mammals that have a well-developed olfactorysystem and a keen sense of smell and thatlargely depend on olfactory cues for operatingin their environment are classified as macros-matic (Moulton, 1967; Rouquier et al., 2000).This group comprises virtually all mammals ex-cept primates. Although some macrosmaticmammals, such as bats and chinchillas, haveother specialized sensory systems for detectingprey or avoiding predators, for the most part,their single most important exteroceptive senseis smell. In the rat and other rodents, the ol-factory epithelium is complex and denselypacked with olfactory sensory neurons, the ol-factory bulbs are large relative to the rest of theforebrain, and projections from the olfactorybulb not only terminate in essentially all of thecortex below the rhinal fissure, a cortical fieldthat is probably the largest sensory cortex inthe rodent brain, but also extend into the amyg-dala and hypothalamus. All of these anatomi-cal features are much reduced in the brain ofprimates that, in contrast to rodents, have awell-developed visual system. To borrow aphrase from Freud, anatomy, in this regard,is destiny and virtually every aspect ofrodent life is guided by and often largely de-pendent on the sense of smell.

The extent to which smell dominates thelife of the rat may not be fully appreciated byinvestigators who use primarily nonolfactorystimuli in studies of learning. Olfaction is thefirst sensory system to become functional, and

even prenatal or perinatal exposure to odor-ants can affect later postnatal behavior (e.g.,Pederson and Blass, 1982; Smotherman, 1982;Hepper, 1990; Hudson, 1993; Abate et al.,2002). Infant rodents are easily conditioned toapproach or avoid odors (Sullivan and Wilson,1991) and quickly learn to prefer the familiarodors of their mother and nesting materialand avoid novel odors, and odors associatedwith foodstuffs very early in life can have long-term effects on food preference. As adults, ratscommunicate largely by odors; they leaveodor marks and trails that identify their age,sex, territory, and dominance status (Rainey,1956; Brown, 1985; Galef and Buckley, 1996).These odor signals are released from special-ized scent glands and are in urine and fecesand even in the animal's breath. Indeed, a con-siderable literature attests to the influence ofbreath odors in social transmission of food pref-erences (see Chapter 34). Perhaps more famil-iar to most investigators are the "Bruce effect,"the "Whitten effect," and the "Vandenberg ef-fect," phenomena primarily demonstrated inmice, that reveal dramatic changes in behav-ior and reproductive status produced by ex-posure to a conspecific's odor stimuli. A num-ber of other behavioral effects largely orentirely dependent on olfactory cues, includ-ing the communication of dominance and sex-ual status, nesting behavior, kin recognition,and, in the neonate, nipple attachment andsuckling, are not graced by the names of in-vestigators but provide additional evidencefor a dominant role of olfaction in rodents.

90

Page 108: 0195162854

Chapter 9. Olfaction 91

Brown (1979) classified these various sig-nals into two categories: those that reflect theemotional status or level or arousal of the in-dividual and those that identify the individual.Emotive odors are those produced or releasedonly in special circumstances such as sexualarousal, maternal behavior, or fear. Identifierodors are those produced by the body's nor-mal metabolic processes that are stable overtime. The latter may be influenced by dietaryconditions (Schellinck et al., 1997) but appearlargely determined by genetic factors. A ma-jor effort in contemporary olfactory researchhas focused on determining both the geneticorigin and the chemical expression of these ol-factory cues. Behavioral, genetic, and bio-chemical analyses have revealed that both themajor histocompatibiity complex (Boyse etal., 1991; Schellinck etal., 1995; Schellinck andBrown, 1999; Schaefer et al., 2002; Beauchampand Yamazaki, 2003) and the major urinaryproteins (Brennan et al., 1999; Hurst et al.,2001; Nevison et al., 2003) appear to be thesource of a unique chemical signature that canbe learned and remembered by rodents.

Recently, a quite different line of olfac-tory studies emerged that were driven by twosets of findings. First, when provided withodor cues, rats showed a remarkable ability tolearn simple and complex discrimination tasks(Dusek and Eichenbaum, 1997; Slotnick,2001). Second, a series of molecular biologicaland anatomical studies identified olfactory re-ceptor genes and the organizational principlesgoverning the projections of olfactory sensoryneurons onto the olfactory bulb (Buck, 1996;Mombaerts et al., 1996). These later findingsprovided the basis not only for a now widelyaccepted view of how odors are coded by thebrain but also the generation of a variety ofgene-targeted mice in which selected featuresof the olfactory system are altered. Behavioralstudies are being designed to exploit thesefindings, to assess potential changes in olfac-tion in genetically modified mice (e.g., Zufalland Munger, 2001), and to test hypothesesgenerated by advances in the molecular biol-

ogy of the olfactory system. Such investiga-tions require the development of sophisti-cated methods for generating and controllingodor stimuli together with behavioral tests forpsychophysical analyses of olfactory function.

SPECIAL PROBLEMS INCONTROLLING THE STIMULUS

The generation, control, and measurement ofodors present the investigator with specialproblems, which are not encountered indealing with stimuli for other modalities(Dravnieks, 1972, 1975). For mammals, odorstimuli are gases, generated by vaporizationfrom some odorant substance. Most naturalodors come from a complex source whosecomponents are generally difficult to identifybut vaporize at different rates, resulting in astimulus whose constituents vary over time.The contribution of each component of thiscomplex will be a function of individualmasses and vapor pressure of each, the partialpressure of the component in the odorantmixture, the substrate on which it is de-posited, relative humidity, and, of course, theextent to which each component changeswith changes in temperature. Unless con-trolled, the vapor itself will diffuse into the at-mosphere, perhaps form "odor plumes" car-ried by prevailing air currents or simplybecome increasingly less concentrated as dis-tance from the point source increases.

Unfortunately, there are no simple de-vices to measure odor concentration or iden-tify the components of a vapor or, for psy-chophysical tests, to generate "pure" odorstimuli or, for that matter, nonodorous stim-uli. Even tubing systems designed to controlvapor flow must contend with a variety ofproblems, including determining optimalflow parameters, potential contamination bycomponent materials, and adsorption of odor-ant molecules on tubing walls. A further com-plication, and often the primary topic of a re-search endeavor, is the biological constraints

Page 109: 0195162854

92 SENSORY SYSTEMS

on the effectiveness of vapors in stimulatingolfactory receptor neurons. Thus, sensory neu-rons maybe relatively unresponsive to the ma-jor component of a vapor complex but be ex-quisitely sensitive to quite minor components.

BEHAVIORAL METHODS: DIFFUSIONOF ODORS FROM A POINT SOURCE

Olfactory studies have used a very wide vari-ety of methods and these lend themselves toa number of potential classificatory schemes.Because of the unique problems associatedwith controlling odor stimuli, we considerseparately methods used in those studies thatexert little or no control of the stimulus, al-lowing the odor to passively diffuse from apoint source, and those that use olfactomet-ric devices in an attempt to generate, control,and deliver defined quantities of the stimulus.

Behavioral methods used in early studiesof olfaction in rats were relatively simple and,for the most part, the effects obtained werequite strong and did not require either precisecontrol of the stimulus or control of stimulussampling by the subject. Such methods con-tinue to be used in studies examining learningand memory in rats. Nonetheless, the oftenidiosyncratic methods and limited control ofthe stimulus used in such studies virtually pre-clude replication from one laboratory to an-other or meaningful comparisons among theiroutcomes. A description of the some of thesetests and their potential limitations follows.

HABITUATION TESTS

These tests are based on the observation thatsuccessive presentations of the same stimulusodor will result in a decrease in investigatorybehavior, that is, habituation. Then, when adifferent odor stimulus is presented to thesame subject, the habituated response will re-occur or will be "dishabituated." This behav-ioral test has been used to determine if thesecond stimulus is perceived as different from

first. The tests are of two types: the habituation-discrimination test and the habituation-dishabituation test.

In the habituation-discrimination test,one odor is presented a number of times, usu-ally in a small odor pot or on a filter paper,and the response to it is measured; then, theinitial stimulus and a second novel odor arepresented simultaneously. If the rat investi-gates the novel odor more often than the orig-inal stimulus, one can conclude that it is ableto discriminate between the two odors. Thus,the test provides a simple assessment of odormemory. The time between the first and sec-ond phase of the test varies depending on thenature of the research question. The test hasbeen used frequently to assess memory forthe odors of conspecifics. In some instances,rather than presenting odors, the whole ani-mal is presented. This form of the test is of-ten referred to as the social recognition test(Bhutta et al., 2001).

The procedure in the habituation-disha-bituation test is similar to that just describedexcept that only one odor is presented at atime in both phases of the task. The habitua-tion phase involves repeated presentations ofone odor, and in the dishabituation phase,samples of a second odor are presented. Insome instances, the test begins with a no-odor-adaptation period during which the subject isput in the test chamber and presented withthe odor vehicle several times (Brown, 1988;Schellinck et al. 1995). Sundberg et al. (1982)devised a standard approach to scoring the be-havioral response of rats during initial inves-tigation and habituation of the odors. Thismethod has been used extensively to provideevidence that congenic strains of rats that dif-fer only in the genes of the major histocom-patibility complex produce discriminably dif-ferent urine odors.

The principal benefit of both of thesetests arises from the simple design, minimalneed for equipment, and speed and ease oftesting. An advantage of the habituation-dishabituation task is that it eliminates the

Page 110: 0195162854

Chapter 9. Olfaction 93

problems associated with creating a mixture oftest odors. Nonetheless, both forms of the testare somewhat time consuming, require con-tinuous observation of the subjects and themeasurement of a not always well-defined "in-vestigative behavior." Moreover, a null out-come is open to several interpretations. Whendishabituation occurs, it seems clear that thetwo odors can be discriminated. The failure ofthe test stimulus to produce dishabituationcould reflect a failure to discriminate, a disin-terest in the test odor, or that the test odor isdiscriminate but not sufficiently different toproduce a clear dishabituation effect.

UNCONDITIONED PREFERENCE TESTS

These tests are analogous to two bottle pref-erence tests used in studies of taste. Quite sim-ply, two or more odors are placed in a testarena, and the frequency and/or duration ofa subject's approach to and investigation ofeach odor is recorded. Usually, the test cham-ber includes a neutral area in which the ani-mal is contained before exposing it to theodors. To reduce the possibility that the sub-ject's responses to the odors will be limitedbecause of either a neophobic response orgeneralized investigation, it is appropriate toinclude a habituation period to the test cham-ber before introducing the odor stimuli(Schellinck et al., 1995). The experimentergenerally records the time the subject spendsinvestigating each odor and thus should beblind with regard to experimental conditions.If possible, the experiment should be video-taped or a computerized tracking system usedso that rater reliability can be assessed. Al-though there is an increasing trend towardtesting animals in their home cages, this is notrecommended because rodents tend to sniff in-discriminately at all novel objects so presented.

Preference tests can be also be used withneonatal and juvenile rats. Because of theirlimited motor abilities, a simple two-choiceapparatus is used for testing pups. Often, thepups are placed on a mesh floor and the odors

diffuse through the mesh from containersplaced below. To avoid the recording of false-positive results because of chance move-ments, a middle neutral zone can be incorpo-rated into the test apparatus. Differences indefining what behavior constitutes a choicemakes it difficult to compare results amongstudies. For example, does turning the headtoward the odor constitute a preference, ormust half of the subject's body extend into thepreference zone?

Except when very strong preference ef-fects are obtained (e.g., Kavaliers and Os-senkopp, 2001), it is often unclear how the re-sults from such tests should be interpreted.Greater investigation of one odor may reflectan absolute or a relative preference for one ofthe test odors, the relative novelty of eachodor, a difference in detectability of the odors,or possibly an aversion to the least-sampledodor (Brown and Wilner, 1983; Amiri et al.,1998). A failure to display a preference mayresult from both odors being equally attrac-tive or aversive. A variety of factors that in-fluence activity, including the subject's circa-dian rhythm and the ambient temperature,may also affect test outcomes. The use of pref-erence tests also has a number of seriousshortcomings regarding the control of odordispersion. When two or more odors are pre-sented simultaneously, their vapors may mix,consequently increasing the difficulty in dis-criminating between them. Moreover, be-cause the odors will reach the subject overtime regardless of whether the animal inves-tigates them, there is a possibility that an an-imal need not approach an odor directly toshow a preference. It is also possible that therat may habituate to the odor before showinga measurable response.

SIMPLE ASSOCIATIVE LEARNING TESTS

The static presentation of odors has also beenused in associative learning paradigms withboth young and adult rats. In these tasks, oneodor is presented with reinforcement and a

Page 111: 0195162854

94 SENSORY SYSTEMS

second odor is presented without any associ-ated reward. A number of daily sessions areusually run in a quasi-randomized order for aspecified number of days. To assess the suc-cess of the task, preference tests may followodor conditioning. If an animal investigates orspends more time in the vicinity of the odorin the absence of reinforcement, this is con-sidered to be an indication of the effectivenessof the conditioning procedure. Johanson andTeicher (1980) paired an odor with oral infu-sions of milk in a warm environment to cre-ate a conditioned odor preference in neonatalrats. Sullivan, Leon, and colleagues (Sullivanet al., 1991, 1994; Johnson and Leon, 1996)used a similar conditioning paradigm to ex-amine the neurobiology of olfactory learning.They exposed neonatal rats to an odor pairedwith tactile stimulation (i.e., stroking with asable brush) that simulates maternal contact.A Y-maze preference test was used to deter-mine if the conditioning was successful.Schellinck and colleagues developed a simplediscrimination task to assess odor learning inadult mice and rats (Brennan et al., 1998; Fair-less and Schellinck, 2001; Forestell et al., 2001;Schellinck et al., 200Ib). The animal is pre-sented with a pot in which finely cut woodchips are diffused with an odor from a lowerlevel. One odor is paired with sugar rein-forcement buried in the woodchips (CS+stimulus) and a second odor is presented alone(CS— stimulus). In a subsequent preferencetest, the rodent is presented with both odorsbut no sugar. The animals only need to befood restricted before the preference test, andthe use of digging rather than sniffing providesan objective and easily scored measure of pref-erence. In the preference test, the subjects digconsistently and almost exclusively in thewoodchips containing the CS+ odor. Micerapidly learn the task and remember the dis-crimination for at least 90 days (Schellinck etal., 200la and unpublished observations).Studies to assess the long-term memory of ratsusing this task are yet to be completed. Theappetitive learning paradigms described here

are easy to set up and are useful for assessingthe neurobiological basis of learning and mem-ory (Wilson and Sullivan, 1994; Forestell et al.,1999, 2002). Given the extraordinary robust-ness of the effect, the test is not particularlyuseful for assessing higher cognitive functions.

Mazes were among the first tests used todetermine if rodents could learn to make a dis-crimination between the correct or reinforcedstimulus and the incorrect or unreinforcedstimulus (Bowers and Alexander, 1967). Theycontinue to be popular. For example, the ra-dial arm and more simple mazes have beenused to examine odor memory and, learning(Staubli et al., 1986; Reid and Morris, 1992;Steigerwald and Miller, 1997) and, in mice,odor-cued maze learning has been used in anextensive series of studies of volatile signals ofthe major histocompatibility complex (Singeret al., 1997; Yamazaki et al., 1979; Beauchampand Yamazaki, 2003). As discussed by Stevens(1975), a number of precautions should be ob-served in using mazes to study odor detection.The reinforcement itself should not be a cueas rats have been shown to discriminate be-tween the odor of a 45 mg food pellet andclean air in a T-maze (Southall and Long,1969). An enclosed start box should be usedto prevent the experimenter from uninten-tionally cueing the subject in any fashion, andthe experimenter should be unaware of the lo-cation of the correct stimulus. Rats have beenshown to track odors of themselves and otherrats (Wallace et al., 2002), so it is essential thatthe arms of the maze are cleaned between tri-als and experimental sessions. Clearly, unlessspecial precautions are taken, mazes are notwell suited for studies of odor learning. Fur-ther, it may be difficult to compare resultsfrom maze studies and those that use morerigorous methods. Thus, the failure of Reidand Morris (1992) to replicate aspects of ol-factory stimulus control originally demon-strated using an olfactometer (Slotnick andKatz, 1974; Nigrosh et al., 1975) was probablydue, in part, to shortcomings in using a mazeto study complex olfactory learning.

Page 112: 0195162854

Chapter 9. Olfaction 95

Rats are quite adept at digging in a sub-strate to obtain a food reward, and variousforms of digging tasks are often used to assessolfaction in both mice and rats (Berger-Sweeney et al., 1998; Zagreda et al., 1999; Mi-halick et al., 2000; Schellinck et al., 2001b). Theparadigm requires that a food-restricted animalfind food reinforcement by digging in a sub-strate of sand, gravel, or bedding material fora food particle. Finding buried food has beenwidely used as a simple measure of whethersome experimental procedure has disrupted oreliminated the sense of smell (Alberts andGalef, 1971; Hendricks et al., 1994; Center etal., 1996). Despite their popularity and simplic-ity, these tests have the various shortcomingscommon to those using uncontrolled diffusionof odors from a point source and, in addition,have proved to be a poor predictor of anosmia(Xu and Slotnick, 1999; Slotnick et al., 2000a).

Examples of better and more interestinguses of digging behavior as a measure of odordetection and discrimination are tests used byEichenbaum and associates. Rats are trainedto dig in a container of odorized sand for afood reward and later tested for their abilityto detect which of several containers has theodor target. Variants of these test methodshave been used to demonstrate complex odor-based learning in rats in which correct re-sponding is dependent on the configuration ofthe stimuli and not simply on associations be-tween individual stimuli and reinforcement(e.g., Dusek and Eichenbaum, 1997; Fortin etal., 2002; Van Elzakker et al., 2003).

Conditioned odor aversion (COA) andodor-cued taste avoidance are two other as-sociative learning methods for assessing odordetection and discrimination. Procedures inproducing a COA are analogous to those usedin the better known conditioned taste aver-sion paradigm. What is perhaps less wellknown is that odors can be effective condi-tioned stimuli for aversion learning and, whenpresented as intraoral stimuli, can supportsingle-trial and long-delay learning (Slotnick etal., 1997). Indeed, COA has been demon-

strated in rats of all ages, including those inthe prenatal and early postnatal stages of life(Rudy and Cheade, 1979, 1983; Smotherman,1982; Smith et al., 1993).

Odor-cued taste avoidance learning re-quires only that rats accustomed to drinkingfrom a spout be exposed on the training trialto a bitter and odorous solution. Learning oc-curs in only one or two trials, and in subse-quent tests, rats sniff at the drinking spout butdo not sample the liquid if the odor is detected(Darling and Slotnick, 1994). These tasks havean advantage over food-finding tasks thathave not been fully exploited: both are easyto implement, allow reasonably good controlof the odor stimulus, and support single-triallearning and long-term memory for odors.

OLFACTOMETRY

CONTROL OF THE STIMULUS

Olfactometers are devices that generate anodor whose concentration, flow parameters,and temporal parameters can be specified,controlled, and varied. When combined withoperant conditioning, olfactometers providepowerful tools for training rats to attend toand differentially respond to selected featuresof the stimulus. Although the term olfactome-ter is often applied to any device that allowssome control over odor flow, including mazesequipped with fans for directing air currents,those meeting these criteria are based largelyon the design principles first described byTucker (1963) and Moulton and Marshall(1976). These devices can generate odors ei-ther by a series of air dilutions of an odorantsaturated vapor or from the headspace of anodorant dissolved in an odorless liquid. Thelatter method is by far the easiest to instru-ment and has the advantage that multiplechannels, each with its own odorant solution,can be incorporated.

Olfactometers for use with rats haveevolved from relatively crude devices (Williams

Page 113: 0195162854

96 SENSORY SYSTEMS

and Slotnick, 1970) or complex multicompo-nent devices (Bennett, 1968; Sakellaris, 1972)to simpler and computer-controlled designsthat are easily cleaned and maintained. Olfac-tometers and training methods designed forresearch with rodents have been describedin detail (Slotnick and Schellinck, 2002). Themultiple-channel liquid dilution system(Bodyak and Slotnick, 2000) is relatively easyto construct and provides the control neededfor studies of odor sampling behavior, simpleodor discrimination, discrimination betweenodor mixtures, absolute and intensity differ-ence detection thresholds for monomolecularodors, odor masking, odor memory, odorquality identification, and other odor sensorytasks that demand reasonably precise presen-tation of the stimulus. As described by Slot-nick and Schellinck (2002), the ease withwhich the multiple-channel liquid dilution canbe cleaned, together with the use of pinchvalves and disposable tubing and odor satura-tor containers, has greatly minimized odorcontamination of components, a problem thatplagues most olfactometer designs.

ODOR CONTROL OF BEHAVIOR

In a series of published and unpublished stud-ies, we assessed a variety of operant trainingprocedures and methods for presenting thestimulus. Initially, rats were trained in a windtunnel and were essentially immersed in thestimulus on each trial (Slotnick and Nigrosh,1974). When it became apparent that ratswere easily trained to sample odor stimuli, weused a simple Plexiglas chamber fitted with avertically oriented glass tube for presentingodors. Traditional operant shaping proce-dures were used to train the rat to sample anodor by inserting its snout into an opening inthe side of the tube. Snout insertions, detectedwith a photobeam, initiated a trial in whichodorized air was added to a constant streamof clean air for 1 to 2 seconds. A discrete tri-als, go, no-go procedure was used in which adesignated response made after sampling the

S+ stimulus was rewarded with water. A re-sponse made after sampling the S— stimuluswas neither reinforced nor punished. The re-sponse "manipulandum" was a stainless steelwater delivery tube located outside of theodor-sampling tube. Rats were required tosample the stimulus (i.e., keep their snout inthe odor sampling tube) for some minimumtime (generally 0.15 second) before respond-ing and shorter samples resulted in immedi-ately aborting the trial and repeating it on thenext trial. The designated response was mak-ing a fixed number of licks (generally 10 ormore) on the water delivery tube.

As described later, this training procedurewas remarkably effective in gaining stimuluscontrol of behavior. Other training methods,including using symmetrical reinforcementwith the go/no-go procedure or using twowater delivery tubes, each associated with adifferent odor, proved far more troublesomeand less efficient. We also rejected the use ofthe more traditional free operant discrimina-tion procedure because the extended expo-sure to the stimulus required to determine aresponse rate might produce adaptation. In aneffort to minimize adaptation effects we ini-tially used a 60- or 90-second intertrial inter-val. It quickly became apparent that a muchshorter intertrial interval was preferred byrats, and they worked far more efficientlywith the minimum interval we thought wasneeded for the clean air stream to clear theodor-sampling tube between trials («*5 sec-onds). With these parameters, well-motivatedrats completed a trial, on average, every 12 to15 seconds and could maintain essentially per-fect performance with suprathreshold stimuliover hundreds of trials. The number of trialsthat could be given in a session was limitedonly by satiation effects and, with reinforce-ment volume of 0.04 ml, 400 to 600 trialscould be run in daily sessions. Such multipletrials sessions are particularly useful for psy-chophysical studies.

Despite the fact that responding on S—trials (false alarms) was not punished, rats

Page 114: 0195162854

Chapter 9. Olfaction 97

quickly acquired odor detection and discrim-ination tasks, and it was clear that nonrein-forcement after responding was sufficientlyaversive to support response inhibition. Er-rors on S+ trials (misses) seldom occurredand, in general, more than 95% of errors werefalse alarms. Thus, acquisition functions werelargely a function of learning to inhibit re-sponding on S— trials.

Acquisition of simple odor discriminationtasks could be remarkably rapid (Nigrosh etal., 1975). For example, after extended train-ing on S+-only trials, subsequent acquisitionwas essentially errorless; rats did not respondor made only one or two responses to the S—stimulus when it was abruptly introduced.Even without initial training on the S+ stim-ulus, rats almost always achieved near perfectperformance on simple two-odor discrimina-tion tasks within 20 to 60 training trials. Ad-ditional evidence for strong stimulus controlby odors was that odor stimuli easily over-shadowed auditory or visual stimuli but thateven extended training on visual stimuli failedto disrupt acquisition of an odor discrimina-tion. Further, when rats were trained equallywell on both auditory and olfactory discrimi-nations, responses in stimulus competitiontests (e.g., pairing the S— auditory stimuluswith the S+ odor stimulus) were almost en-tirely determined by the sign value of the odorstimulus (Nigrosh et al., 1975).

After training on several odor discrimi-nation tasks, rats became quite sophisticatedobservers; when new stimuli were used, theycarefully sampled the odor on each trial andquickly learned to inhibit responding to thestimulus not associated with reward. We ex-ploited this finding in a series of learning-setstudies designed to determine if rats, like pri-mates, could acquire a win-stay/lose-shift re-sponse strategy when presented with a seriesof odor discrimination tasks. In both sequen-tial discrimination reversal tasks and whengiven a series of novel two-odor discrimina-tion tasks, the number of errors to a criterionof 90% correct responding in a block of 20 tri-

als rapidly decreased; by the end of training,most rats achieved criterion performance inthe first block of trials, and some made no er-rors at all (Slotnick and Katz, 1974; Nigrosh etal., 1975; Slotnick, 1984). This level of per-formance was all the more remarkable be-cause it was achieved after training on only 10to 20 problems. Thus, on learning set tasks,rats trained with odors performed at least aswell as did primates trained with visual stim-uli and, like primates, demonstrated the ac-quisition of a response strategy—in Harlow's(1949) terms, they learned to learn. In re-sponse to criticisms of this work by Reid andMorris (1992), Slotnick et al. (2000b) fullyreplicated these learning set outcomes usinga different olfactometer apparatus and a vari-ety of test procedures.

In related studies, it was shown that ratsreadily acquired both matching-to-sample andnon-matching-to-sample tasks when odorswere used (Otto and Eichenbaum, 1992; Lu etal., 1993) and even showed evidence of learn-ing to learn to match to sample (Lu et al.,1993). Their capacity to acquire many differ-ent odor discriminations and to rememberodors was challenged by tasks that requireddiscrimination among eight odors presentedin random order within a training session(Slotnick et al., 1991). Not only did rats quicklysort out which of these odors were associatedwith reinforcement and which were not, butalso, in repeated sessions using novel sets ofeight odors, their performance rapidly im-proved, and by the seventh or eighth such set,most needed only three to five exposures toeach odor to achieve criterion performance.When tested on a reversal task using odorsfrom a set in the middle of the series, theymade many errors, thus demonstrating mem-ory for this earlier set of odors.

OLFACTORY PSYCHOPHYSICS

Determination of absolute detection thresh-old and other measures of odor sensitivity re-quires the precise control of the stimulus con-

Page 115: 0195162854

98 SENSORY SYSTEMS

ferred by olfactometers. Psychophysical stud-ies with rats have generally used either amodified staircase psychophysical procedure(Youngentob et al., 1997) or, more com-monly, a modification of the descendingmethod of limits procedure in which rats re-ceive extensive training in detecting some ini-tial suprathreshold concentrations of thestimulus. Concentration, defined as the percentair or liquid dilution of the odorant source, isthen decreased in quarter-log or half-log steps,and training at each step is continued until theanimal fails to reach some criterion level (gen-erally 75% or 80% correct responding in ablock of 20 or 40 trials) in a fixed number oftrials. As might be expected, lower thresholdsare obtained using small changes in odorantconcentration and by overtraining at eachconcentration. Specifying the precise molecu-lar concentration of the stimulus is fraughtwith a variety of technical and methodologi-cal problems (Dravineks, 1975) and often re-quires assumptions about vapor saturationlevels and other factors that are, at best, diffi-cult to confirm. Fortunately, however, onlythe most demanding sensory studies requireprecise measures of the physical stimulus and,for most behavioral studies, the relative changein sensitivity produced by some experimentalmanipulation provides a perfectly acceptableend point. Thus, for example, Apfelbach et al.(1991) showed the extent to which odor sensi-tivity varied with both age and density of ol-factory receptor neurons, and Slotnick andSchoonover (1993) determined that transec-tion of the lateral olfactory tract in the rat de-creased odor sensitivity to amyl acetate by ap-proximately 2.2 orders of magnitude.

Olfactometers lend themselves to otheruseful measures of odor sensitivity, includingintensity difference threshold (Slotnick andPtak, 1977; Slotnick and Schoonover, 1993)and odor masking (Laing et al., 1989). In thelatter case, the rat may be required to dis-criminate between two odors, A and a com-bination of A and B. In subsequent sessions,the proportion of B in the mixture is gradu-

ally reduced until the discrimination can nolonger be made. Generally, at that point, theconcentration of B alone is well above detec-tion threshold but is masked by the presenceof the A stimulus. Variants of this discrimina-tion task have been used (e.g. Laing et al.,1989; Lu and Slotnick, 1998; Dhong et al.,1999; Slotnick and Bisulco, 2003) and the testcould easily be extended to examine discrim-ination of more complex mixtures of odors.

ODOR QUALITY PERCEPTION

Perhaps one of the most challenging tasks inexperimental studies of olfaction is assessingodor quality perception—determining the per-ceived similarity between odors and whethersome manipulation produces a change in theperceptual quality of an odor. Such issues areparticularly relevant now because contempo-rary molecular biological and anatomicalstudies provide an empirical basis for severaltheories of odor coding (Xu et al., 2000). Achallenge is to develop methods to test thesetheories at the level of behavior. The tradi-tional procedure, and probably still the goldstandard for measuring sensory quality per-ception in animals, is to determine a stimulusgeneralization gradient. But it is unclearwhether this method can be applied in the ol-factory modality because odors differ fromone another along numerous dimensions andit is likely that some, as yet unidentified, com-bination of these dimensions are the determi-nants of odor quality. Alternative methods forassessing odor quality perception in rats havebeen reported, but none are completely satis-factory. One unique approach is that of Youn-gentob and his associates (1990, 1991). Ratswere trained to sample an odor from a cen-tral port and then traverse a runway associ-ated with that odor for a water reward. Fiveodors were used, and each odor signaledwhich of five runways had to be selected. Itwas assumed that errors in response choiceswould reflect the extent to which rats con-

Page 116: 0195162854

Chapter 9. Olfaction

fused one odor with another and, hence, pro-vide a measure of the perceptual similaritybetween odors. However, in practice it wasfound that rats did not learn the task easily butthat once it was acquired, few errors weremade, even between structurally similarodors. A quite different approach is that ofSlotnick and Bodyak (2002), who trained ratsto discriminate among structurally homolo-gous odors and unrelated odors and latertested their memory for the positive and neg-ative assignments of these odors. It was as-sumed that more errors in the memory testwould be made on odors that were similarperceptually. The memory test was per-formed under extinction and rats had no feed-back for correct or incorrect responding. Con-trol rats in this study had near-perfect memoryscores for both the structurally homologousand unrelated odors. Rats with olfactory bulblesions performed almost as well and, hence,failed to show any marked effect of bulbar le-sions on odor quality perception. Althoughthis particular test proved useful for assessingpotential changes in perception resulting frombrain lesions, the excellent performance ofcontrol rats in the Slotnick and Bodyak study(2002) and in the Youngentob et al. studies(1990, 1991) indicate that different and moresensitive methods may be needed to indexodor quality perception in rats.

OLFACTION, ANOSMIA, ANDOTHER CHEMICAL SENSES

Olfactory sensory neurons are not the onlysensory neurons that respond to vapor stim-uli. Macrosmatic mammals possess a well-developed accessory olfactory system and sen-sory neurons in the vomeronasal organ mayhave quite low thresholds for odors (Leinders-Zufall et al., 2000). Indeed, determiningwhether the response to an odor is mediatedby the main or accessory olfactory system (orboth) is nontrivial and a matter of considerableinterest in studies of pheromonal stimuli. There

99

are several experimental methods for disassoci-ating the two systems: the vomeronasal organcan be removed without damage to the mainolfactory epithelium (Wysocki et al., 1991) orthe vomeronasal nerves, which are compactand travel along the medial aspect of the ol-factory bulbs, can be transected (Fleming etal., 1992). It is far more difficult to eliminatethe main olfactory system while leaving theaccessory system intact. The (main) olfactoryepithelium is extensive and an integral part ofthe respiratory system and it would be diffi-cult to completely ablate it without seriouscomplications. Olfactory nerves extend in adiffuse manner through the cribriform plate,making their transection problematical.Costanzo and colleagues (Costanzo, 1985; Yeeand Costanzo, 1995) have, however, describedthe use of a specially constructed knife to tran-sect these fibers, although that transectionmay also interrupt the accessory olfactorynerve. Some caustic agents and toxins may de-stroy olfactory sensory neurons but leave theaccessory olfactory system intact (Setzer andSlotnick, 1998; Slotnick et al., 2000a), but con-firmation of these outcomes requires carefulhistological analyses and effects may be criti-cally dose dependent.

Zinc sulfate is a caustic metallic salt thatdestroys epithelial tissue, and syringing thenasal vault with this agent has been used innumerous studies in an attempt to produceanosmia in the rat. However, as generally prac-ticed, the procedure is only partly effective indestroying the olfactory epithelium, and bothanatomical studies and sensitive olfactometrictests reveal residual unaffected areas of the ep-ithelium as well as considerable residual olfac-tory function in zinc sulfate-treated rats (Slot-nick and Gutman, 1977; Slotnick et al., 2000a).The only completely reliable method for pro-ducing a frank anosmia (complete loss ofsmell) in rodents is surgical removal of the ol-factory bulbs. Complete removal of the bulbsis essential because olfactory function may bemediated by even small remnants of olfactorybulb tissue (Lu and Slotnick, 1998). The sur-

Page 117: 0195162854

100 SENSORY SYSTEMS

gery is somewhat difficult because the poste-rior aspect of the olfactory bulbs lies under thefrontal pole cortex and some care is requiredto excise the bulbs completely without pro-ducing significant ancillary damage to sur-rounding structures. The lesions will neces-sarily invade the anterior olfactory nucleuswhose anterior pole extends into the lateralaspect of the bulb. Completely olfactory bul-bectomized adult rats have no deficits in per-forming the operant olfactometer task but re-spond at chance over hundreds of trials on asimple odor detection problem even whentested with relatively high concentrations ofan odor (Slotnick and Schoonover, 1992). Theemphasis here is on olfactory bulbectomy inthe adult rat because there is evidence for re-generation and functional projections to theforebrain in the olfactory bulbectomizedneonatal rat (Hendricks et al., 1994).

A potential confound in many studies ofolfaction arises from cues that may be medi-ated by trigeminal receptors in the nasal vaultor cornea, and sensory receptors in the tra-chea that can respond to certain vapors. Ofthese, a contribution of corneal receptors andthose in the trachea can be ruled out in de-tecting odors if olfactory bulbectomy resultsin anosmia. However, the nasociliary branchof the opthalamic nerve provides sensory in-put to part of the mucous membrane of thenasal cavity. It travels just lateral to the olfac-tory bulb in the rat and is invariably transectedwhen the bulb is removed. Thus, olfactorybulbectomized rats will also have reduced in-put from trigeminal receptors. Trigeminal re-ceptors respond to irritants and, although theygenerally have a much higher threshold thanolfactory sensory receptors, can respond tohigh concentrations of many commonly usedodorants (Laska et al., 1997). Perhaps the sim-plest way to minimize or eliminate a poten-tial contribution of these nonolfactory recep-tors in an olfactory study is to use odorantconcentrations that are judged as nonirritat-ing by human observers or, better, concen-trations that are well below the known thresh-

old for these receptors. Trigeminal and respi-ratory thresholds have been established formany commonly used odorants (Nielsen etal., 1984; Silver et al., 1986; Silver, 1992;Schaper, 1993; Cometto-Muniz et al., 2002).

CONCLUSIONS

Rats and other rodents live in an olfactoryworld and any attempt to understand rodentbiology must take into account the impor-tance of olfaction for social behavior, feeding,learning, and orientation in the environment.The role of odors in the control of rodent be-havior has long been a primary topic in etho-logically oriented studies of rat behavior, andthe relatively simple tests used in these stud-ies have served to demonstrate the influenceand importance of odors. However, recent ad-vances in odor control of learning and in themolecular biology of olfaction have requiredthe use of more sophisticated test proceduresand better control and understanding of thestimulus. No single test will serve the myriadfacets of olfactory research but all olfactorytests must take into account the unique fea-tures of and problems in dealing with vaporstate stimuli and controlling the attentive orsampling behavior of the subject.

REFERENCES

Abate P, Varlinskaya El, Cheslock SJ, Spear NE, MolinaJC (2002) Neonatal activation of alcohol-related pre-natal memories: impact on the first suckling re-sponse. Alcohol Clinical and Experimental Research26:1512-1522.

Alberts JR, Galef BG Jr (1971) Acute anosmia in the rat:a behavioral test of a peripherally-induced olfactorydeficit. Physiology and Behavior 6:619-621.

Amiri L, Dark T, Noce KM, Kirstein CL (1998) Odorpreferences in neonatal and weanling rats. Devel-opmental Psychobiology 33:157-162.

Apfelbach R, Russ D, Slotnick B (1991) Olfactory devel-opment and sensitivity to odors in rats. ChemicalSenses 16:209-218.

Beauchamp GK, Yamazaki K. (2003) Chemical signalling

Page 118: 0195162854

Chapter 9. Olfaction 101

in mice. Biochemistry Society Tranactions 31(Pt 1):147-151.

Bennett MH (1968) The role of the anterior commissurein olfaction. Physiology and Behavior 3:507-515.

Berger-SweeneyJ, Libbey M, ArtersJ, Junagadhwalla M,Hohmann CF (1998) Neonatal monoaminergic de-pletion in mice (Mus musculus) improves perfor-mance of a novel odor discrimination task. Behav-ioral Neuroscience 112:1318-1326.

Bhutta AT, Rovnaghi C, Simpson PM, Gossett JM,Scalzo FM, Anand KJ (2001) Interactions of inflam-matory pain and morphine in infant rats: long-termbehavioral effects. Physiology and Behavior 73:51-58.

Bodyak N and Slotnick B (2000) Performance of mice inan automated olfactometer: odor detection, dis-crimination and odor memory. Chemical Senses24:637-645.

Bowers JM and Alexander BK (1967) Mice: individualrecognition by olfactory cues. Science 158:1208-1210.

Boyse EA, Beauchamp GK, Bard J, Yamazaki K (1991)Behavior and the major histocompatibility complexof the mouse. In: Psychoneuroimmunology (AderR and Felten D, eds.), pp. 831-846. San Diego: Aca-demic Press.

Brennan PA, Schellinck HM, De La Riva C, KendrickKM, and Keverne KB (1998) Changes in neuro-transmitter release in the main olfactory bulb fol-lowing an olfactory conditioning procedure in mice.Neuroscience 87:583.

Brennan PA, Schellinck HM, Keverne EB (1999) Patternsof expression of the immediate-early gene egr-1 inthe accessory olfactory bulb of female mice exposedto pheromonal constituents of male urine. Neuro-science 90:1463-1470.

Brown RE (1979) Mammalian social odors: a critical re-view. Advances in the Study of Behavior 10:10-162.

Brown RE (1988) Individual odors of rats are discrim-inable independently of changes in gonadal hor-mone levels. Physiology and Behavior 43:359-363.

Brown RE and Mcdonald DW (1985) Social odours inmammals. Oxford: Oxford University Press.

Brown RE and Willner JA (1983) Establishing an "affec-tive scale" for odor preferences of infant rats. Be-havioral and Neural Biology 38:251-256.

Buck LB (1996) Information coding in the vertebrate ol-factory system. Annual Review of Neuroscience19:517-544.

Cometto-Muniz JE, Cain WS, Abraham MH, Gola JM(2002) Psychometric functions for the olfactory andtrigeminal detectability of butyl acetate and toluene.Journal of Applied Toxicology 22:25-30.

Costanzo RM (1985) Neural regeneration and functionalreconnection following olfactory nerve transectionin hamster. Brain Research 361:258-266.

Darling FMC and Slotnick BM (1994) Odor-cued tasteavoidance: A simple and efficient method for as-sessing olfaction in rats. Physiology and Behavior55:817-822.

Dhong HJ, Chung SK, Doty RL (1999) Estrogen protectsagainst 3-methylindole-induced olfactory loss. BrainResearch 824:312-315.

Dravnieks A (1972) Odor measurement. EnvironmentalLetters 3:81-100.

Dravnieks A (1975) Instrumental aspects of olfactome-try. In: Methods in olfactory research (Moulton DG,Turk A, Johnston JW, eds.). New York: AcademicPress.

DusekJA and Eichenbaum HB (1997) The hippocampusand memory for orderly stimulus relations. Pro-ceedings of the National Academy of Science94:7109-7114.

Fairless DS and Schellinck HM (2001) Assessing the ef-fects of 192-saporin lesions on rat olfactory learningand long term olfactory memory using a modifiedodour preference task. Canadian Conference onBrain and Behavioural Science, Laval University,Quebec City.

Fleming AS, Gavarth K, Sarker J (1992) Effects of tran-sections to the vomeronasal nerves or to the mainolfactory bulbs on the initiation and long-term re-tention of maternal behavior in primiparous rats.Behavioural and Neural Biology 57:177-188.

Forestell CA, Schellinck HM, Drumont S, Lolordo VM(2001) Effect of food restriction on acquisition andexpression of a conditioned odor discrimination inmice. Physiology and Behavior 72:559-566.

Forestell CA, Schellinck HM, Lolordo VM, Brown RE,Wilkinson M (1999) Olfactory conditioning and im-mediate early gene expression in mice. ChemicalSenses 24:618.

Fortin NJ, Agster KL, Eichenbaum HB (2002) Criticalrole of the hippocampus in memory for sequencesof events. Nature Neuroscience 5:458-462.

Galef BG Jr and Buckley LL (1996) Use of foraging trailsby Norway rats. Animal Behaviour 51:765-771.

Center MB, Owens DM, Carlone HB, Crofton KM(1996) Characterization of olfactory deficits in therat following administration of 2,6-dichloroben-zonitrile (dichlobenil), 3,3'-iminodipropionitrile, ormethimazole. Fundamentals of Applied Toxicology29:71-77.

Harlow H (1949) The formation of learning sets. Psy-chological Review 56:51-65.

Hendricks KR, Knott JN, Lee ME, Gooden MD, EversSM, Westrum LE (1994) Recovery of olfactory be-havior. I. Recovery after a complete olfactory bulblesion correlates with patterns of olfactory nervepenetration. Brain Research 648:121-133.

Hepper PG (1990) Feotal olfaction. In: Chemical Signals

Page 119: 0195162854

102 SENSORY SYSTEMS

in Vertebrates (Macdonald DW, Muller-SchwarzeD, Natynczuk SE, eds.), pp. 282-286. Oxford: Ox-ford University Press.

Hudson R (1993) Olfactory imprinting. Current Opin-ions in Neurobiology 3:548-552.

Hurst JL, Payne CE, Nevison CM, Marie AD,Humphries RE, Robertson DH, Cavaggioni A,Beynon RJ (2001) Individual recognition in micemediated by major urinary proteins. Nature 414:631-634.

Johanson IB and Teicher MH (1980) Classical condi-tioning of an odor preference in three-day old rats.Behavioral and Neural Biology 29:132.

Johnson BA and Leon M (1996) Spatial distribution of[14C]2-deoxyglucose uptake in the glomerular layerof the rat olfactory bulb following early odor pref-erence learning. The Journal of Comparative Neu-rology 376:557-566.

Kavaliers M and Ossenkopp KP (2001) Corticosteronerapidly reduces male odor preferences in femalemice. Neuroreport 12:2999-3002.

Laing DL, Panhuber H, Slotnick B (1989) Odor maskingin the rat. Physiology and Behavior 45:689-694.

Laska M, Distel H, Hudson R (1997) Trigeminal per-ception of odorant quality in congenitally anosmicsubjects. Chemical Senses 22:447-456.

Leinders-Zufall T, Lane AP, Puche AC, Ma W, NovotnyMV, Shipley MT, Zufall F (2000) Ultrasensitivepheromone detection by mammalian vomeronasalneurons. Nature 405:792-796.

Lu XM and Slotnick BM (1998) Olfaction in rats with ex-tensive lesions of the olfactory bulbs: implicationsfor odor coding. Neuroscience 84:849-866.

Lu XM, Slotnick BM, Silberberg AM (1993) Odor match-ing and odor memory in the rat. Physiology and Be-havior 53:795-804.

Mihalick SM, Langlois JC, Krienke JD, Dube WV (2000)An olfactory discrimination procedure for mice.Journal of the Experimental Analysis of Behavior73:305-318.

Mombaerts P, Wang F, Dulac C, Chai SK, Nemes A,Mendelsohn M, Edmondson J, Axel R (1996) Visu-alizing an olfactory sensory map. Cell 87:675-686.

Moulton DG (1967) Olfaction in mammals. AmericanZoologist 7:421-429.

Moulton DG and Marshall DA (1976) The performanceof dogs in detecting alpha-ionone in the vaporphase. Journal of Comparative Physiology 110:287-306.

Nevison CM, Armstrong S, Beynon RJ, Humphries RE,Hurst JL (2003) The ownership signature in mousescent marks is involatile. Proceedings of the RoyalSociety of London Series B Biological Sciences270:1957-1963.

Nielsen GD, Bakbo JC, Hoist E (1984) Sensory irritation

and pulmonary irritation by airborne allyl acetate,allyl alcohol, and allyl ether compared to acrolein.Acta Pharmacologica Toxicology (Copenhagen)54:292-298.

Nigrosh B, Slotnick BM, Nevin JA (1975) Reversal learn-ing and olfactory stimulus control in rats. Journal ofComparative and Physiological Psychology 80:285-294.

Otto T and Eichenbaum H (1992) Complementary rolesof the orbital prefrontal cortex and the perirhinal-entorhinal cortices in an odor-guided delayed-nonmatching-to-sample task. Behavioral Neuro-science 106:762-775.

Pedersen PE and Blass EM (1982) Prenatal and postna-tal determinants of the first suckling episode in al-bino rats. Developmental Psychobiology 15:349-355.

Reid 1C and Morris RGM (1992) Smells are no surer:rapid improvement in olfactory discrimination isnot due to the acquisition of a learning set. Pro-ceedings of the Royal Society of London Series BBiological Sciences 247:137-143.

Rouquier S, Blancher A, Giorgi D (2000) The olfactoryreceptor gene repertoire in primates and mouse: ev-idence for reduction of the functional fraction in pri-mates. Proceedings of the National Academy of Sci-ences 97:2870-2874.

RudyJW and Cheatle MD (1979) Ontogeny of associa-tive learning: acquisition of odor aversions byneonatal rats. In: Ontogeny of learning and mem-ory (Spear NE and Campbell BA, eds.). New Jersey:Hillsdale.

RudyJW and Cheatle MD (1983) Odor-aversion learn-ing by rats following LiCl exposure: ontogenetic in-fluences. Developmental Psychobiology 16:13-22.

Sakellaris PC (1972) Olfactory thresholds in normal andadrenalectomized rats. Physiology and Behavior9:495-501.

Schaefer ML, Yamazaki K, Osada K, Restrepo D,Beauchamp GK (2002) Olfactory fingerprints formajor histocompatibility complex-determined bodyodors, II: relationship among odor maps, genetics,odor composition, and behavior. Journal of Neuro-science 22:9513-9521.

Schaper M (1993) Development of a database for sen-sory irritants and its use in establishing occupationalexposure limits. American Industrial Hygiene Asso-ciation Journal 54:488-544.

Schellinck HM and Brown RE (1999) Searching for thesource of urinary odours of individuality in rodents.In: Advances in Chemical Signals in Vertebrates(Johnson RE, Muller-Schwarze D, Sorenson PW,eds). 267. New York: Kluwer Academic/PlenumPublishers.

Schellinck HM, Forestell CA, Dill P, Lolordo VM,Brown RE (200 la) The development of a simple as-

Page 120: 0195162854

Chapter 9. Olfaction 103

sociative olfactory test of learning and memory. In:Chemical Signals in Vertebrates 9 (MarchlewskwKoj A, Lepri JJ, Miiller-Schwarze D, ed.). New York:Kluwer Academic Publishers.

Schellinck HM, Forestell C, Lolordo V (2001b) A simpleand reliable test of olfactory learning and memoryin mice. Chemical Senses, 26:663-672.

Schellinck HM, Rooney E, Brown RE (1995) Odors ofindividuality of germfree mice are not discriminatedby rats in a habituation-dishabituation procedure.Physiology and Behavior 57:1005-1008.

Schellinck HM, Slotnick BM, Brown RE (1997) Odors ofindividuality originating from the major histocom-patibility complex are masked by diet cues in theurine of rats. Animal Learning and Behavior 25:193-199.

Setzer AK and Slotnick BM (1998) Disruption of axonaltransport from olfactory epithelium by 3-methylin-dole. Physiology and Behavior 65:479-487.

Silver WL (1992) Neural and pharmacological basis fornasal irritation. Annuals of the New York Academyof Science 641:152-163.

Silver WL, Mason JR, Adams MA, Smeraski CA (1986)Nasal trigeminal chemoreception: responses to n-aliphatic alcohols. Brain Research 376:221-229.

Singer AG, Beauchamp GK, Yamazaki K (1997) Volatilesignals of the major histocompatibility complex inmale mouse urine. Proceedings of the NationalAcademy of Science 94:2210-2214.

Slotnick BM (2001) Animal cognition and the rat olfactorysystem. Trends in Cognitive Sciences 5:216-222.

Slotnick BM (1984) Olfactory stimulus control in the rat.Chemical Senses 9:157-165.

Slotnick BM and Bisulco S (2003) Detection and dis-crimination of carvone enantiomers in rats with ol-factory bulb lesions. Neuroscience 121:451-457.

Slotnick BM and Bodyak N (2002) Odor discriminationand odor quality perception in rats with disruptionsof connections between the olfactory epitheliumand the olfactory bulb. Journal of Neuroscience22:4205-4216.

Slotnick BM, Glover P, Bodyak N (2000a) Does in-tranasal application of zinc sulfate produce anosmiain the rat? Behavioral Neuroscience 114:814-829.

Slotnick BM and Gutman L (1977) Evaluation of in-tranasal zinc sulfate treatment on olfactory dis-crimination in rats. Journal of Comparative andPhysiological Psychology 91:942-950.

Slotnick BM, Hanford S, Hodos W (2000b) Can rats ac-quire an olfactory learning set? Journal of Experi-mental Psychology: Animal Behavior Processes26:399-415.

Slotnick BM and Katz H (1974) Olfactory learning-setformation in rats. Science 185:796-798.

Slotnick BM, Kufera A, Silberberg AM (1991) Olfactory

learning and odor memory in the rat. Physiologyand Behavior 50:555-561.

Slotnick BM and Nigrosh BJ (1974) Olfactory stimuluscontrol evaluated in a small animal olfactometer.Perceptual and Motor Skills 39:583-597.

Slotnick BM and Ptak J (1977) Olfactory intensity dif-ference thresholds in rats and humans. Physiologyand Behavior 19:795-802.

Slotnick BM and Schellinck H (2002) Methods in olfac-tory research with rodents. In: Frontiers and meth-ods in chemosenses (Simon SA and Nicolelis M,eds.), pp. 21-61. New York: CRC Press.

Slotnick BM and Schoonover FW (1984) Olfactorythresholds in normal and unilaterally bulbec-tomized rats. Chemical Senses 9:325-340.

Slotnick BM and Schoonover FW (1992) Olfactory path-ways and the sense of smell. Neuroscience andBiobehavioral Reviews 16:453-472.

Slotnick BM and Schoonover FW (1993) Olfactory sen-sitivity of rats with transection of the lateral olfac-tory tract. Brain Research 616:132-137.

Slotnick BM, Westbrook F, Darling FMC (1997) Whatthe rat's nose tells the rat's mouth: long delay aver-sion conditioning with aqueous odors and potenti-ation of taste by odors. Animal Learning and Be-havior 25:357-369.

Smith FJ, Charnock DJ, Westbrook RF (1983) Odor-aversion learning in neonate rat pups: the role ofduration of exposure to an odor. Behavioral andNeural Biology 37:284-301.

Smotherman WP (1982) Odor aversion learning by therat fetus. Physiology and Behavior 29:769-771.

Southall PF and Long CJ (1969) Odor cues in a maze dis-crimination. Psychonomic Science 16:126-127.

Staubli U, Fraser D, Kessler M, Lynch G (1986) Studieson retrograde and anterograde amnesia of olfactorymemory after denervation of the hippocampus byentorhinal cortex lesions. Behavioral and Neural Bi-ology 46:432-444.

Steigerwald ES and Miller MW (1997) Performance byadult rats in sensory-mediated radial arm maze tasksis not impaired and may be transiently enhanced bychronic exposure to ethanol. Alcohol Clinical andExperimental Research 21:1553-1559.

Stevens DA (1975) Laboratory methods for obtaining ol-factory discrimination in rodents. In: Methods in ol-factory research (Moulton DG, Turk A, JohnstonAW Jr, eds.). New York: Academic Press.

Sullivan RM, McGaugh JL, Leon M (1991) Norepineph-rine-induced plasticity and one-trial olfactory learn-ing in neonatal rats. Developmental Brain Research60:219-228.

Sullivan R and Wilson DA (1991) Neural correlates ofconditioned odor avoidance in infant rats. Behav-ioral Neuroscience 105:307-312.

Page 121: 0195162854

104 SENSORY SYSTEMS

Sullivan RM, Wilson DA, Lemon C, Gerhardt GA (1994)Bilateral 6-OHDA lesions of the locus coeruleus im-pair associative olfactory learning in newborn rats.Brain Research 643:306-309.

Sundberg H, Doving K, Novikov S, Ursin H (1982) Amethod for studying responses and habituation toodors in rats. Behavioural and Neural Biology34:113-119.

Tucker D (1963) Physical variables in the olfactory stim-ulation process. Journal of General Physiology 46:453^89.

Van Elzakker M, O'Reilly RC, Rudy JW (2003) Transi-tivity, flexibility, conjunctive representations andthe hippocampus. I. An empirical analysis. Hip-pocampus 13:292.

Wallace DG, Gorny B, Whishaw IQ (2002) Rats cantrack odors, other rats, and themselves: implicationsfor the study of spatial behavior. Behavioural BrainResearch 131:185-192.

Williams J and Slotnick BM (1970) A multiple-choiceairstream design for olfactory discrimination train-ing of small animals. Behavioral Research Methodsand Instrumentation 2:195-197.

Wilson DA and Sullivan RM (1994) Neurobiologyof associative learning in the neonate: early ol-factory learning. Behavioral and Neural Biology61:1.

Wysocki CJ, Kruczek M, Wysocki LM, Lepri JJ (1991)Activation of reproduction in nulliparous and prim-iparous voles is blocked by vomeronasal organ re-moval. Biology of Reproduction 45:611-616.

Xu F, Greer CA, Shepherd GM (2000) Odor maps in the

olfactory bulb. Journal of Comparative Neurology422:489-495.

Xu W and Slotnick BM (1999) Olfaction and peripheralolfactory connections in methimazole-treated rats.Behavioural Brain Research 102:41-50.

Yamazaki K, Yamaguchi M, Baranoski L, BardJ, Boyse EA,Thomas L (1979) Recognition among mice: evidencefrom the use of a Y-maze differentially scented by con-genie mice of different histocompatibility types. TheJournal of Experimental Medicine 150:755-760.

Yee KK and Costanzo RM (1995) Restoration of olfac-tory mediated behavior after olfactory bulb deaf-ferentation. Physiology and Behavior 58:959-968.

Youngentob SL, Markert LM, Hill TW, Matyas EP,Mozell MM (1991) Odorant identification in rats: anupdate. Physiology and Behavior 49:1293-1296.

Youngentob SL, Markert LM, Mozell MM, Hornung DE(1990) A method for establishing a five odorant iden-tification confusion matrix task in rats. Physiologyand Behavior 47:1053-1059.

Youngentob SL, Schwob JE, Sheehe PR, YoungentobLM (1997) Odorant threshold following methyl bro-mide-induced lesions of the olfactory epithelium.Physiology and Behavior 62:1241-1252.

Zagreda L, Goodman J, Druin DP, McDonald D, Dia-mond A (1999) Cognitive deficits in a genetic mousemodel of the most common biochemical cause ofhuman mental retardation. Journal of Neuroscience19:6175-6182.

Zufall F and Munger SD (2001) From odor andpheromone transduction to the organization of thesense of smell. Trends in Neuroscience 24:191-193.

Page 122: 0195162854

Taste

ALAN C. SPECTOR

The taste world of the common laboratory rat(Rattus norvegicus) appears to be remarkablysimilar to that of the human. Rats and humansqualitatively categorize similarities and differ-ences among taste compounds in comparableways. The rat's robust approach and avoid-ance responses to taste compounds, as well asits innate oromotor reflexes to chemical stim-uli, parallel the hedonic reactions of humans.Perhaps that is why rats have successfullyearned a living as pests in urban dwellings. Tobe sure, there are some interesting excep-tions;—for example, aspartame does not ap-pear to be "sweet" to rats—but on the wholethe similarities between rats and humans faroutweigh the differences in regard to tasteperception. Thus, from the perspective of thefinal output of gustatory system, the rat servesas an excellent and economical animal modelof human taste and affect.

Although the output of the gustatory sys-tem appears to be similar between rats andhumans, the underlying anatomy has somenotable differences. In nonhuman primates(and presumably humans), the second-ordertaste neurons of the nucleus of the solitarytract (NST) project directly to the gustatoryzone of the thalamus, but in rats the taste-re-lated NST output reaches the thalamusthrough an obligate synapse in select subdivi-sions of the parabrachial nucleus (PEN). Inrats, the gustatory zone of the PBN not onlycontibutes to the thalamocortical pathway butalso sends fibers to ventral forebrain, includ-ing the amygdala, substantia innominata, andthe hypothalamus, all of which are structures

associated with motivated behaviors (Nor-gren, 1995). In nonhuman primates, theseventral forebrain structures receive their tasteinput directly from gustatory cortical areassuch as the insula, operculum, and orbito-frontal cortex (Pritchard, 1991). Although thefunctional significance of these anatomical dif-ferences remains unclear, the rat anatomy of-fers some experimental opportunities to se-lectively manipulate the flow of ascendinggustatory information in an effort to discernthe functional organization of the gustatorysystem and to gain insight into related pro-cesses including motivation and affect, andlearning and memory.

Taste function can be heuristically di-vided into three domains (Spector, 2000).First, taste stimuli possess qualitative signa-tures that allow the animal to identify them.Second, taste stimuli have motivational prop-erties in that they can promote or discourageingestion or be neutral. The motivational do-main can be further subdivided into processesassociated with the procurement of the stim-ulus (appetitive behavior) and processes in-volved in the reflex-like oromotor actions trig-gered by the contact of the stimulus with tastereceptors (consummately behavior). Third,some taste stimuli can trigger physiological re-flexes such as salivation that help facilitate di-gestion and assimilation of ingested sub-stances. Depending on the experimentalprocedure, taste stimulus-induced responsescan be classified into one or more of the abovedomains, and it is important to recognize thata given procedure might focus more on one

105

10

Page 123: 0195162854

106 SENSORY SYSTEMS

function than on another. With that in mind,the purpose of this chapter is to review themethodological and conceptual issues associ-ated with some common behavioral proce-dures used to assess taste function. I focus onlaboratory rodents, but many of the principlescan be generalized to other species.

STIMULUS PREPARATION

Chemical stimuli should be prepared as pre-cisely as possible. It is best to use reagent-grade chemicals dissolved in a reasonablypure source of water with minimal contami-nation. Unfortunately, contaminants couldcome from sources beyond the experi-menter's control such as the actual purity ofthe chemical purchased. Nevertheless, excel-lent water distillers and reverse osmosisdeionizers and nitration cartridges are avail-able. Whenever possible, it is best to preparethe solutions fresh. There are, however, situ-ations in which it might be advisable to pre-pare a solution somewhat in advance of itsuse. For example, some organic compoundsundergo mutorotation over time betweenanomeric forms until equilibrium is reached.Potentially this could cause variability in re-sponsiveness if a compound were used beforethe equilibrium was achieved, assuming thatthe different anomers had significantly differ-ent physiological properties. If solutions areprepared and stored under refrigeration, theyshould be allowed to reach room temperaturebefore use. If compounds are light sensitive,solutions can be kept in glassware and bottlesthat are wrapped in aluminum foil to helpminimize light exposure.

In the literature, there are two commonconventions used to define the concentrationof a solution: percentage weight by volume(%w/v) and molarity (M). Researchers whostudy feeding prefer the former because me-tabolizable carbohydrate solutions that areequivalent in their percentage weight by vol-

ume concentration have an identical caloricdensity. For researchers interested in tasteprocesses, however, it is preferable to repre-sent concentration in terms of molarity be-cause isomolar solutions have the same num-ber of molecules per unit volume (seePfafrmann et al., 1954). For example, to saythat a 10% fructose solution is "sweeter" thana 10% sucrose solution would not be mean-ingful from a taste perspective because the for-mer has about twice the number of moleculesper unit volume.

Most taste research is conducted withstimuli presented in liquid phase, because it isdifficult to uniformly mix solid chemicals intofood and to be certain of the concentration thatis reaching taste receptors. Under some cir-cumstances, spillage of powdered food can bea problem. To overcome these practical prob-lems, investigators have dissolved chemicalstimuli into gelatin solutions and then let it so-lidify (e.g., Rowland et al., 2003). For the re-maining portion of this chapter, I discuss pro-cedures involving liquid stimuli, although otherforms of stimulus presentation are available.

Experimenters should be cautious intheir assumptions about the relative intensityof individual qualitative components of mix-tures of chemical compounds. Within limits,the sense of taste appears to be more analyticthan it is synthetic. In other words, whenchemical compounds of different taste quali-ties are mixed in solution, observers can re-port the individual qualitative components inthe mixture, at least with binary and ternarycombinations (see Smith and Theodore, 1984;Laing et al., 2002; Frank et al., 2003). This islike the recognition of notes in a musicalchord and contrasts with the synthetic natureof color vision in which observers are unableto detect the wavelength components of com-bined light sources (e.g., white light arisingfrom the equal mixture of color primaries).Despite the analytical character of taste sen-sations, caution should be exercised regardingassumptions about the sensory properties of

Page 124: 0195162854

Chapter 10. Taste 107

mixtures (Laing et al., 2002; SchifFerstein,2003). Some compounds can interact in mix-tures and affect receptor processes in the pe-ripheral gustatory system in unexpected ways.In such cases, the whole is not the sum of theparts. For example, in hamsters, a weak con-centration of quinine hydrochloride (a "bitter"tasting compound to humans) that does notitself lead to stimulation of the chorda tym-pani nerve when it is placed on the anteriortongue will nonetheless significantly suppressthe responses of this nerve to sucrose whenthe two compounds are presented in mixture(Formaker et al., 1997). In some cases, mixturesuppression could be mediated in the centralnervous system (Lawless, 1979; Travers andSmith, 1984; Vogt and Smith, 1993). Synergiescan also occur. Responses to monosodium glu-tamate, as assessed by chorda tympani nerverecordings in rodents or intracellular Ca2+ con-centration changes in heterologous expressionsystems for taste receptor proteins, are en-hanced when the amino acid salt is mixed withthe 5'-purine nucleotide inosine monophos-phate (e.g., Yamamoto et al., 1991; Li et al.,2002). Various types of mixture interactionshave been identified both electrophysiologi-cally and psychophysically.

INTAKE TESTS

The most common measure used to assess re-sponsiveness to taste stimuli is the intake test.Various procedural incarnations of such testscan be found in the literature, but they all in-volve the measurement of the amount con-sumed from one or more bottles containingchemical solutions. The duration of a test canlast from a few minutes to days. In the two-bottle version, one bottle contains a tastestimulus and the other bottle contains eitherwater or a different taste stimulus. These testsare simple and do not involve specializedequipment, and no extensive training of ani-mals or laboratory personnel is required.

METHODOLOGICAL CONSIDERATIONS

When more than one bottle is available, stim-ulus sampling and position preferences be-come issues. In a short-term test, animals maynever sample one of the alternatives. Sam-pling can be encouraged by first presentingone solution and then presenting the other so-lutions. Once all of the stimuli have been sam-pled, they can be presented simultaneously(Nachman, 1962). This assumes that animalshave been trained on a restricted fluid-accessschedule and are "primed" to ingest solutionswhen they are presented. In both short-termand long-term tests, animals might prefer in-gesting from one bottle over another on thebasis of its position. Researchers have con-trolled for this by switching the positions ofthe bottles halfway through the test. This canbe made difficult by light/dark phase changes,but a 48-hour test avoids the problem. If thereare three or more bottles, this strategy be-comes complicated and preferences actuallyvary with the number of bottles presented(Tordoff and Bachmanov, 2003). Perhaps thisis why researchers rarely use more than twobottles. The order in which various concen-trations of a stimulus are presented can alsoinfluence preference and aversion (e.g., Flynnand Grill, 1988; Fregly and Rowland, 1992).

CONCEPTUAL ANDINTERPRETIVE CONSIDERATIONS

The advantage of the intake test is its relativeease of use compared with other, more in-volved procedures (see later). A drawback ofthis method is its difficulty in dissociatingthe effects of postingestive stimulation fromthose related to taste. For example, it is knownthat rodents display an inverted-U-shapedconcentration-intake function for severaltypes of compounds, including sugar solu-tions. It is thought that the descending limbof such functions is caused by the stimulationof postingestive receptor systems (Davis and

Page 125: 0195162854

108 SENSORY SYSTEMS

Levine, 1977). Indeed, if the ingested contentsare allowed to drain out of an open gastric oresophageal cannula (i.e., sham drinking), in-take monotonically increases with concentra-tion (Mook, 1963; Geary and Smith, 1985). Inanother example, Rabe and Corbit (1973) wereable to recapitulate the inverted-U-shaped con-centration-intake function for NaCl derivedfrom a 1 hour one-bottle test in rats by orallypresenting a single low concentration of NaClwhile simultaneously infusing varying concen-trations of NaCl directly into the stomach. Thisdoes not show that the taste of NaCl cannot alsoproduce a inverted-U-shaped concentration-response function; rats sham drinking NaClshow similar curves to normal-drinking rats.The Rabe and Corbit (1973) study does show,however, that taste is not necessary.

To the extent that the amount ingestedcan be attributed to taste stimulation, intaketests rely on the hedonic characteristics of thechemical stimuli to drive the behavior. Con-sequently, these tests do not necessarily revealmuch about the perceived qualitative identityof taste stimuli (with certain paradigmatic ex-ceptions such as sodium appetite; see Nach-man, 1962). In other words, discriminablestimuli can lead to the same degree of prefer-ence or aversion in an intake test. Moreover,a lack of preference or aversion does not nec-essarily mean that the stimulus is unde-tectable; it simply means that the taste com-pound is hedonically neutral. For example, inone experiment, C57BL/6J (B6) mice gener-ated a rather flat concentration-preferencefunction up to approximately 0.1 M NaCl de-rived from 48-hour two-bottle tests (NaCl ver-sus water; ascending concentration series) andthen displayed progressive degrees of avoid-ance as the concentration was raised further(Eylam and Spector, 2002). However, thesesame animals were able to clearly detect thelow concentrations of NaCl as indicated bytheir performance on a conditioned signal dis-crimination task in an operant response-basedtask (see later). Interestingly, the epithelialsodium channel blocker amiloride, which in-

terferes with one of the sodium taste trans-duction pathways, had no effect on thepreference-aversion function for NaCl butshifted the sensitivity curve measured in theoperant task by close to 1 order of magnitude(Eylam and Spector, 2002).

OROMOTOR AND SOMATICTASTE REACTIVITY

When chemical stimuli contact taste recep-tors, they can elicit stereotypical oromotorand somatic responses from the animal thatvary in a concentration-dependent fashion(Grill and Norgren, 1978a; Grill and Berridge,1985; Grill et al., 1987). These responses arereferred to as taste reactivity and fall into oneof two classes: ingestive responses and aver-sive responses. Ingestive responses are elicitedby normally preferred stimuli such as sucroseand include tongue protrusions, lateraltongue, and mouth movements. Aversive re-sponses are elicited by normally avoided stim-uli such as quinine and include gapes, chinrubs, forelimb flails, and head shakes. Inges-tive responses are normally accompanied byconsumption of the stimulus, whereas aver-sive responses are normally accompanied byfluid ejection from the oral cavity (Fig. 10-1).

METHODOLOGICAL CONSIDERATIONS

Taste reactivity behavior is best quantifiedwhen the stimulus is delivered directly intothe oral cavity under experimenter control. Inrodents, an intraoral cannula can be implantedchronically in anesthetized animals; there area variety of surgical methods to accomplishthis (Grill and Norgren, 1978a; Hall, 1979;Grill et al., 1987; Spector et al, 1988).

After recovery from intraoral surgery (2weeks), the animal is habituated to the test en-vironment, which usually consists of a smallPlexiglas arena, the wall of which sits on poststhat elevate it about 1 cm above the floor. Amirror is positioned on an angle below the

Page 126: 0195162854

Chapter 10. Taste

Figure 10-1. Ventral views of two characteristic taste-elicited oromotor responses. (Top) The gape, hallmark ofaversive taste reactivity, is characterized by a large openingof the mouth accompanied by retraction of its corners, usu-ally exposing the incisors. (Bottom) The tongue protrusion,hallmark of ingestive taste reactivity, is characterized by ex-tension of the tongue across the plane of the incisors (ar-row points to still-frame image of protruding tongue).

transparent floor. A length of PE-160 tubing,that has a small piece of PE-100 tubing at-tached to one end and flanged by heat, is con-nected to the cannula (the heat-flanged end),and the other end of it is attached to a com-mutator on the ceiling of the chamber. Inturn, another length of PE-160 tubing con-nects the commutator to a hyperdermic nee-dle, which in turn is attached to a syringe an-chored in a syringe pump. Fluid is pushedthrough the tubing to eliminate the deadspace before attachment to the rat. When theinfusion is scheduled to start, the pump is ac-tivated; once the fluid reaches the oral cavity,there is a reflexive series of mouth move-ments on which the timer is started.

Some investigators attempt to quantifytaste reactivity during real time, but it is bestto videotape the session for later detailedanalysis. The video camera is aimed at themirror and focused on the ventral view of themouth, and the operator merely follows theanimal as it moves in the chamber. Keepingthe chamber relatively small (e.g., «24 cm di-

109

ameter arena) limits the range of locomotion.The tape is analyzed for different types of tastereactivity behavior. Some responses are dis-tinct brief events that can be counted (e.g.,tongue protrusions); others are longer and aretimed (e.g., passive drip). A description of thedifferent types of taste reactivity behavior thathave been scored can be found elsewhere(Grill and Berridge, 1985; Grill et al, 1987;Spector et al., 1988).

Unfortunately, there is a certain degreeof subjectivity in any scoring method that isinescapable. For example, sometimes gapesare difficult to distinguish from large mouthmovements, and tongue protrusions can bedifficult to distinguish from small mouthmovements depending on the amplitude ofthe response. Accordingly, the scoring can beconducted by a trained observer who is"blind" to the experimental treatment of theanimal. To circumvent the vulnerability ofscoring videotape, some investigators haveimplanted electromyographic electrodes inthe tongue and jaw musculature throughwhich they can measure signature waveformsrepresenting certain types of oromotor re-sponses (e.g., Kaplan and Grill, 1989; Chenand Travers, 2003).

CONCEPTUAL AND INTERPRETIVE ISSUES

There is a growing literature on the theoreti-cal meaning of taste-elicited oromotor and so-matic behaviors (Grill and Berridge, 1985;Breslin et al., 1992; Parker, 1995; Berridge,1996). Taste reactivity experiments are cum-bersome to conduct, and analysis is tedious.Thus, the investigator should have strong jus-tification for taking this methodological path.That said, there are some very strong ratio-nales for the use of this procedure. It is per-haps the best way to quantify purely con-summatory behavior. In neural preparationsthat do not voluntarily eat and drink, such asthe chronic supracollicular decerbrate rat, it isthe only way to behaviorally assess taste re-sponsiveness (Grill andNorgren, 1978b). Even

Page 127: 0195162854

110 SENSORY SYSTEMS

in intact animals, the taste reactivity proce-dure provides a way to circumvent the flooreffects associated with intake and licking testswhen animals no longer approach the drink-ing spout. Moreover, animals can be tested ina nondeprived state, and very small volumesof fluid stimuli can be used and immediate re-sponses measured, minimizing the contribu-tion of postingestive factors. Taste reactivitycan also be modified by learning, as is strik-ingly obvious when rats progressively changetheir initial ingestive responses to 30 secondsucrose infusions delivered every 5 minutesafter an LiCl injection (which causes visceral

malaise), to an aversive profile of behavior(Breslin et al, 1992) (Fig. 10-2).

BRIEF-ACCESS TASTE TEST

The brief-access taste test is designed to assessunconditioned licking responses to taste stim-uli during very short duration trials (Youngand Trafton, 1964; Davis, 1973; Smith et al.,1992; Markison et al., 2000; Glendinning et al.,2002; Spector, 2003). Just as the taste reactiv-ity procedure involves the measurement ofimmediate responses to small volumes of

Figure 10-2. (Left) Mean (± SE) fre-quency of ingestive responses (top)and aversive responses (bottom) to in-traoral infusions (0.5 ml/30 s) of 0.1M sucrose presented immediatelyand every 5 minutes after an intra-peritoneal injection of LiCl (n = 6)or NaCl (n = 8) in rats. The LiCl-injected rats begin to change theirtaste reactivity profile from ingestiveto aversive, presumably as the LiCltakes effect. Naive rats that begin toreceive sucrose after a 20 minute de-lay will not display an aversive pro-file of responding (not shown), sug-gesting that the curves shown in theleft represent a rapid conditioningprocess. (Right) Mean (± SE) fre-quency of ingestive responses (top)and aversive responses (bottom) to0.1 M sucrose during a subsequenttest 4 days later in the same rats. Theinfusions were given 20, 25, and 30minutes after injection of LiCl inboth groups. The rats that receiveda lithium injection 4 days earlier dis-play an aversive profile of respond-ing, whereas the rats that were orig-inally injected with NaCl are stilldisplaying an ingestive profile of re-sponding, although they were in-jected with LiCl 20 minutes earlier.(Reprinted from Spector et al. [1988]with permission. Copyright © 1988by the American Psychological As-sociation.)MINUTES FOLLOWING INJECTION

Page 128: 0195162854

Chapter 10. Taste

taste stimuli, the brief-access taste procedure,pioneered by P. T. Young and his students(e.g., Young and Trafton, 1964), has some ofthe same methodological attributes but re-flects the contribution of both appetitive andconsummately behavior. This procedure re-quires the use of an automated stimulus de-livery and lick monitoring system, commonlyreferred to as a gustometer (e.g., Slotnick, 1982;Spector et al., 1990; Thaw and Smith, 1992;Reilly et al., 1994). Some of these devices arecommercially available, and others have beencustom built. All are capable of deliveringmultiple taste solutions during a session andrecording licking responses from the stimulus-access drinking spout.

METHODOLOGICAL CONSIDERATIONS

In general, each solution from the test arrayis presented for a very brief trial on the orderof seconds («5 to 30 seconds). Normally, thetrial timer starts with the first lick. Solutionpresentations are randomized without re-placement within blocks of trials. Conse-quently, satiation and fatigue factors are uni-formly distributed across the stimuli as thesession progresses. The shorter the trial dura-tion, the more trials that are initiated by theanimal before satiation occurs. On the otherhand, if the trial duration is too brief, then therange of potential variation in lick rate is cur-tailed. In the case of normally preferred stim-uli, animals can be tested in a nondeprivedstate, although they are usually trained ini-tially to sample the stimuli when water-deprived. In the case of normally avoidedstimuli, however, animals must be testedwhile on a water-restriction schedule, pittingthe drive to rehydrate against the aversivenessof the solution. Often times, the concentra-tion of a single chemical solution is varied.The resulting concentration-response curvesare very orderly and sigmoidal regardless ofwhether the stimulus is preferred or avoided.Because this technique is automated and in-volves very little animal training, a reasonably

111

large number of animals can be tested in a sin-gle day. For example, with 30 minute sessionsand 5 gustometers, more than 70 animalscould potentially be tested. Hence, this pro-cedure shows promise for high-throughputtaste screening in mutaginized rodents (seeGlendinning et al., 2002).

CONCEPTUAL AND INTERPRETIVE ISSUES

The basic premise of the brief-access taste testis that the global lick rate (differentiated fromthe local lick rate, which is simply the recip-rocal of the modal fundamental interlick in-terval [ILI]) during a given trial is a measureof the affective potency of the stimulus. De-pending on the experimental manipulation,however, there can be factors other than tastethat can affect licking. For example, a sucroseconcentration-response function would bevirtually flat if tested in a water-deprived rat.Indeed, if one animal is more "thirsty" thananother, this could also influence the degreeof licking even with normally avoided stim-uli. This latter possibility can easily be con-trolled by merely forming a ratio of stimuluslicking to water licking. Thus, differences indrive state or motor faculty can be factoredout. When testing normally preferred stimuliin nondeprived animals, this latter statisticalmanipulation is not meaningful and in fact isinadvisable to use, because minor changes inwater licking can have a disproportional effecton the ratio. One way investigators can adjustfor potential differences in general rate of lick-ing across animals tested in nondeprived con-ditions is to measure the ILI during a watertest under conditions of water restriction andthen use the reciprocal of that value to deter-mine the maximum licks possible in a trial.Thus, the lick rate to each stimulus can bescaled by the maximum possible lick rate foreach subject individually (Glendinning et al.,2002). Although it depends on motivationalproperties of the taste stimulus to drive re-sponsiveness, the brief-access taste test is oneof the few methods available to assess per-

Page 129: 0195162854

112 SENSORY SYSTEMS

ceived intensity in the suprathreshold domainand is becoming a popular alternative to thetwo-bottle preference test.

TASTE STIMULI ASCONDITIONED CUES

All of the procedures discussed up to thispoint assess taste function in the affective do-main. Yet it is possible for two taste solutionsto generate identical values in these meas-ures, because the two stimuli are hedonicallyequivalent under the test conditions, butnonetheless be qualitatively discriminable.Investigators can make inferences about thequalitative properties of chemical solutionsby using the taste stimuli as conditionedcues. Procedures in which rats are trained todiscriminate between two or more tastestimuli provide information about the per-ceived qualitative differences (or lack there-of) of chemical compounds. In other cases,rats can be trained to respond in a specificfashion to a single taste stimulus, and the de-gree to which the animal generalizes its re-sponses to other chemical solutions can pro-vide information on the perceived qualitativesimilarity of the stimuli.

METHODOLOGICAL CONSIDERATIONS

In classical conditioning procedures, tastecompounds can serve as conditioned stimuli(CSs). A common example of this approach isthe application of the conditioned taste aver-sion paradigm. It is well established that whenanimals ingest a novel taste stimulus followedby aversive visceral consequences, usuallycaused by the controlled administration of anemetic such as LiCl (although rats cannotvomit), they will subsequently avoid ingestingthe substance (Riley and Tuck, 1985). Thisprocedure has been exploited to make infer-ences about the perceived taste quality and in-tensity of chemical solutions (Nachman, 1963;Tapper and Halpern, 1968; Nowlis et al., 1980;

Spector and Grill, 1988). For example, inges-tion of 0.3 M NaCl (CS) could be followed byLiCl injection, and on future occasions the de-gree of avoidance to a variety of test stimulicould be evaluated in comparison to that ob-served in unconditioned control animals. Theratio of conditioned avoidance of a test solu-tion relative to the CS (i.e., 0.30 M NaCl) istaken as an index of similarity between thetwo stimuli. Such procedures have been usedwith one- or two-bottle intake tests as the pri-mary measure of avoidance or with gus-tometers (e.g., Spector et al., 1990), which canmeasure lick rates to a variety of stimuli de-livered in a controlled fashion during trials dis-tributed within a single session (i.e., brief-access taste test).

In operant conditioning procedures,taste stimuli can be used as discriminativecues, which signal the opportunity for rein-forcer delivery contingent on the executionof a specific response. These procedures re-quire the use of a gustometer. We have suc-cessfully used a two-response stimulus dis-crimination procedure in which animals aretrained to make one response (e.g., pressright-hand lever) in the presence of a givenstimulus (e.g., NaCl) and another response(e.g., press left-hand lever) in the presenceof a different stimulus (e.g., water). Correctresponses are rewarded with small volumesof water. The animals are tested while wa-ter-deprived, which promotes stimulus sam-pling and potentiates the reinforcing efficacyof water (St. John et al., 1997). We have usedthis procedure to measure detection thresh-olds (Fig. 10-3) and taste quality discrimina-tion (Fig. 10-4) (Spector et al., 1996; St. Johnet al., 1997; St. John and Spector, 1998;Geran and Spector, 2000; Kopka and Spec-tor, 2001; Geran et al., 2002; Spector andKopka, 2002). Other operant conditioningprocedures also have been successfully usedin conjunction with different types of gus-tometers to measure taste sensitivity, dis-crimination, and generalization (see Spec-tor, 2003, for a review).

Page 130: 0195162854

Chapter 10. Taste 113

Figure 10-3. Mean (± SE) percentage of trials with correct responses when stimulus was present, correctedfor the false alarm rate (referred to as Corrected Hit Rate) in a two-response operant procedure in which fivegroups of rats were trained to press one lever when water was presented and to press a different lever whenNaCI was presented. Note that when the chorda tympani nerve (CT) was transected and prevented from re-generating (CTX-7P and CTX-62P), sensitivity to NaCI was profoundly disrupted relative to sham-operatedrats (SHAM-7 and SHAM-62). In contrast, when the CT regenerated (CTX-62R), taste sensitivity to NaCI re-covered completely and the performance-disrupting effect of stimulus adulteration with the epithelial sodiumchannel blocker amiloride also returned to normal. These results highlight the importance of the signals inthe CT, which innervates only about 13% of the total taste buds, and the importance of epithelial sodiumchannels in taste receptor cells in maintaining sensitivity to NaCI. It is noteworthy that, historically, two-bottle tests have not revealed any major effects of CT transection on NaCI taste. (Reprinted from Kopka andSpector [2001] with permission. Copyright © 2001 by the American Psychological Association.)

CONCEPTUAL ANDINTERPRETIVE CONSIDERATIONS

The use of taste stimuli as conditioned cuesto assess gustatory function in the sensory/discriminative domain falls under the rubricof classic animal psychophysics and, as such,engenders standard interpretive issues (Bloughand Blough, 1977; Berkley and Stebbins, 1990;Spector, 2003). First, it is important to demon-

strate and maintain stimulus control of be-havior. In the case of detection threshold de-terminations or multiple tests for avoidanceafter taste aversion conditioning, some ex-tinction can occur. In the former, this canhappen as a result of the presentation of toomany subliminal concentrations, but it is eas-ily discerned by attending to the false alarmrate (proportion of water trials in which theanimal reported the stimulus was present) and

NaCI Concentration (M)

Page 131: 0195162854

114 SENSORY SYSTEMS

Figure 10-4. Overall proportion of correct responses collapsed across stimuli, concentrations, and sessionson an operant conditioning procedure in which rats were trained to discriminate between KCl and NH4Cl(left) or NaCl and NH4Cl (right) before (solid bars) and after (gray liars) combined bilateral transection of chordatympani nerve and greater superficial petrosal nerve (CTX + GSPX). The rats were also presurgically testedduring sessions (AMILORIDE) in which all of the stimuli were mixed in the epithelial sodium channel blockeramiloride hydrochloride (100 ^M). Concentration of the stimuli was varied to render intensity an irrelevantcue. Note that amiloride treatment only affects performance on the NaCl-versus-NH4Cl discrimination, likelybecause of its effect on sodium taste. Also note that transection of the gustatory branches of the seventh cra-nial nerve virtually eliminates the discrimination, despite that close to 70% of the total oral taste buds re-main. Interestingly, transection of the glossopharyngeal nerve, which denervates close to 60% of the totaltaste buds, does not have any effect on this taste discrimination or others, supporting the hypothesis that thegustatory nerves in the rat are functionally specialized (see St. John and Spector [1998]). (Reprinted fromGeran et al. [2002] with permission.)

to the performance on clearly detectable con-centrations. The inclusion of the latter canhelp maintain stimulus control. In taste aver-sion conditioning, in which intake tests areused, some extinction could occur as stimulustesting progresses across days; thus, it is im-portant to include CS probe trials to ensurethat the aversion is still strong.

With gustometers, extraneous cues asso-ciated with stimulus delivery can guide an an-imal's behavior despite the best efforts to min-imize them. At the end of the experiment, weroutinely fill all of the fluid reservoirs in thegustometer with water and arbitrarily assignthem as the "right-lever" or "left-lever" stim-ulus and examine whether the animals canperform the discrimination task without thepresence of the chemical cue.

It is important to also recognize thatchemical stimuli can potentially interact withthe olfactory and trigeminal systems. For ex-ample, as the concentration of sucrose in-

creases, so does its viscosity, and there isevidence that rats can smell sucrose at con-centrations above 0.03 M (Rhinehart-Doty etal., 1994). Thus, if a manipulation in the gus-tatory system fails to alter conditioned re-sponses to a taste stimulus, it might be due tonongustatory influences. In some experi-ments, the contribution of olfactory ortrigeminal cues can be ruled out based on theprofile of results (see Fig. 10-4), but in othercases the possibility remains an interpretivecaveat.

When assessing differences and similari-ties in taste quality among compounds, theperceived intensity of the stimuli must be con-sidered. Animals can use intensity cues to dis-criminate among stimuli. Likewise, general-ization can take place based on either theintensity or the quality of a test compound rel-ative to the training stimulus. Thus, the fail-ure of a test stimulus to generalize to a train-ing stimulus may not necessarily mean that

Page 132: 0195162854

Chapter 10. Taste 115

the two compounds have different taste qual-ities but might merely reflect that the testcompound is relatively weaker. So, it is ad-visable to vary the concentration of test stim-uli to help dismiss this possibility. Alterna-tively, each test stimulus could also serve as atraining stimulus in different experimentalgroups; if generalization between two stimulitakes place asymmetrically (i.e., when onestimulus is trained and the other is tested butnot vice versa), this would raise the concernthat the relative intensity of the taste solutionswas influencing the results.

Finally, in most procedures involving theuse of taste stimuli as conditioned cues, theanimals are placed on food or water restric-tion schedules to provide the motivation tosample the stimuli and perform responses. Al-though this is an effective means of generat-ing substantial behavior, such restrictionschedules have physiological consequencesthat could affect the gustatory system. Thereare examples of hormonal manipulations af-fecting taste-responsive neurons and leadingto upregulation of transduction-related ionchannels in taste receptor cells (Giza andScott, 1987; Giza et al, 1990; Herness, 1992;Gilbertson et al., 1993; Nakamura and Nor-gren, 1995; Tamura and Norgren, 1997; Lin etal., 1999). Thus, a given profile of results maydepend on the physiological state of the ani-mal during behavioral testing.

FINAL REMARKS

This chapter has provided a review of the ma-jor methodological and interpretive issues as-sociated with some common behavioral tech-niques applied in the study of taste processesin laboratory rats. Because of the limitedscope of the chapter, other useful procedures(e.g., taste contrast procedures, progressiveratio, and other reinforcement schedules in-volving taste stimuli as reinforcers) were notdiscussed, but some of the issues addressed

here would be pertinent to these othertechniques. In closing, it is important to rec-ognize that a given manipulation (e.g., ge-netic, anatomical, and pharmacological) couldpotentially affect one taste function whileleaving another unaltered. Accordingly, a va-riety of procedures should be used by investi-gators to comprehensively assess the impactof experimental treatments on taste functionin nonhuman animals.

ACKNOWLEDGMENTS

The author would like to thank Shachar Eylam and Laura C.Geran for providing constructive comments on an earlier ver-sion of this chapter. Part of the work presented here was sup-ported by a grant from the National Institute on Deafness andOther Communications Disorders (R01-DC01628).

REFERENCES

Berkley MA, Stebbins WC (1990) Comparative percep-tion: basic mechanisms. New York: John Wiley &Sons.

Berridge KC (1996) Food reward: brain substrates ofwanting and liking. Neuroscience and Biobehav-ioral Reviews 20:1-25.

Blough D and Blough P (1977) Animal psychophysics.In: Handbook of operant behavior (Honig WK andStraddon JER, eds.), pp. 514-539. Englewood Cliffs,NJ: Prentice-Hall, Inc.

Breslin PA, Spector AC, Grill HJ (1992) A quantitativecomparison of taste reactivity behaviors to sucrosebefore and after lithium chloride pairings: a unidi-mensional account of palatability. Behavioral Neu-roscience 106:820-836.

Brosvic GM and Slotnick BM (1986) Absolute and in-tensity-difference taste thresholds in the rat: evalu-ation of an automated multi-channel gustometer.Physiology and Behavior 38:711-717.

Chen Z and Travers JB (2003) Inactivation of amino acidreceptors in medullary reticular formation modu-lates and suppresses ingestion and rejection re-sponses in the awake rat. American Journal of Phys-iology: Regulatory, Integrative and ComparativePhysiology 285:R68-R83.

Davis JD (1973) The effectiveness of some sugars in stim-ulating licking behavior in the rat. Physiology andBehavior 11:39-45.

Davis JD and Levine MW (1977) A model for the con-trol of ingestion. Psychological Review 84:379-412.

Page 133: 0195162854

116 SENSORY SYSTEMS

Eylam S and Spector AC (2002) The effect of amilorideon operantly conditioned performance in an NaCltaste detection task and NaCl preference in C57BL/6J mice. Behavioral Neuroscience 116:149-159.

Flynn FW and Grill HJ (1988) Intraoral intake and tastereactivity responses elicited by sucrose and sodiumchloride in chronic decerebrate rats. BehavioralNeuroscience 102:934-941.

Formaker BK, MacKinnon BI, Hettinger TP, Frank ME(1997) Opponent effects of quinine and sucrose onsingle fiber taste responses of the chorda tympaninerve. Brain Research 772:239-242.

Frank ME, Formaker BK, Hettinger TP (2003) Taste re-sponses to mixtures: analytic processing of quality.Behavioral Neuroscience 117:228-235.

Fregly MJ and Rowland NE (1992) Comparison of pref-erence thresholds for NaCl solution in rats of theSprague-Dawley and Long-Evans strains. Physiol-ogy and Behavior 51:915-918.

Geary N and Smith GP (1985) Pimozide decreases thepositive reinforcing effect of sham fed sucrose in therat. Pharmacology, Biochemistry and Behavior 22:787-790.

Geran LC, Garcea M, Spector AC (2002) Transecting thegustatory branches of the facial nerve impairsNH4CI vs. KC1 discrimination in rats. AmericanJournal of Physiology: Regulatory, Integrative andComparative Physiology 283:R739-R747.

Geran LC and Spector AC (2000) Sodium taste de-tectability in rats is independent of anion size: thepsychophysical characteristics of the transcellularsodium taste transduction pathway. BehavioralNeuroscience 114:1229-1238.

Gilbertson TA, Roper SD, Kinnamon SC (1993) Protoncurrents through amiloride-sensitive Na+ channelsin isolated hamster taste cells: enhancement by va-sopressin and cAMP. Neuron 10:931-942.

Giza BK and Scott TR (1987) Intravenous insulin infu-sions in rats decrease gustatory-evoked responses tosugars. American Journal of Physiology: Regula-tory, Integrative and Comparative Physiology252:R994-R1002.

Giza BK, Scott TR, Antonucci RF (1990) Effect of chole-cystokinin on taste responsiveness in rats. AmericanJournal of Physiology: Regulatory, Integrative andComparative Physiology 258.-R1371-R1379.

Glendinning JI, Gresack J, Spector AC (2002) A high-throughput screening procedure for identifyingmice with aberrant taste and oromotor function.Chemical Senses 27:461-474.

Grill HJ and Berridge KG (1985) Taste reactivity as ameasure of the neural control of palatability. In:Progress in psychobiology and physiological psy-chology, vol. 11 (Epstein, AN and Sprague J, eds.),pp. 1-61. New York: Academic Press.

Grill HJ and Norgren R (1978a) The taste reactivity test.I. Mimetic responses to gustatory stimuli in neuro-logically normal rats. Brain Research 143:263-279.

Grill HJ and Norgren R (1978b) The taste reactivity test.II. Mimetic responses to gustatory stimuli in chronicthalamic and chronic decerebrate rats. Brain Re-search 143:281-297.

Grill HJ, Spector AC, Schwartz GJ, Kaplan JM, Flynn FW(1987) Evaluating taste effects on ingestive behavior.In: Techniques in the behavioral and neural sciences,vol 1: feeding and drinking (Toates F and Rowland N,eds.), pp. 151-188. Amsterdam: Elsevier.

Hall WG (1979) The ontogeny of feeding in rats: I. in-gestive and behavioral responses to oral infusions.Journal of Comparative and Physiological Psychol-ogy 93:977-1000.

Herness MS (1992) Aldosterone increases the amiloride-sensitivity of the rat gustatory neural response toNaCl. Comparative Biochemistry and Physiology[A] 103:269-273.

Kaplan JM and Grill HJ (1989) Swallowing during on-going fluid ingestion in the rat. Brain Research499:63-80.

Kopka SL and Spector AC (2001) Functional recovery oftaste sensitivity to sodium chloride depends on re-generation of the chorda tympani nerve after tran-section in the rat. Behavioral Neuroscience 115:1073-1085.

Laing DG, Link C, Jinks AL, Hutchinson I (2002) Thelimited capacity of humans to identify the compo-nents of taste mixtures and taste-odour mixtures.Perception 31:617-635.

Lawless HT (1979) Evidence for neural inhibition in bit-tersweet taste mixtures. Journal of Comparative andPhysiological Psychology 93:538-547.

Li XD, Staszewski L, Xu H, Durick K, Zoller M, AdlerE (2002) Human receptors for sweet and umamitaste. Proceedings of the National Academy of Sci-ences of the United States of America 99:4692-4696.

Lin WH, Finger TE, Rossier BC, Kinnamon SC (1999)Epithelial Na+ channel subunits in rat taste cells:localization and regulation by aldosterone. Journalof Comparative Neurology 405:406-420.

Markison S, Gietzen DW, Spector AC (2000) Essentialamino acid deficiency enhances long-term intakebut not short-term licking of the required nutrient.Journal of Nutrition 129:1604-1612.

Mook DG (1963) Oral and postingestional determinantsof the intake of various solutions in rats withesophageal fistulas. Journal of Comparative andPhysiological Psychology 56:645-659.

Nachman M (1962) Taste preferences for sodium saltsin adrenalectomized rats. Journal of Comparativeand Physiological Psychology 55:1124-1129.

Nachman M (1963) Learned aversion to the taste of

Page 134: 0195162854

Chapter 10. Taste 117

lithium chloride and generalization to other salts.Journal of Comparative and Physiological Psychol-ogy 56:343-349.

Nakamura K and Norgren R (1995) Sodium-deficientdiet reduces gustatory activity in the nucleus of thesolitary tract of behaving rats. American Journal ofPhysiology: Regulatory, Integrative and Compara-tive Physiology 269:R647-R661.

Norgren R (1995) Gustatory system. In: The rat nerv-ous system (Paxinos G, ed.), pp. 751-771. Sydney:Academic Press.

Nowlis GH, Frank ME, Pfafrmann C (1980) Specificityof acquired aversions to taste qualities in hamstersand rat. Journal of Comparative and PhysiologicalPsychology 94:932-942.

O'Keefe GB, SchummJ, Smith JC (1994) Loss of sensi-tivity to low concentrations of NaCl following bi-lateral chorda tympani nerve sections in rats. Chem-ical Senses 19:169-184.

Parker LA (1995) Rewarding drugs produce taste avoid-ance, but not taste aversion. Neuroscience andBiobehavioral Reviews 19:143-151.

Pfafrmann C, Young PT, Dethier VG, Richter CP, Stel-lar E (1954) The preparation of solutions for re-search in chemoreception and food acceptance.Journal of Comparative and Physiological Psychol-ogy 47:93-96.

Pritchard TC (1991) The primate gustatory system. In:Smell and taste in health and disease (Getchell TV,Doty RL, Bartoshuk LM, and Snow JB Jr., eds.), pp.109-125. New York: Raven Press.

Rabe EF and Corbit JD (1973) Postingestional control ofsodium chloride solution drinking in the rat. Jour-nal of Comparative and Physiological Psychology84:268-274.

Reilly S, Norgren R, Pritchard TC (1994) A new gus-tometer for testing taste discrimination in the mon-key. Physiology and Behavior 55:401-406.

Rhinehart-Doty JA, Schumm J, Smith JC, Smith GP(1994) A non-taste cue of sucrose in short-term tastetests in rats. Chemical Senses 19:425^131.

Riley AL and Tuck DL (1985) Conditioned food aver-sions: a bibliography. Annals of the New York Acad-emy of Sciences 443:381-437.

Rowland NE, Robertson K, Green DJ (2003) Effect ofrepeated administration of dexfenfluramine on feed-ing and brain Fos in mice. Physiology and Behavior78:295-301.

SchifFerstein HNJ (2003) Human perception of taste mix-tures. In: Handbook of olfaction and gustation (DotyRL, ed.), pp. 805-822. New York: Marcel Dekker.

Smith DV and Theodore RM (1984) Conditioned tasteaversions: generalization to taste mixtures. Physiol-ogy and Behavior 32:983-989.

Smith JC, Davis JD, O'Keefe GB (1992) Lack of an or-

der effect in brief contact taste tests with closely spacedtest trials. Physiology and Behavior 52:1107-1111.

Spector AC (2000) Linking gustatory neurobiology tobehavior in vertebrates. Neuroscience and Biobe-havioral Reviews 24:391-416.

Spector AC (2003) Psychophysical evaluation of tastefunction in non-human mammals. In: Handbook ofolfaction and gustation (Doty RL, ed.), pp. 861-879.New York: Marcel Dekker.

Spector AC, Andrews-Labenski J, Letterio FC (1990) Anew gustometer for psychophysical taste testing inthe rat. Physiology and Behavior 47:795-803.

Spector AC, Breslin P, Grill HJ (1988) Taste reactivity asa dependent measure of the rapid formation of con-ditioned taste aversion: a tool for the neural analy-sis of taste-visceral associations. Behavioral Neuro-science 102:942-952.

Spector AC and Grill HJ (1988) Differences in the tastequality of maltose and sucrose in rats: issues in-volving the generalization of conditioned taste aver-sions. Chemical Senses 13:95-113.

Spector AC, Guagliardo NA, St. John SJ (1996) Amiloridedisrupts NaCl versus KCl discrimination perform-ance: implications for salt taste coding in rats. Jour-nal of Neuroscience 16:8115-8122.

Spector AC and Kopka SL (2002) Rats fail to discrimi-nate quinine from denatonium: implications for theneural coding of bitter-tasting compounds. Journalof Neuroscience 22:1937-1941.

St. John SJ, Markison S, Guagliardo N, Hackenberg TD,Spector AC (1997) Chorda tympani transection andselective desalivation differentially disrupt two-lever salt discrimination performance in rats. Be-havioral Neuroscience 111:450-459.

St. John SJ and Spector AC (1998) Behavioral discrimi-nation between quinine and KCl is dependent on in-put from the seventh cranial nerve: implications forthe functional roles of the gustatory nerves in rats.Journal of Neuroscience 18:4353-4362.

Tamura R and Norgren R (1997) Repeated sodium de-pletion affects gustatory neural responses in the nu-cleus of the solitary tract of rats. American Journalof Physiology: Regulatory, Integrative and Com-parative Physiology 273:R1381-R1391.

Tapper DN and Halpern BP (1968) Taste stimuli: a be-havioral categorization. Science 161:708-710.

Thaw AK and Smith JC (1992) Conditioned suppressionas a method of detecting taste thresholds in the rat.Chemical Senses 17:211-223.

Tordoff MG and Bachmanov AA (2003) Mouse tastepreference tests: why only two bottles? ChemicalSenses 28:315-324.

Travers SP and Smith DV (1984) Responsiveness of neu-rons in the hamster parabrachial nuclei to taste mix-tures. Journal of General Physiology 84:221-250.

Page 135: 0195162854

118 SENSORY SYSTEMS

Vogt MB and Smith DV (1993) Responses of single sters as studied by electrophysiological and condi-hamster parabrachial neurons to binary taste tioned taste aversion techniques. Brain Researchmixtures: mutual suppression between sucrose 451:147-162.and QHC1. Journal of Neurophysiology Young PT and TraftonCL( 1964) Activity contour maps69:658-668. as related to preference in four gustatory stimulus

Yamamoto T, Matsuo R, Kiyomitsu Y, Kitamura R areas of the rat. Journal of Comparative and Physi-(1988) Taste effects of "umami" substances in ham- ological Psychology 58:68-75.

Page 136: 0195162854

Motor Systems III

Page 137: 0195162854

This page intentionally left blank

Page 138: 0195162854

Posture

SERGIO M. PELLIS AND VIVIEN C. PELLIS11

We typically think of behavior as involvingmovement, be it eating, walking, or copulat-ing. But to move effectively, an animal mustsupport its body and make fine postural ad-justments to protect the body's stability (Mar-tin, 1967). One important lesson to be learnedfrom this recognition of the importance ofpostural support for the genesis of movementis that some forms of lack of movement, orimmobility, should still be considered, andstudied, as behavior.

With regard to postural support, thereare a variety of defensive mechanisms used tomaintain or regain stability (Monnier, 1970).Close inspection of postural support responsesreveals them to be complex phenomena,guided by all suitable sensory systems—vestibular, tactile, proprioceptive, and visual(Magnus, 1924). In addition, postural re-sponses are themselves composed of inde-pendent motor output modules (or programs)that can be disassociated during developmentand in pathology (Pellis, 1996).

Studies with rats have been central to theunderstanding of postural support mecha-nisms. Thus, examples from this research onrats are used to illustrate the importance oflooking carefully at states of immobility, iden-tifying the sensory controls guiding posturalresponses, and fractionating such responsesinto their independent modules. For the lat-ter, one type of postural support response,righting, is examined in detail to illustrate theneed for using highly specific tests to evaluatethese independent modules.

IMMOBILITY

When unconscious, such as during sleep, a ratmay appear to have a flaccid body tone. Thisis a state in which not only is the animal un-responsive to its surroundings but also itsbody seems unprepared to deal with environ-mental contingencies. There are drug-inducedstates of immobility that are dramatically dif-ferent from this image of inertness. Both highdoses of morphine, an opioid agonist, andhaloperidol, a dopamine antagonist, producestates of immobility. Although in both cases,the treated rats appear inert, they in fact adoptbodily postures of preparedness for action butnot the same kind of action in both (De Rycket al, 1980; De Ryck and Teitelbaum, 1983).

Haloperidol induces a state in which a ratwill actively maintain static, stable equilib-rium but will not spontaneously move. Thisstate of readiness to act in order to defend itsstable position can be seen even when the ratis left standing by itself on a tabletop (Fig.11-1 A). The limbs are splayed out, and thebody is raised off the ground. If challenged bybeing pushed to one side, it will resist that dis-placement by shifting its body weight in thedirection of the oncoming force (see later).

In contrast, morphine induces a state ofimmobility in which the limbs appear frozen ina step-cycle (Fig. 11-1B). If it is pushed abruptly,the rat will run forward for a few steps beforeagain becoming immobile, but if it is pushedgentry, it can be rolled over onto its side. Thatis, in morphine-induced immobility, the pos-

121

Page 139: 0195162854

122 MOTOR SYSTEMS

Figure 11-1. Immobility induced by different drugs pro-duces states that can be very different with regard to thepostural mechanisms available to the animal. In haloperidol-induced immobility, rats have all of their postural supportmechanisms intact, so that when left standing on a table-top, the head is held off the ground, the body has a curvedconfiguration, and the limbs are pillar-like (A). In contrast,in morphine-induced immobility, the head and body are notsupported off the ground, and the rat's limbs are positionedas if in the middle of a step cycle; this is especially evidentin the visible hind foot (B). (Adapted from Pellis et al.[1986].)

tural support mechanisms are suppressed butthe locomotor mechanisms are not, whereas inhaloperidol-induced immobility, the locomotormechanisms are suppressed but the posturalsupport mechanisms are not. Combined treat-ment with haloperidol and morphine sup-presses both the postural and the locomotormechanisms (Pellis et al., 1986).

Immobility, then, is not just inertness butmay involve states with or without posturalsupport. To properly assess the effects of dif-ferent experimental treatments, any immobil-ity state that is created needs to be carefullyevaluated so as to discern the behavioral ca-pabilities available to the subject. An inadver-tent advantage arising from haloperidol-in-duced immobility (or catalepsy) is that suchdopamine-blocked rats afford the opportunityto study postural support responses inde-pendently of locomotor, exploratory, andother movement systems (Teitelbaum et al.,1982). Two postural responses, bracing andrighting, are described in some detail.

BRACING

Taking protective actions against an horizon-tally displacing force is a quintessential meas-ure to maintain an upright, stable orientation.Shifting one's body weight into the directionof the displacing force can be seen in dynamicsituations as well, such as when one is run-ning fast in a tight circle (Gambaryan, 1974).Naturally occurring shifts in body weight,which buttress the body against a displacingforce, can be evaluated in various tests ofskilled action or locomotion (Whishaw et al.,1994; Miklyaeva et al., 1995).

Bracing responses in their purest form aremost readily analyzed in rats with dopamineblockade or depletion. For example, if a dis-placing force is applied to the side of such arat, it leans laterally toward the displacingforce. Similarly, if it is pushed forward in a"wheelbarrow" configuration (i.e., when it isheld elevated by its rump), it pushes backwardwith its forepaws rather than step in the di-rection of the displacement. An undrugged rattypically steps away or steps forward in thesetests (Schallert et al., 1979; Pellis et al., 1985).

A simple way to test bracing is to placethe rat on an horizontal platform with a mildlyrough surface to allow it to grip and to thenraise its tail end so that gravity begins push-ing the animal forward. An intact rat typicallyresponds with a positive geotaxis responseand turns to face up or slowly walks down asthe slope increases. In contrast, a cataleptic ratpushes its body back (Fig. 11-2A) and so braceagainst the displacing force (Crozier and Pin-cus, 1926; Morrissey et al., 1989; Field et al.,2000).

If the board is tilted further, so that therat can no longer resist the downward forceof gravity, it begins to slide forward. At themoment its postural stability is lost, the ratexplosively jumps forward. On landing, itagain becomes immobile. This test shows thatthe forward jump is strictly a defensive re-sponse to a challenge to the rat's postural sta-bility and not a self-initiated forward move-

Page 140: 0195162854

Chapter 11. Posture 123

Figure 11-2. Bracing in a cataleptic rat can be tested byplacing it on an horizontal board, with a mildly rough sur-face, which allows it to grip. If the tail end of the board isthen gradually tilted, the rat will begin to brace by pushingits limbs forward, shifting its body weight backwards (A).When the rat can no longer resist the downward force ofgravity, it begins to slide forward. As the rat slides forward,it raises its head, by dorsiflexing its neck (B). If the dorsi-flexion of the neck is prevented, by attaching a tablespoononto the rat's back with a harness so that its head is thencapped by the concave bowl of the spoon, the rat will notjump forward (C). (Adapted from Morrissey et al. [1989]and Teitelbaum and Pellis [1992], respectively.)

ment. That the jumping response in catalep-tic rats reflects postural responses discon-nected from other behavior systems is illus-trated by the two triggers that are necessaryfor it to occur.

The initial trigger for jumping is the ratfeeling the loss of stability in its feet. If smallrods are placed on the board so that the ratcan grasp them with its feet, it is able to main-tain its stability, even when the board ap-proaches a slope of 70° or more, as comparedto a typical slope of 50° to 60°. Such stabilitycan be further exaggerated by using a wiremesh, which provides even greater support(Morrissey et al., 1989).

Once the rat feels its feet slide, a secondtrigger is released, when the head is raised bydorsiflexing the neck (see Fig. 11-2B). This

dorsiflexion appears to be necessary to triggerthe forward thrust of the rat's back legs. If atablespoon is attached onto a rat's back witha harness, so that the rat's head is capped bythe concave bowl of the spoon and thus is pre-vented from dorsiflexing (see Fig. 11-2C),then the cataleptic rat will not jump forward.Indeed, as the rat is sliding down, it can beobserved to be making small upward headmovements; however, the presence of thespoon prevents them from being of sufficientmagnitude to trigger the pushing of its hindlegs (Teitelbaum and Pellis, 1992).

When the slope is steep, undrugged ratsoccasionally jump forward. But if the platformwere placed so that its front is flush with theedge of a table, a forward jump would leadthe rat to land somewhere off the table. In thissituation, an undrugged rat orients to one ofthe sides of the board and then jumps ontothe surface of the table. In contrast, thedrugged rat jumps forward and off the table(with a soft cushion being provided as a land-ing pad) (Morrissey et al., 1989). That is, thedrugged rat fails to use visual cues to modifyits jump. The failure to incorporate vision inorganizing protective postural responses iscommon to various behavioral tests when us-ing cataleptic rats (Pellis et al., 1987). Fur-thermore, depending on the test used, whenmarshalling a postural defense, cataleptic rats,unlike undrugged ones, may give precedenceto tactile and proprioceptive information overvestibular information (Pellis et al., 1985; Cor-dover et al., 1993).

As can be seen from this description ofbracing, postural mechanisms can be studiedin a manner that allows the sensory devicesthat regulate their expression to be identified.Another type of postural response, righting, isused to illustrate the motor diversity of thesemechanisms.

RIGHTING

Like many other animals, rats, when sleeping,will "voluntarily" relinquish a stable, upright

Page 141: 0195162854

124 MOTOR SYSTEMS

position. As rat pups suckle beneath an upright,standing mother, to gain access to a nipple,pups typically roll over onto their backs (Eilamand Smotherman, 1998). Similarly, duringfighting, rats may roll over onto their backs soas to defend themselves against attacks fromtheir opponents (Pellis and Pellis, 1987). In con-trast to these behaviors, in righting, a standingposition is regained from a recumbent one.

As demonstrated by Magnus (1926),righting can be elicited by multiple sensory in-puts, each one capable of independently trig-gering righting. He thus labeled each form ofrighting by the type of sensory input involvedas well as by the body part that initiated therighting movement. These different types ofrighting mature independently, and so can bedisassociated developmentally, just as theycan be with some forms of brain damage andsensory deactivation. Furthermore, there arespecific rules of organization that give oneform of righting dominance over the othersin particular functional contexts (Pellis, 1996).

Righting can occur from a static position,such as when a rat is placed supine on theground, and from a dynamic position, such aswhen a rat falls, in a supine position, throughthe air. In either case, different righting re-sponses can be triggered by tactile/proprio-ceptive, vestibular, and visual inputs. Unlikesome other animals, such as cats, in rats, vi-sion can modulate the timing of dynamicrighting, but it cannot initiate such righting(Pellis, 1996). Just as each form of righting in-volves its own distinctive sensory input andmotor output, so each requires a particulartesting paradigm to exclude the activation ofthe other forms of righting.

TESTS OF RIGHTING

In some cases, given that more than one sen-sory system can initiate the righting of thesame part of the body (Magnus, 1926), eithersome test paradigm that restricts input fromthe competing sensory systems must be used,

or the sensory systems must be directlyblocked by some physiological manipulation.In the following tests, emphasis is placed ondisentangling the types of sensory inputs with-out the need to eliminate those systems. Thisapproach makes it easier for researchers toevaluate test subjects for many forms of right-ing without the chronic removal of sensorysystems that may interfere with general mo-tor function, such occurs after labyrinthec-tomy (Chen et al., 1986). Of great advantageis videotaping sequences of righting usinghigh shutter speeds, to allow frame-by-frameinspection of the movements performed.

TACTILE/PROPRIOCEPTIVE

Contact with the surface of the skin of itsflanks or back provides a rat with informationregarding its recumbent position, which al-lows the initiation of righting. Based on thissensory input, one of three forms of rightingcan be triggered from a static, recumbent po-sition. When a rat is in contact with theground but is falling from a stable position,some combination of tactile and propriocep-tive information can trigger righting in this dy-namic context.

Trigeminal RightingWhen the dorsum or the flanks of a rat's headare in contact with the ground, tactile infor-mation via the trigeminal nerve triggers right-ing of the head by a rotation of the neck. Ifthe rat is unrestrained, the righting can pro-ceed cephalocaudally. This was first describedby Troiani et al. (1981) in the guinea pig.

In rats, trigeminal stimulation induces ro-tation of the head, and as the head is rotated,it characteristically maintains firm contactwith the substrate. The only other form ofrighting capable of eliciting rotation of thehead is that triggered by vestibular informa-tion. In rats, vestibularly triggered head rota-tion is a transient feature, present only in earlydevelopment; later vestibular input triggersrotation by the shoulders (Pellis et al., 1991).

Page 142: 0195162854

Chapter 11. Posture 125

Therefore, when placing a rat on theground, in whether the supine or laterally re-cumbent position, the experimenter shouldfirmly apply pressure to the exposed part ofits body. Then, the rat's head should be re-leased after it is pushed down to ensure thatit has contact with the ground; this revealswhether its trigeminal righting program isfunctional.

Body Tactile RightingContact with the surface of either the dorsumor the flank triggers the rat's righting re-sponse. However, there are two distinct right-ing programs that can be elicited in this man-ner. One involves a rotation that begins at theshoulders (body-on-head), and the other be-gins with a rotation of the hindquarters (body-on-body) (Magnus, 1926). To test these formsof righting, the rat can be placed on one of itsflanks, with the experimenter's hand placedover and then pressed down onto its exposedflank. To avoid triggering trigeminal righting,the rat's head should not make contact withthe experimenter's hand, and the other side ofits head should be left to jut out over the edgeof the table's surface.

Under normal circumstances, body-on-head righting is dominant over body-on-bodyrighting. Therefore, to test for the presence ofbody-on-body righting, body-on-head right-ing has to be inactivated. To do this, the ex-perimenter can place one hand over the rat'sshoulders and the other over its pelvis, andthen release the hand over the pelvis. If thebody-on-body program is present, the rat'shindquarters rotate to prone. Evaluating thepresence of body-on-head righting, however,is more difficult because shoulder rotation canalso be triggered vestibularly. In such a case,a labyrinthectomy is necessary to negate theinfluence of the vestibular information ̂ Chenet al., 1986). Even so, if prior testing of vestibu-lar righting (see later) had revealed that suchrighting were not present, then the subjectdoes not need to undergo labyrinthectomy totest for body-on-head righting.

Early in development, body tactile formsof righting involve the rat pushing with itslimbs rather than using axial rotation (Pellis etal, 1991). If an animal reverts to pushing withits limbs to right, it may indicate brain dam-age (Martens et al., 1996). Therefore, whentesting the rat from the laterally recumbentposition, both the body-on-head and body-on-body forms of righting can be evaluated bythe experimenter for whether they demon-strate the adult-typical axial rotation patternor have regressed to the more primitive formof righting involving limb action. The dis-tinction between these two forms can be read-ily seen on videotaped sequences of righting.When righting with axial rotation, the rattucks the paw nearest to the ground close toits body; this prevents the paw from ob-structing the body's rotation to prone (Fig.11-3A). In contrast, when righting using limbaction, the rat places the paw closest to theground beneath its body; the paw is thus in aposition to flip the rat's body to prone bypushing against the ground (Fig. 11-3B).

Tactile/Proprioceptive Dynamic RightingA distinctive, dynamic form of righting is trig-gered in the rat when it falls while in contact

(b) placing

Figure 11-3. When righting with axial rotation, the rattucks the paw that is nearest to the ground close to its body(a). This prevents the paw from obstructing its body's ro-tation to prone. When righting using limb action, the ratplaces the paw that is closest to the ground beneath its body(b). The paw is now in a position to flip the rat's body toprone by pushing against the ground. (Adapted from Pelliset al. [1989b].)

Page 143: 0195162854

126 MOTOR SYSTEMS

with the ground. This form of righting ma-tures later in development than do the staticforms of tactile righting (Pellis and Pellis,1994) and can best be tested by holding the ratin a bipedal position, so that it stands on itshindlimbs. The rat is held by the experimenterfrom the back under its arms and is thenpulled back, onto the ground. If its righting isintact, as the rat begins to fall, it should rotateto face the direction in which it is falling, sothat by the time it lands, it does so in a proneposition. The rotation typically begins in therat's shoulders and progresses caudally. Aswith body-on-head righting, this form of right-ing can also be triggered vestibularly, and soeither pretesting for vestibular function (seelater) or labyrinthectomy is needed to evalu-ate fully whether it is intact.

VESTIBULAR

There are three distinctive ways in whichvestibular input can affect righting. Two in-volve the rat righting its head from a static po-sition based on otolithic function, and thethird involves the rat righting itself whenfalling based on semicircular function (Mon-nier, 1970).

Static AsymmetricalWhen held by the experimenter, laterally in theair, so that both flanks of its body have contactbut its head is left untouched, the rat's head ro-tates toward the ground, so that it is prone rel-ative to gravity. That is, from a situation ofasymmetrical vestibular information, the headturns to a normal prone orientation that hasequal vestibular input on both sides of the head.In this lateral orientation, it is the otoliths onthe side of the rat's head that face the groundthat provide the stimulus for righting.

A developmental dimension to this formof righting may also be used to detect any re-gression to a more primitive functioning. Inearly development, when this form of right-ing first appears, a rat's head rises upward soits snout points skyward and then its head ro-tates toward the ground. However, at later

stages of maturity, the rat's head rotates di-rectly to prone (Pellis et al., 1991).

Static SymmetricalWhen the rat is held by the base of the tailand is lifted into the air so that its head is point-ing downward, the rat's initial response is todorsiflex its head to bring it toward a proneorientation relative to the ground. As the ratbegins to crawl up one of its flanks veryquickly after the initial dorsiflexion of its head,it is useful to videotape this test. Such a con-founding response can be blocked by pre-treatment with a dopamine antagonist such ashaloperidol (see earlier). In the absence ofvestibular function, the rat's head rapidly ven-troflexes (Pellis et al., 1991b).

DynamicThe classic way to test this response is to dropthe rat from a height onto a soft cushion. Therat should be picked up by the experimenterwith one hand under its shoulders and lifted offthe ground and held by the dorsum of thepelvic girdle with the other hand. In this posi-tion, the rat should then be raised to the de-sired height. When the experimenter feels therat to be relaxed, in that it is no longer squirm-ing in the experimenter's grip, the experi-menter should rapidly swing away his or herhands, releasing the rat to fall. Within about 30to 60 milliseconds, an intact rat begins to rightin the manner described earlier, so that on land-ing it is fully prone (Pellis et al., 199la). In theabsence of vestibular input, righting does notoccur, and in cases of partial loss of vestibularfunction, righting may be delayed or incom-plete (Chen et al., 1986; Wallace et al., 2002).

Although some researchers have usedmechanical devices for holding animals in anattempt to standardize the hold-and-drop pro-cedure (Warkentin and Carmichael, 1939;Schonfelder, 1984), we have found that thereis no substitute for holding the animal in one'shands, as the experimenter is then able to usetactile cues to ensure that the rat is droppedonly when fully relaxed (see also Crimieux etal., 1984). Indeed, if the rat struggles or tenses

Page 144: 0195162854

Chapter 11. Posture 127

before its release, its righting can be initiallyinhibited or, due to its struggling, these move-ments may interfere with the progression ofsmooth, cephalocaudal axial rotation.

VISUAL

As noted, vision does not trigger righting in rats,but it can modulate its onset. Modulation canbest be evaluated in the righting-while-falling-in-the-air paradigm. By alternating trials at twodifferent heights (e.g., 60 cm and 30 cm), it canbe determined whether the difference in heightcan influence the onset of the rat's righting. Us-ing videotapes of these trials, the latency of on-set can be counted in terms of the number offrames that have elapsed before the first evi-dence of shoulder rotation is seen. The differ-ence between the two heights can be expressedin milliseconds. Typically, there is a 30 mil-lisecond difference between these two heights;rats begin righting sooner when dropped fromthe lower height (Pellis et al., 1989a; Pellis et al.,1991c). If vision is blocked, rats initiate air right-ing at the same latency, regardless of the heightof the drop (Pellis et al., 1989a, 1996).

CONCLUSION

Postural support is not something that can beleft for kinesiologists and physiologists to dealwith; it is integral to the production of move-ment, and so organized sequences of behav-ior (Martin, 1967). In early development, pos-tural concerns have a more obvious impact onbehavior (Pellis and Pellis, 1997). Even inadulthood, the limitations imposed by the ca-pacity to modify postural support may ac-count for species differences in the types ofmovement strategies used and in the bodilyconfigurations adopted (Berridge, 1990; Pellis,1997). Yet the analysis of postural supportmechanisms is rarely incorporated into analy-ses of movement (for rare exceptions, seeWhishaw et al., 1994; Miklyaeva et al., 1995).

As can be seen from this short, selectivereview, the postural support system of rats is

quite complex, with even seemingly smallsubcomponents (e.g., righting) requiring mul-tiple techniques to be fully assessed. Thus, tofully comprehend rats' behavior, their pos-tural support mechanisms need to be under-stood and that understanding must be incor-porated into studies of their movements.

BIBLIOGRAPHY

Berridge KG (1990) Comparative fine structure of action:rules of form and sequence in the grooming pat-terns of six rodent species. Behaviour 113:21-56.

Chen Y-C, Pellis SM, Sirkin DW, Potegal M, TeitelbaumP (1986) Bandage backfall: labyrinthine and non-labyrinthine components. Physiology and Behavior37:805-814.

Cordover AJ, Pellis SM, Teitelbaum P (1993) Haloperi-dol exaggerates proprioceptive-tactile support re-flexes and diminishes vestibular dominance overthem. Behavioural Brain Research 56:197-201.

Cremieux J, Veraart C, Wanet MC (1984) Developmentof the air righting reflex in cats visually deprived sincebirth. Experimental Brain Research 54:564-566.

Crazier WJ and Pincus G (1926) The geotropic conductof young rats. Journal of Genetic Physiology10:257-269.

De Ryck M and Teitelbaum P (1983) Morphine versushaloperidol catalepsy in the rat: an electromyo-graphic analysis of postural support mechanisms.Experimental Neurology 79:54-76.

De Ryck M, Schallert T, Teitelbaum P (1980) Morphineversus haloperidol catalepsy in the rat: a behavioralanalysis of postural mechanisms. Brain Research201:143-172.

Eilam D and Smotherman WP (1998) How the neona-tal rat gets to the nipple: common motor modulesand their involvement in the expression of earlymotor behavior. Developmental Psychobiology32:57-66.

Field EF, Whishaw IQ, Pellis SM (2000) Sex differencesin catalepsy: evidence for hormone-dependent pos-tural mechanisms in haloperidol-treated rats. Be-havioural Brain Research 109:207-212.

Gambaryan PP (1974) How animals run. New York:Wiley.

Magnus R (1924) Korperstellung. Berlin: Springer.Magnus R (1926) On the co-operation and interference

of reflexes from other sense organs with those ofthe labyrinths. Laryngoscope 36:701-712.

Martin JP (1967) The basal ganglia and posture. London:Pitman Medical Publishing.

Martens DJ, Whishaw IQ, Miklyaeva El, Pellis SM (1996)Spatio-temporal impairments in limb and body

Page 145: 0195162854

128 MOTOR SYSTEMS

movements during righting in an hemiparkinsonianrat analogue: relevance to axial apraxia humans.Brain Research 733:253-262.

Miklyaeva El, Martens DJ, Whishaw IQ (1995) Impair-ments and compensatory adjustments in spon-taneous movement after unilateral dopamine-depletion in rats. Brain Research 681:23-40.

Monnier M (1970) Functions of the nervous system: vol-ume II, motor and sensorimotor functions. Ams-terdam: Elsevier.

Morrissey TK, Pellis SM, Pellis VC, Teitelbaum P (1989)Seemingly paradoxical jumping in cataleptichaloperidol-treated rats is triggered by postural in-stability. Behavioural Brain Research 35:195-207.

Pellis SM (1996) Righting and the modular organizationof motor programs. In: Measuring movement andlocomotion: from invertebrates to humans (Os-senkopp K-P, Kavaliers M, Sanberg P, eds.), pp.116-133. Austin, TX: R.G. Landes Company.

Pellis SM (1997) Targets and tactics: the analysis of mo-ment-to-moment decision making in animal com-bat. Aggressive Behavior 23:107-129.

Pellis SM and Pellis VC (1987) Play-fighting differs fromserious fighting in both targets of attack and tacticsof fighting in the laboratory rat Rattus norvegicus. Ag-gressive Behavior 13:227-242.

Pellis SM and Pellis VC (1994) The development of right-ing when falling from a bipedal standing posture:evidence for the disassociation of dynamic and staticrighting reflexes in rats. Physiology and Behavior56:659-663.

Pellis SM and Pellis VC (1997) The prejuvenile onset ofplay fighting in laboratory rats (Rattus norvegicus).Developmental Psychobiology 31:193-205.

Pellis SM, Chen Y-C, Teitelbaum P (1985) Fractionationof the cataleptic bracing response in rats. Physiol-ogy and Behavior 34:815-823.

Pellis SM, Pellis VC, Teitelbaum P (1987) 'Axial apraxia'in labyrinthectomized lateral hypothalamic-dam-aged rats. Neuroscience Letters 82:217-220.

Pellis SM, Pellis VC, Teitelbaum P (1991a) Air rightingwithout the cervical righting reflex in adult rats. Be-havioural Brain Research 45:185-188.

Pellis SM, Pellis VC, Teitelbaum P (199lb) Labyrinthineand other supraspinal inhibitory controls over head-and-body ventroflexion. Behavioural Brain Re-search 46:99-102.

Pellis SM, Pellis VC, Whishaw IQ (1996) Visual modu-lation of air righting involves calculation of time-to-impact, but does not require the detection of thelooming stimulus of the approaching ground. Be-havioural Brain Research 74:207-211.

Pellis SM, Whishaw IQ, Pellis VC (1991c) Visual mod-ulation of the vestibularly-triggered air-righting inrats involves the superior colliculus. BehaviouralBrain Research 46:151-156.

Pellis SM, de la Cruz F, Pellis VC, Teitelbaum P (1986)Morphine subtracts subcomponents of haloperidol-isolated postural support reflexes to reveal gradientsof their recovery. Behavioural Neuroscience100:631-646.

Pellis SM, Pellis VC, Morrissey TK, Teitelbaum P(1989a) Visual modulation of vestibularly triggeredair-righting in the rat. Behavioural Brain Research35:23-26.

Pellis SM, Pellis VC, O'Brien DP, de la Cruz F, Teitel-baum P (1987) Pharmacological subtraction of thesensory controls over grasping in rats. Physiologyand Behavior 39:127-133.

Pellis SM, Pellis VC, Chen Y-C, Barzci S, Teitelbaum P(1989b) Recovery from axial apraxia in the lateralhypothalamic labyrinthectomized rat reveals threeelements of contact righting: cephalocaudal domi-nance, axial rotation, and distal limb action. Behav-ioural Brain Research 35:241-251.

Pellis VC, Pellis SM, Teitelbaum P (1991) A descriptiveanalysis of the postnatal development of contact-righting in rats (Rattus norvegicus). DevelopmentalPsychobiology 24:237-263.

Schallert T, De Ryck M, Whishaw IQ, Ramirez VD,Teitelbaum P (1979) Excessive bracing reactions andtheir control by atropine and L-DOPA in an animalanalog of Parkinsonism. Experimental Neurology64:33^3.

Schonfelder J (1984) The development of air-righting re-flex in postnatal growing rabbits. Behavioural BrainResearch 11:213-221.

Teitelbaum P, Szechtman H, Sirkin DW, Golani I (1982)Dimensions of movement, movement subsystemsand local reflexes in the dopaminergic systems under-lying exploratory locomotion. In: Behavioral modelsand the analysis of drug action (Spiegelstein MY andLevy A, ed.), pp. 357-385. Amsterdam: Elsevier.

Teitelbaum P and Pellis SM (1992) Towards a syntheticphysiological psychology. Psychological Science3:4-20.

Troiani D, Petrosini L, Passani F (1981) Trigeminal con-tribution to the head righting reflex. Physiology andBehavior 27:157-160.

Wallace DG, Mines DJ, Pellis SM, Whishaw IQ (2002)Vestibular information is required for dead reckon-ing in the rat. Journal of Neuroscience 22:10009-10017.

WarkentinJ and Carmichael L (1939) A study of the de-velopment of the air-righting reflex in cats and rab-bits. Journal of Genetic Psychology 55:67-80.

Whishaw IQ, Gorny B, Tran-Nguyen LTL, CastanedaE, Miklyaeva El, Pellis SM (1994) Making twomovements at once: impairments of movement,posture, and their integration underlie the adultskilled reaching deficit of neonatally dopamine-depleted rats. Behavioral Brain Research 61:65-77.

Page 146: 0195162854

Orienting and Placing

TIM SCHALLERT AND MARTIN T. WOODLEE12

It is not difficult to document the existence ofsensory or motor function asymmetries innormal rats or rats with unilateral damage tothe basal ganglia, sensorimotor cortex, and re-lated systems throughout the central nervoussystem, especially when the deficit on one sideis near maximal. When the unilateral deficitis subtotal, quantifying the extent of asym-metry and changes over time requires uniquetesting methods that directly pit one hemi-sphere against the other.

As an illustration, if a person is slightlyhard of hearing in one ear across all tones,how would you determine which ear is bet-ter and by how much? A simple test would beto put headphones on the person, play thesame level of sound simultaneously in eachear, and determine which side the soundseems to be coming from. Because sound lo-calization is influenced by relative intensity,this method could be used to confirm the ex-istence of a sensory asymmetry. If the soundappears to come from the left, it can be con-cluded that the right ear (or left hemisphere)is impaired relative to the left ear. To deter-mine the magnitude of the asymmetry, youcould then raise the intensity of the sound pre-sented to the relatively impaired ear and/orreduce the intensity of the sound presented tothe better ear until the sound seemed to comefrom neither the left nor right side. The ratioof the sound intensity presented to the im-paired ear relative to that of the sound pre-sented to the better ear would quantify the ex-tent of asymmetry. This two-part method isessentially the approach one can take in as-

sessing sensorimotor asymmetries in rats withpartial unilateral damage to the brain, and inevaluating treatments.

Behavioral deficits in Parkinson's diseaseand stroke often can be traced to both sensoryand movement initiation problems or an im-paired ability to make appropriate motor re-sponses to simple sensory events. In animals,unilateral damage to the sensorimotor cortex,striatum, or nigrostriatal pathway appears tohave the perceptual effect of dulling so-matosensory and proprioceptive sensory inputon one side and, in some cases, enhancing theinput from the other side. Asymmetrical sen-sory deficits, motor reactivity to bilateral sen-sory input, or predominantly motor dysfunc-tions can be examined with tests using atwo-part method in which an asymmetry isfirst identified and then the extent of the asym-metry is quantified.

In animal models, it is important to se-lect sensorimotor tests that are sensitive tothe brain damage and treatment effects. Thischapter describes behavioral tests that havebeen useful for examining the potential clini-cal efficacy of interventions that might be ben-eficial for neurological disorders. It is impor-tant to be able to distinguish whether anintervention promotes brain repair mecha-nisms, saves cells, enhances motor learningand retraining, or reduces the extent of sec-ondary degeneration of tissue. We have cho-sen to include a subset of sensorimotor teststhat we and others have found to be reliable,sensitive, quantitative, and easy to use in ratneurological models. The tests also cover the

129

Page 147: 0195162854

130 MOTOR SYSTEMS

range of cellular degeneration typical of focalischemic injury, nigrostriatal terminal loss,and cervical spinal trauma.

ENVIRONMENTAL ENRICHMENTAND SENSORIMOTOR BEHAVIOR

Most wild rats live in a very complex environ-ment that requires them to navigate obstacles,avoid predators, manipulate objects and cir-cumstances to gain access to food and mates,and so forth, using a wide array of motor skills.By contrast, standard laboratory housing is se-verely lacking in this sort of stimulation, andlaboratory "enriched" environments are stillless complex than the rat's natural habitat(Greenough et al., 1976; Jones et al., 2003;Schallert et al., 2003). Even the most sedentaryof people do not experience as impoverished anenvironment as a rat living in an isolated homecage. Therefore, to study sensorimotor behav-ior in the rat, it may be prudent to make someeffort to house animals so that behaviors anal-ogous to natural rat behavior are encouraged.

BILATERAL TACTILESTIMULATION TEST

Rats compulsively groom themselves and re-spond vigorously to any foreign substancethat becomes stuck to some part of their bod-ies. The adaptive advantages of this behaviormay include thermoregulation and mainte-nance against insects. Somatosensory asym-metries have been effectively determined us-ing a test that involves reacting to, andremoving, small sticky stimuli from the fore-limbs. It is a two-part test; however, few in-vestigators take advantage of both parts,which are needed to evaluate sensory func-tion independent of the motor component.Practice effects and motor learning play a par-tial role in the motor aspects of this test butdo not affect the sensory side, which can beinvestigated independently.

SENSORY ASYMMETRY

Small adhesive paper stimuli (Avery adhesive-backed labels, 113 mm2) are attached to therelatively hairless distal-radial aspect of eachof the rat's forelimbs (Schallert et al., 1982,1983, 2000; Schallert and Whishaw, 1984;Lindner et al., 2003; Fleming et al., 2003) (Fig.12-1). The rat is placed back into its homecage so that it is not distracted by a novel en-vironment, and it quickly uses its teeth to re-move these dots one at a time. In some ani-mals there is a small preoperative bias; in thesecases, the hemisphere selected for injury canbe opposite to the bias. Also, postoperativeoutcome can be compared against baselinevalues for each rat. Rats receiving unilaterallesions to brain areas subserving sensorimo-tor functions, especially those of the fore-limbs, develop an immediate bias for remov-ing adhesive stimuli of similar size from theunimpaired limb first. The order of contact-ing the ipsilateral versus contralateral stimu-lus reflects that there is a bias, but the mag-nitude of the sensory asymmetry requiresfurther evaluation (see later). The latency toremove the stimuli can be used as a measureof motor capacity and is sensitive to practiceeffects, unlike the order of contact (Schallertand Whishaw, 1984).

Each trial ends when the rat removesboth stimuli, or after 2 minutes has elapsed.To avoid habituation to the stimuli, individ-

Figure 12-1. Attaching adhesive stimuli (dots) to a rat's fore-limbs in preparation for the bilateral tactile stimulation test.

Page 148: 0195162854

Chapter 12. Orienting and Placing 131

ual trials should occur at intervals of no lessthan 5 minutes. In addition, the rats usedshould be well handled and have received sev-eral practice trials with the test before preop-erative data are collected. Experience with thetest calms the rats and makes the stimuli eas-ier to apply but does not appear to affect ac-tual performance.

This test is generally used to examine sen-sorimotor integration, although, as indicatedearlier, it is possible to some degree to distin-guish between the sensory and motor com-ponents involved (Schallert et al., 2002). Forexample, a change in the latency between ini-tial contact and subsequent removal of a dot(i.e., how much time it takes to remove thestimulus) can be an index of sensorimotorfunction. As in many of the tests presentedhere, however, it is important that such achange be represented as an asymmetry be-tween the impaired and unimpaired limbs inunilateral lesion models to control for non-motor and nonsensory factors (e.g., motiva-tional state, alertness) that could have a globalinfluence on latencies to contact and removethe dot. The contralateral (impaired limb) mo-tor component of this test is best assessed bydetermining the time point at which the ani-mal makes contact with the stimulus on theimpaired side and scoring how much timeafter that time point it takes to remove thatstimulus. This difference then would be com-pared with a comparable score of intact con-trol animals (i.e., how long after a control ratcontacts a given stimulus before it is removed,again controlling for practice effects by equat-ing extent of experience).

MAGNITUDE OF SENSORY ASYMMETRY

The second part of this test is used as a meansof measuring the degree of sensory asymme-try. In this part, the size of the dot placed onthe impaired limb is progressively increased(by overlapping two dots), while the dot onthe unimpaired limb is made smaller (by cut-ting down one dot). The dot sizes are in-

creased or decreased by 14 mm2, as illustratedin Figure 12-2, allowing for area ratios rang-ing from 1.3:1 to 15:1 between the impairedand unimpaired limbs, respectively. A suffi-cient increase in this ratio leads to a neutral-ization, and even a reversal (with a slightlyhigher ratio), in the bias for the limb that iscontacted first, and the ratio at which this oc-curs is used as the measure of severity of thesensory asymmetry. This measure is corre-lated with the amount of brain damage(Schallert et al., 1983; Schallert and Whishaw,1984; Barth et al., 1990); indeed, a small asym-metry can be detected in rats with simple burrholes in the skull. Animals are started at the2.2:1 ratio (level 3). If the stimulus is removedfrom the unimpaired limb first, animals thenare tested at two levels higher. If the stimulusis removed from the impaired limb first, the an-imal is tested at one level lower. This process iscontinued until the experimenter has deter-mined between which two levels the bias ex-ists and assigned the rat a score that reflects thisratio (e.g., a score of 2.5 is given if the animal'sbias reverses between levels 2 and 3).

Acute and chronic asymmetries on thistest have been demonstrated in models of cor-tical injury and ischemia, parkinsonism, andspinal cord injury (Schallert et al., 2000). Asrecovery occurs, the ratio defining the mag-nitude of asymmetry becomes smaller inde-

Figure 12-2. (Left) Data from an animal model of 6-hydroxydopamine-induced parkinsonism, in which sensoryasymmetries measured with the bilateral tactile stimulationtest improved over time but did not return to sham-operated control levels. (Right) Schematic of the differentarea ratios of the "dots" used in the test.

Dot Ratios Used

Page 149: 0195162854

132 MOTOR SYSTEMS

pendent of how much or how little practiceoccurs. Depending on the degree of striatal ornigrostriatal damage, full recovery can occur,even in hemidecorticate rats (Schallert andWhishaw, 1984). However, small changes inthe testing environment (e.g., partially open-ing the home cage while testing) can partiallyreverse recovery so that the ratio becomeslarger, possibly because the striatum is beingtaxed. This is important because it suggeststhat the testing environment can have amajor influence on measures of functionaloutcome.

LIMB-USE ASYMMETRY("CYLINDER") TEST

The limb-use asymmetry test evaluates theforelimb use of rats placed in a transparentPlexiglas cylinder. It has been used in a widevariety of motor system injury models, includ-ing middle cerebral artery occlusion, spinalcord injury, traumatic brain injury, parkinson-ian models, cortical ablation, and focal corticalischemia (Schallert et al., 2000; Schallert andTillerson, 2000; Tillerson et al., 2001, 2002;Lindner et al., 2003). A notable feature is a highdegree of sensitivity to chronic deficits not no-ticeably masked by postlesion compensatorybehaviors, as well as to chronic sensorimotordeficits that many tests fail to detect. The testis also easy to use and score, has a high inter-rater reliability, is well correlated with the ex-tent of lesions, including a wide range ofdopamine depletion (even 50% or less) (Tiller-son et al., 2001), and is relatively unaffected bypractice effects or, it seems, the compensatorystrategies often adopted by animals after mo-tor system insults (Schallert et al., 2002).

Rats are tireless explorers, in both theirnatural environments and laboratory homecages. They often explore vertical surfaces byrearing up on their hindlimbs and exploring thesurface with their front paws and vibrissae(Gharbawie et al., 2003). The cylinder test takesadvantage of this tendency and of the common

impairment in the initiation of movement andcontrol of static stable equilibrium, especiallycenter of gravity (Schallert et al., 1979, 1992).A rat is placed in an upright Plexiglas cylinder,open at both ends and measuring 30 cm highby 20 cm in diameter, that rests on a tabletop.The number of independent placements ob-served for either the right or left forelimb, aswell as the number of "both limb" (i.e., simul-taneous or near-simultaneous) placements,made onto the inner wall of the cylinder dur-ing rears is recorded. These limb placementsoccur when the rat shifts its weight, touchesthe cylinder wall, or steps to regain center ofgravity during lateral movements along thecylinder wall ("wall stepping").

The data can be recorded over a set pe-riod of time in the cylinder or until a certainnumber of placements has been made. (Weprefer the latter technique because differentrats, and especially different strains, can varywidely in their activity levels in the cylinder.)To film the rat's behavior for later rating, (1)a camera is placed over the cylinder (Fig.12-3A), (2) a camera is positioned to the sideof the cylinder with a mirror angled behindand to the side to enable the experimenter tosee the rat from all angles during live ratingso that no limb movement is missed, or (3)the cylinder is placed atop a raised, transpar-ent surface with a mirror positioned beneathat a 45° angle, with the camera aimed at themirror to film the limb placements from be-low (Fig. 12-3B). Care should be taken so thatthe rats do not habituate to the cylinder lestthey become inactive. This can be avoided bytesting during the dark cycle and by dividinglong trials into shorter segments separated byseveral minutes, during which the rat is placedback in the home cage.

Limb use is scored as the percentage ofleft, right, or both-limb wall placements rela-tive to the total number of placementsobserved. One can also obtain a single limb-use asymmetry score by subtracting the per-cent independent use of the impaired limbfrom the percent independent use of the

Page 150: 0195162854

Setup 1:

Place camera directly beneathcylinder.

Setup 2:

Place a mirror at 45° anglebeneath cylinder and film at 45°angle to mirror.

Figure 12-3. (A) Top-down view of a rat making placements in the cylinder, filmed from a camera placedover the cylinder. (E) Alternate setup that can be used to film the rat from below the cylinder.

133

Page 151: 0195162854

134 MOTOR SYSTEMS

unimpaired limb. Higher numbers indicate agreater bias for use of the unimpaired limb.The former scoring method is advantageousin that it provides more information aboutboth- versus independent-limb use events. Itshould be noted, however, that even with thelatter scoring method, a large number of both-limb use events lower the asymmetry score,albeit not as much as would an equal numberof independent impaired-limb placements. Analternative formula is one that we recentlyadopted because it reduces variability evenfurther and sets a nonbias at 50%:

[(ipsi + l/2 both) divided by(ipsi + contra + both)] X 100

An additional measure can be obtainedfrom this test. The use of a single limb to makelateral weight-shifting movements, independ-ent of the other limb, is reduced in the im-paired limb and enhanced in the nonimpairedlimb and reflects a very high degree of func-tional integrity. That is, when a rat rears andplaces one forelimb on the wall of the cylin-der and then makes a lateral movement to an-other location on the wall during the samerear sequence, this is considered an inde-pendent lateral weight-shifting movement, asopposed to a simple limb placement on thewall (which for the contralateral limb mayshow some recovery). The number of inde-pendent-limb weight-shifting movementsalong the wall for the ipsilateral forelimb canbe compared with that of the contralateralforelimb. After unilateral injury to the senso-rimotor cortex, striatum, or other motor ar-eas, such movements are rarely observed inthe affected forelimb but are commonlychronic in the unaffected forelimb (wherethey are even more frequently observed thanin either limb of control animals, suggestinga reorganization in the intact hemisphere).

Preoperative baseline values should beobtained before animals undergo surgery orother experimental manipulations. Althoughthere is no consistent population bias in limb

preference in the cylinder, some rats do dis-play a predilection for independent use of onelimb. When this occurs, experimental lesionscan be applied contralateral to this preferredlimb so that experimental effects are not con-founded by the preexisting limb-use bias. An-imals without a preoperative bias can be ran-domly assigned the lesion side.

Motivation differs between strains of rats.Long-Evans hooded rats, for example, aremore active and thus might be consideredpreferable as animal models, all other consid-erations being equal. Some rats, especiallySprague-Dawley rats (in our experience), maynot initially engage in an adequate amount ofwall exploratory behavior in the cylinder. Byand large, however, the behavior can be en-couraged in any rat with the use of any num-ber of "tricks" that do not affect the limb-useasymmetry score itself, including the following:

• Momentarily turning out the lights in thetesting room and testing during red light

• Blowing into or tapping the top of thecylinder

• Placing a dark cage cover (especially the rat'sown) over the cylinder

• Placing shavings from the rat's home cageinto the cylinder

• Scooting the cylinder (with the rat inside)gently a few cm along the tabletop

• Lightly touching a pencil eraser or cottontipped applicator to the rat's nose

• Dangling another rat into the cylindermomentarily

• Presenting novel scents or treats at the topof the cylinder

• Picking up the rat and replacing it into thecylinder

• Placing the rat in a new cylinder• Picking up the rat, flipping over the cylinder,

and putting the rat back in

TESTS OF FORELIMB PLACING

Researchers have developed a variety offorelimb-placing tests. Limb placing is usuallytriggered by visual or vestibular cues or by

Page 152: 0195162854

Chapter 12. Orienting and Placing 135

contacting the limb being tested with a sur-face (Wolgin and Kehoe, 1983; Marshall,1982). Rats use their vibrissae to gain bilateralinformation about the proximal environment,and this information is integrated between thehemispheres. When the bottoms of all fourfeet indicate that there is no stable surface forsupport, the rat is motivated to respond to thefirst object that one set of vibrissae contacts.In exploring its natural world, the rat fre-quently encounters surfaces that are unstableor, in the case of a cliff, unsuitable for loco-motion. All four limbs must be able to re-spond to information from either set ofvibrissae.

The test we describe next is the vibrissae-elicited forelimb placing test, which uses stim-ulation of the rat's vibrissae to trigger a plac-ing response (Barth et al., 1990; Schallert et al,2000; Lindner et al., 2003). This is a nice fea-ture in light of the very important role thatthe vibrissae play in the rat's sensory envi-ronment—indeed, they are thought to be oneof the primary tools rats use to explore theirworld. In addition, the test can be adapted toinvestigate neural events in the sensorimotorsystem that occur across the midline (as de-scribed next), a feature that is more difficultto implement using other placing triggers.

In this test, the rat's torso is supported bythe investigator and suspended such that allfour legs hang freely in the air. The experi-menter then brings the rat toward the edge of

a tabletop or another flat surface, taking careto avoid abrupt movements that might trig-ger placing due to a vestibular response. Ifsuch responses are noted, they should be ex-tinguished by taking the rat through the test-ing motions in open space (i.e., away from thetabletop) a few times. In the traditional, same-side version of this test, the rat's vibrissae arebrushed against the table edge on the sameside of the body in which forelimb placing isbeing evaluated. The percentage of trials inwhich the rat successfully places its forepawonto the tabletop is recorded for each side. Inaddition, the triggering stimulus can be pro-vided by moving the rat head on toward thetable edge, thus providing chin-based and/orbilateral vibrissae stimulation, or by holdingthe rat on its side and stimulating the whiskersopposite the limb being evaluated (Fig. 12-4demonstrates these different types of vibrissaestimulation). In all of these testing scenarios,the experimenter should gently restrain thelimb not being tested. Naturally, this requiresa tame rat that has been well handled for sometime before testing, and which ideally has hada chance to acclimate to the test and the ex-perimenter before being introduced to the ex-perimental manipulation. Trials should becounted only when the rat is relaxed and doesnot struggle, and achieving this can require agreat deal of practice on the part of the ex-perimenter. Intact rats will place with 100%success in all variants of this test.

Same-side Cross-midline Head-on

Figure 12-4. Forms of vibrissae-elicited forelimb placing, demonstrating the proper grip and orientation ofthe rat for this test.

Page 153: 0195162854

136 MOTOR SYSTEMS

The same-side version of the test hasbeen used in the evaluation of many centralnervous system injury models (Schallert et al.,2000). At our laboratory, we have begun toinvestigate the recovery of the cross-midlinetype of placing response in rats receiving cor-tical (via middle cerebral artery occlusion orfocal ischemia to the forelimb area of senso-rimotor cortex) or nigrostriatal (via 6-hydrox-ydopamine infusions to the nigrostriatal bun-dle) injury. One striking feature noted here isthat vibrissae stimulation applied to the"good" (i.e., ipsilesional) side of the body isable to trigger a placing reaction in the im-paired forelimb long before stimulation of thecontralesional vibrissae can. In contrast, le-sions to the nigrostriatal system lead to a com-plete failure of placing in the contralesionallimb in this test, consistent with parkinsonianakinesia. Also, the placing deficit recoversover a period of weeks in the cortical injurymodels (the rate of recovery depending on theextent of damage to the forelimb area of thesensorimotor cortex and especially to the ex-tent of striatal damage) but persists chroni-cally in parkinsonian models (Felt et al., 2002;Woodlee et al., 2003).

For example, after middle cerebral arteryocclusion that damages the striatum, the con-tralateral forelimb no longer responds to in-formation from the vibrissae about the loca-tion of stable surfaces, although the ipsilateralforelimb can respond appropriately to infor-mation from the contralateral vibrissae (sug-gesting that the deficit is not due to a puresensory impairment associated with the con-tralateral vibrissae). Moreover, except for se-vere damage to nigrostriatal dopamine termi-nals, in which the contralateral forelimb isakinetic, the contralateral forelimb recoversplacing in response to ipsilateral vibrissae stim-ulation. That is, sensory information sent tothe intact hemisphere can eventually controlmotor function associated with the damagedhemisphere, which is typical of normal rats.

TESTS OF HINDLIMB FUNCTION

Rats do not normally use their hindlimbs toinitiate or execute complex movement. In thisregard, we like to think of rats as being "front-wheel drive," a phenomenon that is illustratedin Figure 12-5, wherein rats supported solely

Figure 12-5. Rats operate primarilyvia "front-wheel drive." This seriesdepicts a rat remaining stationary for10 seconds when supported only onits hindlimbs (rats do not walk ontheir hindlimbs even if given moretime) but proceeding to movebriskly along a beam when flippedover so that it can support its weighton its forelimbs.

Page 154: 0195162854

Chapter 12. Orienting and Placing 137

on the forelimbs or hindlimbs initiate move-ment only when on the forelimbs (Schallertand Woodlee, 2003). In part, this is becausethe vibrissae are not in contact with theground and thus there is essentially a "stop"signal. This makes testing of hindlimb func-tion rather difficult, but some tests have beendeveloped that provide reliable measures ofhindlimb function. This is good news for re-searchers of spinal cord injury, because ex-perimental animal models of these conditionsoften involve lesioning caudal to the thoraciccord so that animals can maintain forelimbfunction and thus continue to care for them-selves postoperatively. Also, research into sci-atic nerve injury, a widely used model of pe-ripheral nerve damage, can benefit fromwell-developed tests of hindlimb function.

A relatively new hindlimb test that we de-veloped is the ledged tapered beam-walkingtest (Schallert et al., 2002). In this test, rats aretrained to traverse an elevated beam that is ta-pered along its extent and has an underhang-ing ledge (2 cm wide, dropped 2 cm below theupper beam surface) that the rat can use as acrutch if it slips (see Figure 12-6 for dimen-sions and setup). Footfaults (slips) made withthe hindlimbs can be measured as an index ofhindlimb function. A footfault can be rated asa half-fault if the paw slips off the upper sur-

face of the beam without falling all the wayto the ledge or as a full fault if the paw is placedfully on the ledge. The difficulty of the rat'straverse increases as it moves along the nar-rowing beam, thus leading to more footfaults.For this reason the beam can be divided intothree "bins" of difficulty along its extent, andthese can be scored separately or weightedrelative to each other to develop a singlescore. We generally run rats for five trials ona given day of testing, with several minutesbetween each trial (during which the rat'scagemate, for example, can be tested) to avoidhabituation to the test.

An important feature of this test is thatthe presence of the ledge allows the rat to dis-play a deficit that it might normally makecompensatory adjustments to hide. Rats arewell known for compensating to overcomelesion-induced deficits, and indeed this canmake the development of good behavioraltests difficult. Some compensatory motor ad-justments appear to be more automatic in thatthey appear immediately in response to an im-pairment, whereas other adjustments requirenew learning. If one wishes to test the directeffect of a therapeutic intervention on the sys-tem in question, it is important to have teststhat will target the deficit directly and be min-imally affected by these compensatory be-

Figure 12-6. The ledged taperedbeam, showing dimensions for con-struction. (Lower right) Rat runningthe beam, making a full left hindlimbfootfault (slip) onto the ledge. * Drawing is not to scale

Page 155: 0195162854

138 MOTOR SYSTEMS

haviors. If the test is influenced by compen-sation, it may not be clear if the therapy is ac-tually ameliorating the deficit per se ratherthan enhancing motor learning mechanismsthat allow for development of the compensa-tory behavior. Beam-walking tasks that do notuse a ledge are plagued by this problem, be-cause rats very quickly learn to make com-pensatory postural adjustments to keep them-selves from falling off the beam. Limbdysfunction may still exist, but the shift inbody weight can hide it. With the ledgedbeam, there is less threat of falling and there-fore compensation is less of a problem. In fact,a beam with a detachable ledge can be usedto measure the ability of the rat to learn com-pensatory skills. For example, even severalweeks after the insult, rats sustaining braindamage due to middle cerebral artery occlu-sion (a commonly used stroke model) con-tinue to show a stable deficit on the ledgedbeam test (Schallert et al, 2002). If the ledgeis removed, however, the rats learn to com-pensate over the course of just a few trials un-til they are running the beam successfully withno footfaults. This does not necessarily indi-cate recovery of limb function, though, be-cause rats will begin to make foot faults againif the ledge is subsequently replaced. Thespeed with which rats are able to shift be-tween displaying a deficit on the ledged beamand learning to compensate in the ledge's ab-sence may be reflective of the level of im-pairment and the capacity of motor learningcircuits that may or may not have been af-fected by the lesion.

Some hints make the use of the beammore successful. Preoperatively, rats must betrained to run the beam without fault, andpreferably without stopping to explore thebeam or its surroundings during the run.There is no prescribed number of trainingtrials needed to achieve this result; each ratcan simply be trained until this criterion isreached. Good training eases the testingphase, because stopping to encourage the ratto traverse the beam becomes less necessary.

When setting up the beam, the experimentermay want to place the rat's home cage at theend to serve as a reinforcer. The cage may alsobe covered with a dark cloth to make it moreenticing. During the initial training, the ex-perimenter can encourage the rat to run bytapping on the beam in front of the rat, pick-ing up the rat's tail from behind to encourageit to move away, or "tucking" the rat'shindquarters with the experimenter's hands.On early trials, the rat frequently stops to sniffthe beam or have a look around the testingroom, but this generally ceases during thecourse of pretraining. Objects should not beplaced to the side or below the beam becausethese tend to distract the animal. A compre-hensive review of the setup, use, and scoringof the beam can be found in Schallert et al.(2002).

Other opportunities exist for testinghindlimb function. Although rats rely prima-rily on their forelimbs for most movement,the hindlimbs are used in behaviors such asjumping, swimming (in which the forelimbsusually stay stationary as the hindlimbs pad-dle [Whishaw et al., 1981; Kolb and Tomie,1988; Stoltz et al., 1999]), and backing out oftunnels or other tight areas in which the ratcannot turn around. They can also use thehindlimbs to maintain balance during a rear;one can also quantify hindlimb stepping in thecylinder test (see earlier) as an index ofhindlimb function, if the cylinder is set up tobe filmed from below (Fleming et al., 2002).In the cylinder, rats with unilateral nigrostri-atal system damage mimicking a hemiparkin-sonian state tend to leave the impairedhindlimb planted in one place and pivotaround this akinetic limb by stepping with theunimpaired limb.

CONCLUSION

The sensorimotor tests described earlier arecertainly not the only ones that should be con-sidered useful, but in our experience these

Page 156: 0195162854

Chapter 12. Orienting and Placing 139

qualify as among the best for assessing func-tional outcome after unilateral focal ischemicinjury, nigrostriatal degeneration, traumatichead injury, damage to the intrinsic neuronsof striatum, and cervical spinal hemisection.It is possible to use aspects of these tests alongwith others to determine the location and ex-tent of injury and the degree of improvementover time. With practice, investigators can re-liably and rapidly evaluate treatment effects.Our Web site (http://www.schallertlab.org)has downloadable videos and informationthat can help new researchers adopt these andrelated tests of sensory and motor function.

REFERENCES

Earth TM, Jones TA, Schallert T (1990) Functional sub-divisions of the rat somatic sensorimotor cortex. Be-havioural Brain Research 39:73-95.

Felt BT, Schallert T, Shao J, Liu Y, Li X, Barks JD (2002)Early appearance of functional deficits after neona-tal excitotoxic and hypoxic-ischemic injury: Fragilerecovery after development and role of the NMDAreceptor. Developmental Neuroscience 24:418-425.

Fleming SM, Delville Y, Schallert T (2003) An intermit-tent, controlled-rate, slow progressive degenerationmodel of Parkinson's disease suitable for evaluatingneuroprotective therapies: Effects of methylphe-nidate. Behavioural Brain Research, in press.

Fleming SM, Woodlee MT, Schallert T (2002) Chronichindlimb and forelimb deficits differ between ratmodels of Parkinson's disease and stroke: A motionpicture poster. 2002 Abstract viewer, program No.885.4. Orlando, FL: Society for NeuroscienceConference.

Gharbawie OA, Whishaw PA, Whishaw IQ (2003) Thetopography of three-dimensional exploration: A newquantification of vertical and horizontal exploration,postural support, and exploratory bouts in the cylin-der test. Behavioural Brain Research, in press.

Greenough WT, Pass B, DeVoogd T (1976) The influ-ence of experience on recovery following braindamage in rodents: Hypotheses based on develop-mental research. In: Environments as therapy forbrain dysfunction (Walsh RN and Greenough WT,eds.), pp. 10-50. New York: Plenum Press.

Jones TA, Bury SD, Adkins-Muir DL, Luke LM, AllredRP, Sakata JT (2003) Importance of behavioral ma-nipulations and measures in rat models of braindamage and brain repair. ILAR Journal 44:144-152.

Kolb B and Tomie JA (1988) Recovery from early corti-cal damage in rats. IV: Effects of hemidecorticationat 1, 5, or 10 days of age on cerebral anatomy andbehavior. Behavioural Brain Research 28:259-274.

Lindner MD, Gribkoff VK, Donlan NA, Jones TA (2003)Long-lasting functional disabilities in middle-agedrats with small cerebral infarcts. Journal of Neuro-science 23:10913-10922.

Marshall JF (1982) Sensorimotor disturbances in the ag-ing rodent. Journal of Gerontology 37:548-554.

Schallert T, De Ryck, M, Whishaw IQ, Ramirez VD,Teitelbaum P (1979) Excessive bracing reactions andtheir control by atropine and L-dopa in an animalanalog of parkinsonism. Experimental Neurology64:33-43.

Schallert T, Fleming SM, Leasure JL, Tillerson JL, BlandST (2000) CNS plasticity and assessment of forelimbsensorimotor outcome in unilateral rat models ofstroke, cortical ablation, parkinsonism, and spinalcord injury. Neuropharmacology 39:777-787.

Schallert T, Norton D, Jones TA (1992) A clinically rel-evant unilateral rat model of parkinsonian akinesia.Journal of Neural Transplantation and Plasticity(currently Neural Plasticity) 3, 332-333.

Schallert T and Tillerson JL (2000) Intervention strate-gies for degeneration of dopamine neurons inparkinsonism: Optimizing behavioral assessment ofoutcome. In: CNS diseases: Innovate models of CNSdiseases from molecule to therapy (Emerich DF,Dean RLI, Sanberg PR, eds.), pp. 131-151. Totowa,NJ: Humana Press.

Schallert T, Upchurch M, Lobaugh N, Farrar SB, Spir-duso WW, Gilliam P, Vaughn D, Wilcox RE (1982)Tactile extinction: Distinguishing between sensori-motor and motor asymmetries in rats with unilat-eral nigrostriatal damage. Pharmacology, Biochem-istry, and Behavior 16:455-462.

Schallert T, Upchurch M, Wilcox RE, Vaughn DM(1983) Posture-independent sensorimotor analysisof inter-hemispheric receptor asymmetries in neos-triatum. Pharmacology, Biochemistry, and Behav-ior 18:753-759.

Schallert T and Whishaw IQ (1984) Bilateral cutaneousstimulation of the somatosensory system in hemi-decorticate rats. Behavioral Neuroscience 98:518-540.

Schallert T and Woodlee MT (2003) Brain-dependentmovements and cerebral-spinal connections: Keytargets of cellular and behavioral enrichment inCNS injury models. Journal of Rehabilitation Re-search and Development 40(1S):9-18.

Schallert T, Woodlee MT, Fleming SM (2002) Disen-tangling multiple types of recovery from braininjury. In: Pharmacology of cerebral ischemia(Krieglstein J and Klumpp S, eds.), pp. 201-216.Stuttgart: Medpharm Scientific Publishers.

Page 157: 0195162854

140 MOTOR SYSTEMS

Schallert T, Woodlee MT, Fleming SM (2003) Experi-mental focal ischemic injury: Behavior-brain inter-actions and issues of animal handling and housing.ILAR Journal 44:130-143.

Stoltz S, Humm JL, Schallert T (1999) Cortical injuryimpairs contralateral forelimb immobility duringswimming: A simple test for loss of inhibitory mo-tor control. Behavioural Brain Research 106:127-132.

Tillerson JL, Cohen AD, Caudle WM, Zigmond MJ,Schallert T, Miller GW (2002) Forced nonuse in uni-lateral parkinsonian rats exacerbates injury. Journalof Neuroscience 22:6790-6799.

Tillerson JL, Cohen AD, Philhower J, Miller GW, Zig-mond MJ, Schallert T (2001) Forced limb-use effectson the behavioral and neurochemical effects of

6-hydroxydopamine. Journal of Neuroscience 21:4427-4435.

Whishaw IQ, Schallert T, Kolb B (1981) An analysis offeeding and sensorimotor abilities of rats afterdecortication. Journal of Comparative and Physio-logical Psychology (currently Behavioral Neuro-science) 95:85-103.

Wolgin DL and Kehoe P (1983) Cortical KC1 reinstatesforelimb placing following damage to the internalcapsule. Physiology and Behavior 31:197-202.

Woodlee MT, Choi SH, Zhao X, Aronowski J, Grotta JC,Chang J, Hong JJ, Lin T, Redwine GG, Schallert T(2003) Distinctive behavioral profiles and stages of re-covery in animal models of stroke and Parkinson's dis-ease. 2003 Abstract viewer, program No. 947.5. NewOrleans, LA: Society for Neuroscience Conference.

Page 158: 0195162854

Grooming

J. WAYNE ALDRIDGE 13

Natural grooming in rodents is an advanta-geous behavioral model useful for studyingthe organization and neural mechanisms ofmovement sequences. Grooming consists ofcomplex strings of movements to dean andmaintain the fur and skin of the body; thesemovements include wiping, licking, andscratching. Grooming is natural and ubiqui-tous. It is observed readily; rats spend up to halfof their time during waking hours engaged ingrooming (Bolles, 1960). Most grooming boutsare initiated by paw-licking or face-washingmovements that proceed to grooming of thefur around the head, neck, and body in acephalocaudal stepwise pattern (Richmondand Sachs, 1978). Unitary grooming actionssuch as scratching or direct contact with thetrunk are emitted on their own in some in-stances; however, the cephalocaudal succes-sion of grooming actions across the body sur-face is most frequent and well established.

Kent Berridge, John Fentress, and theircolleagues and students made critical break-throughs in our understanding of the func-tional organization of grooming sequences(Berridge et al., 1987; Berridge and Fentress,1987a, 1987b; Berridge and Whishaw, 1992;Aldridge et al., 1993; Cromwell and Berridge,1996). They demonstrated that grooming pat-terns have lawful relationships among the in-dividual components and that the basal gangliaplay a key role in sequence implementation(Berridge et al., 1987). By meticulous tran-scriptions of the timing and serial order of in-dividual movements from continuous video-tape recordings and by evaluations of the

statistical predictability of sequence elementsand the probabilities of sequential patterns,they showed that the temporal structure ofgrooming actions has predictable organiza-tional features. Grooming is not random;rather, it exhibits a marked serial dependence(Berridge et al., 1987; Berridge, 1990).

Most grooming sequences consist offlexibly ordered mixtures of strokes, licking,scratches, etc.; however, occasionally ratsemit a fixed pattern, or "chain," of groomingactions. These occasional chain sequences arecomposed of the same movements as flexiblegrooming patterns, but the serial structure ofchains is relatively fixed in order and time.The chains are consistent and repeatable incontrast to "nonchain" grooming patterns,which are more flexible in their sequentialcomposition and serial structure.

The stereotyped grooming sequence hasapproximately 25 contiguous movements last-ing approximately 5 seconds in total. Thischain sequence has a stable serial order of fourphases (Berridge et al., 1987; Berridge, 1990)(Fig. 13-1, top). Phase 1 consists of 5-9 rapidelliptical strokes over the nose and mystacialvibrissae lasting for about 1 second. Phase 2 isshort (0.25 second) and consists of small asym-metrical strokes of increasing amplitude.Phase 3 consists of large bilateral strokes thattake 2 to 3 seconds for the animal to complete.The chain concludes with phase 4, which con-sists of a postural turn followed by a period(1 to 3 seconds) of body licking directed to theflank. The last phase varies more in lengththan other phases and often ends by blending

141

Page 159: 0195162854

142 MOTOR SYSTEMS

CHAIN GROOMING

Figure 13-1. Syntactic and nonchain grooming. (Top) The four syntactic phases—elliptical strokes, unilateralstrokes, bilateral strokes, and body licking. The choreographical timeline has forepaw movement as distancefrom the midline (right indicates up; left, down) as a function of time (x-axis, tics = 1 second) for a typicalsyntactic chain (left paw represented by line below the axis, right paw represented by line above the axis).(Bottom) Nonchain grooming consists of the same movements. Unlike chain grooming, the serial order ofgrooming actions is flexible and movements occur in varying combinations.

into subsequent nonchain grooming in whichchains are embedded. For practical purposes,the signature rapid elliptical strokes of phase1 provide a reliable marker for the stereotypedsyntactic chain. In flexible nonchain groom-ing, elliptical strokes are usually unitary.

Thus, rodent grooming has two notablesequence patterns: (1) a fixed chain sequenceand (2) nonchain sequences in more variableand flexible patterns (Fig. 13-1, bottom). Both

grooming patterns are composed of the sameactions, except the rigidity of the sequencestructure differs (Fig. 13-1). Stereotyped chaingrooming is less frequent, occurring at rates of2 to 15 chains per hour and comprising a totalof approximately 10 to 75 seconds of groom-ing. In contrast, nonchain grooming is morefrequent with total durations up to or morethan 20 times greater than chain grooming.

In a seminal paper, Karl Lashley (1951)

Page 160: 0195162854

Chapter 13. Grooming

noted the importance of action sequences andtheir implications for probing underlying neu-ral mechanisms. All behavior is sequential, butsome sequences exhibit syntax. A syntactic se-quence follows rules that determine the tem-poral progression of its elements. These rulesimpart lawful predictability to the sequence.Language has syntax. Given an arbitraryword, one can predict with some level ofprobability the next word in a sequence ofwords. Other behaviors can be described ashaving properties of syntax if one can demon-strate lawful sequential dependencies. Chainsequence grooming, as noted by Berridge andcolleagues (1987), exhibits syntax. It followspredictable rules for its components and struc-ture. Once the stereotyped grooming chainbegins, each remaining phase can be predictedwith greater than 90% accuracy. This entiresyntactical chain occurs with a frequency thatis more than 13,000 times greater than couldbe expected by chance (based on the relativeprobabilities of the component 25 actions ob-tained from grooming outside of this syntacticchain). A comparative phylogenetic analysis ofrat, mouse, hamster, gerbil, guinea pig, andground squirrel grooming patterns (Berridge,1990) has demonstrated that grooming syntaxis a basic biological trait conserved across relatedspecies.

Syntactic grooming chains were discov-ered through laborious moment-by-momenttranscriptions of behavior from continuousvideotape records. An alternative and, indeed,the more common method for cataloging be-havior relies on sampling. In sampling, the ac-tion in which the animal is engaged at everymeasurement interval, typically 15 seconds orlonger, is recorded to build a distribution ofbehavioral events in time. Sampling methodshave clear advantages in many applications.More animals can be studied per session be-cause a single observer can scan multiplecages. In contrast, continuous tabulation re-lies on tedious examination of all movementson a one-by-one basis. Sampling methodshave been particularly productive in identify-

143

ing the basic elements of grooming behavior(Bolles, 1960; Spruijt et al., 1992) and the ef-fects of drug manipulations on groomingbehavior (Spruijt et al., 1986; Molloy andWaddington, 1987).

Sampling methods, however, have dis-advantages for reconstructing detailed se-quential organization patterns. Unless thetime between samples is extremely short, thedetailed temporal structure of grooming maybe missed. To capture the multiple strokes ofphase 1 or the phase 2 transition in syntacticgrooming chains, for example, sampling in-tervals would need to be less than 1 second.Syntactic grooming chains typically have aduration of 5 seconds for four phases of up to25 strokes. To even make "hits" on chain se-quences, sampling must occur at less than5 second intervals. Because chain and nonchaingrooming strokes are similar, the special se-quential properties of grooming chains mightbe missed with a sampling method approach.Without information about preceding and fol-lowing actions, it would not be clear whethera grooming movement occurred within a syn-tactic chain or a flexible nonchain sequence.Under some circumstances, the effects ofdrugs or other manipulations on groomingmight be inadvertently confounded acrossthe two sequence types. Flexible nonchaingrooming dominates normal grooming withsyntactic chains interspersed irregularly atrates of about 10 chains per 2 hour session innormal, undisturbed animals Q. W. Aldridge,unpublished observations). Thus, observationperiods and recordings of 1 to 2 hours are typ-ical minimum durations necessary to exposethe syntactic chain and its properties. Al-though sampling procedures are useful forsome behavioral investigations, unless theyare fast enough to approximate continuousrecording, they may result in syntactic chainsbeing missed altogether or under counted.

A particular advantage of rodent groom-ing as a model system for evaluating the func-tional organization of behavior sequences andunderlying neural mechanisms is the fact the

Page 161: 0195162854

144 MOTOR SYSTEMS

grooming actions of the syntactical chain se-quence occur in unpredictable order and flex-ible combinations outside of the syntacticalchain sequence. Thus, the same movementscan be studied in two different sequence con-texts: (1) the syntactical chain and (2) flexiblecombinations of nonchain grooming. Thesimilarity of individual grooming actions andsequences of actions among individuals andeven across species (Berridge, 1990) facilitatescomparisons of kinematically similar move-ments in syntactic and nonsyntactical, flexiblesequences. In addition, grooming sequencesdo not depend on learning and memory.Learned sensorimotor sequences, which arecommonly used in behavioral neuroscience,are valuable and productive tools for complexcognitive testing, however, in some instances,it may be difficult to dissociate coding for asequence function from a memory function.In contrast, innate movement sequences suchas grooming, do not depend on explicit train-ing and thus provide a window on behavioralorganization and neuronal mechanisms inde-pendent of memory and explicit training.

SYNTACTICAL GROOMING:IMPLICATIONS FOR NERVOUS

SYSTEM FUNCTION

Berridge and colleagues have shown that syn-tactic grooming chains had a neural substratein the basal ganglia and brain stem (Berridgeand Fentress, 1987b; Berridge and Whishaw,1992; Aldridge et al., 1993; Cromwell andBerridge, 1996). The basal ganglia play a crit-ical role in the organization of grooming syn-tax, as evidenced by the fact that striatal le-sions have a potent impact on syntacticgrooming. The number of grooming chains"completed" drops by more than 50%(Berridge and Fentress, 1987a) after a striatallesion, even while the number of chains initi-ated is not reduced. Rats appeared unable toimplement the syntactic rule despite "at-tempts" to do so. The crucial regions of the

striatum were "mapped" by small bilaterallesions (^1 mm) (Cromwell and Berridge,1996). Cromwell and Berridge found thatsmall lesions in the dorsolateral quadrant ofthe neostriatum led to sequence deficits assignificant as large striatal lesions. This dor-solateral quadrant is part of the dorsal "mo-tor" circuit defined by Nauta and others(Nauta and Domesick, 1984; Alheid andHeimer, 1988). It is noteworthy that groom-ing chains are still attempted after striatal le-sions even though they fail to be implementedeffectively. This finding suggests that the stria-tum may play a role in implementing or fa-cilitating the sequence rather than a role in se-quence initiation.

The pontine hindbrain also contributesto generating sequential patterns. Decere-brated rats still occasionally generate the ba-sic pattern of syntactic chains (Berridge,1989a). Complete syntactic chains were neverseen in myelencephalic decerebrates, whichlacked a pons and cerebellum and had only amedulla remaining. Myelencephalic decere-brates still emit grooming actions with pos-tural support but show grossly degraded se-quential organization; the syntax of groomingis lost. Mesencephalic decerebrates (intact mid-brain and hindbrain) and metencephalic decer-ebrates (hindbrain only) still produce occa-sional syntactic chains with structuralsequence errors. Decerebrates complete lessthan half of the chains that they begin, sug-gesting that the basal ganglia may control im-plementation and completion of syntacticchain grooming sequences. In contrast, thebrain stem may have a role in generating thekinematic details of the individual actions andcontributing to the sequence in a rudimentaryform. Normal execution of syntactic groom-ing chains requires an intact neostriatum.

The critical importance of the basal gan-glia to syntactic grooming sequences wasaffirmed by comparisons of lesions of neo-striatum, motor cortex, and cerebellum andcomplete decortication (Berridge and Whishaw,1992). Only neostriatal lesions produced perma-

Page 162: 0195162854

Chapter 13. Grooming 145

nent impairment of the serial organization of syn-tactic grooming chains. All other lesions (sec-ondary motor cortex, combined primary andsecondary motor cortex, complete decortica-tion, cerebellar ablation) produced only mi-nor temporary disruption of sequential or-ganization, related to nonsequential motordeficits, such as deficits of fine coordinationof forelimb trajectories, movement timing, orposture. The basic structure of the sequenceremained intact if the neostriatum wasspared. The failure of primary and secondarymotor cortex lesions to induce sequentialdeficits in grooming syntax is interesting,given the "parallel loops" between cortex-striatum-thalamus-cortex. The contributionof motor cortical regions to grooming be-havior remains to be determined in futurework.

The implication from this work is thatthe striatum and basal ganglia are crucial tothe implementation of motor sequences. Thefact that decerebrate animals emit occasionalsyntactic chains, albeit poorly structured, sug-gests that the neostriatum is not the "centralpattern generator" of syntactic groomingchains. Instead, the pattern generator circuitrymust be contained to a considerable extentwithin the brainstem. The neostriatum mustbe intact to implement the pattern on the flowof normal behavior. This special role in im-plementing sequential behavior may haveevolved to include learned behavioral se-quences (Rapoport, 1989; Aldridge et al.,1993), and it may account for the characteris-tic breakdown in sequence organization inhuman basal ganglia disorders in humans,such as occurs in Parkinson's disease. Otherrecent studies in animals and model systemsmake it clear that sequencing of action andcognition into syntactic patterns of serial or-der is an important behavioral function me-diated by the basal ganglia (Kermadi andJoseph, 1995; Mushiake and Strick, 1995;Beiser and Houk, 1998; Berns and Sejnowski,1998; Matsumoto et al., 1999; Lieberman,2000).

NEURAL SUBSTRATES FORCODING OF GROOMING SYNTAX

The activation profiles of neurons in the neos-triatum and substantia nigra pars reticulata(SNpr) during grooming further support abasal ganglia role in coding and implementinggrooming syntax (Aldridge and Berridge,1998; Meyer-Luehmann et al., 2002). Whetherneuronal activation was correlated to groom-ing movements depended on the sequentialcontext in which the movement occurred (syn-tactic chain versus nonchain flexible groom-ing sequences). Furthermore, the dorsolateralstriatum site crucial to syntactic grooming hadstronger activation than did ventromedial re-gions (Aldridge and Berridge, 1998).

These findings are based on neuronalrecordings from rats implanted with perma-nent multisite recording electrodes in neostria-tum (dorsolateral or ventromedial) or SNpr.The implant, which is connected by a flexiblecable to a commutator, allows animals tomove about and to groom freely with norestraint (Aldridge and Berridge, 1998; Meyer-Luehmann et al., 2002). Spontaneous behaviorfor 1 or more hours of normal groomingand free movement is recorded on time-synchronized videotape from under a glassfloor with simultaneous recording of neuronalspike activity on a computer. A frame-by-frameanalysis of the videotaped grooming sequencesis done off-line (Aldridge and Berridge, 1998;Meyer-Luehmann et al., 2002) to demarcatesyntactic chain grooming and flexible nonchaingrooming and to determine the onset and endtimes of syntactic grooming phases and indi-vidual grooming movements within phases.Changes in neuronal activity in relation togrooming actions is assessed by perievent timehistograms to average neural spike activityover 5 to 10 repetitions of each movement.Recording sites are verified histologically afterthe completion of recording.

The activity of 41% of striatal cells waspreferentially related to the sequential patternof syntactic chains; that is, neurons were acti-

Page 163: 0195162854

146 MOTOR SYSTEMS

Nonchain Bilateral Stroke Figure 13-2. Neural activation dur-ing syntactic grooming. (A) Dorso-lateral striatal neuron illustrating anincrease of activity associated withthe onset of phase 3 (bilateralstrokes) during chain grooming (left,arrow). The same neuron was not ac-tivated during nonchain bilateralstrokes (right, arrow). (B) Nigral neu-ron is activated during the ellipticalstrokes of phase 1 during chaingrooming (left, arrow) but not duringthe elliptical strokes in nonchaingrooming (right, arrow).

vated during movements in the context ofsyntactic chain grooming but not duringequivalent movements emitted in flexiblenonchain grooming (Aldridge and Berridge,1998) (Fig. 13-2A). Only 14% of neurons hada pattern of activation suggesting they couldcode simple motor properties of groomingmovements—that is, activated during move-ment in both contexts, inside and outside ofsequential chains.

Regional differences in neuronal activa-tion patterns were apparent (Aldridge andBerridge, 1998). Neurons in the dorsolateralstriatal site crucial for syntactic grooming in-creased their firing rates by 116% during syn-tactic chains compared with 30% in the ven-tromedial region. Dorsolateral neurons also

seemed to code phase-specific aspects of thesyntactic sequence of grooming. In the crucialdorsolateral region, more neurons respondedduring multiple phases of the sequence thanin the ventromedial region (18% of dorsolat-eral neurons and 5% of ventromedial neu-rons).

To determine if activation during groom-ing was strictly a correlate of movement, theactivation patterns of neurons responsive dur-ing syntactic chains were examined duringsimilar movements emitted during flexiblenonchain (nonsyntactic) grooming. Most neu-rons exhibited context sequence-dependentactivation patterns (see Fig. 13-2A). Few neu-rons (16%) could be categorized as simplymovement related, that is, as neurons active

Chain Phase 3

Page 164: 0195162854

Chapter 13. Grooming 147

during movement without regard to the se-quence in which it occurred. Neuronal firingrates affirmed the important influence of se-quence context. During syntactic chaingrooming, rates were significantly higher thanduring nonchain grooming (17%; paired t test,p < 0.001) or quiet resting (44%; p < .001).

The SNpr is an important output struc-ture of the basal ganglia. The SNpr receivesmassive projections from dorsolateral stria-turn (Deniau et al., 1996); it is not surprisingthat neuronal activity profiles of SNpr, like thestriatum, reflect a dependence on the contextof sequential syntax (Meyer-Luehmann et al.,2002). Fifty-five percent (n = 26) of SNpr neu-rons were active during grooming behavior,with most activation (73%, 19 of 26 respon-sive neurons) occurring during the first twophases of the syntactic sequence. The onset ofthe syntactic grooming sequence is a domi-nant feature with vigorous activation duringphase 1 (96%, 25 of 26 responsive neurons).The importance of sequential context is againclear. Many neurons (36%) responsive duringphase 1 did not respond at all during similarnonchain grooming movements (see Fig.13-2B). Even the neurons that were activeduring elliptical strokes of nonchain groom-ing had significantly faster firing rates whenthe strokes occurred in the context of syntac-tic grooming (50 versus 28 spikes per second).

In addition to phase 1, SNpr neurons, likethe striatum, exhibited sequence-dependentactivation patterns during subsequent groom-ing phases. In contrast to the striatum, how-ever, SNpr had more vigorous activation dur-ing some nonchain grooming strokes thanduring kinematically similar syntactic groom-ing. For example, the proportion of activatedneurons was higher during bilateral strokes ofnonchain grooming (65% of neurons) and fir-ing rates were faster during phases 2 and 3than during kinematically similar strokes ofsyntactic chain grooming. Thus, SNpr neu-rons preferentially coded the onset of the syn-tactic sequence (phase 1) and then seemed todiminish or inhibit activation related to mo-

tor parameters of phases 2 and 3. Althoughthe direction of neuronal activation differs inthe nigra compared with the striatum, theseresults demonstrate that neurons at the out-put of the basal ganglia, like the striatal inputregion, are modulated by the sequence con-text during the performance of instinctivegrooming movements. The specificity ofSNpr activation to the onset of the syntacticchain pattern and relatively diminished activ-ity to later phases of the chain may reflect asequence-dependent balance between excita-tory subthalamic nucleus and inhibitory stri-atal inputs.

ROLE OF DOPAMINE INSEQUENTIAL CHAIN GROOMING

Evidence suggesting a role for dopaminergicneural mechanisms in syntactic chain groomingincludes the following. Lesions of dopaminer-gic afferents to the striatum with the neuro-toxin 6-hydroxydopamine (6-OHDA) (Berridge,1989b) have consequences as severe as de-struction of the striatum itself. In the normalontogeny of rats, the motor components ofgrooming appear much earlier than syntacticchain sequence even though it uses the samemovements. The appearance of syntacticgrooming develops in parallel with the devel-opment of the dopaminergic markers in thematuring striatum (Colonnese et al., 1996). Fi-nally, dopamine Dl receptor agonists enhancethe quantity of grooming (Starr and Starr, 1986);this is supported by evidence that dopamine Dlagonists given either systemicaUy or intra-ventricularly (Berridge and Aldridge, 2000a;Berridge and Aldridge, 2000b) enhance syntac-tic grooming over and above the increase inoverall grooming.

Dopamine agonists have a potent effecton syntactic grooming. In general, Dl ago-nists enhanced and D2 agonists impaired syn-tactic grooming sequences (Berridge andAldridge, 2000a). Corticotropin (ACTH),which is known to elicit excessive grooming

Page 165: 0195162854

148

(Dunn, 1988; Dunn and Berridge, 1990), alsoreduced the amount of grooming and the like-lihood that the sequential pattern would becompleted. Most interesting, Dl activationspecifically enhances stereotyped groomingover and beyond increases in movement itself(superstereotypy). Both the relative frequencyof syntactic chain emission and the percent-age rate of syntactic chain completion wereincreased relative to changes in nonchaingrooming. The efficacy of different Dl ago-nists depends on the route of administrationand whether the agent is a full or partial ago-nist; however, the main effects were clear. Dlactivation enhances syntactic grooming overand above the increase in flexible grooming.This stands in contrast to the decrease associ-ated with D2 activation and other peptides byany route or dose (Berridge and Aldridge,2000a).

CONCLUSION

Rodent grooming behavior is a particularlyuseful model system with which to study theorganization and neuronal mechanisms of be-havioral sequences. Because it has syntacticaland nonsyntactical modes and each containsessentially the same component actions, ro-dent grooming is an ideal system for dissoci-ating motor from sequence control proper-ties. It is possible to compare, on the sameneurons, coding mechanisms related to move-ments in stereotyped and syntactical se-quences versus flexible, less rigidly structuredsequences. The clinical importance of thesebrain regions for normal and pathological se-quence control cannot be overstated. Parkin-son's disease, Huntington's disease, andTourette syndrome all have basal ganglia dis-turbances associated with movement se-quences. A better understanding of the neu-ral mechanisms related to sequence controlmight lead to new therapeutic tactics for neu-rological treatments.

MOTOR SYSTEMS

REFERENCES

Aldridge JW and Berridge KG (1998) Coding of serial or-der by neostriatal neurons: A "natural action" ap-proach to movement sequence. Journal of Neuro-science 18:2777-2787.

Aldridge JW, Berridge KG, Herman M, Zimmer L (1993)Neuronal coding of serial order: Syntax of groom-ing in the neostriatum. Psychological Science4:391-395.

Alheid GF and Heimer L (1988) New perspectives inbasal forebrain organization of special relevance forneuropsychiatric disorders: The striatopallidal,amygdaloid, and corticopetal components of sub-stantia innominata. Neuroscience 27:1-40.

Reiser DG and HoukJC (1998) Model of cortical-basalganglionic processing: Encoding the serial orderof sensory events. Journal of Neurophysiology 79:3168-3188.

Berns GS and Sejnowski TJ (1998) A computationalmodel of how the basal ganglia produce sequences.Journal of Cognitive Neuroscience 10:108-121.

Berridge KG (1989a) Progressive degradation of serialgrooming chains by descending decerebration. Be-havioural Brain Research 33:241-253.

Berridge KG (1989b) Substantia nigra 6-OHDA lesionsmimic striatopallidal disruption of syntactic groom-ing chains—A neural systems analysis of sequencecontrol. Psychobiology 17:377-385.

Berridge KG (1990) Comparative fine structure of action:Rules of form and sequence in the grooming pat-terns of six rodent species. Behaviour 113:21-56.

Berridge KG and Aldridge JW (2000a) Super-stereotypy.I: Enhancement of a complex movement sequenceby systemic dopamine Dl agonists. Synapse 37:194-204.

Berridge KG and Aldridge JW (2000b) Super-stereotypy.II: Enhancement of a complex movement sequenceby intraventricular dopamine Dl agonists. Synapse37:205-215.

Berridge KG and Fentress JC (1987a) DeafFerentationdoes not disrupt natural rules of action syntax. Be-havioural Brain Research 23:69-76.

Berridge KC and Fentress JC (1987b) Disruption of nat-ural grooming chains after striatopallidal lesions.Psychobiology 15:336-342.

Berridge KC, Fentress JC, Parr H (1987) Natural syntaxrules control action sequence of rats. BehaviouralBrain Research 23:59-68.

Berridge KC and Whishaw IQ (1992) Cortex, striatum,and cerebellum: control of serial order in a groom-ing sequence. Experimental Brain Research 90:275-290.

Bolles RC (1960) Grooming behavior in the rat. Journal

Page 166: 0195162854

Chapter 13. Grooming 149

of Comparative and Physiological Psychology 53:306-310.

Colonnese MT, Stallman EL, Berridge KG (1996) On-togeny of action syntax in altricial and precocial ro-dents: Grooming sequences by rat and guinea pigpups. Behaviour 113:1165-1195.

Cromwell HC and Berridge KC (1996) Implementationof action sequences by a neostriatal site: A lesionmapping study of grooming syntax. Journal of Neu-roscience 16:3444-3458.

Deniau JM, Menetrey A, Charpier S (1996) The lamel-lar organization of the rat substantia nigra parsreticulata: Segregated patterns of striatal afferentsand relationship to the topography of corticostriatalprojections. Neuroscience 73:761-781.

Dunn AJ (1988) Studies on the neurochemical mecha-nisms and significance of ACTH-induced grooming.Annals of the New York Academy of Sciences525:150-168.

Dunn AJ and Berridge CW (1990) Physiological and be-havioral responses to corticotropin-releasing factoradministration: Is CRF a mediator of anxiety orstress responses? Brain Research—Brain ResearchReviews 15:71-100.

Kermadi I and Joseph JP (1995) Activity in the caudatenucleus of monkey during spatial sequencing. Jour-nal of Neurophysiology 74:911-933.

Lashley KS (1951) The problem of serial order in be-havior. In: Cerebral mechanisms in behavior (Jef-fress LA, ed.), pp. 112-146. New York: Wiley.

Lieberman P (2000) Human Language and Our Reptil-ian Brain: The Subcortical Bases of Speech, Syntax,and Thought. Harvard University Press, Cambridge.

Matsumoto N, Hanakawa T, Maki S, Graybiel AM,Kimura M (1999) Nigrostriatal dopamine system inlearning to perform sequential motor tasks in a pre-

dictive manner. Journal of Neurophysiology 82:978-998.

Meyer-Luehmann M, Thompson JF, Berridge KC,Aldridge JW (2002) Substantia nigra pars reticulataneurons code initiation of a serial pattern: Implicationsfor natural action sequences and sequential disorders.European Journal of Neuroscience 16:1599-1608.

Molloy AG and Waddington JL (1987) Assessment ofgrooming and other behavioural responses to theD-l dopamine receptor agonist SK & F 38393 andits R- and S-enantiomers in the intact adult rat. Psy-chopharmacology (Berl) 92:164-168.

Mushiake H and Strick PL (1995) Pallidal neuron activ-ity during sequential arm movements. Journal ofNeurophysiology 74:2754-2758.

Nauta WJH and Domesick VB (1984) Afferent and ef-ferent relationships of the basal ganglia. In: Func-tions of the basal ganglia. Ciba Foundation Sympo-sium 107, pp. 3-29. London: Pitman.

Rapoport JL (1989) The boy who couldn't stop washing.New York: Penquin Books.

Richmond G and Sachs BD (1978) Grooming in Norwayrats: The development and adult expression of acomplex motor pattern. Behaviour 75:82-96.

Spruijt BM, Cools AR, Ellenbroek BA, Gispen WH(1986) Dopaminergic modulation of ACTH-inducedgrooming. European Journal of Pharmacology120:249-256.

Spruijt BM, VanHooffJARA, Gispen WH (1992) Ethol-ogy and neurobiology of grooming behavior. Phys-iological Reviews 72:825-852.

Starr BS and Starr MS (1986) Differential effects ofdopamine Dl and D2 agonists and antagonists onvelocity of movement, rearing and grooming in themouse. Implications for the roles of Dl and D2 re-ceptors. Neuropharmacology 25:455-463.

Page 167: 0195162854

Locomotion

GILLIAN MUIR

Locomotion is one of the most common be-haviors in which rats engage, and thus the as-sessment of locomotor abilities is an essentialcomponent of many behavioral analyses. Thisis particularly true for rodent models of dis-eases such as spinal injury and stroke, wherethe recovery of locomotion is the ultimategoal of experimental therapies. In addition,many of the more complicated behavioraltasks that are assessed in the laboratory re-quire that rats move over ground and be ableto use their limbs well. Results from suchanalyses may be confounded if rats have lo-comotor impairments that affect their per-formance on behavioral tasks. The measure-ment of locomotion in rats is difficult,however, because of their small size and therapidity with which they can move and ma-neuver. A thorough understanding of both themusculoskeletal and neural requirements tomove over ground is necessary to properly as-sess locomotor behavior in these animals.This chapter provides an overview of the cur-rent knowledge on the mechanics and neuralcontrol of locomotion in the rat and outlinesthe methods available for measuring locomo-tor abilities in this species.

MECHANICS OF LOCOMOTION

To move over ground, rats, like all terrestrialanimals, need to support their body weightagainst gravity, maintain their posture andequilibrium, and provide propulsion in the di-rection of their movement (Grillner, 1975). In

addition, rats need to alter their speed andnavigate uneven terrain (Grillner, 1975,1981).This section discusses the mechanical aspectsof overground locomotion, including themovements of the individual limbs and the co-ordination between limbs that allows rats tomove over ground at different speeds.

THE STEP CYCLE

During locomotion in all limbed terrestrial an-imals, including rats, individual limbs moverepeatedly through a step cycle that consists ofa stance phase and a swing phase. During thestance phase, the limb is in contact with theground surface and thus contributes to weightsupport and propulsion of the body relative tothe ground (left forelimb in Fig. 14-1A). Limbaction during the stance phase consists of aninitial flexion of the limb joints as the limbyields under the weight of the body (Philip-son's E2 phase) (Grillner, 1975); then the limbjoints extend, pushing the rat forward (E3).During the swing phase, the limb is flexed andmoved forward relative to the motion of thebody (Philipson's F phase), and in the finalstages of the swing phase, the limb joints areextended to prepare for the subsequent stancephase (EO (left forelimb in Fig. 14-1B). Eachlimb moves through the step cycle only oncefor each stride, which is defined as the com-plete pattern of limb movements that beginswith the onset of ground contact of an indi-vidual limb and ends with the subsequentground contact of the same limb.

Stride parameters, like almost all mea-

150

14

Page 168: 0195162854

Chapter 14. Locomotion 151

Figure 14-1. Rat trotting unrestrained for a food reward. During each stride, diagonal limb pairs make con-tact with the ground sequentially (A, right forelimb plus left hindlimb; C, left forelimb plus right hindlimb),separated by two brief aerial phases (B and D). For each limb, the stride is composed of a stance phase (e.g.,right forelimb in A) and a swing phase (e.g., right forelimb in B through D). (Photograph credit: LauraTaylor.)

surements related to locomotion, vary greatlywith the forward speed of movement. Speed,by definition, is a function of the stride lengthand the stride duration. Rats, like other ani-mals, increase speed both by lengthening thestride and by shortening the stride duration.Because the durations of both the stance phaseand the swing phase make up the stride du-ration, these parameters also change withthe speed of locomotion. Importantly, theydo not change similarly—the stance phaseshortens dramatically as velocity increases,whereas the duration of the swing phasechanges very little (Fig. 14-2). The shorteningof the stance phase has important conse-quences for the events that occur during thisphase, particularly the forces that must be ex-erted on the ground to support and propel therat forward. These forces, the ground reactionforces, must be exerted over increasinglyshorter time periods as the speed of locomo-tion increases (Fig. 14-3).

The actual movement of the limbs andlimb segments throughout the step cycle isproduced by action of the limb muscles,which are activated in a particular pattern soas to rotate the limb segments on each other.The characteristic movement pattern of thelimb segments has been accurately measuredin the rat for different speeds of locomotion(Fischer et al., 2002). In general, locomotorlimb action in rats is similar to that of other

small mammals in that they maintain acrouched limb posture at all speeds, provid-ing increased maneuverability and stabilitycompared with larger animals with more up-right limbs (Biewener, 1989, 1990, 1983; Fis-cher et al., 2002).

In addition to the dependence of individ-ual limb parameters on speed, the coordina-tion between the limbs (i.e., gait) also changeswith speed of locomotion. There are manydifferent gaits that quadrupedal animals use,and these are well described for species suchas the horse (Adams, 1987). Nevertheless,walking, trotting, and galloping are the basicgaits used by most quadrupeds, including rats.The following discussion describes the ener-getics and limb dynamics of these gaits.

WALKING

During walking, the limbs are used as rigidstruts. During the first half of the stance phasefor each limb, the body initially rises up overeach limb and then falls during the latter halfof stance. At the same time, during the firsthalf of stance, the limb produces a brakingforce on the body; then, as the body movesforward of the limb, propulsive force is pro-duced which accelerates the body. In this way,the rat is rising and falling, decelerating andaccelerating with each step. This oscillatingpattern provides a means for reducing the en-

Page 169: 0195162854

152 MOTOR SYSTEMS

0 20 40 60 80 100 120

Speed (ems'1)

Figure 14-2. Effect of speed on stride, stance, and swing du-ration during rat locomotion. Stride duration decreases withincreasing locomotor speed (A). This decrease is largely dueto the reduction in stance duration (B) as swing duration re-mains relatively constant over a large speed range (Q. Dif-ferent symbols represent data from different individuals.(From Gillis and Biewener, 2001, with permission.)

ergetic costs of locomotion. Forward kineticenergy is initially converted to potential en-ergy during the first half of the stance phase,as the rat slows down and rises up. In the lasthalf of the stance phase, potential energy istransferred to forward kinetic energy as therat falls and speeds up (Cavagna et al, 1977).This alternating transfer can reduce the ener-getic cost of overground locomotion at thewalk by up to 75% (Heglund et al., 1982).

During walking in quadrupeds, eithertwo or three limbs contact the ground at anyone time. The walking gait used by most ratsis referred to as a lateral walk (Gillis andBiewener, 2001). During this gait, the order oflimb contact in one stride (beginning with theleft hindlimb as reference) is left hindlimb, leftforelimb, right hindlimb, and right forelimb.On a treadmill, rats walk to move at speeds be-tween 0 and 55 cm/sec (Gillis and Biewener,2001). Over ground, rats walk at the samespeed range (0 to 55 cm/sec), but it is difficultto record consistent walking for several strides,because rats moving at these slow speeds areusually exploring at the same time and fre-quently stop and start and change directions.

TROTTING

During walking, the maximum stride lengthattainable, and therefore the maximum speedattainable, is limited by leg length, so animals,including rats, must incorporate an aerialphase into the gait to further increase speed.This requires a switch to trotting inquadrupeds, or to running in bipeds. Duringthese gaits, the limbs act as springs rather thanas struts. During the first half of the stancephase, the body is slowing down as the limbproduces a braking force on the body, just asfor walking (see Fig. 14-3, fore-aft force). Un-like during walking, however, the body is alsofalling during early stance and the forward ki-netic energy of the body is converted not topotential energy but to elastic energy in ten-dons and ligaments as the limb is loaded. Thiselastic energy is then released as forward ki-

Page 170: 0195162854

Chapter 14. Locomotion 153

Figure 14-3. The effect of locomo-tor speed on ground reaction forcesin rats during trotting at 55 and 90cm /sec. As speed increases, peakground reaction forces also increasebecause these forces are exerted foran increasingly shorter time spanwith each ground contact. Forces inthree orthogonal directions (vertical,fore-aft, mediolateral) are expressedin Newtons per kilogram of bodyweight. Each force recording showsforces exerted by a forelimb (stanceduration indicated by black horizon-tal bar) followed by the ipsilateralhindlimb (gray horizontal bar). Verti-cal forces demonstrate that the bodyweight is borne relatively equallybetween the forelimbs and thehindlimbs. Fore-aft forces demon-strate that the forelimbs producemost of the braking force, whereasthe hindlimbs produce most ofthe propulsive forces. Mediolateralforces are mainly directed laterally.Force recordings are from a singleindividual.

netic energy again as the rat springs off thelimb in the latter half of stance.

During trotting, only two limbs contactthe ground at any one time—the rat movesfrom one diagonal limb pair (e.g., right fore-limb and left hindlimb) to the next diagonalpair (left forelimb and right hindlimb). Atfaster trotting speeds, the rat may actuallyjump from one diagonal limb pair to theother, so that there would be two phases dur-ing a trotting stride when the rat is not incontact with the ground (see Fig. 14-1). Asdemonstrated in Figure 14-1, the order oflimb contacts during a trotting stride wouldbe right forelimb and left hindlimb (aerialphase), left forelimb and right hindlimb (ae-rial phase). Although the forelimb andhindlimb diagonal pair may contact theground simultaneously, often the forelimb orthe hindlimb contacts the ground slightly ear-lier (i.e., about 20 milliseconds) than its diag-onal partner. Which limb makes contact first

depends in part on the speed of movement aswell as differences between individual rats.On a treadmill, rats trot to move at speeds be-tween 55 and 80 cm/sec (Gillis and Biewener,2001). Rats that have been trained to moveover ground for a food reward generally usea trotting gait and move at speeds from 50 to90 cm/sec (Muir and Whishaw, 1999b, 2000;Webb and Muir, 2002, 2003a, 2003b).

GALLOPING

Galloping or bounding gaits occur at the high-est speeds in quadrupedal animals, includingrats. During these gaits, stride length is fur-ther extended by incorporating movements ofthe torso, which extends and flexes alternatelythroughout the stride. The hindlimbs andforelimbs act more in synchrony during gal-loping than during slower gaits. At the begin-ning of the galloping stride, the trunk is flexedas hindlimbs are brought forward and placed

Page 171: 0195162854

154 MOTOR SYSTEMS

on the ground. The torso then extends as therat stretches forward onto the forelimbs andthen flexes again as the hindlimbs leave theground and are brought under the body. Anaerial phase is incorporated into the gait as therat moves off the forelimbs and before thehindlimbs contact the ground for the nextstride. In bounding gaits, there is a second aer-ial phase during each stride, which occurs asthe trunk is extending and the rat jumps fromthe hindlimbs onto the forelimbs. The ener-getics of galloping gaits combine the strategiesof both walking and trotting, in that forwardkinetic energy is converted to, and recoveredfrom, gravitational potential energy and elas-tic strain energy at different phases in thestride (Cavagna et al., 1977).

During galloping gaits, the precise coor-dination of the limbs differs at differentspeeds, and there are either one, two, or nolimbs on the ground at any one time duringthe stride. At slower gallops, the hindlimbsand the forelimbs do not act completely insynchrony; that is, one hindlimb makes con-tact with the ground earlier than the oppositehindlimb. The disparity between contacttimes is generally greater for the forelimb pairthan for the hindlimbs. For each limb pair, thelimb that makes contact with the groundslightly earlier than the other is referred to asthe trailing limb, and the opposite limb is theleading limb. The order of limb contact in aslow gallop would be trailing hindlimb, lead-ing hindlimb, trailing forelimb, leading fore-limb. In some instances, the leading hindlimband the trailing forelimb may contact theground simultaneously (a gait referred to as acanter in horses). At faster gallops, the fore-limbs as well as the hindlimbs move increas-ingly in synchrony, until the rat begins tobound from the forelimbs onto the hindlimbsand then onto the forelimbs again. On a tread-mill or over ground, most rats begin to gallopat speeds greater than 80 cm/sec (Muir andWhishaw, 2000; Gillis and Biewener, 2001).For both treadmill and over-ground locomo-tion, there is a wide range of speeds, from ap-

proximately 70 to 100 cm/sec, at which ratseither trot or gallop.

NEURAL CONTROLOF LOCOMOTION

OUTPUT FROM SPINAL CIRCUITRY CANPRODUCE THE BASIC STEPPING PATTERN

During locomotion, limb muscles need to beactivated in a particular pattern such that thisactivity can (I) move limbs in a cyclical manner(i.e., through the step cycle), (2) provide alter-nation of right and left limbs, and (3) coordinateforelimbs and hindlimbs according to the gaitbeing used. It is well established in vertebratesthat the circuitry required to produce the neu-ral output for the first two of these tasks—movement of the limbs through the step cycleand alternation of right and left limbs—is con-tained completely within the spinal cord (Grill-ner and Wallen, 1985). Cats with a completespinal transection are able to produce steppingmovements that are identical to those in the in-tact animal (Grillner and Zangger, 1979; Be-langer et al., 1988). Even after removal of all sen-sory feedback from the limb, the limb is still ableto move through the step cycle, indicating thatthe spinal cord itself is able to generate the os-cillating output required to activate muscles inan pattern appropriate for stepping.

Similar experiments have not been re-peated in the adult rat, although there is muchinformation on spinal pattern generation inthe neonatal rat (Cazalets et al., 1995; Cowleyand Schmidt, 1997; Kiehn and Kjaerulff, 1998;Ballion et al., 2001). The spinal locomotor cir-cuitry for the hindlimbs in rats appears to bedistributed throughout the lumbar enlarge-ment and the lower thoracic cord, and likelyconsists of many rhythm generators that con-trol different muscles or joints, although themost excitable rhythmic activity is located inthe rostralmost part of the lumbar enlarge-ment (Kiehn and Kjaerulff, 1998). For the fore-limbs, the locomotor circuitry responsible for

Page 172: 0195162854

Chapter 14. Locomotion 155

generating rhythmic activity is located in thelower cervical and uppermost thoracic spinalsegments (Ballion et al, 2001).

Of course, to produce functional locomo-tion, spinal locomotor circuits normally requireinput from two important sources. First, seg-mental afferent feedback from the limb acts toconstantly regulate and reinforce ongoing mus-cle activity as well as to provide information tocontrol the transitions between the stance andswing phases of the step cycle. Second, supra-spinal inputs from the brain stem and higherregions of the brain are required for the initia-tion and ongoing control of locomotion. Theinfluence of both of these sources of input isdiscussed in more detail. Much of the researchin this area has been acquired in the cat model,but specific information regarding the rat issupplied when available.

NORMAL STEPPING REQUIRES PERIPHERALAFFERENT INPUT FROM THE LIMB

Segmental afferent feedback arises from sev-eral receptors distributed throughout the limb,including muscle spindles, Golgi tendon or-gans, and receptors located in the skin andjoint capsules. Muscle spindles provide infor-mation regarding muscle length and velocityof length changes, whereas Golgi tendon or-gans provide information about tendon forces.Input from both of these receptors have beenshown to influence the level of extensor mus-cle activity during the stance phase, such thatloss of this input substantially reduces thestrength of extensor muscle activity (Pearsonand Collins, 1993; Guertin et al., 1995; Hiebertand Pearson, 1999). Inputs from cutaneous re-ceptors, especially those on the footpads, canalso result in an increase in extensor activityduring the stance phase (Duysens and Pearson,1976). Importantly, the influence of both mus-cle and cutaneous receptors on limb muscleactivity is not constant but instead is modu-lated throughout the step cycle (Forssberg,1979; Drew and Rossignol, 1985, 1987). Dur-ing the swing phase, for example, stimulation

of cutaneous receptors of the paw results inlimb flexion rather than limb extension (Drewand Rossignol, 1985, 1987).

Another major role of afferent input is inthe regulation of transitions between stanceand swing (for a review, see Pearson et al.,1998). During the latter half of stance, the limbbegins to move caudally and bears less weight,and there is increased activity in muscle spin-dles within hip flexor muscles as well as a de-crease in Golgi tendon organ activity withinlimb extensors. Both of these inputs con-tribute to the decreased activity in extensorsand an activation of limb flexors, resulting inlimb flexion and thus the onset of the swingphase (Pearson et al., 1998).

SUPRASPINAL INPUT IS REQUIREDFOR INITIATION AND ONGOINGCONTROL OF LOCOMOTION

The other major source of input onto spinallocomotor circuits arises from the brain.Spinal circuits receive direct input from thecerebral cortex, red nuclei, vestibular nuclei,and numerous nuclei in the pons and medulla,including the locus coeruleus and raphe nu-clei. In addition, there are many areas in thebrain that are involved in locomotor controlbut do not project directly to the spinal cord;these include the cerebellum, the basal gan-glia, and several areas collectively known aslocomotor regions. These are areas in the brainthat, when stimulated with electric current,are able to initiate locomotion in decerebrateanimals, including rats (Atsuta et al., 1990,1991). These areas are located in the mesen-cephalon (mesencephalic locomotor region),the hypothalamus, and the deep cerebellar nu-clei. For an in-depth discussion of the possibleroles of these locomotor regions during vari-ous forms of locomotion, see Jordan (1998).

In the rat, the locomotor contributions ofbrain regions with direct spinal input have beenexamined by lesioning various spinal funiculi.Axons arising from nuclei in the pons andmedulla travel in the ventral half of the spinal

Page 173: 0195162854

156 MOTOR SYSTEMS

cord. These inputs likely play an important rolein locomotor control, as large ventral lesionsproduce severe impairment of overground lo-comotion in the rat (Loy et al., 2002; Schuchtet al., 2002). There appears to be some func-tional redundancy in these ventral pathways, inthat small lesions involving only portions of theventral funiculi had mild effects on overgroundlocomotion (Loy et al., 2002). Of course, theaxons affected by ventral lesions arise frommany different nuclei in the brain stem, and itis possible that more detailed information onthe contributions of individual nuclei can be ob-tained using more sensitive techniques withwhich to measure locomotor abilities.

Descending inputs located in the dorsalhalf of the spinal cord arise largely from thecerebral cortex and the red nucleus. These in-puts are thought to be primarily required forskilled locomotion, such as making adjust-ments for uneven terrain or avoiding obstacles.Large dorsal lesions do not greatly affect grosslocomotor abilities overground in the rat butdo produce large deficits in skilled locomotion,such as ladder walking (Schucht et al., 2002).

More specific lesions of dorsal pathwayshave distinguished the contributions of thecorticospinal and rubrospinal tracts. The cor-ticospinal tract does not appear to contributeto over ground locomotion but is required forskilled locomotion (Metz et al., 1998; Muirand Whishaw, 1999a; Metz and Whishaw,2002; Whishaw and Metz, 2002). The rubro-spinal tract also contributes to skilled loco-motion but in addition has been shown to playsome role during overground locomotion inrats, in that unilateral lesions of the red nu-cleus or of the tract itself result in a perma-nent locomotor asymmetry (Muir andWhishaw, 2000; Webb and Muir, 2003b).

MEASURING LOCOMOTIONIN THE LABORATORY

Much of what we know regarding quantifica-tion of locomotion in rats has arisen from re-search focusing on central nervous system dis-

orders such as spinal cord injury or stroke.There are several reviews that have outlinedappropriate methods for measuring locomo-tor recovery in rats after spinal cord injury(Goldberger et al., 1990; Kunkel et al., 1993;Metz et al., 2000; Muir and Webb, 2000).These methods can be applied to locomotoranalysis in normal rats and in rat models ofmany different diseases. The following sectiondescribes the methods available for measuringlocomotion in rats. For further details on spe-cific techniques, readers are referred to the as-sociated references.

LOCOMOTOR RATING SCALES

Locomotor rating scales, such as the BBB scaleand the Tarlov scale, are ordinal rating scalesdevised to assess locomotor movements ofthe hindlimbs after thoracic spinal injury(Basso et al., 1995; Fehlings and Tator, 1995).Animals are observed in a open area and givena score based on criteria such as the move-ments of the hindlimbs, the tail position, andthe coordination between forelimbs andhindlimbs. Such assays are quick to perform,require a minimum of equipment, and aregeneral enough to include animals with awide range of functional abilities. Neverthe-less, it is important to note that locomotorscales are relatively specific for the type of in-jury for which they were designed and maynot accurately measure recovery in differentinjury models or after different treatments.For example, the BBB scale is useful for as-sessing locomotion after thoracic contusioninjuries or thoracic dorsal hemisections, but isless successful at describing the functionalstates of animals with other spinal cord lesions(Metz et al., 2000; Loy et al., 2002; Schucht etal., 2002; Webb and Muir, 2002). Finally, a lim-itation of any locomotor assay that relies onthe observation of freely moving animals is thedependence on the motivation of individual an-imals; rats have been shown to display somefunctional "deficits" such as toe-dragging dur-ing exploratory behavior that completely dis-appear when the same animals are perform-

Page 174: 0195162854

Chapter 14. Locomotion 157

ing a motivated behavior such as trotting ona runway for a food reward (Webb and Muir,2002). The following sections describe themeasurements that are best to apply to ani-mals that have been trained to complete a lo-comotor task.

KINEMATIC MEASUREMENTS

Kinematic measurements encompass a widerange of measures, including distances, an-gles, velocities, and accelerations of the body,of the limbs, and of the limb segments. Stridecharacteristics, such as stride and step lengths,and stride durations are considered to be kine-matic measurements, as are limb joint angles.Kinematic measurements are essentially aquantitative and detailed description of theanimal's movements. Many kinematic mea-sures, as for other locomotor parameters, nor-mally vary with the speed and gait of theanimal, so it is important that speed of move-ment is recorded along with the measure(s) ofinterest.

Step and stride characteristics can be ob-tained in several ways. Animals can be video-taped as they move over ground or on a tread-mill. During overground locomotion, thecamera may be positioned for a lateral view(see Fig. 14-1) or a caudal view or, with thehelp of a transparent floor surface and a mir-ror placed at 45° beneath the floor, can pro-vide a ventral view that clearly shows theplacement of the footpads (Cheng et al., 1997;Webb and Muir, 2003a). Animals should betrained to run in a straight path so as to im-prove the accuracy of the measurements, par-ticularly the distance measurements. Frame-by-frame analysis of the videotape can thenprovide step and stride lengths and durations,as well as durations of the swing and stancephases. Even more simply, use of an inkpadand paper allows the measurement of steplengths and foot placement, although tempo-ral measurements are not available (Kunkeland Bregman, 1990).

Analysis of limb joint angles, such asshoulder, elbow, or knee angles, requires the

use of video or digital cameras and frame-by-frame analysis. Specialized computer equip-ment is available to assist in capturing videoor digital frames, thus supplying a time seriesof limb positions during locomotor move-ments. Three-dimensional analysis, using atleast two cameras simultaneously, is also pos-sible, although normal movement of limbs ina single plane during locomotion makes two-dimensional analysis acceptable.

A significant problem with the measure-ment of joint angles during locomotion arisesfrom the inability to accurately identify limbsegment positions. Markers placed on the skinmove as the skin moves over the limbs dur-ing normal locomotion. This movement in-troduces systematic errors in the location ofthe joint positions, particularly the proximaljoints, which are compounded in small mam-mals such as rats that possess a crouched limbposture. The only solution is cineradiography,which provides a clear view of the limb bonesduring movement. Normal kinematics usingthis technique have been recorded for the rat(Fischer et al., 2002). Although cineradiogra-phy is impractical to use for most studies, theinformation obtained from such studies couldbe used to calculate correction factors for skinmovements as has been done for other species(van den Bogert et al., 1990).

KINETIC MEASUREMENTS

Muscles move limbs by exerting forces onlimb segments; the limbs then exert forcesagainst surfaces to move the animal. Kineticsis the measurement of these forces. Force orstrain transducers, either implanted within thelimb in muscle tendons or on bone surfaces,or positioned on external surfaces, can be usedto provide a sensitive measure of the ways inwhich animals move (Gillis and Biewener,2002; Biewener and Blickhan, 1988; Bieweneret al., 1988; Biewener and Taylor, 1986; Muirand Whishaw, 1999a, 1999b, 2000).

Ground reaction forces are the forces ex-erted through the limb on the ground duringlocomotion (see Fig. 14-3). They are mea-

Page 175: 0195162854

158 MOTOR SYSTEMS

sured with force platforms and provide a sen-sitive, quantitative, and noninvasive methodfor measuring locomotion. This method is es-pecially useful for the analysis of rat locomo-tion because even though rats are capable ofmaking quick maneuvers and adjustments totheir gait, the slightest changes are manifestthrough the ground reaction forces. Record-ing of these forces thus provides unique andquantifiable information that is not obtainablewith other methods. Measurement of groundreaction forces has demonstrated the subtleand characteristic methods that rats use tocompensate for different central nervous sys-tem lesions (Muir and Whishaw, 1999a,1999b, 2000; Webb and Muir, 2002, 2003b).Measurement of these forces, however, re-quires specialized equipment that must becustomized or custom-built to provide theappropriate size and sensitivity for studies inthe rat.

ELECTROMYOGRAPHY

Recording of muscle electrical activity duringlocomotion provides valuable information onhow muscles are used to move the limbsthrough the step cycle. Simultaneous record-ing from many muscles during locomotion ispossible with available techniques, althoughmost studies in the rat have been limited tothe measurement of one or two muscles (Co-hen and Cans, 1975; Loeb and Cans, 1986;Roy et al., 1991; de Leon et al., 1994; Gorassiniet al., 1999, 2000; Gramsbergen et al., 2000;Gillis and Biewener, 2001; Kaegi et al., 2002;Schumann et al., 2002). Deviations from thenormal pattern after lesioning or treatmentscan help determine the ways in which thenew movement pattern differs from normal(Gramsbergen et al., 2000; Kaegi et al., 2002).

Importantly, however, there is a widerange of normal muscle activity patterns. Thisis due in part to the large number of musclesthat span each joint, so that the same move-ment can be produced by through the re-cruitment of several different combinations of

muscles. In addition, electromyography doesnot provide direct information about thestrength of muscle contraction because of thecomplex relationship between individualmuscle fiber activation and force production(Basmajian and De Luca, 1985). Recording ofelectromyographic data entails some disad-vantages because of the invasive nature of theelectrodes and associated recording devices,which may alter the animars normal loco-motor behavior. Implantation of electrodesinto the muscle also requires surgery as wellas careful postoperative care to prevent infec-tion and maintain the electrodes in positionfor the duration of the study.

LOCOMOTOR TASKS

The previous section discussed measurementtechniques applicable to overground locomo-tion, but of course rats can be trained to lo-comote in a number of situations, each ofwhich is amenable to measurement usingthese same methods. Rats can be trained tolocomote on a treadmill and, because the an-imal is essentially stationary, kinematic andelectromyographic measurements are some-what easier to obtain compared with over-ground locomotion. Treadmill speed can becontrolled precisely, as can the level of incline,so that uphill or downhill locomotion can beexamined (Gillis and Biewener, 2002).

Rats can also be trained in more skilledlocomotor tasks, such as ladder or beam walk-ing. In addition to the kinematic, kinetic, andelectromyographic measurements previouslydiscussed, other measures of locomotor abil-ity can be recorded for these tasks. These in-clude counts of footslip errors and scoring ofthe paw positions on the ladder rungs orbeams (Muir and Webb, 2000; Metz andWhishaw, 2002). To alter the degree of diffi-culty of skilled locomotor tasks, it is possibleto vary or randomize the spacing between lad-der rungs or to vary the width of the beam(Metz et al., 2000; Metz and Whishaw, 2002).Rats can also be trained to climb inclines or

Page 176: 0195162854

Chapter 14. Locomotion 159

ropes, behaviors that also lend themselves tothe measurement techniques previously dis-cussed (Thallmair et al., 1998; Ramon-Cuetoet al., 2000). Their natural motor abilities canbe exploited for the investigation of a varietyof tasks that, in combination with differentmeasurement techniques, can provide a com-prehensive analysis of locomotor skills in thisspecies.

REFERENCES

Adams OR (1987) Natural and artificial gaits. In: Adam'slameness in horses (Stashak TS, ed.), pp. 834-839.Philadelphia: Lea & Febiger.

Atsuta Y, Garcia-Rill E, Skinner RD (1990) Characteris-tics of electrically induced locomotion in rat in vitrobrain stem-spinal cord preparation. Journal of Neu-rophysiology 64:727-735.

Atsuta Y, Garcia-Rill E, Skinner RD (1991) Control of lo-comotion in vitro: I. Deafferentation. Somatosen-sory Motor Research 8:45-53.

Ballion B, Morin D, Viala D (2001) Forelimb locomotorgenerators and quadrupedal locomotion in theneonatal rat. European Journal of Neuroscience14:1727-1738.

Basmajian JV and De Luca CJ (1985) EMG signal am-plitude and force. In: Muscles alive: Their functionsrevealed by electromyography, pp. 187-200. Balti-more: Williams and Wilkins.

Basso DM, Beattie MS, Bresnahan JC (1995) A sensitiveand reliable locomotor rating scale for open fieldtesting in rats. Journal of Neurotrauma 12:1-21.

Belanger M, Drew T, Rossignol S (1988) Spinal loco-motion: A comparison of the kinematics and theelectromyographic activity in the same animal be-fore and after spinalization. Acta Biologica Hungar-ica 39:151-154.

Biewener AA (1983) Locomotory stresses in the limbbones of two small mammals: The ground squirreland chipmunk. Journal of Experimental Biology103:131-154.

Biewener AA (1989) Scaling body support in mammals:Limb posture and muscle mechanics. Science 245:45-48.

Biewener AA (1990) Biomechanics of mammalian ter-restrial locomotion. Science 250:1097-1103.

Biewener AA and Blickhan R (1988) Kangaroo rat loco-motion: Design for elastic energy storage or acceler-ation? Journal of Experimental Biology 140:243-255.

Biewener AA, Blickhan R, Perry AK, Heglund NC, Tay-

lor CR (1988) Muscle forces during locomotion inkangaroo rats: Force platform and tendon bucklemeasurements compared. Journal of ExperimentalBiology 137:191-205.

Biewener AA and Taylor CR (1986) Bone strain: A de-terminant of gait and speed? Journal of Experimen-tal Biology 123:383^00.

Cavagna GA, Heglund NC, Taylor CR (1977) Mechan-ical work in terrestrial locomotion: Two basicmechanisms for minimizing energy expenditure.American Journal of Physiology 233:R243-R261.

Cazalets JR, Borde M, Clarac F (1995) Localization andorganization of the central pattern generator forhindlimb locomotion in newborn rat. Journal ofNeuroscience 15:4943^4951.

Cheng H, Almstrom S, Gimenez LL, Chang R, Ove OS,Hoffer B, Olson L (1997) Gait analysis of adult para-plegic rats after spinal cord repair. ExperimentalNeurology 148:544-557.

Cohen AH and Cans C (1975) Muscle activity in rat lo-comotion: Movement analysis and electromyogra-phy of the flexors and extensors of the elbow. Jour-nal of Morphology 146:177-196.

Cowley KG and Schmidt BJ (1997) Regional distributionof the locomotor pattern-generating network in theneonatal rat spinal cord. Journal of Neurophysiol-ogy 77:247-259.

de Leon R, Hodgson JA, Roy RR, Edgerton VR (1994)Extensor- and flexor-like modulation within motorpools of the rat hindlimb during treadmill locomo-tion and swimming. Brain Research 654:241-250.

Drew T and Rossignol S (1985) Forelimb responses tocutaneous nerve stimulation during locomotion inintact cats. Brain Research 329:323-328.

Drew T and Rossignol S (1987) A kinematic and elec-tromyographic study of cutaneous reflexes evokedfrom the forelimb of unrestrained walking cats.Journal of Neurophysiology 57:1160-1184.

Duysens J and Pearson KG (1976) The role of cutaneousafferents from the distal hindlimb in the regulationof the step cycle of thalamic cats. ExperimentalBrain Research 24:245-255.

Fehlings MG and Tator CH (1995) The relationshipsamong the severity of spinal cord injury, residualneurological function, axon counts, and counts ofretrogradely labeled neurons after experimentalspinal cord injury. Experimental Neurology 132:220-228.

Fischer MS, Schilling N, Schmidt M, Haarhaus D, WitteH (2002) Basic limb kinematics of small therianmammals. Journal of Experimental Biology 205:1315-1338.

Forssberg H (1979) Stumbling corrective reaction: Aphase-dependent compensatory reaction during lo-comotion. Journal of Neurophysiology 42:936-953.

Page 177: 0195162854

160 MOTOR SYSTEMS

Gillis GB and Biewener AA (2001) Hindlimb musclefunction in relation to speed and gait: In vivo pat-terns of strain and activation in a hip and knee ex-tensor of the rat (Rattus norvegicus). Journal of Ex-perimental Biology 204:2717-2731.

Gillis GB and Biewener AA (2002) Effects of surfacegrade on proximal hindlimb muscle strain and acti-vation during rat locomotion. Journal of AppliedPhysiology 93:1731-1743.

Goldberger ME, Bregman BS, Vierck-CJ J, Brown M(1990) Criteria for assessing recovery of function af-ter spinal cord injury: Behavioral methods. Experi-mental Neurology 107:113-117.

Gorassini M, Bennett DJ, Kiehn O, Eken T, Hultborn H(1999) Activation patterns of hindlimb motor unitsin the awake rat and their relation to motoneuronintrinsic properties. Journal of Neurophysiology82:709-717.

Gorassini M, Eken T, Bennett DJ, Kiehn O, Hultborn H(2000) Activity of hindlimb motor units during lo-comotion in the conscious rat. Journal of Neuro-physiology 83:2002-2011.

Gramsbergen A, IJkema-Paassen J, Meek MF (2000) Sci-atic nerve transection in the adult rat: AbnormalEMG patterns during locomotion by aberrant in-nervation of hindleg muscles. Experimental Neu-rology 161:183-193.

Grillner S (1975) Locomotion in vertebrates: Centralmechanisms and reflex interaction. Physiology Re-view 55:247-304.

Grillner S (1981) Control of locomotion in bipeds,tetrapods and fish. In: Handbook of physiology, sec-tion I: The nervous system (Brooks VB, ed.), pp.1179-1236. Bethesda: American Physiological Society.

Grillner S and Wallen P (1985) Central pattern genera-tors for locomotion, with special reference to ver-tebrates. Annual Review of Neuroscience 8:233-261.

Grillner S and Zangger P (1979) On the central genera-tion of locomotion in the low spinal cat. Experi-mental Brain Research 34:241-261.

Guertin P, Angel MJ, Perreault MC, McCrea DA (1995)Ankle extensor group I afferents excite extensorsthroughout the hindlimb during fictive locomotionin the cat. Journal of Physiology 487(Pt 1):197-209.

Heglund NC, Cavagna GA, Taylor CR (1982) Energet-ics and mechanics of terrestrial locomotion. III. En-ergy changes of the centre of mass as a function ofspeed and body size in birds and mammals. Journalof Experimental Biology 97:41-56.

Hiebert GW and Pearson KG (1999) Contribution ofsensory feedback to the generation of extensor ac-tivity during walking in the decerebrate cat. Journalof Neurophysiology 81:758-770.

Jordan LM (1998) Initiation of locomotion in mammals.

Annals of the New York Academy of Science 860:83-93.

Kaegi S, Schwab ME, Dietz V, Fouad K (2002) Elec-tromyographic activity associated with spontaneousfunctional recovery after spinal cord injury in rats.European Journal of Neuroscience 16:249-258.

Kiehn O and Kjaerulff O (1998) Distribution of centralpattern generators for rhythmic motor outputs inthe spinal cord of limbed vertebrates. Annals of theNew York Academy of Science 860:110-129.

Kunkel BE and Bregman BS (1990) Spinal cord trans-plants enhance the recovery of locomotor functionafter spinal cord injury at birth. Experimental BrainResearch 81:25-34.

Kunkel BE, Dai HN, Bregman BS (1993) Methods to as-sess the development and recovery of locomotorfunction after spinal cord injury in rats. Experi-mental Neurology 119:153-164.

Loeb GE and Cans C (1986) Electromyography for ex-perimentalists. Chicago: University of Chicago Press.

Loy DN, Magnuson DS, Zhang YP, Onifer SM, MillsMD, Cao QL, Darnall JB, Fajardo LC, Burke DA,Whittemore SR (2002) Functional redundancy ofventral spinal locomotor pathways. Journal of Neu-roscience 22:315-323.

Metz GA, Dietz V, Schwab ME, van de Meent MH(1998) The effects of unilateral pyramidal tract sec-tion on hindlimb motor performance in the rat. Be-havioural Brain Research 96:37-46.

Metz GA, Merkler D, Dietz V, Schwab ME, Fouad K(2000) Efficient testing of motor function in spinalcord injured rats. Brain Research 883:165-177.

Metz GA and Whishaw IQ (2002) Cortical and subcor-tical lesions impair skilled walking in the ladder rungwalking test: A new task to evaluate fore- andhindlimb stepping, placing, and co-ordination. Jour-nal of Neuroscience Methods 115:169-179.

Muir GD and Webb AA (2000) Mini-review: Assessmentof behavioural recovery following spinal cord injuryin rats. European Journal of Neuroscience 12:3079-3086.

Muir GD and Whishaw IQ (1999a) Complete locomo-tor recovery following corticospinal tract lesions:Measurement of ground reaction forces duringoverground locomotion in rats. Behavioural BrainResearch 103:45-53.

Muir GD and Whishaw IQ (1999b) Ground reactionforces in locomoting hemi-parkinsonian rats: A de-finitive test for impairments and compensations. Ex-perimental Brain Research 126:307-314.

Muir GD and Whishaw IQ (2000) Red nucleus lesionsimpair overground locomotion in rats: A kineticanalysis European Journal of Neuroscience 12:1113-1122.

Pearson KG and Collins DF (1993) Reversal of the in-

Page 178: 0195162854

Chapter 14. Locomotion 161

fluence of group Ib afferents from plantaris onactivity in medial gastrocnemius muscle duringlocomotor activity. Journal of Neurophysiology70:1009-1017.

Pearson KG, MisiaszekJE, Fouad K (1998) Enhancementand resetting of locomotor activity by muscle af-ferents. Annals of the New York Academy of Sci-ence 860:203-215.

Ramon-Cueto A, Cordero MI, Santos-Benito FF, Avila J(2000) Functional recovery of paraplegic rats andmotor axon regeneration in their spinal cords by ol-factory ensheathing glia. Neuron 25:425—435.

Roy RR, Hutchison DL, Peirotti DJ, Hodgson JA, Edger-ton VR (1991) EMG patterns of rat ankle extensorsand flexors during treadmill locomotion and swim-ming. Journal of Applied Physiology 70:2522-2529.

Schucht P, Raineteau O, Schwab ME, Fouad K (2002)Anatomical correlates of locomotor recovery fol-lowing dorsal and ventral lesions of the rat spinalcord. Experimental Neurology 176:143-153.

Schumann NP, Biedermann FH, Kleine BU, StegemanDF, Roeleveld K, Hackert R, Scholle HC (2002)Multi-channel EMG of the M. triceps brachii in ratsduring treadmill locomotion. Clinical Neurophysi-ology 113:1142-1151.

Thallmair M, Metz GA, Z'Graggen WJ, Raineteau O,

Kartje GL, Schwab ME (1998) Neurite growth in-hibitors restrict plasticity and functional recoveryfollowing corticospinal tract lesions. Nature andNeuroscience 1:124-131.

van den Bogert AJ, van Weeren PR, Schamhardt HC(1990) Correction for skin displacement errors inmovement analysis of the horse. Journal of Biome-chanics 23:97-101.

Webb AA and Muir GD (2002) Compensatory locomo-tor adjustments of rats with cervical or thoracicspinal cord hemisections. Journal of Neurotrauma19:239-256.

Webb AA, Gowribai K, Muir GD (2003a) Fischer (F-344)rats have different morphology, sensorimotor andlocomotor abilities compared to Lewis, Long-Evans, Sprague-Dawley and Wistar rats. Behav-ioural Brain Research 144:143-156.

Webb AA and Muir GD (2003b) Unilateral dorsal col-umn and rubrospinal tract injuries affect over-ground locomotion in the unrestrained rat. Euro-pean Journal of Neuroscience 18:412-422.

Whishaw IQ and Metz GA (2002) Absence of impair-ments or recovery mediated by the uncrossed py-ramidal tract in the rat versus enduring deficits pro-duced by the crossed pyramidal tract. BehaviouralBrain Research 134:323-336.

Page 179: 0195162854

Prehension

IAN Q. WHISHAW 15

Following Peterson's (1932) description ofreaching in the rat, rats have been trained toreach through slots, down tubes, onto rotat-ing tables, onto conveyor belts, off shelves,and through bars to get food. They have per-formed tasks of manipulating puzzle, latches,pushing force transducers, and pressing bars(Whishaw and Middyaeva, 1996). They havegrasped and snapped pasta to measure reachlength, force, and strength (Ballermann et al.,2000, 2001; Remple et al., 2001). Their pawmovements have been observed as they pickup and manipulate food pellets, variouslyshaped pieces of pasta, nuts, and fruits(Whishaw and Coles, 1966). Their limb move-ments have even been analyzed in predatoryacts of catching, manipulating, and eatingcrickets (Ivanco et al., 1996).

The movements made by rats when theyuse their digits, paws, and forelimbs for catch-ing, manipulating, and holding objects arecalled skilled movements. Early views of theevolutionary origins of skilled paw move-ments proposed that they evolved in the pri-mate lineage by modifications of movementsused by the forelimbs for grasping branches.The rat's dexterity refutes this notion. Skilledmovements are widely used by terrestrial ver-tebrates and have been lost in some mam-malian orders and have been elaborated inother mammalian orders (Iwaniuk andWhishaw, 2000). The almost 2000 rodentspecies, which comprise approximately half ofall mammalian species, are members of an or-der with well-developed skilled movements.The laboratory rat is a worthy representativeof this order.

162

Skilled movements are special with re-spect to their neural control. For an animal touse its forelimbs to reach for and manipulateobjects, the limbs must be released from theirfunction of supporting the body against grav-ity. The neural structures that control fore-limb use must be partially different from thoseused for supporting body weight and loco-motion (Metz et al., 1998; Muir and Whishaw,1999). Thus, the forebrain not only containscircuits whose function is in part to suppressmovement subsystems of postural supportand walking but also must contain circuitsthat allow the limbs to be used for manipu-lating objects. Because skilled movementshave a long phylogenetic history, they areprobably conserved across mammalianspecies. For this reason, the skilled move-ments of the laboratory rat are of interest inlines of research directed toward modelinghuman neurological disorders, many of whichmanifest themselves with compromisedskilled movements.

SKILLED MOVEMENTS AND LIMBSTRUCTURE AND MOVEMENT

The skeletal and muscular structures of the ratforelimb are illustrated by Green (1963). Therat forepaw has five digits. The first digit(thumb) is small, but it has a nail, whereas dig-its 2 through 5 have claws. Nails typically in-dicate that a digit is used for skilled move-ments. Rats can move digit 1 medially towardthe palm in a precision grasping pattern forholding an object between the thumb pad and

Page 180: 0195162854

Chapter 15. Prehension 163

the pads of the other digits (Whishaw andColes, 1996). For example, rats use this pincergrip to hold spaghetti (Fig. 15-1). Rats havewell-developed pads on the tips of the digitsand on the palm, and when they grasp and

Figure 15-1. Grasp patterns during pasta eating in the rat.(A) The rat is presenting pasta to its mouth with its rightpaw and pushing the food with the left paw. (B) The grasppattern displayed by the right paw involves holding thepasta between the pads of digit 1 (thumb) and digit 2. (C)Digit 1 has a nail, whereas digits 2 through 5 have claws.The presence of a nail on digit 1 suggests that it may re-ceive considerable use in grasping objects. (Based onWhishaw and Coles, 1996.)

manipulate objects, they typically do so usingthe digit pads, although larger objects will beheld against the palm. Rats have a few sinushairs on the medial surface of the lower armthat project toward the palm of the paw. Thesinus hairs are likely useful for detecting ob-jects an animal is about to grasp and are es-pecially useful for sensing movements of a liveprey object that is held in the paws.

The degrees of freedom of movement ofthe arm of the rat are similar to that of primates,but there are differences in the way that free-dom of movement is achieved. Although hu-mans have a ball-and-socket joint at the shoul-der that allows the upper arm a wide range ofmovement, the rat has a scapula that is teth-ered by muscles, thus permitting almost thesame range of movement. Pronation andsupination of the hand of humans are achievedby rotating the radius and ulna with respect toeach other. The radius and ulna of the rat arefused. Humans cannot rotate the hand aroundthe wrist, but the rat is able to do so. Thus, therat limb has almost the same range of move-ment as the human limb, but the mechanismsunderlying movement freedom are partiallydifferent (Whishaw and Micklyaeva, 1966).

Limb musculature of rats and humans isalso similar, as rats have intrinsic muscles inthe paw but its extensor-flexor movementsand opening-closing movements are con-trolled by forearm muscles. As is the case forother animals, each forearm muscle is con-trolled by a column of spinal cord motor neu-rons, with more distal musculature controlledby more caudally located motor neuroncolumns (McKenna et al., 2000). Wise andDonoghue (1986) reviewed the neuroanatomyof brain structures that control motor neu-rons, and in the rat they are similar to thoseof primates.

FOOD HANDLING

Rats consume a wide range of foodstuffs, in-cluding grasses, nuts, leaves, roots, and almostany type of food discarded by humans. They

Page 181: 0195162854

164

pick up food using a characteristic sequenceof five movements (Whishaw et al., 1992). (1)When a rat encounters food, it first sniffs andpalpates the food with its vibrissae and perio-ral receptors. (2) It then grasps the food in itsmouth and sits back onto its haunches andtransfers the food to its paws. (3) To take thefood from the mouth with the paws, the fore-limbs are positioned so that the elbows arebrought toward the midline of the body andthe palms of the forelimbs are rotated so thatthey face medially. Then, by adduction of theupper arms, both paws are moved medially tograsp the food. (4) As the paws approach thefood, the digits are adjusted so that their aper-ture is appropriate to the size of the piece offood that is being grasped. (5) Food is graspedwith the tips of the digits and is manipulatedwith the tips of the digits for eating. The dig-its may adopt a large number of postures (in-cluding different postures for each paw) de-pending on the size and shape of the fooditem. The five components of spontaneouseating appear similar in different rodent spe-cies and so may comprise a "rodent-commonpattern."

There is a modification to this pattern ofmovements for predation (Ivanko et al., 1996).When catching crickets, rats catch the cricketwith a forepaw. They then sit back and holdthe cricket in the forepaws to prepare it foreating. They manipulate the cricket with thedigits while plucking the wings, limbs, andhead from the cricket, and then they rotatethe ventrum of the cricket toward theirmouth for eating.

SENSORY CONTROL

Rats use olfactory and tactile information tolocate food (Whishaw and Tomie, 1989). Forexample, they sniff food before picking it upwith their mouth, they locate prey items withtheir vibrissae before grasping with a forepaw,and they use tactile information to shape thedigits for grasping (Whishaw et al., 1992).

MOTOR SYSTEMS

When transferring food from the mouth tothe paws, the anticipatory shaping of the dig-its must be commanded by sensory informa-tion from the perioral region (Fig. 15-2). Pri-mates use vision to locate items and to shapethe digits for grasping. It is interesting in thisregard that insectivores have large olfactorybulbs and a small cortex relative to prosimi-ans, which have relatively smaller olfactorybulbs and a larger cortex. Perhaps the trans-fer of sensory control of skilled movementsfrom olfaction and tactile systems to the vi-sual system required a massive rewiring of theforebrain to the visual system in prosimiansand their descendents, with the consequent in-creased growth of the cortex relative to therest of the brain (Whishaw, 2003).

Olfaction is also used by rats to locatefood before reaching for it with a forepaw(Whishaw and Tomie, 1989). If the eyes arepatched, a rat continues to locate food asquickly and as accurately as it did whensighted. If olfaction is eliminated, however, arat acts as if blind, systematically reaching

Figure 15-2. Digits are preshaped to take food from themouse according to food size. Food: top, rice; middle, 500mg food pellet; bottom, laboratory chow. (Based onWhishaw, Dringenberg, and Pellis, 1992.)

Page 182: 0195162854

Chapter 15. Prehension 165

through each set of bars, often beginning atone end of the cage and working toward theother.

The guidance of the limb to food is notunder olfactory control. To reach, a rat mustlift its nose away from the target. Thus, theadvance of the limb must be under centralrather than olfactory control. It is interestingin this respect that humans, although awareof the target for which they are reaching, areunaware of the movements made by theirforearm. Guidance of the limb of rodents andprimates may be different in another way.Many neurons in the motor cortex of primatesrespond in relation to the direction of forearmmovement, whereas other neurons respondto the force and torque of the movement. Ifthe directionally sensitive neurons representvisual control of the limb, it is unlikely thatthey will be present in the rat, which alwaysreaches to the former location of its nose.

REACHING MOVEMENTS

Most animals show asymmetries in the selec-tion of a limb at the individual level, and someshow asymmetries at the population level. Al-though humans display limb dominance, inthat the right limb is favored in about 90% ofthe population, rats display only individualasymmetries (Peterson, 1932; Whishaw, 1992).About 13% of rats are ambidextrous, and theremainder is almost equally left and rightlimbed. Ambidexterity itself may not be somuch an absence of limb preference as a lackof motor skill. Limb preference is displayed assoon as a rat begins to reach and may be anaccident of learning rather than reflecting lat-eralization of central control. There are no sexdifferences.

Skilled reaching movements have the ap-pearance of an action pattern and so are rec-ognizable (Metz and Whishaw, 2000). Eshkol-Wachman movement notation (EWMN) canbe used to describe the relations of body seg-ments (Whishaw and Pellis, 1990), whereas

Laban movement analysis can be used todescribe qualitative aspects of movement(Whishaw et al, 2003). These analyses indicatethat reaching can be subdivided into a num-ber of movements.

The stepping movements made by a ratas it approaches and reaches are central to itsability to advance its limb to the food. As a ratbegins to reach, its reaching forelimb and itscontralateral hindlimb move forward to-gether. The diagonally coupled movementnot only advances the forelimb but also bringsthe contralateral hindlimb beneath the bodyso that the rat can sit back onto its hauncheswith the food that it retrieves.

The movement of the forelimb in reach-ing is composed of 10 movement subcompo-nents (Fig. 15-3).

1. Digits to the midline. Using mainly the upperarm, the reaching limb is lifted from the floorso that the tips of the digits are aligned withthe midline of the body.

2. Digits flexed. As the limb is lifted, the digits areflexed, the paw is supinated, and the wrists arepartially flexed.

3. Elbow in. Using an upper arm movement, theelbow is adducted to the midline while the tipsof the digits retain their alignment with themidline of the body.

4. Advance. The limb is advanced directlythrough the slot toward the food target.

5. Digits extend. During the advance, the digitsextend so that the digit tips are pointing to-ward the target.

6. Arpeggio. When the paw is over the target, thepaw pronates from digit 5 (the outer digit)through to digit 2, and at the same time, thedigits open.

7. Grasp. The digits close and flex over the food,with the paw remaining in place, and the wristis slightly extended to lift the food.

8. Supination I. As the paw is withdrawn, the pawsupinates by almost 90°.

9. Supination II. Once the paw is withdrawn fromthe slot to the mouth, the paw further supinatesby about 45° to place the food in the mouth.

10. Release. The mouth contacts the paw and the pawopens to release the food. In order to evaluate

Page 183: 0195162854

MOTOR SYSTEMS

Figure 15-3. Movement components that comprise a reach. Lifting involves raising the forepaw, supinatingthe paw so that the palm faces inward and the digits are aligned with the midline of the body. Aiming in-volves bringing the elbow to the midline while the digits maintain their midline alignment. Advance of thelimb is produced by a movement at the shoulder and the digits are extended. Pronation is achieved with anadduction of the elbow and rotation around the wrist, and the digits are opened. Grasp supination occurs intwo stages: to withdraw the paw and to present the food to the mouth (Based on Whishaw, 2000.)

reaching performance, movements can bescored as being present/absent or present andimpaired (Whishaw et al., 1993; Whishaw, 2000).

Many aspects of the reaching movementare complex in that they require fixations: hold-ing one portion of the limb in a fixed bodywiseor topographic position while other portions ofthe limb are moved. For example, after the dig-its are brought to the midline of the body, ad-justment in the limb must occur to keep themat this location while the elbow moves to themidline location (movements of the elbowwould otherwise displace the digits). When thehead is raised to allow the limb to advance tothe food, compensatory adjustments must takeplace in the advancing limb so that it continuestoward the target and is not carried away bymovement of the head and trunk. After the foodis retrieved, the paw must maintain a fixed mid-line position so that the mouth can retrieve thefood, otherwise the limb would move away asthe mouth turned toward it. Electromyo-

graphic recording from forelimb muscles showthat a surprising number of muscles are activeduring all phases of reaching (Hyland andReynolds, 1993). It is likely that much of thisconcurrent activity of the musculature is relatedto fixations. After nervous system damage,breakdown in movement is seen in fixations.

ARPEGGIO MOVEMENT

High-speed video recording (60 frames/sec onnormal replay and 120 fields/sec on replay)show that the rat grasps food with an arpeg-gio movement (Whishaw and Gorny, 1994).As the paw is pronated to grasp a food pelletfrom a shelf, the fifth digit is placed onto theshelf first, followed in succession by digits 4,3, and 2 in an arpeggio pattern (Fig. 15-4). Aseach digit is placed onto the shelf, it is opened(spread away) from the preceding digit, andso covers a wide area of the shelf. The paw is

166

Page 184: 0195162854

Chapter 15. Prehension 167

Figure 15-4. Relatively independent use of the digits in grasping a 20 mg food pellet. After the paw is pronatedover the food pellet (A), the arpeggio movement brings digit 4 into contact with the food pellet (B). The foodis then grasped between digit 4 and digit 5, with digit 5 moving medially in a relatively independent move-ment as might a primate thumb. As the paw is supinated, the other digits eventually close. (Based on Whishawand Gorny, 1994.)

then pushed down onto the shelf in a palpat-ing motion. If food is not present, the paw iswithdrawn with closing but not flexure of thedigits and then the reach is repeated. Whenthe food is present, the pattern of grasping de-pends in part on the size of the food pellet.Larger pieces of food are contacted by the padof digit 3 and grasped against the palm by dig-its 3 and 4. Smaller food pellets are contactedby the pad of digit 4 and grasped betweendigits 4 and 5.

There is some degree of independentdigit movement during grasping. When largefood pellets are grasped, flexure of digits 3 and4 appears to precede flexion of the other dig-its, and when smaller food pellets are grasped,flexure of digits 4 and 5 appears to precedeflexure of the other digits. The most strikingindependent digit movement is made by digit5, which turns medially to grasp, thus playinga role like a human thumb. (There are manyother situations in which independent digitmovements might occur, including climbing,handling prey or food, and grasping the furwhen grooming, but there are no studies ofthese movements.)

STRAIN DIFFERENCES

Although there are many strains of laboratoryrats, there are only a couple of studies of straindifferences in skilled movements. Nikkhah etal. (1998) describe strain differences in successin reaching down a staircase for food. Outbredalbino Sprague-Dawley rats were among themost successful of the strains studied. In thesingle pellet-reaching task, impairments inreaching success have been described in al-bino and inbred (brother-sister matings)Fisher 344 rats relative to Long-Evans rats(VandenBerg et al., 2002). Other work showsthat Sprague-Dawley and Long-Evans ratshave equivalent reaching success on the sin-gle pellet-reaching task, but the movementsused by the strains are different (Whishaw etal., 2003). Similarly, we have observed that al-bino Wistar rats display movements that aresimilar to those of Sprague-Dawley rats.These results suggest that there may be neu-ral differences in the motor system of albinorats, perhaps consisting of misrouting offibers, as has been described in the visualsystem.

Page 185: 0195162854

168 MOTOR SYSTEMS

SKILLED MOVEMENT INNEUROINVESTIGATION

AND DISEASE

There are a number of studies relating centralnervous system structure and function toskilled movements. Damage to a surprisinglylarge number of structures impairs skilledreaching and changes the movements used;structures include the motor cortex (Whishawet al., 2000), pyramidal tract (Whishaw et al.,1993), dorsolateral caudate nucleus (Piza andCry, 1990), red nucleus (Whishaw and Corny,1992; Whishaw et al., 1998), and dorsalcolumns (McKenna et al., 1999). For this rea-son, skilled reaching is useful for the study ofmany disease conditions, including stroke(Whishaw, 2000), spinal cord injury (McKennaand Whishaw, 1999; Ballermann et al., 2001)and Parkinson's disease (Metz et al., 2001). Se-lective damage to these structures does not

completely abolish skilled movements, butmovement quality changes. What do thesechanges reveal about the organization of themotor system?

If success measures are used to evaluatenervous system injury, changes in successscores can be obtained, but for many kinds ofinjury, rat, with practice, will return to pre-operative success. An analysis of the move-ments used during reaching shows that suc-cess is not regained by true recovery butthrough the use of compensatory behaviors(Whishaw et al., 1991, 1993, 2000). For exam-ple, the movements of pronation and supina-tion usually produced by intrinsic limbmovement can also be achieved by body ro-tation: pronation with contraversive body ro-tation and supination with ipsiversive bodyrotation (Fig. 15-5). Dorsal and ventral move-ments usually made by the limb can likewisebe assisted or achieved with movements of

Control Motor Cortex

Figure 15-5. Reaching in a controlrat and the use of compensatorymovements in a rat reaching withthe paw contralateral to a motor cor-tex lesion. Note the absence of bodyrotation in the control rat versus ro-tation in the motor cortex rat (ip-siversive on limb advance and con-traversive to release the food to themouth) and the wide base of sup-port. (Based on Whishaw et al.,1991.)

Page 186: 0195162854

Chapter 15. Prehension 169

the trunk. If a rat is unable to achieve the fix-ation of holding the paw to the mouth to re-trieve the food pellet, the unimpaired paw canbe used to hold the impaired paw. Descriptivemovement scoring can provide a good as-sessment of compensatory movements. Manyof the pathways of the motor system arecrossed, and so studies of both paws in skilledreaching can provide insights into the contri-butions of crossed pathways. For example,damage to the pyramidal tract crossed path-way produces no detectable impairment inskilled reaching (Whishaw and Metz, 2002),whereas damage to crossed dopaminergicprojections does produce impairments (Ver-gara et al., 2003).

Because skilled reaching is an acquiredmotor act, the behavior can be used for stud-ies of motor learning. According to the princi-ple of proper mass, the topographic represen-tation of movement in motor cortex shouldcorrespond to movement ability, and studiesdo find that with the acquisition of motor skillthere are changes in representation of the dis-tal relative to the proximal cortical representa-tions of the limb. Similarly, skill acquisition isaccompanied by morphological changes insuch structures as the motor cortex, includingchanges in dendritic arbor, synapse number,and synaptic function (Kleim et al., 2002; Kolbet al., 2003).

ACKNOWLEDGMENTS

This research was supported by The Canadian Stroke Networkof Canada and the Natural Sciences Engineering Council ofCanada.

REFERENCES

Ballermann M, Metz GA, McKenna JE, Klassen F,Whishaw IQ (2001) The pasta matrix reaching task:A simple test for measuring skilled reaching dis-tance, direction, and dexterity in rats. Journal ofNeuroscience Methods 30:39^5.

Eshkol N and Wachmann A (1958) Movement notation.London: Weidenfeld and Nicholson.

Green EC (1963) Anatomy of the rat. New York: HafnerPublishing.

Hyland BJ and Reynolds JN (1993) Pattern of activity inmuscles of shoulder and elbow during forelimbreaching in the rat. Human Movement Science12:51-70.

Ivanco TL, Pellis SM, Whishaw IQ (1996) Skilled fore-limb movements in prey catching and in reachingby rats (Rattus norvegicus) and opossums (Mon-odelphis domestica): Relations to anatomical differ-ences in motor systems. Behavioural Brain Research79:163-181.

Kleim JA, Barbay S, Cooper NR, Hogg TM, Reidel CN,Remple MS, Nudo RJ (2002) Motor learning-dependent synaptogenesis is localized to function-ally reorganized motor cortex. Neurobiology ofLearning and Memory 77:63-77.

McKenna JE and Whishaw IQ (1999) Complete com-pensation in skilled reaching success with associatedimpairments in limb synergies, after dorsal columnlesion in the rat. Journal of Neuroscience 19:1885-1894.

McKenna JE, Prusky GT, Whishaw IQ (2000). Cervicalmotoneuron topography reflects the proximodistalorganization of muscles and movements of the ratforelimb: A retrograde carbocyanine dye analysis.Journal of Comparative Neurology 419:286-296.

Metz GA, Dietz V, Schwab ME, van de Meent H (1998).The effects of unilateral pyramidal tract section onhind limb motor performance in the rat. Behav-ioural Brain Research 96:37—46.

Metz GA and Whishaw IQ (2000) Skilled reaching anaction pattern: Stability in rat (Rattus norvegicus)grasping movements as a function of changing foodpellet size. Behavioural Brain Research 116:111-122.

Miklyaeva El, Woodward NC, Nikiforov EG, TompkinsGJ, Klassen F, loffe ME, Whishaw IQ (1997) Theground reaction forces of postural adjustments dur-ing skilled reaching in unilateral dopamine-depletedhemiparkinson rats. Behavioural Brain Research88:143-152.

Muir GD and Whishaw IQ (1999) Complete locomotorrecovery following corticospinal tract lesions: Mea-surement of ground reaction forces during over-ground locomotion in rats. Behavioural Brain Re-search 103:45-53.

Nikkhah G, Rosenthal C, Hedrich HJ, Samii M (1988)Differences in acquisition and full performance inskilled forelimb use as measured by the 'staircasetest' in five rat strains. Behavioural Brain Research92:85-95.

Peterson GM (1932-1937 Mechanisms of handedness inthe rat. Comparative Psychological Monographs9:21-43.

Pisa M and Cyr J (1990) Regionally selective roles of the

Page 187: 0195162854

170 MOTOR SYSTEMS

rat's striatum in modality-specific discriminationlearning and forelimb reaching. Behavioural BrainResearch 37:281-292.

Remple MS, Bruneau RM, VandenBerg PM, GoertzenC, Kleim JA (2001) Sensitivity of cortical movementrepresentations to motor experience: Evidence thatskill learning but not strength training induces cor-tical reorganization. Behavioural Brain Research123:133-141.

VandenBerg PM, Hogg TM, Kleim JA, Whishaw IQ(2002) Long-Evans rats have a larger cortical topo-graphic representation of movement than Fischer-344 rats: A microstimulation study of motor cortexin naive and skilled reaching-trained rats. Brain Re-search Bulletin 59:197-203.

Whishaw IQ (1992) Lateralization and reaching skill re-lated: Results and implications from a large sampleof Long-Evans rats. Behavioural Brain Research52:45-48.

Whishaw IQ (2000) Loss of the innate cortical engramfor action patterns used in skilled reaching and thedevelopment of behavioural compensation follow-ing motor cortex lesions in the rat. Neuropharma-cology 39:788-805.

Whishaw IQ (2003) Did a change in sensory controlskilled movements stimulate the evolution of theprimate frontal cortex. Behavioural Brain Research146:31-41.

Whishaw IQ and Coles BL (1996) Varieties of paw anddigit movement during spontaneous food handlingin rats: Postures, bimanual coordination, prefer-ences, and the effect of forelimb cortex lesions. Be-havioural Brain Research 77:135-148.

Whishaw IQ, Dringenberg HC, Pellis SM (1992) Spon-taneous forelimb grasping in free feeding by rats:Motor cortex aids limb and digit positioning. Be-havioural Brain Research 48:113-125.

Whishaw IQ and Gorny B (1994) Arpeggio and frac-tionated digit movements used in prehension byrats. Behavioural Brain Research 60:15-24.

Whishaw IQ and Gorny B (1996) Does the red nucleusprovide the tonic support against which fractionatedmovements occur? A study on forepaw movementsused in skilled reaching by the rat. Behavioural BrainResearch 74:79-90.

Whishaw IQ, Gorny B, Sarna J (1998) Paw and limb usein skilled and spontaneous reaching after pyramidal

tract, red nucleus and combined lesions in the rat:Behavioral and anatomical dissociations. Behav-ioural Brain Research 93:167-183.

Whishaw IQ, Gorny B, Foroud A, Jeffrey A. Kleim JA(2003) Long-Evans and Sprague-Dawley rats havesimilar skilled reaching success and topographiclimb representations in motor cortex but use dif-ferent movements as assessed by EWMN andLaban movement analysis. Behavioural BrainResearch 145:221-232.

Whishaw IQ and Metz GA (2002) Absence of impair-ments or recovery mediated by the uncrossed py-ramidal tract in the rat versus enduring deficits pro-duced by the crossed pyramidal tract. BehaviouralBrain Research 134:323-336.

Whishaw IQ and Miklyaeva El (1996) A rat's reach shouldexceed its grasp: Analysis of independent limb anddigit use in the laboratory rat. In: Measuring move-ment and locomotion: From invertebrates to humans(Ossenkopp K-P, Kavaliers M, Sandberg RP, eds.), pp.130-146. New York: RG Landes Co.

Whishaw IQ, O'Connor RB, Dunnett SB (1986). Thecontributions of motor cortex, nigrostriatal do-pamine and caudate-putamen to skilled forelimbuse in the rat. Brain 109:805-843.

Whishaw IQ and Pellis SM (1990) The structure ofskilled forelimb reaching in the rat: Proximallydriven movement with a single distal rotatory com-ponent. Behavioral Brain Research 41:49-59.

Whishaw IQ, Pellis SM, Gorny B, Kolb B, Tetzlaff W(1993) Proximal and distal impairments in rat fore-limb use in reaching follow unilateral pyramidaltract lesions. Behavioural Brain Research 56:59-67.

Whishaw IQ, Pellis SM, Gorny BP, Pellis VC (1991) Theimpairments in reaching and the movements ofcompensation in rats with motor cortex lesions: Anendpoint, videorecording, and movement notationanalysis. Behavioural Brain Research 42:77-91.

Whishaw IQ, Pellis SM, Pellis VC (1992) A behavioralstudy of the contributions of cells and fibers of pas-sage in the red nucleus of the rat to postural right-ing, skilled movements, and learning. BehaviouralBrain Research 52:29^t4.

Wise SP and Donoghue JP (1986) Motor cortex of ro-dents. In: Jones EJ, Peters A, eds. Sensory-motor ar-eas and aspects of cortical connectivity: CerebralCortex, Vol 5. New York: Plenum, pp 243-265.

Page 188: 0195162854

Locomotor andExploratory Behavior 16ILAN GOLANI, YOAV BENJAMINI,ANNA DVORKIN, DINA LIPKIND,AND NERI KAFKAFI

Rat exploratory behavior includes motor, lo-comotor, motivational, and cognitive aspects;it consists of a stimulating combination of sto-chastic and lawful elements. As technologyimproves, it becomes increasingly more ac-cessible for data acquisition and analysis. Withthe advancement of statistical and computa-tional data analyses, ethological knowledge,previously the exclusive property of experi-enced observers, becomes widely accessibleby being captured by formal algorithms ap-propriate for automated analysis.

The study of rat exploratory behavior cancommence at two levels of resolution: (1) thatof the trajectory traced by the animal's wholebody and (2) that in which the animal is con-ceived as a linkage of body parts moving inrelation to each other.

Studies relating to the animal's trajectoryin the environment and relating to interlimbcoordination are reviewed. In each section,we start from the stage of automated data ac-quisition and proceed through the isolation ofpatterns of movement to global regularities.

TRAJECTORY ANALYSIS

SETUP

To highlight intrinsic aspects of behavior, weuse a large circular empty arena. To obtain a

slow, gradual buildup of behavior, however,one might introduce a shelter into the arena,so as to induce polarity between it and thelarge open area. Repeated exposures, each ofa long duration, further extend the buildupprocess.

DATA ACQUISITION

To allow the computation of velocity and ac-celeration and to ensure the capture of short-duration stops (which last as little as 0.2 sec-ond), rats have to be tracked at a rate of atleast 25 to 30 frames/sec.

DATA PREPARATION AND ANALYSIS

SmoothingTo obtain a smooth path that would allowa meaningful derivation of moment-to-moment velocities without eliminating short,but behavioraUy meaningful, arrests, we usetwo different statistical smoothing tools: one forthe arrests and one for the progression segments(Hen et al., 2004). First, we capture the arrestsby using a running medians robust smoothingmethod (Tukey, 1977). Only after smoothingthe arrests to zero velocity, and indexing them,we use another method of Robust Lowess(Cleveland, 1977) to smooth the remainder ofthe location time series. In this way we end upwith a record that has both the richness of the

171

Page 189: 0195162854

172 MOTOR SYSTEMS

original time series and the smoothness neces-sary for the computation of velocities (whichare computed and stored for each data point).To visualize the smoothing process see http://www.tau.ac.il/ilan99/see/help.

Segmentation into Lingering Episodesand Progression SegmentsThe velocity and acceleration of the animal arethe outcome of all of the concurrent "forces"that act on it. Conversely, the attraction or re-pulsion exerted by a wall, a cliff, an edge, or anattractive place is revealed by the momentaryvalues of these parameters. The momentary ve-locity of the rat can tell us whether it "thinks"it is running away or toward a familiar place.The locations and the kinematics of stops (closeto zero velocity episodes), during which scan-ning occurs, may betray perceptual and evencognitive aspects of the behavior.

The rat's path (Fig. 16-1 A) is punctuatedby arrests. We classify the interarrest intervalsby the maximal velocity attained in them andobtain a density plot of these velocities. Thisplot is typically composed of one Gaussian oflow maximal speed segments (Fig. 16-1B left)and one Gaussian consisting of high maximalspeed segments (Fig. 16-1B right). With theaid of a Gaussian mixture model and an ap-propriate algorithm (the expectation-maxi-mization (EM) algorithm) (Everitt, 1981), weestablish a cutoff value at the deep betweenthe two Gaussians. The segments to the leftof the cutoff value are defined as lingeringepisodes (episodes of stops or staying in placebehavior), and the segments to the right of itare defined as progression segments. Treatingthe path as a string of discrete building blocksrather than a continuous series of coordinates(Fig. 16-lC) allows a more straightforward

40000

Figure 16-1. Principles of SEEanalysis: The time series of the rat'slocation in the arena is automaticallytracked in a rate of 25 to 30 Hz andsmoothed. (A) The path in a three-dimensional representation of X, Y,and TIME. (B) The distribution ofspeed peaks (thin line) is used toparse the data into segments of twotypes: slow local movements ("lin-gering" [L], in black) and progression(P, in gray). (C) The Data can betreated as a string of these discretebehavioral units. (D,E) The path plotand speed profile of two progressionsegments (Pi and P2) separated byone lingering episode (Ll) is demon-strated. The typical properties ofthese units are used to quantify thebehavior. For example, the "seg-ment acceleration" (Fig. 16-2, bot-tom) is estimated by dividing the seg-ment's speed peak by its duration.(Adapted from Benjamini and Kafkafiet al., submitted)

Page 190: 0195162854

Chapter 16. Locomotor and Exploratory Behavior 173

analysis of each of the pattern types (Drai andGolani, 2001; Kafkafi et al., 2001, 2003a,2003b). Once the relatively smooth progres-sion segments are separated from the jitterylingering episodes, the road is open for ameaningful computation of velocity, acceler-ation, heading, curvature, and other measuresof progression segments (Fig. 16-1D,E).

The behavior of an animal, a strain, or aspecies may be characterized by the cutoff valuethat distinguishes between lingering episodesand progression segments. After segmentation(Drai et al., 2000), both segment types can becharacterized by simple quantitative measures(termed endpoints in behavior genetics) such astheir length, duration, maximal speed, acceler-ation, and other measures derived from these(an example is shown in Fig. 16-2).

Separation between Wall (or Edge)and Center BehaviorHaving at hand a sequence of progression andlingering segments defined only in terms oftheir kinematics, we can now examine the re-lationship of these patterns to the environ-ment. The first obvious reference for locationin the environment is the arena's wall or edge.Instead of using the arbitrary criterion of 15to 25 cm from wall to distinguish wall fromcenter behavior, we customize the cutoffpoint to the particular species, strain, and evenindividual animal, by using radial speed anddistance from wall as classifying criteria. Since

lingering progression

Figure 16-2. Boxplot graphs of medians of maximal speedin lingering and progression segments in a 2.5 m diametercircular arena.

Figure 16-3. Results of the wall/center separation proce-dure in three 1/z h sessions of three male Long-Evanshooded rats. Plots of all data points for progression seg-ments in each of the animals. Black: Movement along thewall; Gray, movement in the center.

the radial speeds of an animal running alongthe wall are distributed around and close tozero cm/sec, it is possible to separate move-ment along the wall from movement in thecenter by applying the segmentation proce-dures described above to radial speed and dis-tance from wall (Lipkind et al., in press).

The end product of the separationprocess for a 30 minute session of 3 Long-Evans hooded rats is presented in Figure16-3. The procedure supplies a number ofnew behavior patterns—segments of progres-sion along the wall, segments of progressionin the center, lingering episodes along thewall, lingering episodes in the center and in-cursions (forays into the center). These pat-terns can be characterized by length, duration,maximal speed, acceleration, and other prop-erties (see endpoints 22 to 27 in Table 16-1).

Clustering of Lingering Episodesinto Operational PlacesThe coordinates acquired by the tracking sys-tem specify a topographical, but not neces-sarily an ethologically significant, location forthe animal. We reserve the term place for aneighborhood of x,y locations with a unifiedoperational significance for the animal. Clearly,although the animal is always located in aspecified x,y location, it is not necessarily vis-iting a place that is meaningful to it.

Because lingering episodes are definedkinematically, we can examine the locationsin which they occur. In rats, even in a barearena devoid of objects, the x,y locations oflingering episodes often tend to be clustered

Page 191: 0195162854

174 MOTOR SYSTEMS

Table 16-1. Locomotor and Exploratory Behavior

General Endpoints

12

34567891011

12131415161718192021

Distance traveled (cm)Lingering mean speed (cm /sec)Proportion of time spent more than 15 cm away from wallProportion of lingering time spent away from wallNo. of progression segments (segments)Median spatial spread of lingering episodes (cm)Median length of progression segments (cm)Median duration of lingering episodes (sec)Quantile 95 of duration of progression segments (sec)Quantile 95 of progression segment maximum speed (cm /sec)Median segment acceleration to maximum speed (cm /sec2)Dart (see Kafkafi et al.', 2003b)Latency to maximum half speed (sec)Center activity proportionCenter rest proportionTime proportion of lingering episodesActivity proportion of lingering episodesMaximum spatial spread of progression segments (cm)No. of stops per distance (segments /cm)Lingering progression threshold speed (cm /sec)No. of stops per excursion quantile 90 (stops /excursion)

Mean ± SE

8312.28 ± 967.8321.31668 ± 0.169496

0.0189796 ± 0.005861720.0084797 ± 0.0024573

93.9167 ± 11.90656.22084 ± 0.58027445.892 ± 4.78906

3.08667 ± 0.2937265.41 ± 0.466115

68.3786 ± 3.6806224.6988 ± 1.332661.30355 ± 0.066485433.2467 ± 10.2184

0.0872066 ± 0.02804490.0263737 ± 0.006465640.884496 ± 0.01536240.269275 ± 0.028709230.391 ± 5.36781

0.0115714 ± 0.00060224224.3093 ± 1.4072112.1667 ± 1.42931

Wall-Center Endpoints

2223242526

Incursion maximum distance from wall (cm)Median incursion length (cm)Ratio of mean speed to and from center per incursionRatio of center to wall mean speedWall ring thickness (cm)

13.4907 ± 1.1498831.5759 ± 4.86886

0.765158 ± 0.04751051.11172 ± 0.06521218.27086 ± 0.366949

Twenty-six selected behavioral measures (endpoints) characterize male Long-Evans hooded rat locomotor and exploratory behavior in a2.5 m diameter walled circular and empty arena. The endpoints were computed with the SEE software (see for example, Kafkafi et al., 2003b; Lip-kind et al., in press; Benjamini and Kafkafi et al., submitted; http://www.tau.ac.il/ilan99/see/help).

in relatively circumscribed neighborhoods.Such clusters of lingering episodes have beentermed by Tchernichovski et al. (1996) as prin-cipal, or preferred, places.

The Home Base. There are one or two neigh-borhoods of locations that stand out from allthe other neighborhoods in the arena in termsof the cumulative time spent in them and num-ber of visits (stops, lingering episodes) paid tothem. These neighborhoods, which were for-merly demonstrated by manual scoring of thevideo (Eilam and Golani, 1989), are now pickedup by using an algorithm that calculates the cu-mulative amount of time spent in different sec-tions of the arena and the number of lingeringepisodes performed per section (see Fig. 16-4).

In an arena devoid of any objects or places thatsuggest shelter, each rat establishes its homebase in a different place, early in the session(Eilam and Golani, 1989). This suggests thathome base location involves the use of spatialmemory. If, however, there is an object pres-ent that is clearly more conspicuous, most ratsestablish their home base near it. This propertyhas been used to standardize home base loca-tion (Tchernichovski and Golani, 1995; Tcher-nichovski et al., 1998).

Because of its relative stability, the homebase is a natural candidate for an origin of axesof a frame of reference for the examination ofthe rat's path. Visits to the home base are thusused to partition the path into excursions—sequences of stop-and-go behavior that start

Page 192: 0195162854

Chapter 16. Locomotor and Exploratory Behavior 175

Figure 16-4. (A) The height of the graph at a particular location in the circular 2.5 m diameter arena repre-sents: cumulative dwell time of a Lewis rat during lingering episodes at that location in the course of a 30minute session (left) and number of lingering episodes (stops) in that location (right). As shown, one neigh-borhood in the empty arena (at 11 o'clock) stands out in the magnitude of these two measures. This neigh-borhood is the rat's home base. (B) Time series of the angular component of polar coordinates of a Long-Evans hooded rat's 30 minute session in same arena, highlighting home base location (gray band), lingeringepisodes (black dots near-wall, gray dots away from wall), and excursions performed from home base (line).Thick line depicts near-wall behavior, and thin line, away-from-wall behavior. Note the gradual buildup inexcursion length. (Q (Left) Histogram and estimated density function of the pulled data (n = 15), of wild rats,1/2 h sessions, on number of stops per round trip in rats that established a single home base, divided by therat session's maximum. (Right) Quantile plot of the (same) pooled data on number of stops per round-trip inrats having a single home base. Straight line implies a uniform distribution. (From Golani et al., 1993.)

and end at the home base. Rats establishingmore than one base have both roundtrips tothe same home base and excursions that startat one home base and end at another. In sin-gle home base rat sessions, excursions havetwo basic features: (1) their outbound portionis slow and intermittent, whereas their in-bound portion is fast (Eilam and Golani, 1989;Tchernichovski and Golani, 1995), and (2) thenumber of stops per excursion is constrained(Golani et al., 1993).

Number of Stops per ExcursionWe have noted that after a rat leaves the homebase, it never performs more than about 12stops before returning home. This apparentupper bound could, however, be a result ofvery different modes of behavior: excursionscould contain a typical number of, say, 8 stops.In such case, each excursion would containabout 8 stops, the frequency distribution of thenumber of stops per excursion would havea bell-shaped form with a peak frequency at

Page 193: 0195162854

176 MOTOR SYSTEMS

8 stops, and the number of stops would rarelyexceed 12 stops. A very large sample of ex-cursions would nevertheless yield from timeto time a higher number, revealing that theupper bound is not real.

Another option is that after each stop, therat decides whether to go back home or toproceed to the next stop. With a probabilityof returning home of, say, 1:2, 50% of the ex-cursions would include one stop, 25% (half ofthe remaining half) would include two stops,12.5% would include three stops, and so on;the frequency distribution would steeply trailoff, again creating the false impression of anupper bound.

Still another option is that the probabil-ity of returning home would increase aftereach stop. Only in this mode would the num-ber of excursions containing 1, 2, 3 ... n stopsbe similar, thus yielding a uniform distribu-tion. We have found that rats use the third op-tion. The maximal number of stops per ex-cursion does not further increase, regardlessof how much we increase the number of ex-amined rat sessions, even when the arena wasincreased from 4 m2 to 64 m2 (Golani et al.,1993). Scaling interstop distances so as to fitinto increasingly larger arenas has been repli-cated in voles (Eilam, 2003; Eilam et al., 2003).

Other Stop-Related Findings. The "hyperactiv-ity" attributed to rats with Fimbria-Fornix le-sions is due to the abundance of stops of shorterduration (and thus more movement segments)(Whishaw et al., 1994). The distribution ofstops and the number of home bases can bemodified with D-amphetamine (Eilam andGolani, 1994; Cools et al., 1997, Gingras andCools, 1997) and with quinpirole (Eilam andSzechtman, 1997). Both excursion length andnumber of stops per excursion are shortenedand consolidated into topographically stereo-typed chunks that are performed en bloc un-der D-amphetamine (Eilam and Golani, 1990).

Buildup of Arena OccupancyWhen rats were exposed daily for a 1 hour ses-sion to a large (6.5 m diameter) empty out-

door arena, they manifested a gradual buildupof arena occupancy. Excursion length is de-fined in Figure 16-5A as the maximal angulardistance along the wall reached by the ratwithin an excursion. The figure presents asummary of excursion length values for allrats as a function of their temporal order ofperformance, for each session separately, ses-sion by session. As shown, excursion lengthincreases both within and across sessions. Themost prominent increase occurs during the

Figure 16-5. (A) The average values of the excursion lengthswithin the first eight sessions. Data are smoothed and arepresented together with SD bars, computed at 10 parts ofeach session. The x-axis represents the temporal order of ex-cursions, and the y-axis represents angular distance from thehome base. The numbers below the x-axis represent the tem-poral sequence of sessions. (B) Smoothed average attractiondistance (AD, represented by higher series of curves) and av-erage repulsion distance (RD, represented by lower series ofcurves) within the first eight sessions. Data are computed at10 parts of each session. The x-axis is as in A, and the y-axisrepresents angular distance from home base.

Page 194: 0195162854

Chapter 16. Locomotor and Exploratory Behavior 177

first half of the session, whereas later thelength stabilizes or even decreases.

Examination of the high variability (re-flected in the high standard deviation values inFigure 5A) reveals that during the first stage ofthe first session each rat has its own character-istic length of excursions, but all rats share a sim-ilar rate of excursion growth both within andacross sessions. The exploratory process is ad-ditive from one session to the next, and the rateof increase is constant and independent of theprevious experience of the rat. An additional ef-fect is that of progression within a session. Thedynamics of excursion length might reflect in-trinsic constraints on the amount of novelty arat can handle per excursion. It may also be anexpression of the buildup observed in manyspecies-specific behaviors (Lorenz, 1937).

Velocity across the Buildup Process. In thelarge outdoor arena the rats mostly pro-gressed back and forth along the wall. This al-lowed us to compare, for each excursion ateach location, the outbound velocity with theinbound velocity at that location. A negativevalue of the algebraic sum indicated that atthat location the inbound speed was higherthan the outbound speed, whereas a positivevalue indicated that at that location the out-bound speed was higher. In this way, the ex-cursion was divided into portions with nega-tive values and portions with positive values.When outbound speed is slow and inboundspeed is high (negative sum), it is as thoughthe rat moved upstream on the way out,against the attraction exerted on it by thehome base, whereas on the way in, it moveddownstream, in cooperation with that attrac-tion. Conversely, when outbound speed washigh and inbound speed low (positive sum), itwas as though the rat moved on the way outdownstream, in cooperation with a repulsionexerted on it by the home base, whereas onthe way in, it moved against that repulsion.

By computing dwell time at each location,we found that high home base attraction char-acterized low exposure (i.e., unfamiliar) por-tions of the arena circumference, whereas low

attraction or even repulsion characterized highexposure (familiar) portions. By computing theportions of the arena circumference fromwhich the home base became attractive and thecorresponding portions from which it becamerepulsive, we found that the velocity pattern ofthe rat changed concurrently with the increasein excursion length and in correlation with thefamiliarity of places. The primitive velocity pat-tern consisted of a slow outbound and a fast in-bound pattern. During exposure, the asymme-try in velocity was inverted. The inversionspread across successive excursions from thehome base out. Both excursion length and theattraction-repulsion dynamics might reflectthe same intrinsic constraints on the amount ofnovelty a rat could handle per and across ex-cursions (see Fig. 16-5B). An analytic model de-veloped by Tchernichovski and Benjamini(1998) explains both the progressive increase inexcursion length and the dynamics of velocity.

The progressive nonmonotonic expansionof rat movement from excursion to excursionis a general characteristic of species-specific be-havior patterns (Lorenz, 1937). This growth pat-tern shows several similarities to the searchingbehavior of the desert ant (Cataglyphis spp.)(Wehner, 2003) and other invertebrates (Hoff-mann, 1978,1983). Although Tchernichovski etal. (1998) and Tchernichovski and Benjamini(1998) attribute the expansion of excursionlength and the dynamics of velocity to a famil-iarization process, they also point out that itcould well develop on an idiothetic basis.

An idiothetic basis during early phases ofexploration has been suggested in severalstudies (e.g., McNaughton et al., 1996). It hasalso been indicated that increased velocity inthe inbound portion of excursions has an id-iothetic basis (Whishaw et al., 2001). Controland fimbria-fornix rats had similar outwardvelocities in segments starting from homebase, but the return paths of the fimbria-fornixrats were significantly slower, more cir-cuitous, and more variable. This was inde-pendent of light or dark conditions. The lackof dependence on allothetic cues thus sug-gested that rats used dead reckoning naviga-

Page 195: 0195162854

178

tional strategies to initiate the homeward por-tion of excursions (see Chapter 38).

A Selection of AutomaticallyMeasured Parameters (Endpoints)Having cleaved the behavior into the ethologi-cally distinct patterns described in this reviewand having defined these patterns algorithmi-cally, one can readily compute the durations,spatial properties, velocities, frequencies, etc.,characterizing these patterns (a selection ofendpoints is given in Table 16-1; download SEEat http: //www.tau.ac.il/ ~ilan99 / see /help).

MULTI-SEGMENTALMOVEMENT ANALYSIS

Capturing the relations and changes in rela-tion between the parts of the body in freelymoving rats is followed by the stages of acloser examination, analysis, and integration

MOTOR SYSTEMS

of the multiple measurements into few keykinematic parameters. The values taken bythese parameters can subsequently be charac-terized in various situations and preparationsrelated to locomotor and exploratory behav-ior. High-resolution video-based tracking ofmultiple points on the animal's body, algo-rithmic analysis combined with EW move-ment notation analysis (Eshkol and Wach-mann, 1958; Eshkol, 1990) and a dynamicalsystems approach (Kafkafi et al., 1996; Kafkafiand Golani, 1998) are combined to yield an ar-ticulated description of the organization of ratlocomotor and exploratory behavior.

DATA ACQUISITION

A bottom view of a freely moving rat walk-

ing on a glass platform provides access to therat's trunk and all legs. A sampling rate of50-60 Hertz provides an appropriate tempo-ral resolution (Fig. 16-6A).

Figure 16-6. (A) A bottom view of an automatically tracked rat with markers attached to five points on its trunkand six points on its feet. Hindpaw and forepaw positions are respectively measured throughout the step cyclein reference to a frame whose origin of axes is schematically placed at the point lying between the hip (for hind-paws) and the shoulder joints (for forepaws). The frame's vertical axis coincides with the direction of progressionof the midsaggital axis of the respective trunk part. (B) A snapshot taken from an animated reconstruction of therelations and changes in relation between the parts of the trunk, the feet, and their respective footprints duringforward progression. Gray disks represent tracked markers; fall discs on feet stand for feet in support phase, andopen circles stand for feet in swing phase. The footprint trail of each foot is shown in shades of gray, with'lightgray representing establishment of contact, and dark gray release of foot contact with substrate. Arrows indicateheel-to-digits direction of hindfeet (see http://www.tau.ac.il/~ilan99/see/multilimb).

Page 196: 0195162854

Chapter 16. Locomotor and Exploratory Behavior

DATA ANALYSIS

The process of compression of the multiplemeasurements into fewer key parameters ofwhole-body unrestrained locomotor and ex-ploratory behavior is described in several stud-ies (Kafkafi et al., 1996; Kafkafi and Golani,1998; Golani et al., 1999). Here we highlightthree key parameters whose interaction gen-erates a behavioral growth pattern, reminis-cent of that described for excursions from thehome base, which involves a progressive in-crease in the amplitude of movements, and aprogressive increase in the animal's behav-ioral repertoire.

Warm-upInfant rats respond to a novel environment bybecoming immobile and then showing a

179

process of motorial expansion called warm-up.Starting from immobility, horizontal move-ment, forward movement, and finally verticalmovement are successively incorporated intothe behavior. Within each of these movementparameters separately, the parts of the trunkand the legs are recruited into the movementin a cephalocaudal order, head first andhindlegs last. The head and neck never moveforward unless they already moved horizon-tally (laterally), and they never move up un-less they already moved forward. The samerule applies to the chest and to the pelvis (Figs.16-7 and 8). Concurrent with the increase inthe animal's repertoire (and hence in theunpredictability of behavior) and concurrentwith a repetition of movements that were per-formed earlier, there is an increase in the am-plitude of movements. This results in a grad-

Figure 16-7. The infant rat's trunkas a linkage of articulated axes.Columns represent the three mobil-ity gradient key parameters, and hor-izontal rows represent the most cau-dal part of the trunk that moved.The axis of the most caudal part thatmoved is represented by a thick bar.

Page 197: 0195162854

180 MOTOR SYSTEMS

Figure 16-8. Timing of first appear-ance of movement of the parts of thetrunk along the three spatial parame-ters. One example was selected at ran-dom for each developmental day.Within each moment-to-momentwarm-up sequence, the parts of thebody are recruited along the three spa-tial parameters (lateral, forward, andvertical) in a cephalocaudal order.(From Eilam and Golani, 1988.)

ual expansion of the explored portion of theenvironment (Golani et al., 1981; Eilam andGolani, 1988).

The warm-up process is a particular man-ifestation of a more general "mobility gradi-ent" that characterizes the transition fromimmobility to increasing complexity andunpredictability of behavior (Golani, 1992). Inparticular, a progression in the opposite di-rection, with decreasing spatial complexityand increased stereotypy, that is termed "shut-down" occurs under the influence of the non-selective dopaminergic drugs apomorphineand amphetamine and in part the selectivedopaminergic agonist quinpirole (Szechtman

et al., 1985; Eilam et al., 1989; Adani et al., 1991;Golani, 1992). The mobility gradient behaviorappears to be mediated by a family of basalganglia-thalamocortical circuits and their de-scending output stations (Golani, 1992).

The observation that, after pronouncedimmobility, the first forward movement of apart of the trunk must be preceded by a lat-eral movement of that part, suggested that theperformance of one type of movement en-ables (potentiates) the performance of thenext type, which enables the next, and so on,without necessarily eliciting that next type(Golani, 1992). Chevalier and Deniau (1990)attributed an enabling function to the stria-

Page 198: 0195162854

Chapter 16. Locomotor and Exploratory Behavior 181

turn. A reciprocal relationship between hori-zontal scanning and forward progression hasbeen observed in rats: as soon as forward pro-gression stops, lateral head scans appear (dur-ing lingering episodes) (Drai et al., 2000).More recently, it has been shown that thereis an inverse relationship between steppingand scanning at the level of the central nerv-ous system. This relationship is mediated bythe hypothalamus (Sinnamon et al., 1999).

CONCLUSION

A "mobility gradient" involving increasingcomplexity and unpredictability unfolds in ratlocomotor and exploratory behavior at thelevels of the path (location), trajectory (ve-locity), and interlimb coordination. Comput-erized data acquisition, appropriate prepara-tion of these data for analysis, and cleavageof the stream of behavior into intrinsicallydefined parameters and patterns make thisglobal growth pattern more accessible forstudy.

ACKNOWLEDGMENTS

This study was supported by a grant from the Israel Academyof Sciences, Israel Science Foundation, and by National Insti-tutes of Health grant 5-R01-NS040234-03.

REFERENCES

Adani N, Kirtati N, Golani I (1991) The description ofrat drug-induced behavior—kinematics versus re-sponse categories. Neuroscience and BiobehavioralReviews 15:455-460.

Benjamini Y, Kafkafi N, Sakov A, Elmer G, Golani I(submitted) Genotype-environment interactions inmouse behavior: A way out of the problem.

Chevalier G and Deniau JM (1990) Disinhibition as a ba-sic process in the expression of striatal functions.Trends in Neuroscience 13:277-280.

Cleveland WS (1977) Robust locally weighted regressionand smoothing scatterplots. Journal of AmericanStatistical Association 74:829-836.

Cools AR, Ellenbroek BA, Gingras MA, Engbersen A,

Heeren D (1997) Differences in vulnerability andsusceptibility to dexamphetamine in Nijmegen highand low responders to novelty: A dose-effect analy-sis of spatio-temporal programming of behavior.Psychopharmacology 132:181-187.

Drai D, Benjamini Y, Golani I (2000) Statistical discrim-ination of natural modes of motion in rat ex-ploratory behavior. Journal of Neuroscience Meth-ods 96:119-131.

Drai D and Golani I (2001) SEE: A tool for the visuali-zation and analysis of rodent exploratory behavior.Neuroscience and Biobehavioral Reviews 25:409-426.

Eilam D and Golani I (1989) Home base behavior of rats(Rattus norvegicus) exploring a novel environment.Behavioural Brain Research 34:199-211.

Eilam D and Golani I (1988) The ontogeny of ex-ploratory-behavior in the house rat (Rattus rattus)the mobility gradient. Developmental Psychobiol-ogy 21:679-710.

Eilam D and Golani I (1990) Home base behavior in am-phetamine-treated tame wild rats (Rattus norvegi-cus). Behavioural Brain Research 36:161-170.

Eilam D and Golani I (1994) Amphetamine-inducedstereotypy in rats: its morphogenesis in locale spacefrom normal exploration. In: Ethology and psy-chopharmacology. (Cooper SJ and Hendrie CA,eds.). New York: John Wiley & Sons.

Eilam D, Golani I, Szechtman H (1989) D-2-agonistquinpirole induces perseveration of routes and hy-peractivity but no perseveration of movements.Brain Research 490:255-267.

Eilam D (2003) Open-field behavior withstands changesin arena size. Behavioural Brain Research 142:53-62.

Eilam D, Dank M, Maurer R (2003) Voles scale loco-motion to the size of the open-field by adjusting thedistance between stops: A possible link to path in-tegration. Behavioural Brain Research 141:73-81.

Eilam D and Szechtman H (1997) A plausible rat modelof obsessive-compulsive disorder: Compulsivechecking behavior is induced in rats chronicallyinjected with quinpirole. Neuroscience letters48:S16-S16.

Eshkol N and Wachmann A (1958) Movement Notation.London: Weidenfield & Nicholson.

Eshkol N (1990) Angles and Angels. Tel-Aviv: TheMovement Notation Society.

Everitt BS (1981) Finite mixture distributions. London:Chapman and Hall.

Gingras MA and Cools AR (1997) Different behavioraleffects of daily or intermittent dexamphetamine ad-ministration in Nijmegen high and low responders.Psychopharmacology 132:188-194.

Golani I (1992) A mobility gradient in the organizationof vertebrate movement: The perception of move-

Page 199: 0195162854

182 MOTOR SYSTEMS

ment through symbolic language. Behavioral andBrain Sciences 15:249-308.

Golani I, Benjamini Y, Eilam D (1993) Stopping be-havior: Constraints on exploration in rats (Rattusnorvegicus). Behavioural Brain Research 53:21-33.

Golani I, Bronchti G, Moualem D, Teitelbaum P (1981)"Warm-up" along dimensions of movement in theontogeny of exploration in rats and other infantmammals. Proceedings of the National Academy ofSciences USA 78:7226-7229.

Golani I, Kafkafi N, Drai D (1999) Phenotyping stereo-typic behaviour: Collective variables, range of vari-ation and predictability. Applied Animal BehaviorScience 65:191-220.

Hen I, Sakov A, Kafkafi N, Golani I, Benjamini Y (2004)The dynamics of spatial behavior: How can robustsmoothing techniques help? Journal of Neuro-science Methods 133:161-172.

Hoffrnann G (1978) Experimentelle und theoretischeanalyse eines adaptiven Orientierungsverhaltens: die'optimale' Suche der Wustenassel Hemilepistus reau-muri, Audouin und Savigny (Crustacea, Isopoda, Onis-coidea) nach ihrer Hohle. PhD thesis, Regensburg

Hoffinann G (1983) The random elements in the system-atic search behavior of the desert isopod Hemilepis-tus reaumuri. Behavioral Ecology and Sociobiology13:81-92.

Kafkafi N and Golani I (1998) A traveling wave of lat-eral movement coordinates both turning and for-ward walking in the ferret. Biological Cybernetics78:441-453.

Kafkafi N, Levi-Havusha S, Golani I, Benjamini Y (1996) Astereotyped motor pattern as a stable equilibrium in adynamical system. Biological Cybernetics 74:487-495.

Kafkafi N, Lipkind D, Benjamini Y, Mayo CL, Elmer GI,Golani I (2003a) SEE locomotor behavior test dis-criminates C57BL/6J and DBA/2J mouse inbredstrains across laboratories and protocol conditions.Behavioral Neuroscience 117:464-477.

Kafkafi N, Mayo C, Drai D, Golani I, Elmer G (2001)Natural segmentation of the locomotor behavior ofdrug-induced rats in a photobeam cage. Journal ofNeuroscience Methods 109:111-121.

Kafkafi N, Pagis M, Lipkind D, Mayo CL, Bemjamini Y,Golani I, Elmer GI (2003b) Darting behavior: Aquantitative movement pattern designed for dis-crimination and replicability in mouse locomotorbehavior. Behavioural Brain Research 142:193-205.

Lipkind D, Sakov A, Kafkafi N, Elmer GI, Benjamini Y,Golani I (in press) New replicable anxiety-related

measures of wall versus center behavior of mice inthe open field. Journal of Applied Physiology.

Lorenz KZ (1937) Uber die Bildung des Instinktbegriffes.Naturwissenschaften 25:289-331.

McNaughton BL, Barnes CA, Gerrard JL, Gothard K,Jung MW, Knierim JJ, Kudrimoti H, Qin Y, SkaggsWE, Suster M, Weaver KL (1996) Deciphering thehippocampal polyglot: The hippocampus as a pathintegration system. Journal of Experimental Biology199:173-185.

Sinnamon HM, Karvosky ME, Ilch CP (1999) Locomo-tion and head scanning initiated by hypothalamicstimulation are inversely related. Behavioural BrainResearch 99:219-229.

Szechtiman H, Ornstein K, Teitelbaum P, Golani I(1985) The morphogenesis of stereotyped behaviorinduced by the dopamine receptor agonist apo-morphine in the laboratory rat. Neuroscience 14:783-798.

Tchernichovski O, Benjamini Y, Golani I (1996) Con-straints and the emergence of free exploratory be-havior in rat ontogeny. Behaviour 133:519-539.

Tchernichovski O, Benjamini Y, Golani I (1998) The dy-namics of long-term exploration in the rat—Part I.A phase-plane analysis of the relationship betweenlocation and velocity. Biological Cybernetics 78:423^32.

Tchernichovski O and Benjamini Y (1998) The dynam-ics of long-term exploration in the rat—Part II. Ananalytical model of the kinematic structure ofrat exploratory behavior. Biological Cybernetics78:433-440.

Tchernichovski O and Golani I (1995) A phase plane rep-resentation of rat exploratory behavior. Journal ofNeuroscience Methods 62:21-27.

Tukey JW (1977) Exploratory data analysis. Boston:Addison-Wesley.

Wehner R (2003) Desert ant navigation: How miniaturebrains solve complex tasks. Journal of ComparativePhysiology A 189:579-588.

Whishaw IQ, CasselJC, MajchrzakM, et al. (1994) Short-stops in rats with fimbria fornix lesions—Evidencefor change in the mobility gradient. Hippocampus4:577-582.

Whishaw IQ, Hines DJ, Wallace DG (2001) Dead reck-oning (path integration) requires the hippocampalformation: Evidence from spontaneous explorationand spatial learning tasks in light (allothetic) anddark (idiothetic) tests. Behavioural Brain Research127:49-69.

Page 200: 0195162854

Orcadian Rhythms

MICHAEL C. ANTLE AND RALPH E. MISTLBERGER 17

The day-night cycle caused by the rotation ofthe Earth about its axis poses many survivalchallenges. Not surprisingly, most organismshave evolved specializations to exploit tem-poral niches in their environment. For exam-ple, diurnal (day-active) animals tend to havevisual systems adapted for acuity at the ex-pense of sensitivity, whereas the visual sys-tems of nocturnal (night-active) animals, suchas the rat, typically sacrifice acuity in the in-terests of sensitivity. For these and other spe-cializations to be adaptive, animals must havea mechanism to appropriately coordinatetheir behavior and physiology with theday-night cycle. The mechanism that hasevolved is a system of circadian (from theLatin circa, which means "about," and dies,which means "day") oscillators, located in thebrain and in other organs, that generates dailyrhythms and that actively synchronizes theserhythms with environmental cycles. Circa-dian rhythms are ubiquitous in rat behaviorand physiology and, since the pioneeringwork of Richter (1922), the rat has been an im-portant animal model for elucidating the func-tional properties and neural mechanisms ofcircadian regulation.

MEASUREMENT AND ANALYSIS

The tool of choice for measuring circadian be-havioral rhythmicity in rats and other rodentshas been the running wheel. Rats are avid run-ners, and wheels are simple to install and rel-atively cheap. Traditionally, running wheels

were connected to Esterline Angus pen andpaper recorders (Slonaker, 1908; Richter,1922). Today, computers and specialized sen-sors are available by which to measure circa-dian rhythms in activity and other functions.The most commonly used devices are me-chanical rnicroswitchs (for running wheelsand tilt-floors), photobeams (for activity orfeeding), electrical contact circuits (for eatingor drinking meters), and implantable radio-telemetry transmitters (for general activity,body temperature, blood pressure, and heartrate). Sleep-wake states can be measured elec-trophysiologically by implanted skull elec-trodes connected to cables or radiofrequencytransmitters.

Wheel running data are conveniently dis-played in the "actogram" format (Fig. 17-1).In this format, consecutive days of activity arealigned vertically, permitting easy visual de-termination of two fundamental parametersof a daily rhythm: its phase (position within acycle) and its period (average duration of a cy-cle, the reciprocal being frequency). As illus-trated in Figure 17-1, wheel running activityin rats exhibits a very robust daily rhythm. Ina standard light-dark (LD) cycle (12 hourlight: 12 hour dark) running is concentrated atnight, and the onset of running is typicallyabrupt and predictable near the time of lights-ofF. The onset of running thus is an easily ob-servable phase of this activity rhythm. The pe-riod of the rhythm can then be quantified bymeasuring the average interval between suc-cessive onsets. This can be done with linearregression (the slope of a line fit to a set of

183

Page 201: 0195162854

184

Figure 17-1. Wheel running activity in single-housed male (A and B) and female (Q rats maintained in a 24hour light-dark cycle (dark indicated by shading) with free access to a 33 cm running wheel. Each line rep-resents 1 day, with time plotted from left to right in 10 minute bins and consecutive days aligned vertically.Time bins in which wheel revolutions were detected are indicated by small vertical deflections (creating aheavy line when running is continuous). (A and C) Nocturnal concentration of wheel running typical of rats,and the persistence of this daily rhythm with an approximately 24 hour periodicity during 1 week in constantdark. The female rat (C) also shows a prominent increase of running every 4 to 5 days, marking the night ofbehavioral estrous. (B) Dramatic effect of restricting food to the daytime. This rat runs almost exclusively atnight when food is available ad libitum but exhibits a prominent bout of running in anticipation of mealtimewhen food is restricted to 3 hours in the middle of the lights-on period (indicated by the open rectangle). Inthis example, the entire circadian profile appears inverted, and this is sustained during the last 2 days, whenfood was omitted completely. More commonly, rats express both daytime food anticipatory running and apersisting but lower level of nocturnal activity. Ablation of the hypothalamic suprachiasmatic nucleus elim-inates circadian rhythms entrained to light-dark cycles or expressed in constant dark but does not eliminatethe daily rhythm of food anticipatory running that emerges if meals are provided at circadian intervals.

consecutive onsets is the average deviationfrom 24 hours). In an LD cycle, the averageperiod is 24 hours, and the phase is stable (on-sets occur near lights-off). Thus, the LD cyclecontrols the phase and period, a phenomenonknown as entrainment. The phase of entrain-ment may differ by individual or by age; thatis, some rats may begin running just beforelights-off (phase advanced relative to LD), andothers may begin slightly after lights-off(phase delayed; e.g., Fig. 17-1B). Phase may

be altered by aging, reproductive status (e.g.,estrous, Fig. 17-1C), and other factors.

In the absence of an LD cycle or othersignificant periodic environmental stimuli("constant conditions"), the daily wheel run-ning rhythm persists, indicating that it is gen-erated by an internal timekeeping device (aself-sustaining oscillator). The oscillator canmarch to the beat of the LD cycle, but in theabsence of periodic environmental signals, it"free-runs" and expresses its own intrinsic pe-

MOTOR SYSTEMS

Page 202: 0195162854

Chapter 17. Circadian Rhythms 185

riodicity. In rats (as in humans), activity on-sets in constant conditions tend to occur a bitlater each "day" (i.e., relative to local time),indicating that the intrinsic period of the in-ternal oscillator is slightly longer than 24hours (thus the designation "circadian"). Theperiod is not immutable and exhibits individ-ual differences, changes with age and hor-monal status, and a dependence on light in-tensity; the brighter the constant light, theslower the circadian oscillator runs. The free-running phase also can be shifted (advancedor delayed) by those stimuli capable of en-training circadian rhythms (see later). Phaseshifts can be quantified by comparing regres-sion lines fit to activity onsets before and af-ter the stimulus.

Other variables, typically those that varycontinuously such as core body temperatureor levels of some hormones, are more easilyvisualized as waveforms. Simple waveformscan be described by their period and their am-plitude. The amplitude is the difference be-tween the peak (or trough) value and themean value (the range is the difference be-tween the peak and the trough). Period andamplitude can be quantified by curve-fittingprocedures. The peak (acrophase) of a fittedsine wave is a definable phase that can be usedfor regression analyses. Besides linear regres-sion, the two most common approaches toquantifying circadian period are the fastFourier transform and the chi-squared peri-odogram (Refinetti, 1993).

CIRCADIAN REGULATION OFBEHAVIOR AND PHYSIOLOGY

SLEEP AND WAKE

Rats sleep predominantly during the daily lightperiod (about 80% of a 12 hour light period)but also significantly at night (about 30% of a12 hour dark period; Fig. 17-2). Two primarystages of sleep are recognized in all mammals:rapid eye movement (REM) sleep and non-

REMS (NREM) sleep. Because eye movementsare generally not measured in rat sleep studies,the REM stage is often referred to as "para-doxical" sleep, in recognition of its paradoxicaldefining characteristics of a waking-like elec-troencephalogram (EEC) combined with higharousal threshold and active inhibition of anti-gravity muscles. NREM sleep is often referredto as slow-wave sleep (SWS), but this is a mis-nomer because the cortical EEC during muchof NREM sleep may have few slow waves (i.e.,those in the 1 to 4 Hz range). SWS should re-fer only to that portion of NREM with a highconcentration of slow waves. SWS typically ismaximal early in the sleep period and declinesexponentially. It is also greatly enhanced byprior sleep loss and is considered to be a cor-relate of sleep intensity or important sleep re-covery processes. REM sleep, by contrast, grad-ually increases in bout duration and frequencyas the sleep period progresses. NREM sleep andREM sleep tend to alternate at about 20 minuteintervals in rats (90 minutes in humans), butthis "ultradian" sleep cycle is weak and variablein the rat.

The strength of the circadian modulationof sleep can be further appreciated in conflictexperiments, in which rats are sleep deprivedand then allowed to recover during their usualwake period (Mistlberger et al., 1983; Fig.17-2). Sleep (particularly SWS and REM) issignificantly elevated relative to the normalamount of sleep at this phase (taken as evi-dence for homeostatic regulation of sleep),but total sleep in 2 hour blocks throughoutthe night remains much lower by comparisonwith sleep during the usual lights-on period.If recovery begins during the light period,there is little elevation of total sleep time, in-dicating that the circadian clock maintainssleep at a physiological ceiling during theusual sleep phase.

LOCOMOTION

In rats, spontaneous wheel running is morestrictly nocturnal than is waking. Thus, al-

Page 203: 0195162854

186 MOTOR SYSTEMS

Figure 17-2. Group mean wave forms of polygraphically recorded sleep-wake states in rats maintained in a12 hour light-dark cycle for 2 days before and 2 days after 24 hours of total sleep deprivation on a slowly ro-tating drum over water. Some sleep (heavy line) is evident at all times of day, but the total amount is great-est during the day. Total sleep time is elevated by prior sleep deprivation but also is constrained by circadianregulation. Non-rapid eye movement sleep is divided into high (HS2) and low (HS1) electroencephalographicamplitude substages. These vary inversely with time of day and with time after sleep deprivation; HS2 is con-sidered to be functionally high-intensity sleep. Paradoxical sleep (PS also known as rapid eye movement sleep)is maximal at the end of the usual sleep period but significantly elevated immediately after total sleep loss.(Reprinted from Sleep, Vol 6, Mistlberger et al., pp. 217-233, Copyright 1983, with permission from Associ-ated Professional Sleep Societies.)

though rats do engage in some waking activ-ities during the day, wheel running is mini-mal. At night, the pattern of running activityvaries by individual and by strain. Membersof some rat strains tend to run primarily earlyin the night and others late in the night, andsome exhibit bimodal or trimodal patterns,with two or three peaks at 12 or 6 hour in-tervals (Wollnik, 1991). These patterns are ev-ident in sleep-wake states and general lo-comotion and are not secondary to the bioen-ergetic consequences of wheel running.

The amount, latency, and spatial distri-bution of locomotor activity in certain situa-tions are often used as metrics for inferredpsychological states such as anxiety or de-pression. There is some evidence that the re-sponse of rats in these situations (e.g., an openfield or elevated plus maze) may vary withtime of day and that this circadian influencemay vary by metric (e.g., Jones and King,2001; Andrade et al., 2003). It is thus recom-

mended that in all such tests, time of day beconsidered as a potentially significant inde-pendent variable.

FEEDING AND DRINKING

Rats with free access to food eat in discretebouts. These bouts are both longer and morefrequent at night, when about 75% of totaldaily food intake occurs (Rosenwasser et al.,1981). Often the daily feeding rhythm is bi-modal, with food intake concentrated at thestart and the end of the dark period, but thismay vary with strain of rat (Glendinning andSmith, 1994). Circadian regulation of feedingis also evident in the compensatory responseto food deprivation. Thus, if a rat is food de-prived for 42 hours, the size of the first "re-covery" meal is significantly smaller if it oc-curs during the day rather than during thenight (Bellinger and Mendel, 1975). However,rats are opportunistic feeders, and if food is

Page 204: 0195162854

Chapter 17. Circadian Rhythms 187

restricted to a particular time of day, the feed-ing rhythm adjusts. If food is available onlyfor a few hours in the daytime, the amount offood eaten during this time gradually in-creases over about 1 week, so that initialweight loss is reversed. Rats also exhibit ac-tivity in anticipation of the mealtime (Mistl-berger, 1994; e.g., Fig. IT-IB). This anticipa-tory activity is circadian clock driven; once itis established, it reappears at the same timeeach day even if the animal is food deprived.Adaptive plasticity in the circadian program-ming of food seeking and ingestive behavioris undoubtedly one of the reasons that rats aresuch a widely distributed and successfulspecies.

Drinking is also controlled by the circa-dian clock and tends to be more strictly noc-turnal, with 85% of fluid intake occurring atnight. Compensatory drinking in response toan osmotic challenge (e.g., hypertonic salineinjection) is significantly greater at night(Johnson and Johnson, 1991). Rats also showcircadian anticipatory activity to a daily op-portunity to drink, but this is much less promi-nent than is food-anticipatory activity (Mistl-berger, 1994).

THERMOREGULATION

In an LD cycle, body temperature (Tb) in ratsvaries from a daytime mean of about 37.3° Cto a nighttime mean of about 38.1° C. Al-though Tb is influenced by behavioral state,the rhythm is not secondary to daily variationsin behavior. First, Tb begins to rise about 2hours before the daily active period, with thesteepest increase observed from 30 minutesbefore to 60 minutes after waking (Refinettiand Menaker, 1992). Second, in constant light(LL), circadian rhythms in rats damp out overa period of weeks to months, but the Tbrhythm typically persists longer than does theactivity rhythm (Eastman and Rechtschaffen,1983).

Thermoregulation in the rat is accom-plished by both autonomic and behavioral

means. Notably, rats show a circadian rhythmin self-selected ambient temperature, prefer-ring about 28° C during their usual rest pe-riod, when the Tb is low, and about 22° to24° C during their usual active phase, whenTb is high. Thus, the circadian rhythm of pre-ferred ambient temperature is in antiphasewith the circadian rhythm of Tb. This indi-cates that the circadian rhythm of Tb is notdue to a circadian modulation in "setpoint,"as is the case with fever (Refinetti andMenaker, 1992).

REPRODUCTIVE BEHAVIOR

Many aspects of reproductive behavior exhibita circadian rhythm. In the presence of a re-ceptive female, male rats copulate more fre-quently and with shorter latency at night thanduring the day (Beach and Levinson, 1949).When tested at various times throughout thenight, the greatest number of intromissions(Harlan et al, 1980) and the shortest latencyto mounting, intromission, and ejaculation(Dewsbury, 1968) are observed during the lasthalf of the dark phase. A circadian rhythm inspontaneous seminal emissions (producingseminal plugs) has also been reported (Kihl-strom, 1966; Stefanick, 1983).

Female rats also exhibit circadian rhythmsin reproductive behavior and physiology. Fe-males become sexually receptive and exhibittheir highest level of mating behavior (be-havioral estrus) on the night of ovulation,which occurs every 4 to 5 days in rats en-trained to an LD cycle (Ball, 1937). The es-trous cycle remains coupled to circadianrhythms in constant conditions, with behav-ioral estrous recurring every four or five cir-cadian cycles.

Although rats are not a photoperiodicspecies (i.e., breeding is not dependent on daylength), the length of the estrous cycle is af-fected by photoperiod. In an LD cycle with a12 hour light period, about 70% of rats havea 4-day estrous cycle, whereas only 10% havea 5-day cycle. If lights are on for 16 hours per

Page 205: 0195162854

188 MOTOR SYSTEMS

day, 21% of rats have a 4-day cycle, whereas46% have a 5-day cycle (Hoffmann, 1968).

Timing of birth is also regulated by thecircadian clock and tends to occur during thedaytime on gestational day 22 or 23, whetherin an LD cycle or in constant dark (DD) (Lin-coln and Porter, 1976). This rhythm of partu-rition is eliminated by either ablation of thematernal circadian clock in the suprachias-matic nucleus (SCN) or removal of the brainsof the fetuses during late gestation (Reppertet al., 1987).

Daily rhythms may be problematic forscientists engaged in animal husbandry, par-ticularly if access to pups immediately afterbirth is desired. The time of peak breeding ef-ficiency is in antiphase with the time of par-turition. If the animals are housed in a reverseLD cycle, breeding can be done during regu-lar 9 AM-to-5 PM work hours, but births willoccur during the rats' "day," which will be theresearcher's night. Using a normal LD cycle,births will occur conveniently for the investi-gator, but the time of peak mating efficiencywill occur at night. Mayer and Rosenblatt(1997) have devised a husbandry protocol thatovercomes these issues. Animals are bred in areverse LD cycle. After 1 week, the pregnantdams are transferred at mid-dark phase to aroom with a normal LD cycle. Not only areall pups born during the daytime on gesta-tional day 22 but also 75% of births occur dur-ing the middle third of the day. Although thisprotocol may simplify the work schedule of aresearcher, it should be noted that invertingthe LD cycle of a pregnant dam may havesome effect on the pups' physiology, such thatthey may differ from pups produced under amore conventional protocol.

LEARNING AND MEMORY

In rats, circadian rhythms are relevant to learn-ing and memory in at least three ways. First,there are circadian variations in acquisition orrecall on some tasks. Second, phase shifts ofcircadian rhythms can disrupt learning and

memory. Third, circadian oscillators can pro-vide time of day cues that can be used by ratsto enable time-place learning; that is, circadianoscillators can be used as continuously con-sulted clocks to recognize time of day in theabsence of environmental time cues.

The effect of time of day on learning, re-call, and extinction are varied and may be taskdependent. Performance of rats in a passiveavoidance task was better during the light pe-riod, although testing and training occurred atthe same phase, making it impossible to dis-tinguish acquisition from recall (Davies et al.,1973). Rats appear to be better at acquiring anactive avoidance task during the night undersome situations. Extinction also appears to bemore rapid at night under some situations(Novakova et al., 1983). Another study foundno rhythm in acquisition of a free-operantavoidance task but did find that rats weremore efficient responders during the night(Ghiselli and Patton, 1976). A more recentstudy noted that old rats, but not young rats,had performance deficits when tested duringthe late night, compared with the early night,on both inhibitory avoidance and delayed al-ternation tasks (Winocur and Hasher, 1999).

Disruption of circadian rhythms has vary-ing effects on learning in rats. LD phase shiftsor exposure to LL impairs performance on apassive avoidance task learned before the lightcycle change (Tapp and Holloway, 1981;Fekete et al., 1985). LD shifts also impair re-call, but not acquisition, on a water-maze task(Devan et al., 2001). By contrast, LD phaseshifts may facilitate extinction on an activeavoidance task (Fekete et al., 1985) and haveno affect on social memory (Reijmers et al.,2001).

Rats appear to perform best on passiveand active avoidance tasks when tested at thesame circadian phase at which they weretrained (Holloway and Wansley, 1973a,1973b; Wansley and Holloway, 1975). Thismay indicate that time of day (circadian phase)is encoded in the memory for these tasks. En-coding of circadian phase is also suggested by

Page 206: 0195162854

Chapter 17. Circadian Rhythms 189

evidence that rats can discriminate time ofday. Rats can learn to press one lever at onetime of the day and another lever at anothertime of the day for food access (Boulos andLogothetis, 1990; Mistlberger et al., 1996).Time-place learning may be more difficult todemonstrate on other types of tasks such asplace preferences, water-maze escape routes,and radial arm mazes (Thorpe et al., 2003).

PHARMACOLOGY

Circadian rhythms in the response to drugs areto be expected, due to rhythms in theabsorption/clearance ratio and in the suscepti-bility of the target tissue. For orally adminis-tered drugs, absorption is influenced by stom-ach content, which changes over the course ofthe day according to the circadian rhythm offeeding behavior. Absorption is also influencedby intestinal enzyme activity, gastric activity,and rate of glucose uptake into the circulatorysystem, all of which exhibit circadian rhythms.Clearance is a product of excretion and me-tabolism/inactivation, all of which again havea circadian rhythm (for a review, see Moore-Ede et al., 1982). There is a rhythm in the ac-tivity of liver enzymes. The pH of urine is alsorhythmic, which influences the movement ofdrugs from the blood into urine.

There are many anecdotal reports ofresearchers inadvertently discovering a circa-dian rhythm in the action of a drug (Moore-Ede et al., 1982). In these cases, large vari-ability in response over trials could beexplained only when time of administrationwas taken into account. For those testing newPharmaceuticals, it is imperative that time ofday be taken into account. For some drugs,peak efficacy occurs at one phase while peaktoxicity occurs at a different phase. Appropri-ate timing of administration may thereforemaximize clinical effect while minimizing ad-verse effects.

Two classes of drugs of particular inter-est to those working with rats are anestheticsand analgesics, both of which have circadian

rhythms in efficacy. The concentration ofhalothane required to maintain anesthesia inrats is lower during the day than it is duringthe early night (Munson et al., 1970). How-ever, the lethal concentration of halothane islower during the early night (Matthews et al.,1964). The efficacy/toxicity phase-responsecurves for the anesthetic pentobarbitol aresimilarly inverted (Moore-Ede et al., 1982).These two examples illustrate how the thera-peutic index of a drug can have a circadianrhythm. For drugs such as pentobarbital, itmay be advisable to restrict use to the phaseswhen the therapeutic index is at its broadest(i.e., the last half of the day in this case).

Surgical protocols frequently require theinclusion of an analgesic in addition to theanesthetic. Unfortunately, the trough in thecircadian rhythm of analgesia provided by agiven dose of morphine coincides with thephase when the therapeutic index for pento-barbital is at its broadest. This means that thetime of day when pentobarbital is the safestto use is the same time of day when morphineprovides the least pain relief.

ENVIRONMENTAL ANDBEHAVIORAL INFLUENCES

Circadian rhythms influence many physiolog-ical systems relevant to phenomena of inter-est to behavioral neuroscientists. Thus, time ofday is usually an important methodologicalconsideration. Given that circadian rhythmsare entrained by environmental time dues, at-tention to environmental factors that may al-ter circadian timing is vital.

LIGHT

Light affects behavioral rhythmicity in rats intwo ways. First, acute light exposure inhibitsactivity and promotes sleep. Constant lighttonically suppresses activity, but sleep dura-tion does not remain enhanced (sleep is self-limiting). The acute affect of light is often re-

Page 207: 0195162854

190 MOTOR SYSTEMS

7:00 19:00

Figure 17-3. Wheel running activity of rats subjected to (A) an 8 hour delay in the light/dark cycle, simu-lating travel west, or (B) an 8 hour advance of the light/dark cycle, simulating travel east. Reentrainment tothe delay shift is accomplished more quickly. In addition to direction of shift, many other factors can affectthe rate of reentrainment (see text). (Modified from Journal of Neuroscience, Vol 23, Nagano et al., pp.6141-6151, Copyright 2003, with permission from Society for Neuroscience.)

ferred to as masking, because it may mask thetrue phase of the rat's circadian rest-activitycycle. The second major effect of light is itsrole as a dominant entraining stimulus (a zeit-geber, which is a German word for "time-giver") for the circadian clock. Light entrainsthe clock because of its ability to induce phaseshifts, which compensate each day for the dis-crepancy between the circadian period of theclock and the 24 hour period of the LD cycle.Light exposure in the evening or early nightnormally phase delays the clock, whereas lightexposure in the morning or late night phaseadvances the clock. Light during the day haslittle effect. Thus, the circadian clock has a cir-cadian rhythm of sensitivity to light. The netresult is that any LD cycle with a period inthe circadian range will "capture" the free-running clock and prevent it from drifting bydaily phase adjustments opposing the direc-tion of natural drift. For more detail on theformal mechanism of entrainment, see Mistl-berger and Rusak (2000). Suffice it to say, turn-ing on the light during the usual night can al-ter the timing of circadian rhythms in rats.Moreover, the light need not be bright or pro-longed; even a few seconds each day can en-train free-running rhythms in nocturnal ani-mals otherwise maintained in DD.

In LL, rats initially free-run with a periodthat is proportional to light intensity. Overweeks to months, bright LL can attenuate andmay eliminate circadian rhythms. Circadianrhythmicity can be restored by one cycle ofLD (Eastman and Rechtschaffen, 1983).

The time required to re-entrain to ashifted LD cycle, such as may occur when an-imals are shipped from a breeding facility toa research facility, depends on a number offactors, including the direction of shift (lesstime if the LD cycle is delayed rather than ad-vanced; e.g., see Fig. 17-3), the magnitude ofshift, and the housing conditions (e.g., lesstime if light is brighter or if rats have a run-ning wheel in their home cage). A rule ofthumb is to allow 1 day for every hour of LDcycle shift, but complete reentrainment cantake 3 to 4 weeks.

FEEDING

Along with light, the timing of food intake isthe most important determinant of circadiantiming in rats. As noted earlier, rats can adaptto a scheduled daily meal by increasing activ-ity in anticipation of mealtime and by a grad-ual increase in meal size. Mealtime is the dom-inant zeitgeber controlling the phase of daily

Page 208: 0195162854

Chapter 17. Circadian Rhythms 191

rhythms of digestive processes and metabo-lism. Feeding schedules are widely used in be-havioral studies of rats. For example, to mo-tivate appetitive learning, rats are typicallymaintained at some reduced percentage offree-feeding weight. Typically, rats areweighed and fed at a similar time each day,and this can be expected to affect circadiantiming of behavior and physiology (see Mistl-berger, 1990, 1994).

LOCOMOTOR ACTIVITY

In DD or LL, the period of free-runningrhythms in rats is shortened if they are pro-vided ad libitum access to a running wheel(Yamada et al., 1986). Some physiological cor-relate of spontaneous behavioral activity canfeed back to alter the rate at which the circa-dian clock cycles. Rats in DD can also be en-trained by a daily bout of forced exercise ona treadmill (Mistlberger, 1991). However, in-duction of locomotor activity in the middle ofthe usual sleep phase does not induce phaseshifts in rats, as has been shown for hamsters.

SOCIAL INFLUENCES

Rats that cohabitate in a seminatural envi-ronment have been reported to organize lo-comotor and feeding activity according to so-cial status, with subordinates being forced tofeed more during the light period. It is notknown whether this represents a change inphase of the subordinate's circadian clock (fora review, see Mistlberger and Skene, 2004).

STRESS AND AROUSAL

Social defeat is considered a highly stressfulstimulus for rats. A single social defeat can de-crease the amplitude of locomotor, tempera-ture, eating, drinking, and heart rate rhythmsfor a number of weeks (Meerlo et al., 2002).Other types of stress, such as surgical stress,chronic mild stress, forced swimming, re-straint, and foot shocks, have similar effects.

However, these stimuli do not phase shift theclock. Furthermore, attenuation of circadianamplitude in overt rhythms results fromprocesses downstream from the circadianclock and is not caused by attenuation of theamplitude of the circadian clock itself (Meerloet al., 2002).

NEURAL MECHANISMS OFCIRCADIAN RHYTHMS

A MASTER CIRCADIAN PACEMAKER ISLOCATED IN THE HYPOTHALAMUS

Curt Richter, a pioneer of circadian rhythmsresearch, spent many decades searching forthe organ responsible for generating circadianrhythms in rats. He used a wide range of in-terventions, including brain lesions and re-moval of glands (e.g., adrenals, gonads, pitu-itary, thyroid, pineal, and pancreas). Onlyventral hypothalamic lesions eliminated free-running rhythms in activity, drinking andfeeding (Richter, 1967). The site of a mastercircadian pacemaker in rats was later localizedto the SCN of the ventral anterior hypothala-mus (Moore and Eichler, 1972; Stephan andZucker, 1972). Much of the convergent evi-dence that the SCN is the site of the circadianclock comes from studies of rats. The SCN ex-presses rhythms in metabolic and single-unitneural activity, and individual rat SCN cells indissociated cell cultures can maintain circa-dian rhythmicity for weeks at a time (Welshet al., 1995). Transplantation of fetal SCN intoan SCN-lesioned rat can restore behavioralrhythmicity (e.g., Lehman et al., 1987). TheSCN receives direct and indirect projectionsfrom the retina, and many of its neurons re-spond to retinally mediated light. The SCN isthus considered to be the site of a master cir-cadian clock that mediates light-entrainablecircadian rhythms. It receives a variety ofother inputs, notably from the median rapheand thalamic intergeniculate leaflet, that maymediate feedback effects of behavioral activ-

Page 209: 0195162854

192 MOTOR SYSTEMS

ity (for a review, see Mistlberger et al., 2000).However, it is not necessary for entrainmentof circadian rhythms to feeding schedules.SCN-ablated rats are arrhythmic when fed adlibitum, but they develop robust food antici-patory rhythms if feed once or twice per day(Mistlberger, 1994).

CIRCADIAN CLOCK GENES HAVE BEEN

IDENTIFIED AND ARE EXPRESSED

IN MANY TISSUES

Circadian rhythms in single SCN neurons aredriven by autoregulatory transcription-trans-lation feedback loops, involving a set of so-called circadian clock genes and their proteinproducts (Reppert and Weaver, 2001). These

Figure 17-4. A simplified conceptual model of the mam-malian circadian system. At the heart of the system is thesuprachiasmatic nucleus (SCN), which consists of a het-erogeneous population of circadian clock cells that gener-ate circadian rhythms and that are entrained to environ-mental light/dark cycles via a retinohypothalamic tract(RHT) originating in photoreceptive retinal ganglion cells.The SCN drives daily activity rhythms but also is subjectto feedback from activity, which can alter its circadian pe-riod or control its phase in the absence of light/dark cycles.A food-entrainable circadian pacemaker (FEP) is located atan unknown site outside of the SCN. Other circadian os-cillators exist in peripheral organs and tissues. These arealso entrainable by food, via unknown pathways, and pos-sibly also by the SCN. In rats, light directly inhibits activityand promotes sleep. (Adapted from Cell, Vol 111, Schiblerand Sassone-Corsi, pp. 919-922 Copyright 2002, with per-mission from Elsevier.)

genes are also expressed in a variety of othertissues, including elsewhere in the brain andin tissues in skeletal muscle, heart, lungs, andliver (Yamazaki et al., 2000). In vitro, circadianoscillations in these tissues damp out in about4 days but can be reinitiated by changing theculture media. In vivo, the phase of peripheraloscillators (but not the SCN) is controlled byfeeding time. The rat (and mammalian) circa-dian system can thus be characterized asmultioscillatory, anatomically distributed, andsensitive to photic and nonphotic entrainingstimuli (Fig. 17-4).

REFERENCES

Andrade MM, Tome MF, Santiago ES, Lucia Santos A,de Andrade TG (2003) Longitudinal study of dailyvariation of rats' behavior in the elevated plus-maze.Physiology and Behavior 78:125-133.

Ball J (1937) A test for measuring sexual excitability inthe female rat. Comparative Psychology Mono-graphs 14:1-37.

Beach FA and Levinson G (1949) Diurnal variation inthe mating behavior of male rats. Proceedings of theSociety for Experimental Biology and Medicine72:78-80.

Bellinger LL and Mendel VE (1975) Effect of deprivationand time of refeeding on food intake. Physiologyand Behavior 14:43-46.

Boulos Z and Logothetis DE (1990) Rats anticipate anddiscriminate between two daily feeding times. Phys-iology and Behavior 48:523-529.

Davies JA, Navaratnam V, Redfern PH (1973) A 24-hourrhythm in passive-avoidance behaviour in rats. Psy-chopharmacologia 32:211-214.

Devan BD, Goad EH, Petri HL, Antoniadis EA, HongNS, Ko CH, Leblanc L, Lebovic SS, Lo Q, RalphMR, McDonald RJ (2001) Circadian phase-shiftedrats show normal acquisition but impaired long-term retention of place information in the watertask. Neurobiology of Learning and Memory75:51-62.

Dewsbury D (1968) Copulatory behavior of rats—Vari-ations within the dark phase of the diurnal cycle.Communications in Behavioral Biology 1:373-377.

Eastman C and Rechtschaffen A (1983) Circadian tem-perature and wake rhythms of rats exposed to pro-longed continuous illumination. Physiology andBehavior 31:417-427.

Fekete M, van Ree JM, Niesink RJ, de Wied D (1985)

Page 210: 0195162854

Chapter 17. Circadian Rhythms 193

Disrupting circadian rhythms in rats induces retro-grade amnesia. Physiology and Behavior 34:883-887.

Ghiselli WB and Patton RA (1976) Diurnal variation inperformance of free-operant avoidance behavior ofrats. Psychology Reports 38:83-90.

GlendinningJI and Smith JC (1994) Consistency of mealpatterns in laboratory rats. Physiology and Behav-ior 56:7-16.

Harlan RE, Shivers BD, Moss RL, Shryne JE, Gorski RA(1980) Sexual performance as a function of time ofday in male and female rats. Biology Reproduction23:64-71.

Hoffmann JC (1968) Effect of photoperiod on estrouscycle length in the rat. Endocrinology 83:1355-1357.

Holloway FA and Wansley RA (1973a) Multiple reten-tion deficits at periodic intervals after active and pas-sive avoidance learning. Behavioral Biology 9:1-14.

Holloway FA and Wansley RA (1973b) Multiphasic re-tention deficits at periodic intervals after passive-avoidance learning. Science 180:208-210.

Johnson RF and Johnson AK (1991) Drinking after os-motic challenge depends on circadian phase in ratswith free-running rhythms. American Journal ofPhysiology 261:R334-R338.

Jones N and King SM (2001) Influence of circadian phaseand test illumination on pre-dinical models of anx-iety. Physiology and Behavior 72:99-106.

Kihlstrom JE (1966) Diurnal variation in the sponta-neous ejaculations of the male albino rat. Nature209:513-514.

Lehman MN, Silver R, Gladstone WR, Kahn RM, Gib-son M, Bittman EL (1987) Circadian rhythmicity re-stored by neural transplant. Immunocytochemicalcharacterization of the graft and its integration withthe host brain. Journal of Neuroscience 7:1626-1638.

Lincoln DW and Porter DG (1976) Timing of the pho-toperiod and the hour of birth in rats. Nature260:780-781.

Matthews JH, Marte E, Halberg F (1964) A circadian sus-ceptibility-resistance cycle to fluothane in male Blmice. Canadian Anaesthetists' Society Journal 11:280-290.

Mayer AD and Rosenblatt JS (1997) A method for regu-lating the duration of pregnancy and the time of par-turition in Sprague-Dawley rats (Charles River CDstrain). Developmental Psychobiology 32:131-136.

Meerlo P, Sgoifo A, Turek FW (2002) The effects of so-cial defeat and other stressors on the expression ofcircadian rhythms. Stress 5:15-22.

Mistlberger RE (1990) Circadian pitfalls in experimentalparadigms employing food restriction. Psychobiol-ogy 18:23-29.

Mistlberger RE (1991) Effects of daily schedules of forced

activity on free-running rhythms in the rat. Journalof Biological Rhythms 6:71-80.

Mistlberger RE (1994) Circadian food-anticipatory ac-tivity: Formal models and physiological mecha-nisms. Neuroscience and Biobehavioral Reviews18:171-195.

Mistlberger RE and Rusak B (2000) Circadian rhythmsin mammals: Formal properties and environmentalinfluences. In: Principles and practice of sleep med-icine, 3rd Edition (Kryger MH, Roth T, DementWC, eds.), pp. 321-333. Philadelphia: WB Saunders.

Mistlberger RE and Skene DJ (2004) Social influences oncircadian rhythms in man and animal. BiologicalReviews, in press.

Mistlberger RE, de Groot MH, Bossert JM, Marchant EG(1996) Discrimination of circadian phase in intactand suprachiasmatic nuclei-ablated rats. Brain Re-search 739:12-18.

Mistlberger RE, Antle MC, Glass JD, Miller JD (2000)Behavioral and serotonergic regulation of circadianrhythms. Biological Rhythm Research 31:240-283.

Misdberger RE, Bergmann BM, Waldenar W, Rechtschaf-fen A. (1983) Recovery sleep following sleep depri-vation in intact and suprachiasmatic nuclei-lesionedrats. Sleep 6:217-233.

Moore RY and Eichler VB (1972) Loss of a circadian ad-renal corticosterone rhythm following suprachias-matic lesions in the rat. Brain Research 42:201-206.

Moore-Ede MC, Sulzman FM, Fuller CA (1982) The clocksthat time us: Physiology of the circadian timing sys-tem. Cambridge, Mass: Harvard University Press.

Munson ES, Martucci RW, Smith RE (1970) Circadianvariations in anesthetic requirement and toxicity inrats. Anesthesiology 32:507-514.

Nagano M, Adachi A, Nakahama K, Nakamura T,Tamada M, Meyer-Bernstein E, Sehgal A, Shige-yoshi Y. (2003) An abrupt shift in the day/night cy-cle causes desynchrony in the mammalian circadiancenter. Journal of Neuroscience 23:6141-6151.

Novakova V, Sterc J, Knez R (1983) The active avoidancereaction of laboratory rats: Differences between ex-periments carried out in the phase of motor activ-ity and inactivity. Physiologica Bohemoslovaca32:38-44.

Refinetti R (1993) Laboratory instrumentation and com-puting: Comparison of six methods for the deter-mination of the period of circadian rhythms. Phys-iology and Behavior 54:869-875.

Refinetti R and Menaker M (1992) The circadian rhythmof body temperature. Physiology and Behavior51:613-637.

Reijmers LG, Leus IE, Burbach JP, Spruijt BM, van_ReeJM (2001) Social memory in the rat: Circadian vari-ation and effect of circadian rhythm disruption.Physiology and Behavior 72:305-309.

Page 211: 0195162854

194 MOTOR SYSTEMS

Reppert SM and Weaver DR (2001) Molecular analysisof mammalian circadian rhythms. Annual Reviewof Physiology 63:647-676.

Reppert SM, Henshaw D, Schwartz WJ, Weaver DR(1987) The circadian-gated timing of birth in rats:Disruption by maternal SCN lesions or by removalof the fetal brain. Brain Research 403:398-402.

Richter CP (1922) A behavioristic study of the activityof the rat. Comparative Psychology Monographs1:1-55.

Richter CP (1967) Sleep and activity: Their relation tothe 24-hour clock. Research Publications—Associa-tion for Research in Nervous and Mental Disease45:8-29.

Rosenwasser AM, Boulos Z, Terman M (1981) Circadianorganization of food intake and meal patterns in therat. Physiology and Behavior 27:33-39.

Schibler U and Sassone-Corsi P (2002) A web of circa-dian pacemakers. Cell 111:919-922.

Slonaker JR (1908) Description of an apparatus forrecording the activity of small mammals. TheAnatomical Record 2:116-122.

Stefanick ML (1983) The circadian patterns of spontaneousseminal emission, sexual activity and penile reflexesin the rat. Physiology and Behavior 31:737-743.

Stephan FK and Zucker I (1972) Circadian rhythms indrinking behavior and locomotor activity of rats areeliminated by hypothalamic lesions. Proceedings ofthe National Academy of Science USA 69:1583-1586.

Tapp WN and Holloway FA (1981) Phase shifting cir-cadian rhythms produces retrograde amnesia. Sci-ence 211:1056-1058.

Thorpe CM, Bates ME, Wilkie DM (2003) Rats havetrouble associating all three parts of the time-place-event memory code. Behavioral Processes 63:95-110.

Wansley RA and Holloway FA (1975) Multiple retentiondeficits following one-trial appetitive training. Be-havioral Biology 14:135-149.

Welsh DK, Logothetis DE, Meister M, Reppert SM(1995) Individual neurons dissociated from ratsuprachiasmatic nucleus express independentlyphased circadian firing rhythms. Neuron 14:697-706.

Winocur G and Hasher L (1999) Aging and time-of-dayeffects on cognition in rats. Behavioral Neuro-science 113:991-997.

Wollnik F (1991) Strain differences in the pattern and in-tensity of wheel running activity in laboratory rats.Experientia 47:593-598.

Yamada N, Shimoda K, Takahashi K, Takahashi S (1986)Change in period of free-running rhythms deter-mined by two different tools in blinded rats. Physi-ology and Behavior 36:357-362.

Yamazaki S, Numano R, Abe M, Hida A, Takahashi R,Ueda M, Block GD, Sakaki Y, Menaker M, Tei H(2000) Resetting central and peripheral circadian os-cillators in transgenic rats. Science 288:682-685.

Page 212: 0195162854

Regulatoty •%/Systems

Page 213: 0195162854

This page intentionally left blank

Page 214: 0195162854

Eating

PETER G. CLIFTON

18

For many, perhaps most, of those who studythe rat in a laboratory context, feeding is ameans to an end. A hungry rat rapidly learnsto discriminate two tones, acquire spatial in-formation in an eight-arm radial maze, or rundown an alley for food reward. The behaviorthat allows the rat to ingest that food may onlybe "observed" in proxy form as the deliveryof a food pellet into an operant chamber. Yet,feeding in the rat has a richness and a com-plexity that is worth study for its own sake,with relevance to those whose interest in therat has a quite different focus. In addition, therat is widely used in applied studies of feedingthat extend our understanding of obesity inhumans. My intention in this chapter is to givea brief description of the salient features of ratfeeding behavior relevant to laboratory-basedstudies of this species.

LIFETIME PATTERNS OF INTAKE

Young rats, like all mammals, receive their ini-tial nutrition from their mother in the formof milk. In a laboratory context, they typicallybegin to take solid food from day 16 of life(Thiels et al., 1990) and are usually weanedfrom the mother by day 21. However, if wean-ing is not enforced, the young continue tosuckle, although progressively less frequently,until day 34 (Thiels et al., 1990). Food intakerelative to body mass is high during early life,gradually decreasing as the rate of increase inbody weight declines after sexual maturity.Asymptotic body weight and food consump-

tion show a clear sex difference, with malesbeing 1.2 to 1.5 times the weight of femalesas adults. In addition, there are considerabledifferences in asymptotic body weight be-tween commonly used laboratory strains.Among pigmented animals, which are usedmore often in behavioral studies, typical adultweights vary from 300 grams in the darkagouti strain to 550 grams in the Lister hoodedstrain. Daily food intake is sensitive to envi-ronmental variables and especially to reduc-tions in environmental temperature (Leungand Horwitz, 1976).

Food intake in sexually mature femalerats varies significantly over the estrus cycle.On the night of estrus, when estradiol levelsare high, the female reduces food intake andbecomes more active. In mice, the intake ofregular chow follows the same pattern, butthey ingest more highly palatable food at es-trus if it is made available (Petersen, 1976).The evidence in rats is less clear. For exam-ple, estradiol reduces the intake of palatablesucrose solutions as well as chow (Geary etal., 1995), but taste reactivity studies (see later)suggest a slightly different picture. Food in-take increases markedly in response to thephysiological demands of pregnancy and lac-tation. Both have been intensively studied,particularly in relation to their physiology(Hansen and Ferreira, 1986; Linden, 1989).

Although ad libitum feeding of rats isvery common in laboratory settings, it maynot be the ideal for longer-term experimentalstudies. Moderate food restriction reducesobesity, increases longevity, and decreases the

197

Page 215: 0195162854

198 REGULATORY SYSTEMS

incidence of neoplasms (Koolhaas, 1999) andmay be justified on welfare grounds for a widerange of studies. Standard rodent diets maycontain a higher proportion of fat and proteinthan is optimum, with additional variation informulation, even from the same supplier, de-pending on geographical area.

DAILY PATTERNS OF INTAKE

Rats are crepuscular feeders, although their in-take patterns are highly maleable. Thus, whenfood and water and freely available, lightingis on a 12 hour schedule, and disturbance isminimal, food intake is typically high aroundthe period of lights-ofT and for 2 to 3 hoursthereafter and then again before lights-on.Regular scheduling of food availability, or theprovision of palatable food at particular times,leads to rapid adaptation of diurnal patternsof intake. For example, rats provided with apalatable mash made by mixing standard pow-dered chow with water will eat up to 10 grams(dry weight) over a 40 minute period duringa time of the day when intake is usually min-imal; this represents about 50% of total dailyintake, and no food deprivation is required toproduce this response. Total daily intake isalso typically enhanced by this manipulation.In a similar way, rats can be provided with adlibitum access to chow in their living envi-ronment and the opportunity to visit an ex-tremely cold area (—15 °C), away from theircage, to forage for palatable food items at par-ticular times. They will take a substantial pro-portion of their total daily calorie intake whenthe palatable food is made available in thisway (Cabanac and Johnson, 1983). Substantialchanges in the diurnal patterning of food in-take are associated with physiological adapta-tion in hormonal rhythms within the hypo-thalamo-pituitary-adrenal axis. For example,corticosterone levels are usually maximal inthe early dark phase of the photoperiod butwill shift to anticipate feeding time when ratsare fed on a daily schedule (Gallo and Wein-

berg, 1981). It is therefore important to allowsufficient time during the habituation phase ofan experiment for such physiological changesto stabilize.

Intakes of water and food are well corre-lated over the diurnal cycle. This is likely tobe promoted by both behavioral and physio-logical mechanisms. Rats, especially whenhoused alone for detailed feeding studies, maybe relatively inactive for much of the time.Thus, all active behavior patterns are likely tobe mutually correlated over time. Rats mayalso learn to drink while feeding on dry foodto avoid the unpleasant consequences of dryfood within the stomach (Lucas et al., 1989).In addition, there are specific physiologicalmechanisms that may stimulate drinking as aconsequence of the presence of food in thestomach (Kraly, 1983).

MEAL PATTERNS

A tendency to feed at high intensity over rel-atively short periods is characteristic of manymammalian species from groups as diverse asprimates, ruminants, carnivores, and rodents.These periods of intense feeding behavior areusually termed "meals." It might be hypothe-sized that meals are simply epiphenomena re-sulting from patchy food availability in theenvironment. However, rats in a laboratorysetting with food and water continuouslyavailable nevertheless structure their intakeinto a sequence of meals.

Studies of meal patterning first requiresome method of continuously monitoring theintake of food, and preferably water intake aswell. This can be achieved in a variety of ways.One technique involves the use of an auto-matic dispenser delivering single 45 mg pel-lets to a feeding niche where their presence ismonitored until they are removed by the rat(Kissileff, 1970). Alternatively, the availabilityof cheap, but accurate, strain gauges allowsonline measures of the weight of a food hop-per. The pellet dispenser has the advantage of

Page 216: 0195162854

Chapter 18. Eating 199

greater temporal precision and the delivery ofnonvarying food items with identical handlingcharacteristics but makes it more difficult tovary the diet constituents. By contrast, directweighing tends to reduce temporal resolutionbecause of the requirement for the rat to moveaway from the food hopper but does allow foreasier variation in diet constituents. The choiceof technique depends on the particular ques-tions of interest to the experimenter.

Once a series of feeding observationsover time is available, a number of decisionshave to be made before meals can be definedwithin the record. For example, is there a rel-atively clear drop in the probability of feedingagain as the elapsed time since consumptionof the last food item increases? The log-sur-vivor technique represents an easy graphicalapproach to this problem and provides asimple, if implausible, associated null model(Clifton, 1987). A point close to the maximuminflection provides one estimate of a criterionthat can be used to separate within-meal frombetween-meal intervals (Lester and Slater,1986). However, if the modeling of these dis-tributions is of primary interest, then the ex-perimenter will wish to explore alternatives tothis approach, fitting Weibull, log-normal, ora variety of other distributions (Sibly et al.,1990; Yeates et al., 2001). For those who sim-ply wish to extract meal structure, it may besufficient to choose a criterion and then repeatthe analyses with other meal criteria that varyaround that chosen to indicate the robustnessof the conclusions (Castonguay et al., 1986).The criteria chosen in recent studies vary from2 minutes to 10 minutes. Some earlier studiesused a criterion as long as 30 minutes (Le Mag-nen and Tallon, 1966), which is likely to con-flate individual meals.

After a minimum interval has been cho-sen, a data set can be processed as series ofmeals and the intervals between them. Onefurther decision may be critical. Do meals de-fined in this way consist of two qualitativelydifferent types of feeding behavior in whichthere are "snacks" (very short meals) and true

meals? Some authors exclude meals belowsome criterion size (e.g., 0.1 gram), and thisdecision may be especially important whenthe raw data are obtained from weighing sys-tems that respond to both actual feeding andto exploratory behavior that may includeclimbing over the food container.

A number of parameters can be used todescribe particular features of the meal struc-ture. Within-meal feeding rate slows slightlyduring a meal (Clifton, 2000) and may be oneuseful index of satiety. However, it is impor-tant to note that the changes in feeding rateduring a freely initiated meal are minimal bycomparison with those seen in deprived ratsworking for food in an operant session ordrinking a palatable solution (see later). Instudies involving drug manipulations, feedingrate can be a very sensitive measure of motorimpairment. Meal size is often taken a as ameasure of within-meal satiation, whereas in-termeal intervals may reflect the between-meal increase in hunger. The ratio of meal sizeto intermeal interval is termed the satiety ra-tio and may provide a more sensitive index ofthe satiating capacity of food (Clifton, 2000).

Although stable meal patterns are foundin rats with free access to food, these patternsare also greatly affected by food availability.An extensive series of studies by George Col-lier provides some of the most striking exam-ples of this type of effect (Collier et al., 1972;Collier, 1987). Rats were required to work foraccess to food by operating a conventional op-erant lever and lived continuously in this sit-uation, rather than simply being exposed to itfor a short period each day. The results de-pended on whether the rats either had to workfor each food item or simply had to initiate ameal. In the latter case, once a single food itemhad been earned, subsequent items were freeprovided that feeding did not cease for 10 min-utes or longer. This led to a substantial changein meal patterns in which, at extreme ratios(e.g., fixed ratio [FR] 5120), the rats took a sin-gle meal each day. Collier provided a strongfunctional interpretation of these data, sug-

Page 217: 0195162854

200 REGULATORY SYSTEMS

gesting that the rats responded to food avail-ability by minimizing the work necessary forsurvival. However, such schedules may alsolock an animal into repeated cycles of fooddeprivation, involving prolonged periods ofoperant responding, substantial consumptionwhen food becomes available, and a long suc-ceeding period of nonresponding. Althoughthe effects of imposing a work requirementfor each item of food are less dramatic thanthose produced by having to work for the ini-tiation of a meal, they are clear. Meal size de-creases as the work requirement for each fooditem within the meal is increased (Clifton etal, 1984; Timberlake et al, 1988). It seemsmost likely that this effect arises from inter-ference with positive feedback processes thatserve to reinforce feeding as a meal begins.

BEHAVIOR THATANTICIPATES FEEDING

Ethologists have long made the distinction be-tween an initial flexible appetitive phase inmotivated behavior and the subsequent morestereotyped consummatory phase (Craig,1918). Casual observation of a rat that expectsfood to arrive suggests general behavioral ac-tivation coupled with response to cues thatidentify the location of food or the exact timeat which it will arrive. Such behavior can bemeasured in a variety of test situations. Forexample, Blackburn et al. (1987) describe asimple Pavlovian conditioning paradigm inwhich rats come to expect a liquid diet rewardafter presentation of a compound light /buzzerstimulus with a duration of 150 seconds.While the conditioned stimulus is present, therat makes anticipatory nose-pokes into theaperture where the diet will be delivered atthe offset of the stimulus. A low dose of thedopamine D2 receptor antagonist pimozide(0.4 mg/kg) substantially attenuated this an-ticipatory nose-poke response but had no ef-fect on intake of the liquid diet when this wasprovided ad libitum during a 20 minute test

session (Blackburn et al., 1987). In a similarvein, Gallagher et al. (1990) described a sim-ple food conditioning paradigm in which theorientation response of a rat to a light cue thatpredicts the arrival of a food pellet is meas-ured. In addition, they measured the rats' in-vestigatory responses at the location wherefood is actually delivered. Although both re-sponses develop as the rat is exposed to in-creasing numbers of conditioning trials, itturns out that the underlying neural mecha-nisms that subserve the two responses arerather different. Lesion studies suggest thatthe central nucleus of the amygdala makes animportant contribution to the development ofresponding to the light cue but not to the in-vestigatory responses at the food cup.

Runways provide another way of meas-uring appetitive responses to food. Speed ofrunning from a start to a goal box can easilybe measured and declines with repeated trialsduring a single test session. Several drugs that,at least under some circumstances, enhancefood intake also enhance running speed in theearly trials of session. They include the atyp-ical antipsychotic olanzapine, whose clinicaluse is associated with the development of obe-sity (Thornton-Jones et al., 2002). Runningspeed can also be enhanced by cues that pre-dict a particular trial is to be rewarded. Al-though dopamine antagonists reduced instru-mental responding for food, the increasedrunning speed in the presence of appropriatestimuli is unaffected by pretreatment with thedopamine antagonist haloperidol (McFarlandand Ettenberg, 1998).

Conditioned place preference is anothertechnique that is often used to study appeti-tive aspects of feeding. Although the paradigmhas been generally used to assess drug reward,it is also appropriate for natural reinforcerssuch as food (Perks and Clifton, 1997) or sex(Everitt, 1990). Typically the task involvespairing particular contextual cues with thepresence or absence of food. In test sessions,the rat is allowed to choose between the twoenvironments, and the duration of time spent

Page 218: 0195162854

Chapter 18. Eating 201

in each, or an appropriate ratio of these, pro-vides a convenient measure of preference forfood-associated cues. The task has a consider-able advantage over T-maze studies that at-tempt to demonstrate the same phenomenon,in that spontaneous alternation between trialsis not a problem. Preference measured in thisway is sensitive to devaluation of the food re-inforcer by sickness and to motivational state(Perks and Clifton, 1997). The paradigm hasalso been used to separate the role of differ-ent brain circuits in the approach to food-related contextual cues. For example, discon-nection of the basolateral amygdala and ven-tral striatum impairs the expression of a food-reinforced place preference (Everitt et al.,1991).

More conventional operant studies havealso distinguished the neurochemical systemsthat underlie the appetitive and consummatoryphases of response to food. The selective do-pamine D2 receptor antagonist raclopride, atapproximately the same doses (approximately0.5 mg/kg), strongly reduces lever pressing forfood (Nakajima and Baker, 1989) but stimulatesfeeding on such pellets when they are availablead libitum (Clifton et al., 1991). In a more re-cent set of studies using similar drug manipu-lations, Salamone and colleagues (Cousins etal., 1994) provided rats with a concurrentchoice of either working for 45 mg pellets oreating larger pieces of regular chow scatteredon the floor of the operant cage. Untreated an-imals mostly work for pellets, whereas drug-treated animals switch to eating the less pre-ferred chow on the cage floor.

FEEDING AND THE BEHAVIORTHAT FOLLOWS FEEDING

The details of food handling and drinking arecovered elsewhere in this volume (Chapters15 and 19). However, studies of taste reactiv-ity and the microstructure of ingestion of liq-uid diets have been very influential in thestudy of feeding behavior.

The first group of taste reactivity studies,recently reviewed by Berridge (2000), empha-size the detailed analysis of facial expressionsthat rats exhibit while ingesting solutions withdifferent sensory or nutritive characteristics.Although rats show such behavior patterns inmany of the paradigms described elsewherein this review, they are often difficult to ob-serve and score accurately. In a typical tastereactivity study, the test solution is infusedinto the mouth through a previously im-planted oral catheter and the behavior isrecorded from below through a clear-bottomcage. Ingestion of a palatable solution is asso-ciated with a cluster of behavior patterns, in-cluding rhythmic tongue protrusion, paw lick-ing, and lateral tongue movements. Infusionof unpalatable solutions evokes very differentbehavior patterns, including gapes, headshakes, face wipes, and chin rubs. This para-digm has been especially valuable in separat-ing hedonic and incentive components offeeding. For example, pretreatment of ratswith a dopamine antagonist, such as pi-mozide, has no effect on the proportion ofingestive and aversive responses elicited by asucrose solution (Pecina et al., 1997). Inter-estingly, taste reactivity measures also suggestthat female rats, at a point in the estrus cyclewhen estradiol levels are high, are more re-sponsive to both attractive and unattractivetaste cues (Clarke and Ossenkopp, 1998).

A second group of studies was pioneeredby Jack Davis (1998). A rat presented with anutritive or palatable solution during a 30minute test session may ingest the same vol-ume of solution in quite different ways. Thus,a highly palatable solution that is clearedslowly from the gut will be ingested rapidlyat the beginning of the session, but then in-take will quickly decline. A less-palatablesolution that clears more quickly will be in-gested more uniformly throughout the ses-sion. Davis discusses the use of negative ex-ponential fitting techniques to characterizethese effects. The method has been widely tostudy both enhancement of and reduction in

Page 219: 0195162854

202 REGULATORY SYSTEMS

intake. For example, the selective serotoninreuptake inhibitor fluoxetine produces reduc-tions in intake that are similar to those ob-served after slowed gut emptying rather thanreduced palatability, a finding that is consis-tent with the hypothesis that the drug en-hances satiation processes (Lee and Clifton,1992). Davis and Smith (1992) described a sec-ond method for analyzing this type of drink-ing record that relies on the way in which therat performs short bursts of licking while in-gesting a palatable solution. For example, anincrease in the concentration of a sucrose so-lution increased the size of bouts of licking,whereas sham feeding increased the numberof bouts rather than the size of bouts. In fact,Davis and Smith distinguished between twolevels of organization, which they termedbursts and dusters. Subsequent authors haveoften used a single level of description (Spec-tor et al, 1998).

The duration and intensity of feeding be-havior may be strongly influenced by extrin-sic as well as interoceptive cues. For example,presentation of cues previously associatedwith food intake may enhance feeding by ap-parently sated rats (Weingarten, 1984). Thisprocedure has also been used to investigatethe role of amygdala nuclei in the processingof conditioned stimuli that may facilitate feed-ing (Petrovich et al., 2002). Hungry rats wereinitially exposed to a situation in which theylearned that one light conditioned stimuluspredicted food, whereas the presentation of asecond, different conditioned stimulus wasnot correlated with food presentations. Sub-sequently, the same rats were given con-sumption tests while sated. The conditionedstimulus previously associated with food con-sumption strongly enhanced eating in this sit-uation. However, the effect was not seen inrats in which the basolateral amygdala and lat-eral hypothalamus had been disconnected.

Cues associated with specific features ofthe diet may also influence the quantity offood consumed in a meal or test session. Thephenomenon of conditioned satiety provides

a clear example (Booth, 1972). In this study,rats were given repeated daily training ses-sions in which a low-calorie diet was associ-ated with one specific flavor, alternating withsessions in which a high-calorie diet was as-sociated with a different flavor. In one condi-tion of the subsequent test sessions, the ratswere tested on an isocaloric diet of interme-diate value to those used in training. The dietwas flavored with either the flavor associatedwith low- or high-caloric content during train-ing. The rats consumed more of the diet thathad been flavored with the low-calorie flavor,despite their current identical calorific value,illustrating that consumption had come undercontrol of the conditioned flavor cue.

Casual observation of rats that have justbeen fed suggests that a relatively stereotypedsequence behavior follows the cessation ofeating. A rat may first appear quite activemoving around the cage. It then settles downto a prolonged bout of grooming (see Chap-ter 13), beginning with the whiskers and faceregion and moving on down the body. Withina few minutes, the rat is likely to be quiet andinactive in one corner of the cage. These reg-ularities in behavior were noted in early stud-ies (Bolles, 1960), and it was then suggestedthat this sequence of behavior was character-istic of satiety in the rat and might be used todistinguish between experimental treatmentsthat enhanced satiety from those that reducedfeeding for other reasons (Antin et al., 1975).Since then, the so-called behavioral satiety se-quence has been widely used to characterizechanges in feeding behavior after pharmaco-logical or neural manipulations. For example,drugs that primarily stimulate 5-hydroxy-tryptamine (serotonin) (5-HT)2c receptors ap-pear to advance the satiety sequence in the ratbut preserve its overall form. The evidencesupports this contention for some drugs thathave nonspecific effects on serotonin systems,including the 5-HT releaser fenfluramine(Halford et al., 1998) and the selective sero-tonin reuptake inhibitor fluoxetine (Clifton etal., 1989), as well as more selective 5-HT2c re-

Page 220: 0195162854

Chapter 18. Eating 203

ceptor agonists such as mCPP (Halford et al.,1998). By contrast, drugs such as DOI, whichhave substantially greater action at 5-HT2A re-ceptors in addition to their effects at 5-HT2creceptors, produce increases in locomotor ac-tivity and a disruption of the normal satietysequence (Simansky and Vaidya, 1990).

There are several comprehensive re-views of studies and methodology in this area(Clifton, 1994; Halford et al., 1998). The latterauthors correctly place an emphasis on the ad-vantages that may be gained from a completevideo transcription of behavior from individ-ual rats. However, depending on the particu-lar purpose of the study, investigators maywish to consider using a time-sampling pro-cedure that eliminates video records and al-lows a number of animals to be scored si-multaneously (Clifton et al., 1989). In manycircumstances, the decreased time requiredfor such sampling permits an increase in sta-tistical power by adding extra subjects, drugdoses, or nondrug control conditions to thestudy.

DIET CHOICE

Although it is often assumed that rats willself-select a nutritionally appropriate diet incafeteria-style experiments, the experimentalevidence is less than compelling (Galef, 1991)and is not reviewed here. There has been con-siderable interest in the way in which rats se-lect different proportions of protein, fat, andcarbohydrate, the three macronutrient con-stituents of the diet. A variety of paradigmshave been used to study such dietary selec-tion. In one version, the rat is allowed tochoose from three almost pure macronutrientsources. Each is supplemented with an ap-propriate mix of minerals and vitamins so thata range of diet choices will not impair generalhealth. In one widely cited group of studies ofthis type, Liebowitz and her colleagues re-ported that treatment with drugs, such as fen-fluramine, that enhance serotonergic neuro-

transmission reduced overall food intake butspared consumption of protein relative to thatof carbohydrate (Shor-Posner et al., 1986;Weiss et al., 1990). However, there are sub-stantial methodological issues in conductingsuch studies. If fat is provided as lard or a veg-etable fat, protein as casein powder, and car-bohydrate as dextrin or starch, then macronu-trient content is confounded with many otherfactors, including taste, smell, texture, andwater content. As a consequence, preferencemay vary substantially across the different di-ets. In addition, individual rats may show sta-ble but idiosyncratic differences in the pro-portions of each diet that they consume.Other studies, which appear superficially sim-ilar to these, have generated very differentpatterns of results. For example, it has beenreported that fenfluramine selectively sup-presses fat consumption and spares carbohy-drate consumption (Smith et al., 1998). Asimilar result has been obtained using the se-lective serotonin reuptake inhibitor fluoxetine(Heisler et al., 1999). It seems likely that theseand similar inconsistencies arise from thecomplex nature of the diet selection paradigmin which a number of factors vary withmacronutrient content.

In an attempt to disentangle these factors,a number of alternative diet selection tech-niques have been used. For example, rats maybe provided with a carbohydrate supplementof polycose in addition to a standard chow diet(Lawton and Blundell, 1992). Under theseconditions, the effects of fenfluramine variedmarkedly with the water content of the twodiet components. Fenfluramine suppressedthe consumption of dry polycose relative tochow presented as a wet mash, yet sparedconsumption of either sucrose or polycosesolutions relative to dry chow. Data such asthese reinforce the critical point made in a re-cent review of studies in this area (Thibaultand Booth, 1999). Results for any single dietselection paradigm do not allow conclusionsthat can be discussed in terms of general ef-fects on macronutrient selection. Instead, they

Page 221: 0195162854

h

must be interpreted solely as effects on theparticular test diets that have been chosen forthe study.

NEOPHOBIA AND DIET VARIETY

As an opportunistic omnivore, it might be ex-pected that rats would sample potentiallyvaluable, but novel, food items. Although thisdoes occur, rats also show a strong neopho-bia to novel foods, even to those that areclosely related to familiar foods (Barnett,1963). Thus a first presentation of "palatable"mash made by soaking standard laboratorychow with water will be sampled but may alsoshow signs of rejection (pushed to the back ofthe cage, covered with bedding material). Af-ter several days, intake increases rapidly, es-pecially in situations that allow social facilita-tion between individuals (Galef et al., 1997).Once the items of a diet have become famil-iar, provision of several foods that vary inmacronutrient content or sensory character-istics may enhance intake and, with chronicexposure, increase body weight. Such effectscan be produced by presenting rats with a se-quence of "courses" within a meal that varyin a single sensory characteristic (Treit et al.,1983) or even by simple alternation of two dif-ferently flavored courses (Clifton et al., 1987).

IS THERE A ROLE FORSIMPLE INTAKE TESTS?

Simple intake tests over short periods of timeremain the most common form of data pre-sented in studies that explore feeding behav-ior in the rat. When used with care, they pro-vide valuable preliminary data. However, it isalways important to consider possible floorand ceiling effects when designing such an ex-periment. Tests using regular chow in non-deprived rats are not likely to reveal suppres-sion of intake. Moderate food deprivation, orprovision of more palatable food, is likely to

be more successful. Equally, an increase infood intake in nondeprived animals may beharder to demonstrate when either a palatablefood or a fresh supply of regular chow is pro-vided as the test meal. Both procedures arelikely to take food intake to ceiling levels.Within-subject designs using scheduled foodavailability may promote the development oftolerance with repeated drug administrationand reduce statistical power even when ap-propriate balancing is used. More generally,after completing initial studies using simpleintake measures, consider the experimentalhypotheses that are under consideration. Dothey predict differential effects on the appeti-tive and consummately phases of feedingbehavior? Do they suggest that basic intakemight remain similar but that the behavioraltrajectory that leads to similar intake might bequite different? If the answer to these or sim-ilar questions is "yes," then consider adoptingone or more of the paradigms described herefor further investigation.

CONCLUSION

Feeding behavior in the rat has a complex-ity that should be mirrored in the experi-mental paradigms that are used for its ob-servation and measurement. A combinationof approaches, combining the detailed ob-servational techniques of ethology with tra-ditional behavioral analysis derived fromexperimental psychology, provides a goodstarting point. Appropriate choice of test sit-uation, in conjunction with techniques thatmeasure or manipulate aspects of the rat'sphysiology or neural function, has the ca-pacity to reveal a good deal about the rela-tionships between brain, physiology, and be-havior in this species.

ACKNOWLEDGMENT

The author is grateful to Dr. Liz Somerville for her insightfulcomments on an earlier draft of this chapter.

REGULATORY SYSTEMS

Page 222: 0195162854

Tt

REFERENCES

Antin J, Gibbs J, Holt J, Young RC, Smith GP (1975)Cholecystokinin elicits the complete behavioral se-quence of satiety in rats. Journal of Comparative andPhysiological Psychology 89:784-790.

Barnett S (1963) A study in behaviour. London:Methuen.

Berridge KC (2000) Measuring hedonic impact in ani-mals and infants: microstructure of affective tastereactivity patterns. Neuroscience and BiobehavioralReviews 24:173-198.

Blackburn JR, Phillips AG, Fibiger HC (1987) Dopamineand preparatory behavior: I. Effects ofpimozide. Be-havioral Neuroscience 101:352-360.

Bolles R (1960) Grooming behaviour in the rat. Journalof Comparative and Physiological Psychology53:306-310.

Booth DA (1972) Conditioned satiety in the rat. Journalof Comparative and Physiological Psychology 81:457-471.

Cabanac M and Johnson KG (1983) Analysis of a conflictbetween palatability and cold exposure in rats. Phys-iology and Behavior 31:249-253.

Castonguay TW, Kaiser LL, Stern JS (1986) Meal pat-tern analysis: Artifacts, assumptions and implica-tions. Brain Research Bulletin 17:439^143.

Clarke SN and Ossenkopp KP (1998) Taste reactivity re-sponses in rats: Influence of sex and the estrouscycle. American Journal of Physiology 274:R718-R724.

Clifton P (1987) Analysis of feeding and drinking pat-terns. In: Feeding and drinking (Toates F and Row-land N, eds.). Oxford: Elsevier.

Clifton PG (1994) The neuropharmacology of mealpatterning. In: Ethology and psychopharmacology(Cooper SJ and Hendrie CA, eds.). Chichester: Wiley.

Clifton PG (2000) Meal patterning in rodents: psy-chopharmacological and neuroanatomical studies.Neuroscience and Biobehavioral Reviews 24:213-222.

Clifton PG, Popplewell DA, Burton MJ (1984) Feedingrate and meal patterns in the laboratory rat. Physi-ology and Behavior 32:369-374.

Clifton PG, Burton MJ, Sharp C (1987) Rapid loss of stim-ulus-specific satiety after consumption of a secondfood. Appetite 9:149-156.

Clifton PG, Barnfield AM, Philcox L (1989) A behav-ioural profile of fluoxetine-induced anorexia. Psy-chopharmacology (Berlin) 97:89-95.

Clifton PG, Rusk IN, Cooper SJ (1991) Effects ofdopamine Dl and dopamine D2 antagonists on thefree feeding and drinking patterns of rats. Behav-ioral Neuroscience 105:272-281.

Collier G (1987) Operant methodologies for studyingfeeding and drinking. In: Feeding and drinking(Toates F and Rowland N, eds.). Oxford: Elsevier.

Collier G, Hirsch E, Hamlin PH (1972) The ecologicaldeterminants of reinforcement in the rat. Physiol-ogy and Behavior 9:705-716.

Cousins MS, Wei W, Salamone JD (1994) Pharmaco-logical characterization of performance on a con-current lever pressing/feeding choice procedure:Effects of dopamine antagonist, cholinomimetic,sedative and stimulant drugs. Psychopharmacology(Berlin) 116:529-537.

Craig W (1918) Appetites and aversions as constituents ofinstincts. Biological Bulletin of Woods Hole 34:91-107.

Davis J (1998) A model for the control of ingestion—20Years on. Progress in Psychobiology and Physio-logical Psychology 17:127-173.

Davis JD and Smith GP (1992) Analysis of the mi-crostructure of the rhythmic tongue movements ofrats ingesting maltose and sucrose solutions. Be-havioral Neuroscience 106:217-228.

Everitt BJ (1990) Sexual motivation: A neural and behav-ioural analysis of the mechanisms underlying appet-itive and copulatory responses of male rats. Neuro-science and Biobehavioral Reviews 14:217-232.

Everitt BJ, Morris KA, O'Brien A, Robbins TW (1991)The basolateral amygdala-ventral striatal systemand conditioned place preference: Further evidenceof limbic-striatal interactions underlying reward-related processes. Neuroscience 42:1-18.

Galef BG Jr (1991) A contrarian view of the wisdom ofthe body as it relates to dietary self-selection. Psy-chological Reviews 98:218-223.

Galef BG Jr, Whiskin EE, Bielavska E (1997) Interactionwith demonstrator rats changes observer rats' af-fective responses to flavors. Journal of ComparativePsychology 111:393-398.

Gallagher M, Graham PW, Holland PC (1990) The amyg-dala central nucleus and appetitive Pavlovian condi-tioning: Lesions impair one class of conditioned be-havior. Journal of Neuroscience 10:1906-1911.

Gallo PV and Weinberg J (1981) Corticosterone rhyth-micity in the rat: interactive effects of dietary re-striction and schedule of feeding. Journal of Nutri-tion 111:208-218.

Geary N, Trace D, Smith GP (1995) Estradiol interactswith gastric or postgastric food stimuli to decreasesucrose ingestion in ovariectomized rats. Physiol-ogy and Behavior 57:155-158.

Halford JC, Wanninayake SC, Blundell JE (1998) Be-havioral satiety sequence (BSS) for the diagnosis ofdrug action on food intake. Pharmacology, Bio-chemistry, and Behavior 61:159-168.

Hansen S and Ferreira A (1986) Food intake, aggression,and fear behavior in the mother rat: Control by neu-

205

Page 223: 0195162854

h

ral systems concerned with milk ejection and ma-ternal behavior. Behavioral Neuroscience 100:64-70.

Heisler LK, Kanarek RB, Homoleski B (1999) Reductionof fat and protein intakes but not carbohydrate in-take following acute and chronic fluoxetine in fe-male rats. Pharmacology, Biochemistry, and Be-havior 63:377-385.

KissileffHR (1970) Free feeding in normal and "recov-ered lateral" rats monitored by a pellet-detectingeatometer. Physiology and Behavior 5:163-173.

Koolhaas J (1999) The laboratory rat. In: The care andmanagement of laboratory animals (Poole T, ed.),pp. 313-330. London: Blackwell.

Kraly FS (1983) Histamine plays a part in induction ofdrinking by food intake. Nature 302:65-66.

Lawton CL and Blundell JE (1992) The effect of d-fenfluramine on intake of carbohydrate supple-ments is influenced by the hydration of the testdiets. Behavioural Pharmacology 3:517-523.

Le Magnen J and Tallon S (1966) La periodicite sponta-nee de la prise d'aliments ad-libitum du rat blanc.Journal of Physiology (Paris) 58:323-349.

Lee MD and Clifton PG (1992) Partial reversal of fluox-etine anorexia by the 5-HT antagonist metergoline.Psychopharmacology (Berlin) 107:359-364.

Lester NP and Slater PJB (1986) Minimising errors in split-ting behaviour into bouts. Behaviour 79:153-161.

Leung PM and Horwitz BA (1976) Free-feeding patterns ofrats in response to changes in environmental temper-ature. American Journal of Physiology 231:1220-1224.

Linden A (1989) Role of cholecystokinin in feeding andlactation. Acta Physiologica Scandinavica Supple-mentum 585:i-vii, 1-49.

Lucas GA, Timberlake W, Gawley DJ (1989) Learningand meal-associated drinking: meal-related deficitsproduce adjustments in postprandial drinking. Phys-iology and Behavior 46:361-367.

McFarland K and Ettenberg A (1998) Haloperidol doesnot affect motivational processes in an operant run-way model of food-seeking behavior. BehavioralNeuroscience 112:630-635.

Nakajima S and Baker JD (1989) Effects of D2 dopaminereceptor blockade with raclopride on intracranialself-stimulation and food-reinforced operant behav-iour. Psychopharmacology (Berlin) 98:330-333.

Pecina S, Berridge KG, Parker LA (1997) Pimozide doesnot shift palatability: Separation of anhedonia fromsensorimotor suppression by taste reactivity. Phar-macology, Biochemistry, and Behavior 58:801-811.

Perks SM and Clifton PG (1997) Reinforcer revaluationand conditioned place preference. Physiology andBehavior 61:1-5.

Petersen S (1976) The temporal pattern of feeding overthe oestrous cycle of the mouse. Animal Behaviour24:939-955.

Petrovich GD, Setlow B, Holland PC, Gallagher M(2002) Amygdalo-hypothalamic circuit allows learnedcues to override satiety and promote eating. Jour-nal of Neuroscience 22:8748-8753.

Shor-Posner G, Grinker JA, Marinescu C, Brown O, Lei-bowitz SF (1986) Hypothalamic serotonin in thecontrol of meal patterns and macronutrient selec-tion. Brain Research Bulletin 17:663-671.

Sibly R, Nott HMR, Fletcher DJ (1990) Splitting behav-iour into bouts. Animal Behaviour 39:63-69.

Simansky KJ and Vaidya AH (1990) Behavioral mecha-nisms for the anorectic action of the serotonin (5-HT) uptake inhibitor sertraline in rats: Comparisonwith directly acting 5-HT agonists. Brain ResearchBulletin 25:953-960.

Smith BK, York DA, Bray GA (1998) Chronic d-fenflu-ramine treatment reduces fat intake independentof macronutrient preference. Pharmacology, Bio-chemistry, and Behavior 60:105-114.

Spector AC, Klumpp PA, Kaplan JM (1998) Analytical is-sues in the evaluation of food deprivation and su-crose concentration effects on the microstructure oflicking behavior in the rat. Behavioral Neuroscience112:678-694.

Thibault L and Booth DA (1999) Macronutrient-specific di-etary selection in rodents and its neural bases. Neuro-science and Biobehavioral Reviews 23:457-528.

Thiels E, Alberts JR, Cramer CP (1990) Weaning in rats:II. Pup behavior patterns. Developmental Psy-chobiology 23:495-510.

Thornton-Jones Z, Neill JC, Reynolds GP (2002) Theatypical antipsychotic olanzapine enhances inges-tive behaviour in the rat: A preliminary study. Jour-nal of Psychopharmacology (Oxford, England)16:35-37.

Timberlake W, Gawley DJ, Lucas GA (1988) Time hori-zons in rats: The effect of operant control of accessto future food. Journal of the Experimental Analy-sis of Behavior 50:405-417.

Treit D, Spetch ML, Deutsch JA (1983) Variety in theflavor of food enhances eating in the rat: A con-trolled demonstration. Physiology and Behavior30:207-211.

Weingarten HP (1984) Meal initiation controlled bylearned cues: Basic behavioral properties. Appetite5:147-158.

Weiss GF, Rogacki N, Fueg A, Buchen D, Leibowitz SF(1990) Impact of hypothalamic d-norfenfluramineand peripheral d-fenfluramine injection on mac-ronutrient intake in the rat. Brain Research Bulletin25:849-859.

Yeates MP, Tolkamp BJ, Allcroft DJ, Kyriazakis I (2001)The use of mixed distribution models to determinebout criteria for analysis of animal behaviour. Jour-nal of Theoretical Biology 213:413-425.

RE GU LA TO RY SYA TEMS

Page 224: 0195162854

Drinking

NEIL E. ROWLAND

19

Food and fluid intakes often are considered asexamples of behaviors with a homeostaticfoundation. This chapter is organized arounda construct ofhydromineral homeostasis and dis-cusses only water and mineral consumption.Other liquids that are often used in rat stud-ies include liquid diets, sugars, and alcoholicdrinks; these are not mentioned explicitly, butthe same general principles and proceduresmay be applied to them. Water, sometimeswith dissolved trace minerals, is the only nat-urally occurring fluid source. Sodium appetiteis also much studied in the laboratory, as aninnate and specific appetite. Sodium is inex-tricably related to water in the body and isthe primary mineral in "hydromineral." Mostlaboratory studies use solutions of sodiumsalts, as is discussed, but it is likely that natu-ral sources of sodium are not in fluid form.

Mammals have no mechanism for fluidstorage, so states of physiological fluid needmust engage powerful behavioral mecha-nisms (e.g., motivation, thirst) that drive theanimal to fluid. Most laboratory studies de-liberately minimize this motivational compo-nent and instead offer the fluids without ef-fort in a safe environment; in this case thedrinking is more reflexive than motivated.There is a relative lack of fluid studies in morenaturalistic or effortful environments (Mar-wine and Collier, 1979; Quartermain et al.,1967). Under noneffortful conditions, need-related drinking is termed primary or homeosta-tic and is in contrast to secondary or nonhomeo-static drinking that occurs without identifiedneed (Fitzsimons, 1979). The latter could in

principle fall within a category of predictivehomeostasis (Rowland, 1990), but there is nodirect behavioral evidence that rats can pre-dict future fluid needs (Strieker et al., 2003).

PHYSIOLOGY OF FLUID BALANCE

This section provides a brief overview of themain body fluid compartments, of how fluidsare gained and lost by those compartments,and of relevant neural and hormonal signals.Design of experiments in fluid intake requiresa working knowledge of these homeostaticprinciples. Thirst is caused by multiple factors,so the choice of specific stimulus to use in anexperiment is of great theoretical importance.

INTRACELLULAR ANDEXTRACELLULAR FLUID

COMPARTMENTS

Fluids compose about 69% of a rat's total bodyweight. Approximately two-thirds of thesebody fluids are contained inside cells (intracel-lular), and one-third, outside cells (extracellu-lar). Extracellular fluid is distributed betweenvascular (blood plasma) and interstitial (tissue)subcompartments in about a 1:3 ratio (Fig.19-1). Solutes dissolved in these fluids giverise to osmotic pressure; net flow of wateracross cell walls is driven by associated differ-ences in osmotic pressure. Under conditionsof perfect balance (euhydration), intracellularand extracellular compartments have the

207

Page 225: 0195162854

208

Figure 19-1. Schematic representation of the relative sizesof intracellular and extracellular body fluid compartmentsin states of hydromineral balance (euhydration) and threetypes of dehydration. For simplicity, vascular and intersti-tial extracellular fluids (^1:3 ratio) are not distinguished.Arrows indicate the initial net movement or loss of fluid.Deprivation of fluid produces compound dehydration,which involves loss of extracellular fluid and movement ofcellular water to the extracellular space.

same osmotic pressure (about 290 millios-moles per liter, or isotonic) and there is no netwater movement between compartments.However, the solutes producing osmolar load(osmolytes) differ substantially between theintra- and extracellular compartments, asshown in Table 19-1. For this discussion, themain extracellular solute is sodium chloride.An isotonic solution of NaCl is thus approxi-mately 0.15 molar (M).

Table 19-1. Body Fluid Compartments

and Constituents

Property

Volume (% body wt)

Na+ (mEq/L)

K+ (mEq/L)

Ca2+ (mEq/L)

Cl~ (mEq/L)

HCO3~ (mEq/L)

Phosphates (Pj, mEq/L)

Intracellular

4612

1500.001

512

100

Extracellular

23*145

45

105252

*BCF is =75% interstitial fluid and =25% plasma.

REGULATORY SYSTEMS

INTRACELLULAR DEHYDRATION THIRST

Intracellular dehydration occurs when theconcentration of extracellular solute is in-creased above isotonic (a hypertonic condi-tion) and water is then pulled from inside cellsuntil the osmotic pressure is again equalizedon the two sides of the cell wall (Fig. 19-1).This causes physical shrinkage of cells; cellscalled osmoreceptors have stretch receptors thattransduce stretch into biological signals. Bothperipheral (e.g., gut, liver) and central (e.g.,forebrain) osmoreceptors play roles in hy-dromineral balance in rats. The principalway to produce intracellular dehydration isthrough the administration of hypertonicNaCl, because the extra sodium ions arelargely trapped outside the cells. The admin-istration of impermeable hypertonic solu-tions, like NaCl, produces water intake in pro-portion to the cell shrinkage; comparablehypertonic solutions of permeable solutes(e.g., glucose, urea) do not cause drinking. Sig-nals from cell shrinkage are integrated in thebrain to produce the state of thirst that in turnmotivates water-seeking behavior. The termosmotic thirst is commonly used, but intracel-lular dehydration thirst is accurate.

EXTRACELLULAR DEHYDRATION THIRST

Extracellular dehydration occurs when iso-tonic extracellular fluid is lost without changein osmotic pressure (hypovolemia); there is nonet fluid movement across cell membranes.The vascular and interstitial (tissue) extracel-lular compartments are in rapid exchangeequilibrium; hypovolemia thus reduces bloodvolume. Serious losses compromise the deliv-ery of adequate blood to tissues and can rap-idly become life threatening. Low-pressure(venous) vessels of the circulatory systemhave elastic walls that allow for these vesselsto change diameter in response to increasedor decreased blood volume. Stretch receptorsor mechanoreceptors in the walls of these ves-sels transduce this volume status into a neu-

Page 226: 0195162854

Chapter 19. Drinking

Figure 19-2. Major components of renin-angiotensin sys-tems. In the circulation, renin from the kidney is the rate-limiting step in synthesis of the decapeptide angiotensin I,which is very rapidly cleaved to the principal biologicallyactive form, angiotensin II (an octapeptide). AngiotensinI-converting enzyme (ACE) inhibitors (e.g., captopril) slowthis cleavage. Angiotensin II activates specific receptors inmany locations, including in select brain regions such as thesubfbrnical organ.

ronal signal that generates local (reflexive) andcentral (e.g., thirst) responses. Reduced bloodpressure also causes release of renin from thekidney into the circulation (Fig. 19-2); reninthen catalyzes the synthesis of angiotensin II(Ang II), a peptide whose circulating concen-tration is related to hypovolemia (Fitzsimons,1998). Rats exhibit extracellular dehydrationthirst and drinking (Strieker 1968); this is alsoknown as volumetric thirst.

COMPOUND DEHYDRATION

Many naturally occurring situations that causephysiological dehydration are not purely intra-cellular or purely extracellular but are insteada mixture (see Fig. 19-1). Experimentally im-posed water deprivation, whether acute or ona daily schedule, is such a compound stimulus.The principal stimulus of thirst during depri-vation of water is the amount and type of foodconsumed during that time. Rats that are de-prived of both food and water have minimalfluid needs. The usual laboratory food is com-mercial rat chow, which has a relatively high(approximately 0.5%) content of NaCl. When

209

food is consumed, there are transient changesas fluid is secreted into the gastrointestinaltract, intermediate changes as the solutes areabsorbed and cause intracellular dehydration,and late changes as the NaCl and other wasteproducts from the food are excreted in urinecausing hypovolemia. Additionally, as durationof deprivation increases, physiological anorexiaoccurs (Watts, 2000), and so the further intakeof solutes is slowed. Under many conditions,intracellular and extracellular signals combineto produce an integrated thirst signal (see Fig.19-3), observations that form the basis of adual-depletion model of thirst (Rowland, 2002).

HORMONAL SIGNALS

Both intracellular and extracellular dehydra-tion stimulate the release of vasopressin fromnerve terminals in the posterior pituitarygland. These are the terminals of neurosecre-tory magnocellular neurons in the supraopticand paraventricular nuclei of the hypothala-mus. These cells have osmoreceptor proper-ties and have afferents from peripheral os-moreceptor and baroreceptor elements. Therelevant receptors for circulating vasopressinare of the VI subtype in the kidney, and theiractivation causes the retention of water. Thisis a primary reason that hypovolemic rats ex-crete little or no urine (anuria). However, inthe case of osmotic sodium loads, this is coun-teracted by atrial natriuretic factor (secretedfrom the atrium of the heart), which causesthe excess NaCl load to be excreted in urine(natriuresis) and coincidental loss of some wa-ter because there is an upper limit on the con-centrating capacity of rat kidney. Thus, afterthe injection of NaCl, osmolality first in-creases and then, even if water intake is notallowed, decreases as natriuresis occurs. Thisloss of fluid leaves a hypovolemic condition,so if a delay is imposed between injection ofNaCl and access to water, the dehydration iscomplex. In fact, the amount of drinking thatoccurs is less than theoretically needed to di-lute the salt load to isotonicity, except if uri-

Page 227: 0195162854

210 REGULATORY SYSTEMS

Additivity of osmotic and volumetric stimuli

Figure 19-3. Schematic of fluid intake as a function of in-creasing plasma osmolality and decreasing blood volume.The expected intake is approximately the algebraic sum ofthe intake produced by each component alone. The shadedarea indicates the region in which sodium appetite isobserved.

(packed cell volume), and plasma can be re-moved for measurement of protein concen-tration (a hand refractometer is simple to use),sodium concentration (flame photometer orion electrode), and/or osmolality (freezingpoint depression). Each of these requires onlysmall volumes of plasma. Plasma sodium con-centration and osmolality are indices of os-motic imbalance. Fractional increases in pro-tein and hematocit are approximations ofvolume depletion. Plasma hormones such asaldosterone and vasopressin can be measuredusing radioimmunoassay, as can renin activ-ity, which, because renin is the rate-limitingstep in Ang II synthesis, is correlated with AngII concentration. Urinary volume and sodiumand potassium concentrations in urine areuseful measures; for these, metabolic cages orstands are needed that allow separation ofurine from feces and/or spilled food.

nary excretion is prevented (e.g., by nephrec-tomy, surgical removal of the kidneys).

The peptide hormone Ang II also is madefrom renin that is released during hypov-olemia (Fig. 19-2). Ang II has many effects, in-cluding increasing blood pressure through di-rect actions on receptors in blood vessels, butalso gives rise to a central signal via Ang IItype 1 (ATI) receptors in the subfornical or-gan, a brain circumventricular organ that isaccessible to circulating Ang II. Although AngII alone can stimulate thirst, it also causes re-lease of the sodium-retaining hormone aldos-terone from the adrenal cortex, which, to-gether with Ang II, produces sodium appetite.

DOCUMENTATION OFPHYSIOLOGICAL EFFECT

Simple blood assays are often advisable to doc-ument the efficacy of treatments. Small vol-umes can be taken from a tail nick (a localanesthetic agent may be necessary) into a cap-illary tube. These samples may then be cen-trifuged for measurement of hematocrit ratio

SPECIFIC PROCEDURESAND STIMULI TO INDUCEWATER INTAKE IN RATS

INTRACELLULAR DEHYDRATION

Administration of hypertonic NaCl reliablycauses drinking, often within a few minutes.The amount of water that would need to beconsumed to dilute the salt load to isotonicityis equal to the solute load (in milliosmoles) di-vided by the initial plasma osmolality (in mil-liosmoles per liter). Thus, an injection of1 ml of 1 mol/L NaCl would require an intakeof 5 to 6 ml of water to dilute to isotonicity.However, as noted earlier, observed intake istypically less than 50% of this amount becauseof concurrent sodium excretion. Intake is re-lated to the dose of NaCl, and the threshold risein plasma osmolality for drinking to occur is ap-proximately 2%. To study thresholds, the bestmethod for the administration of hypertonicNaCl is through an indwelling catheter in thegeneral circulation, such as the jugular vein(Fitzsimons, 1963). Although this method re-

Page 228: 0195162854

Chapter 19. Drinking 211

quires surgery, it allows remote and painless in-fusions into freely moving rats. A catheter mayalso be placed into the hepatic portal vein(which drains the intestinal field into the liverand so is the natural route of entry of ingestedsolutes), and this method has been used tostudy the contribution of hepatic osmorecep-tors to thirst. An elegant dual-catheter method,in which NaCl is infused into the portal vein atthe same time that water is infused into thegeneral circulation, allows the experimenter toeffectively restrict the stimulus to the liver(Morita et al., 1997). Infusion procedures usu-ally are performed in short (e.g., 1 to 2 hourslong) sessions.

More conveniently, acute intraperitonealor subcutaneous injections of hypertonicNaCl induce reliable drinking in rats, althoughsuch injections appear to be temporarilypainful. Two procedures can be done to min-imize the potential distress, which can inter-fere with the expression of drinking behaviors.First, the rats should be handled and accus-tomed to injection procedures. Second, unlessit will interfere with the experimental goal, asmall amount of local anesthetic (e.g., bupi-vacaine, lidocaine) may be added to the in-jection. Under ideal conditions, rats start todrink within 10 to 20 minutes of injection anddrinking is complete within 60 to 90 minutes.

A chronic version, producing a sustainedosmotic load, can be simply realized by addingNaCl to the food: for example, adding 3%NaCl to powdered food produces an approx-imately 50% increase in daily water intake with-out significant anorexia. Higher concentrationsof NaCl are tolerated but may be associatedwith reduced food intake. This mode of ad-ministering the stimulus will preferentiallystimulate visceral osmoreceptors (Strieker etal., 2003; see also "Meal-Associated Drinking").

EXTRACELLULAR DEHYDRATIONDRINKING

The most direct way of reducing blood vol-ume is hemorrhage; indeed, wounded people

with high blood loss often experience intensethirst. In the laboratory, this should be facili-tated using an indwelling vascular catheter.However, this method removes critical bloodconstituents, leaving a weakened animal, so itis rarely used to study drinking.

The method of choice to produce extra-cellular dehydration is injection of the colloidpolyethylene glycol (PEG). This sequestersisotonic filtrate of plasma at the injection site,visible as an edema, for several hours. PEG isbest if administered subcutaneously, in theloose skin of the scapular region, as a solutionin water or isotonic saline. High formulaweight PEG (>20,000) should be used, typi-cally in 20% or 30% solution (weight/volume)and at a dose of 1% to 2% body weight(Strieker, 1968). Solutions should be warmedto body temperature to dissolve and for in-jection. These solutions are very viscous, so alarge-diameter needle is needed for injection.Unlike hypertonic saline, injection of PEG isnot painful, but to obtain the best edema thebolus should be palpated gently to spreadfrom the injection site. Brief gas anesthesiamay be used for this, but it is not necessary inwell-handled rats.

The edema, and consequent hypov-olemia, takes 1 to 2 hours to develop fully, andthe onset of thirst is correspondingly slow. Thehypovolemia is accompanied by anuria, so anywater consumed dilutes the extracellular fluid(hyponatremia), which is an inhibitory signalfor drinking (Strieker, 1969). Thus, water intakein this model is self-limiting because of dilu-tional hyponatremia. Typically, water intakestarts 2 to 3 hours after injection; salt intake isdelayed. The best control for this procedure isa sham injection rather than saline (which con-stitutes a volume expansion).

Hypovolemia also may be produced bynatriuretic agents such as furosemide. Func-tionally, these cause the kidney to lose a rela-tively large amount of near-isotonic urine, andthe attendant loss of sodium causes hypo-volemia. Restoration of extracellular volumerequires the intake of both water and NaCl,

Page 229: 0195162854

212 REGULATORY SYSTEMS

and so the relationship of water to salt intakeis of theoretical importance. Although diuret-ics can cause thirst, they are more often usedto stimulate sodium appetite (see "Appetite").

As discussed earlier, Ang II is a dipsogen(Fitzsimons, 1998). Because it has a short bio-logical half-life, the best route by which to ad-minister Ang II is intravenous infusion. Thethreshold dose for drinking under normallaboratory conditions is approximately 100ng/kg body weight per minute and a drink-ing latency of about 10 minutes. Procedurallysimpler, but not useful for questions aboutthresholds, subcutaneous bolus administra-tion at doses of greater than 50 /Ag/kg alsocauses a robust drinking response. Other pu-tative dipsogenic substances can be screenedusing these procedures.

Ang II is also dipsogenic when adminis-tered either acutely or chronically into thebrain. For such injections, a cannula must besurgically implanted into either into the cere-bral ventricles or specific brain regions. Drink-ing often occurs within a few seconds ofinjection.

WATER DEPRIVATION

As noted earlier, the primary cause of physi-ological dehydration during water deprivationis the concurrent intake of food. Thus, forstudies involving water deprivation, it is ad-visable to record concurrent food intake be-cause differences (say, between experimentalgroups) could produce a change in drinkingthat is only secondary to changes in food in-take. Some studies may call for more rigorousmeasurement of intakes and urinary excretionusing metabolic cages. Water deprivation inexcess of 24 hours is not normally approvedby Institutional Animal Care and Use Com-mittees (lACUCs).

MEAL-ASSOCIATED DRINKING

Under conditions of continuous access to foodand water, rats take discrete meals (usually

about 10 per day) that are either interruptedor followed immediately by episodes of drink-ing. In fact, about 80% of spontaneous waterintake occurs in this prandial manner. Directmeasurement of prandial drinking requirescontinuous recording of food and water in-take, either qualitatively using sensors ofwhen the commodities are accessed (e.g., lick-ometers, photobeams) or quantitatively usingweighing devices. A different aspect of meal-associated drinking can be measured moresimply by forcing a single meal, such as byprior food restriction, and then measuring vol-umetrically the water intake that occurs withthat meal. In this case, the water-to-food ratio(given in, for example, milliliters per gram) isa useful derivative measure.

THE TEST ENVIRONMENT

Not only are the choice and mode of admin-istration of the dipsogenic stimulus important,but so is the test environment. Rats are natu-rally apprehensive of novel environments andcommodities. They should be adapted to thecage in which drinking will occur and to thefluid(s) to be presented.

THE ENVIRONMENT

The choice of environment is usually betweeneither home cage or a test arena. If using thelatter, the rats should be allowed at least one(preferably more) previous drinking episodein that environment. If a behavior more com-plex than reflexive licking (e.g., a lever pressoperant) is required, more training is needed.A second important variable is time of day:rats naturally do most of their drinking atnight, when certain thresholds may be func-tionally lower. However, the intake of food(also mostly nocturnal) may have an un-wanted or uncontrolled effect on drinking.For this reason, most short-term drinkingstudies are performed either during the lightperiod when food intake is infrequent or by

Page 230: 0195162854

Chapter 19. Drinking 213

removing food 1 to 2 hours before the drink-ing test. A third variable is the temperature ofboth fluids and environment. Most animalsare housed and tested at temperatures that arecomfortable to humans, but if rats are movedto a separate room for testing, it should be atthe same temperature, and the drinking fluidequilibrated at that temperature. A fourthvariable is social factors. Most studies exam-ine intake of individual rats; either direct orindirect interaction with conspecifics could in-fluence drinking and under normal conditionsshould be avoided. Interaction with humansis an unavoidable aspect of most drinkingstudies, ranging from direct handling such asgiving injections or placing in a test cage toindirect influences such as other human ac-tivity in the room. Rats are extremely adapt-able to a range of conditions, and their drink-ing behavior is usually robust, but a consistentroutine is essential.

The approved standard for rat housinghas changed over the past decade from stain-less steel mesh to plastic tubs with soft bed-ding. These latter may better reproduce nat-ural burrow material, although they do nottypically allow these nocturnal animals anytype of shade. Drinking studies in these cagesnormally require sipper spouts that protrudethrough the metal grill top of the cage (thisavoids the spout from making contact withthe bedding and leaking fluid). Steel meshcages allow the spout to protrude through thewall of the cage. Fluid intake studies also arelinked to measurement of urinary output, andthese have to be performed in metabolic cageswith mesh floors. As noted earlier, rats shouldbe adapted to whichever caging is chosen; ifthat is considered "nonstandard," then ap-proval for the exception should be requestedfrom the IACUC.

THE SOLUTION AND ITS PRESENTATION

In the study of water intake, one decision thatshould be made is between using either deion-ized or tap water. This decision depends in

part on the experimental question and thequality of tap water in a particular laboratory.In general, if the tap water is known to havehigh or variable mineral content and/or hasan odor to humans, it is advisable not to useit. The rats should be adapted to the waterchosen for several days before the experiment.If solutes (e.g., NaCl) are to be added, thesame choice of solvent applies. If in doubt, useeither deionized/distilled water or, if that isnot readily available, commercially availablebottled water. Added solutes should be of thehighest purity available.

If rats are purchased from a commercialvendor, they are most likely accustomed todrinking from the nipple of an automatic wa-ter system, and this may also be used in yourvivarium. However, such systems are unsuit-able for measuring water intake, so most lab-oratory experiments use water bottles withmetal sipper spouts to which rats must be ac-customed. These can be purchased commer-cially. Spouts differ considerably in their char-acteristics, and one of the major and oftenoverlooked sources of variance within an ex-periment comes from spout topography. Ratslick in bouts separated by pauses; the lickingrate within a bout is about seven licks per sec-ond. Thus, the amount of fluid consumed perunit time depends both on the volume per lickand on the pause duration(s). The former de-pends on the diameter at the orifice of thespout and on whether a ball bearing is pres-ent. It is therefore best to use only one typeof spout in your laboratory: in that way, everyanimal has the same type of spout and thereis no possibility of day-to-day variance. Also,if more than one fluid is offered, both spoutswill be identical. If the spouts are used withrubber stoppers or washers, as is usually thecase, ensure that the shaft is pushed all theway through the stopper to avoid the forma-tion of airlocks.

In the real world, rats lick water frompuddles or other open surfaces. Richter tubesare drinking cylinders that have horizontalsurface drinking troughs. Glass models can be

Page 231: 0195162854

214 REGULATORY SYSTEMS

purchased commercially, but they are gener-ally more expensive and harder to clean thantubes with spouts and are used relatively lit-tle. One exception is the use of fluid dippers,which are small (e.g., 0.1 ml) troughs or cups,to deliver standard volume fluid reinforce-ments in operant procedures; rats readilyadapt to drinking from them.

TEST DURATION

Most adult rats with standard food availabledrink 30 to 50 ml/day. Thus, the measure-ment of 24-hour intakes to the nearest ± 1 mlusually is adequate. I use 50 or 100 ml plasticgraduated cylinders (with the top lip cut offby using a hacksaw) with rubber one-holestoppers and metal sipper spouts. The startand finish volume graduations are read di-rectly. An alternative, for which graduatedtubes are not necessary, is to weigh the tubeor bottle before and after. This has the ad-vantage that most electronic scales allow theweights to be sent directly to a computerspreadsheet.

Intakes of about 10 ml or less, such asstimulated by acute dipsogens, need to berecorded with greater accuracy (±0.1 ml). Al-though gravimetric measurement (as earlier)is viable, the likelihood of losing a few dropsof fluid when the bottle is placed on or re-moved from the cage is quite high. For thisreason, I recommend direct volumetric meas-urement while the tube is on the cage. Forsuch measurements, I use either 25 ml grad-uated plastic or glass pipets with each end cutoff so that a sipper spout (in a collar of plastictubing) can be firmly wedged in one end anda small rubber stopper in the other.

Volume consumed is the principal de-pendent variable in many studies of fluid in-take, but the pattern of intake is also ofimportance in some applications. Severalcomputer-linked lick sensors are availablecommercially for this purpose. One design hasa small infrared beam across the tip of thespout, which itself is slightly recessed so that

normally the protruded tongue will break it.Another design is a contact sensor in whichthe rat completes an electric circuit when itlicks from the spout. The currents involvedare too small to be appreciable to the rat.These may be used in acute drinking studies,including examination of the effect of tastantson behavior, but in conjunction with sensorsof food intake, they can be used to study tem-poral relationships between food and fluid in-takes over long periods.

SODIUM PREFERENCEAND APPETITE

PREFERENCE

It is conceptually important to separate pref-erence from appetite. Preference, in this casefor sodium solutions, is exhibited under con-ditions of sodium balance (need free) and isdetermined by comparing intake of sodiumsalt with a reference solution (e.g., water).This may be conducted either in separate butotherwise identical sessions called one-bottletests or in sessions with both fluids availablesimultaneously, called two-bottle tests. Long-duration (e.g., 24 hours) preference tests maybe affected by the rat learning about postinges-tional consequences of a particular flavor.Short-duration preference tests can overcomethis pitfall but usually require fluid deprivationto induce the intake, so the possible interactionof need state with preference must be consid-ered in the experimental design. More com-plex designs use more than two bottles orchoices of fluid; however, the number ofavailable options may influence choice be-havior. Many rat strains show a spontaneouspreference for NaCl over water in the rangeof approximately 0.05 to 0.2 mol/L.

APPETITE

Appetite is defined as intake in excess of thatduring normal or based conditions, and that

Page 232: 0195162854

Chapter 19. Drinking 215

has motivational characteristics. The mostcommon way of studying sodium appetite inrats is to offer a hypertonic NaCl solution (0.3to 0.5 mol/L) that is above the spontaneouslypreferred range. Studies on the taste speci-ficity have shown that sodium appetite pro-duced by the methods to be discussed is spe-cific for the cation sodium but is insensitiveto changes in the associated anion. Mineral-deficient mammals most likely do not en-counter salt solutions in their natural habitat:rather, they obtain their minerals from foodor mineral-enriched soil deposits. Curiously,very few studies have successfully foundsodium appetite for salty foods in rats. How-ever, in recent work in our laboratory, wefound that salt gels work remarkably well.These are made by mixing concentrated saltsolutions (we have used 0.5 to 1.5 mol/L)with gelatin powder (5% w/v) and then allowthem to solidify in glass jars for presentation.Rats ingest minimal amounts of 1 mol/L orabove under need-free conditions but exhibita robust intake during sodium depletion.

PRODUCING SODIUM DEPLETION

The acute injection of a rapid-acting loop di-uretic such as furosemide (also called frusemide)causes a dose-related loss of sodium and wa-ter in urine and hypovolemia. Subcutaneoussingle injection of 2 mg/kg or more causes anear-maximal loss of about 2 mEq sodium inan adult rat within 1 to 2 hours. This hypo-volemia is associated with a sodium appetite(see Fig. 19-2), but this develops relativelyslowly over the next 12 to 24 hours. Thus, oneof the most used protocols involves injectionof furosemide followed by a 24 hour periodwithout available sodium. To accomplishthis reliably, a fresh cage or bedding shouldbe provided, along with distilled water and alow- or no-sodium diet. At the end of thistime, rats will ingest several milliliters of hy-pertonic NaCl, often in substantial excess oftheir 2 mEq deficit.

A chronic version of this protocol may also

be used. Either daily injections of furosemideor die addition of the diuretic hydrochloro-thiazide to the low-sodium food produces arobust, sustained sodium appetite (Rowlandand Colbert, 2003).

OTHER STIMULI OF SODIUM APPETITE

Several other procedures will produce sodiumappetite; most involve the use of Ang IIand/or aldosterone (Fregly and Rowland,1985). Under natural conditions, these hor-mones probably work in synergy to producesodium appetite. However, activation of ei-ther hormonal system alone is sufficient.Adrenalectomy, which removes the endoge-nous source of aldosterone, produces a high-Ang sodium appetite. Conversely, the admin-istration of high doses of deoxycorticosterone,a mineralocorticoid hormone, produces so-dium appetite with concurrent suppression ofAng II formation. Thus, just as the study ofthirst involves a choice between stimuli of dis-crete component systems, a similar situationapplies to sodium appetite.

MOTIVATION AND THE STRUCTUREOF SODIUM APPETITE

Several authors have made the point thatsodium appetite is both innate and motivated.In regard to the latter, it has been shown thatrats given some of the above stimuli will per-form operant tasks in discrete sessions to ob-tain sodium solutions (Quartermain et al.,1967). Under free access conditions, rats show-ing an appetite for concentrated NaCl solutionconsume it in discrete bouts in close tempo-ral proximity to spontaneous meals and pran-dial water (Strieker et al., 1992). Recently, us-ing a combination of operant and free accessprotocols and standard rat operant cages, wehave found that rats temporally structuresodium "meals" in accordance with relativecost and need. Thus, although rats will takewater and salt together, this is not physiolog-ically obligatory.

Page 233: 0195162854

216 REGULATORY SYSTEMS

REFERENCES

Fitzsimons JT (1963) The effects of slow infusions of hy-pertonic solutions on drinking and drinking thresh-olds in rats. Journal of Physiology 167:344-354.

Fitzsimons JT (1979) The physiology of thirst andsodium appetite. Monographs of the PhysiologicalSociety #35, Cambridge University Press.

Fitzsimons JT (1998) Angiotensin, thirst, and sodium ap-petite. Physiology Review 78:583-686.

Fregly MJ and Rowland NE (1985) Role of renin-angiotensin-aldosterone system in NaCl appetite ofrats. American Journal of Physiology Regulatory,Integrative, and Comparative Physiology 248:R1-Rll.

Marwine A and Collier G (1979). The rat at the water-hole. Journal of Comparative Physiology and Psy-chology 93:391-402.

Morita H, Yamashita Y, Nishida Y, Tokuda M, HataseO, Hosomi H (1997). Fos induction in rat brain neu-rons after stimulation of the hepatoportal Na-sensi-tive mechanism. American Journal of PhysiologyRegulatory, Integrative, and Comparative Physiol-ogy 272:R913-R923.

Quartermain D, Miller NE, Wolf G (1967) Role of ex-perience in relationship between sodium deficiencyand rate of bar pressing for salt. Journal of Com-parative Physiology and Psychology 63:417—420.

Rowland NE (1990) On the waterfront: Predictive andreactive regulatory descriptions of thirst and sodiumappetite. Physiology and Behavior 48:899-903.

Rowland NE (2002) Thirst and sodium appetite. In:Stevens' handbook of experimental psychology, 3rdedition, vol. 3: Learning, motivation and emotion(Pashler H and Gallistel CR, eds.), pp. 669-707. NewYork: Wiley.

Rowland NE and Colbert CL (2003). Sodium appetiteinduced in rats by chronic administration of a thi-azide diuretic. Physiology and Behavior 79:613-619.

Strieker EM (1968) Some physiological and motivationalproperties of the hypovolemic stimulus for thirst.Physiology and Behavior 3:379-385.

Strieker EM (1969) Osmoregulation and volume regu-lation in rats: Inhibition of hypovolemic thirst bywater. American Journal of Physiology 217:98-105.

Strieker EM, Gannon KS, Smith JC (1992) Salt appetiteinduced by DOCA treatment or adrenalectomy inrats: Analysis of ingestive behavior. Physiology andBehavior 52:793-802.

Strieker EM, Hoffmann ML, Riccardi CJ, Smith JC (2003)Increased water intake by rats maintained on highNaCl diet: Analysis of ingestive behavior. Physiol-ogy and Behavior 79:621-631.

Watts AG (2000) Understanding the neural control ofingestive behaviors: Helping to separate cause fromeffect with dehydration-associated anorexia. Hor-mones and Behavior 37:261-283.

Page 234: 0195162854

Foraging

IAN Q. WHISHAW20

Rats forage on a wide variety of foodstuffs.Some food is eaten where it is found and someis carried to havens, where it is eaten or leftfor later (Lore and Flannelly, 1978; Takahashiand Lore, 1980; Whishaw and Whishaw,1996). With the exception of lactating females,rats do not cache (hoard) food. Food carryingmay help rats avoid food theft, avoid preda-tion, and redistribute food stores for the ben-efit of colony members. The susceptibility ofa foraging rat to theft by conspecifics is docu-mented by observations made of wild andsemi-wild colonies (Barnett and Spencer,1951). Chitty (1954) observes large rats catchand overturn smaller rats to steal grain fromtheir mouths, and Whishaw and Whishaw(1996) observe that large "dominant" rats areless likely to carry food or be attacked by con-specifics than are smaller rats.

Optimal foraging theory proposes that for-aging behavior represents a tradeoff betweenstrategies for obtaining food and strategies foravoiding attack and predation. Rules govern-ing eating, food theft and protection, and foodcarrying illustrate innovative ways that the be-havior of the rat has been sculptured by opti-mizing principles. The following sections de-scribe eating behavior, food protection andtheft, food carrying, and some aspects of theirneural control.

EATING TIME

A rat can optimize food acquisition by eatingquickly. Rats vary their eating speed in re-sponse to exposure, time of day, food depri-

vation, and previous previous deprivationhistory. They display individual differences ineating speed (Whishaw et al, 1992). Increas-ing eating speed, however, can have digestivecosts through reduced chewing and salivawetting (Morse, 1985).

The largest influence on eating behavioris in the time required to eat. Obviously a ratwill eat a small piece of food more quicklythan a large piece of food, but many other fac-tors influence eating speed. The time taken toeat a piece of food of a given size is influencedby location. In the open, such as on an opentable or in a cage without a cover, eating ismore rapid than it is in a shelter. In the open,rats also make many head scans, during whichthey continue chewing, whereas head scansare rare in sheltered environments. Eating isfaster in a novel location than in a familiar lo-cation. Eating speed also varies as a functionof the amount of food eaten. As the numberof food pellets eaten increases, so does theaverage eating time per pellet. Thus, eatingspeeds and head scanning behavior suggestthat eating rats are vigilant and sensitive to thepossibility of predation or attack as well as tonutrient need.

Time of day and personal history affecteating speed (Fig. 20-1). Rats eat more quicklyduring the dark phase of their 24 hourday-night cycle and eat more quickly if theyhave been, or are, food deprived. Lightingduring the day-night cycle also influences eat-ing speed, as rats eat more slowly when thelights are off, especially in the light portion ofthe cycle. Finally, some rats eat very quicklyand others eat more slowly independent of

217

Page 235: 0195162854

218 REGULATORY SYSTEMS

0 2 4 6 8 10 12 14 16 18 20 22 24

Time (hour)

Figure 20-1. Mean time to eat a 1 g food pellet as a func-tion of time of day and food deprivation schedule. Inset:Mean eating time as a function of deprivation schedule.Black bars, lights off. (Based on Whishaw et al., 1992.)

their deprivation level and previous feedinghistory. The individual differences could berelated to nursing success during infancy or toprenatal and genetic influences.

Rats retain a retrospective knowledge oftheir feeding time (Whishaw and Gomy, 1991).After eating a food pellet, a rat scans its sur-roundings by sniffing and running its vibrissaeover the ground. The size of the area that itscans is proportional to the size of the food pel-let that it has just eaten, with larger scans fol-lowing consumption of larger food pellets. Iffood hardness is varied independent of foodsize, the area that is scanned is best predicted bythe time required to eat a food item. Althoughit is likely that scans are directed toward findingdropped crumbs, rats that are tested on mesh,through which crumbs fall, still scan.

FOOD WRENCHING AND DODGING

The immobile feeding posture of the rat, inwhich it sits on its haunches with the food heldin its paws, leaves it vulnerable to attack fromother rats. This vulnerability is amplified whenit is considered that other rats are excited by afeeding rat, investigate it, pick up bits of foodthat it has dropped, sniff its snout, and lickcrumbs from its lips (Barnett and Spencer, 1951;

Galef, 1983; Galef and Wigmore, 1983; Posadas-Andrews and Roper, 1983).

Rats are artists in food theft and theftavoidance (Whishaw, 1988; Whishaw andTomie, 1987). A rat attempting to steal foodapproaches a feeding rat from the rear, walksalong the side of the victim, and reaches un-der its snout to wrest the food from its paws(Fig. 20-2). It may grasp the victim's paw sothat it can expose the food and/or knock it

Figure 20-2. Food wrenching attempt (left) and a success-ful dodge (right). (Based on Whishaw and Tomie, 1987.)

Page 236: 0195162854

Chapter 20. Foraging 219

free. A victim evades the robber by dodging(Whishaw and Tomie, 1987; Whishaw, 1988).A dodge consists of a turn of the head fol-lowed by steps with the hindlimbs that turnthe rat away from the robber. The maneuverleaves the victim time to continue eating.

This description of an average dodgeshould not belie the variations in the move-ment. A rat might run forward, dodge back-ward over an approaching rat, or simply twistits head and pivot. In some dodges, the foodis held in both front paws so that the rat caneasily continue eating, but the food may beheld in one front paw while the other assiststurning, or the food may be transferred to themouth so that the rat can use both front pawsto increase the size of the dodge. The fooditem is transferred back to the paws for eat-ing at the completion of the dodge. Dodgesmay be assisted with a hop, end with a hop,or be attached to short runs.

There is seldom overt aggression betweenthe robber and victim. The robbing-dodging in-teraction can occur repeatedly, as long as onemember of the pair has food to eat. If a rob-ber wrenches the food away from a victim, itwill in turn dodge, and the former victim willreplace it as the robber. Experience inrobber-victim interactions contributes bothto the effectiveness of the victim's dodges andto the skill and aggression of the robber.

EATING TIME INFLUENCESDODGE SIZE

The victim of a theft is conservative in themovements made to avoid the robber(Whishaw and Gorny, 1994). If a rat has asmall food pellet (20 to 94 mg), it picks it upby mouth and quickly chews and swallows it,and so there is no possibility of theft. If a foodpellet is larger (>190 g), it is held in the pawsas the rat adopts a sitting posture to eat, thusallowing the robber an opportunity to steal.Dodge distance and dodge angle in responseto an attempted theft increase with increases

in the size of food. A rat with a small food pel-let may avert its head; with a larger food pel-let, it may make a partial turn, and with a verylarge pellet, it may turn completely (about 180degrees) and sometimes run a short distance.The victim also tracks the size of the food thatit is eating and adjusts its dodges to theamount food it has yet to eat. When begin-ning to eat a large piece of food (e.g., 1 g pel-let), the victim makes a maximum sized dodgeand as size of the food diminishes, the size ofdodge diminishes, and eventually as the foodbecomes smaller, small dodges are replacedby a head turn or no movement at all.

If dodging were simply a function of thesize of the food being eaten, a rat might havedifficulty gauging how to protect a large, easy-to-eat piece of food relative to a small, hard-to-eat piece of food. Rats solve this problemby calibrating food size in terms of anticipatedeating time (Whishaw and Gorny, 1994).When rats are given food pellets baked to dif-ferent hardnesses, victims make larger dodgeswith the harder food. To further determinethe role of food size and eating time in deter-mining dodge size, we gave rats a comparisonseries of foods and a test series of foods. Thecomparison series consisted of 10 different-sized, round, commercial food pellets weigh-ing between 20 mg and 1000 mg. The test se-ries of foods included barley, wheat, Mungbeans, and Azuki beans. In terms of size, thetest series was comparable to the low end ofcomparison series, and in terms of time re-quired to eat, the test series fell in the upperhalf of the comparison series. When dodgeprobability and dodge size are measured as afunction of the different foods that a victimwas eating, these measures were closely re-lated to the time required to eat the food itemand not to the size of the food item.

Eating time is arguably the easiest way togauge susceptibility to food loss given thewide range of food items that a rat might con-sume. Of course, a rat must learn how long itwill take to eat various kinds of food. Whenwe gave a grain of rice, which is about the size

Page 237: 0195162854

220 REGULATORY SYSTEMS

of a 20 mg food pellet, to rats, they did notdodge on the first approach of a robber butdid on the second approach. Thus, one eatingexperience is sufficient for the rat to learn thata small piece of hard food takes much longerto eat than an equal-sized soft piece of food.

SEX DIFFERENCES IN DODGING

The dodging movements made by male and fe-male rats are different (Field et al., 1996,1997a).Female rats move their snout through a greaterspatial curvature, and the snout achieves agreater velocity, relative to the pelvis, than oc-curs for males. Stepping movements of femalesare also simpler. They step away first with thecontraversive, to the robber, hind limb andthen with the ipsiversive hind limb. Males firststep with ipsiversive, to the robber hind limb,a movement that brings the rear of the victimtoward the rear of the robber. The victim thensteps away from the robber with the contra-versive hind limb. Thus, the male gives therobber "the hip" before pivoting its forequar-ters. There is some variation in the steppingmovements of both sexes, but in general thefemale pivots around a point more posterior onthe body than a male (Fig. 20-3).

Sex differences in dodging, in which thefemale's dodge propels it away from the rob-ber while the male's dodge initially propels itinto the robber, may reflect the more generaldifferences in behavior of the sexes. Dodging

Figure 20-3. Difference in dodges by female and male rats:The female pivots around its pelvis, and the male pivotsaround the mid body. (Based on Field and Pellis, 1998.)

is exhibited by female rats during mating tocorrect ineffective approaches by a male rat(Whishaw and Kolb, 1985). The "give him thehip" behavior of the male rat is similar to themale rat's threat display (Pellis and Pellis,1987). Thus, a "dodge" is akin to a word in abehavioral lexicon in that it can be used in dif-ferent behavioral contexts while retaining theaccent of the performer.

That dodging movements are sex relatedsuggests that they are determined by hor-monal influences. The hormonal influencemust occur very early in development be-cause castration of juvenile rats does not af-fect the male-typical patterns of movement.Neonatal castration does feminize the male'spattern of movement, whereas the early ad-ministration of testicular hormones to femalesmasculinizes their movement (Field et al.,1997a, 1997b). Thus, it is likely that prenataland perinatal hormonal influences affect notonly the later reproductive roles of rats butalso the movements that they make in thenonreproductive behavior of dodging.

FOOD CARRYING

Food carrying is an elaboration of dodging(Fig. 20-4). When food pellets of varying sizeare presented through a small aperture to ahungry rat in an alley, the rat produces afood size-related escalation in movement(Whishaw et al., 1990; Whishaw and Tomie,1989). Small food pellets are swallowed as therat grasps them with its mouth (eat). Inter-mediate-sized food pellets are transferredfrom the mouth to the paws and are eaten asthe rat adopts a sitting posture (sit). As foodsize increases further, rats dodge farther awayfrom the aperture (dodge) until, with largefood pellets, they run to the far end of the al-ley with the food before eating (carry}.

When given a choice of food sizes, rats areselective with respect to the size of food itemsthat they choose. If a pile of variously sized foodpellets is available, a rat first scans the pellets

Page 238: 0195162854

Chapter 20. Foraging 221

Figure 20-4. Relation between food size and movement. In the series 1 through 8, the rat receives succes-sively larger food pellets; as it does so, it makes larger movements that lead it to eventually dodge away fromthe food source. (Based on Whishaw and Tomie, 1989.)

and then chooses the largest food pellets andcarries them. In addition, it will attempt tocarry more than one pellet of a cartable size.We have observed a rat stuffing three 1 g foodpellets into its mouth while attempting to carryan additional food pellet in a forepaw, as it ranon three legs. Surprisingly, if only small foodpellets are available, a rat will only eat and willnot attempt to carry a number of them.

THE HOME BASE

If a refuge or covered area is available, it be-comes the home base for a rat's foraging be-havior, although the rat's behavior is still cal-

ibrated by food size (Fig. 20-5). A rat will onlyleave the refuge after a "stop-sniff-look" in-vestigation of its surroundings, and it leavesthe refuge with a cautious walk, with the bodyheld low. The eat, sit, and dodge behaviorsstill occur in the open if it finds smaller foodpellets. If it finds a large piece of food, it gal-lops back to the refuge. The latency to initi-ate a carry response and the travel speed homeincrease with increases in the size of a foodpellet. In addition, the latency to return aftereating increases as a function of the size of afood pellet just eaten.

A refuge can be a source of potentialproblems (Whishaw, 1991). If another rat alsouses the home base as a refuge, it might be

Page 239: 0195162854

222

Figure 20-5. Behaviors displayed by rats in a food-carryingtask. (A) Stop-sniff-look before leaving the refuge. (B) Cau-tious approach to food source. (C) Eat response in which asmall food pellet is swallowed. (D) Sit response in which arat sits on its haunches with an intermediate-sized piece offood held in the paws. (E) Carry response in which a ratturns back to the refuge with a large piece of food. (Basedon Whishaw and Oddie, 1989.)

thought that a food-carrying rat will avoid thehome base to reduce the possibility of foodtheft. They do not. Rats given three differentrefuges (covered boxes with an entrance)quickly adopt a preferred refuge. Placing ahungry rat in the preferred refuge, however,results in little change in preference. If the

REGULATORY SYSTEMS

preferred refuge is modified by removing itscovers (black paper taped onto the Plexiglasboxes), the rats will quickly adopt a newrefuge. Thus, the refuge appears to be a refugefrom predation and not from conspecific com-petition. This finding seems consistent withthe observations of wild rats (Whishaw andWhishaw, 1996). When provided with a largesupply of peanuts, some of the rats carriedpeanuts back to their home territory only tohave them repeatedly stolen by other rats.This confirms that the carrying rats fail to an-ticipate and avoid theft.

Rats may avoid carrying food to a homebase if it is their nest area. Rats given accessto a number of refuges quickly adopt onerefuge as a "home" in which they sleep whilethey carry the food to a different refuge foreating.

EATING TIME INSTRUCTSFOOD CARRYING

A central prediction of optimal foraging the-ory is that a foraging animal should attemptto maximize its food intake while minimizingits exposure to attack/predation. An effectiveway of making optimal judgments is to usetime as a measure of exposure and nutrientintake. Thus, rats should follow the rule:"Carry if eating time exceeds return trip time"(Whishaw, 1990). Using an alley with food ac-cess at one end and a refuge at the other end,we provided rats with 10 different-sized foodpellets and a number of natural foods, in-cluding wheat, pearl barley, Mung beans, andAzuki beans. Three experiments were per-formed. In the first experiment, the 10 differ-ent-sized food pellets were either soft or hard(produced by increased baking). In the secondexperiment, the rats received soft food pelletsand the natural foods. In both of these exper-iments, food carrying was predicted by eatingtime and not food size or hardness (Fig. 20-6).

In a third study, rats were given an adap-tation experience, modeled on human psy-

Page 240: 0195162854

Figure 20-6. Food-carrying probability (mean and standarderror) with different kinds of food plotted against eatingtime. Anticipated eating time rather than food size or hard-ness predicts the probability of food carrying. (Based onWhishaw, 1990.)

chophysical experiments. One group of ratsreceived experience only with small pellets(sizes 1 to 7), and another group received ex-perience only with large food pellets (sizes 4to 10). After training, all rats were given a testwith the full range of pellets. If the rats wereinfluenced by context, intermediate-sizedfood pellets should be more likely to be car-ried by the group receiving the small com-parison series and less likely to be carried bythe group receiving the large comparison se-ries. If the rats were responding to internalcues, such as, time, the contextual trainingshould not affect food-carrying behavior. Theresults indicated that the rats were not influ-enced by context, which suggests that theywere using eating time. Thus, these experi-ments demonstrated that rats did obey therule: "Carry if eating time exceeds return triptime/'

Optimal foraging theory also predictsthat the distance to the refuge influences foodcarrying (Whishaw, 1993). If the distance isshort, it will be worthwhile carrying smallfood pellets to the refuge, whereas if the dis-tance is long, a larger piece of food should berequired to initiate a return. When rats wereallowed to forage over distances varying froma few centimeters to over 600 cm, variationsin distance produced significant changes in

223

food-carrying probability that were related totravel time. The probability that food was car-ried decreased linearly with travel distance.

Other manipulations of travel time alsoaffected the probability of carrying food. If ratsare required to walk across a short narrowbeam, which increases travel time and intro-duces the risk of falling, they display a reducedtendency to carry food of a size that theywould otherwise carry. If a frank risk is intro-duced, such as, the smell of a cat, the animalsarrest all foraging behavior.

IMPLICATIONS FORBRAIN FUNCTION

Because of its role in optimizing behavior inits use of time, space, and different motor acts,food handling provides a rich behavior forexamining the nervous system function. Be-ginning with Wolfe's (1939) description of"hoarding" in laboratory rats, there has beeninterest in the neural control of food carrying(Mark, 1950; Munn, 1955; Ross et al., 1955).Food-carrying behavior has also been of in-terest in assays for neural injury (Whishawand Tomie, 1988; Whishaw and Oddie, 1989;McNamara and Whishaw, 1990) and for ap-plied problems such as modeling anxiety(Dringenberg et al, 1998; 2000).

Damage to limbic structures, includingthe medial frontal cortex, hippocampus, andnucleus accumbens, reduces hoarding. Inter-est in control exerted by these structures isheightened by two observations. First, con-text is important (Whishaw, 1993). Animalswith hippocampal damage that have stoppedfood carrying may begin again if activated be-haviorally (e.g., being startled while eating).They may also stop carrying food if theirrefuge is moved, but, again, carrying can berestored with training. If the distance betweenfood and the refuge is increased, however, ratswith hippocampal lesions will stop carrying atshorter distances than that needed to arrestcarrying in control rats.

Eating Time (sec)

Chapter 20. Foraging

Page 241: 0195162854

224 REGULATORY SYSTEMS

Early studies on food carrying focused onthe influence of food deprivation on food-carrying behavior. Bindra's (1978) demonstra-tion that both food deprived and non-food-deprived rats "hoard" suggests that not onlywill rats hoard food for nutrition, they alsowill hoard because of the food's incentivevalue. Whishaw and Kornelesen (1993) con-firm this finding by dissociating food carryingand food hoarding with nucleus accumbenslesions. Both food-deprived control rats andrats with neurotoxic damage to the cells of thenucleus accumbens carried food from an openarea to a refuge. As the rats consumed thefood and became sated, the nucleus accum-bens group stopped carrying food, whereasthe control rats continued to carry. Thus, con-trol rats responded to both the nutrient andincentive value of the food, whereas the nu-cleus accumbens group rats responded only toits nutrient value.

Taken together, the rich array of behav-iors related to food handling in the rat pro-vides a challenge to investigators of the neu-ral control of behavior. In addition, eatingbehavior can be a useful measure in a widerange of studies, including those related toaddiction, eating disorders, individual differ-ences, and spatial and temporal behavior.

REFERENCES

Barnett SA and Spencer MM (1951) Feeding, social be-haviour and interspecific competition in wild rats.Behavior 3:229-242.

Bindra D (1978) How adaptive behaviour is produced:a perceptual-motivational alternative to responsereinforcement. Behavioural Brain Sciences 1:41-91.

Chitty D (1954) The control of rats and mice, Vols 1 and2: Rats. Oxford: Clarendon Press.

Dringenberg HC, Wightman M, Beninger RJ. (2000)The effects of amphetamine and raclopride on foodtransport: Possible relation to defensive behavior inrats. Behavioral Pharmacology 11:447-454.

Dringenberg HC, Kornelsen RA, Pacelli R, Petersen K,Vanderwolf CH (1998) Effects of amygdaloid le-sions, hippocampal lesions, and buspirone on black-white exploration and food carrying in rats. Behav-ioural Brain Research 96:161-172.

Field EF and Pellis SM (1998) Sex differences in the or-ganization of behavior patterns: Endpoint measuresdo not tell the whole story. In: (Ellis L and EbertzL, eds.). West Point, Conn: Praeger.

Field EF, Whishaw IQ, Pellis SM (1996) A kinematicanalysis of evasive dodging movements used duringfood protection in the rat (Rattus norvegicus): Evi-dence for sex differences in movement. Journal ofComparative Psychology 119:298-306.

Field EF, Whishaw IQ, Pellis SM (1997a) Organizationof sex-typical patterns of defense during food pro-tection in the rat: The role of the opponent's sex.Aggressive Behavior 23:197-214.

Field EF, Whishaw IQ, Pellis SM (1997b) A kinematicanalysis of sex-typical movement patterns used dur-ing evasive dodging to protect a food item: The roleof testicular hormones. Behavioral Neuroscience111:808-815.

Galef BG Jr (1983) Utilization by Norway rats (R.norvegicus) of multiple messages concerning dis-tant foods. Journal of Comparative Psychology97:364-371.

Galef BG Jr and Wigmore SW (1983) Transfer of infor-mation concerning distant foods: A laboratory in-vestigation of the "information-center" hypothesis.Animal Behaviour 31:748-758.

Lore RK and Klannelly K (1978) Habit selection and bur-row construction by wild Rattus norvegicus in a land-fill. Journal of Comparative and Physiological Psy-chology 92:888-896.

Marx MH (1950) Stimulus-response analysis of hoardinghabit in the rat. Psychological Review 57:80-94.

McNamara RK and Whishaw IQ (1990) Blockade ofhoarding in rats by diazepam: an analysis of the anx-iety and object value hypotheses of hoarding. Psy-chopharmacology 101:214-221.

Munn ML (1933) Handbook of psychological researchon the rat. Boston: Houghton Mifflin.

Pellis SM and Pellis VC (1987) Play-fighting differs fromserious attack in both target of attack and tactics offighting in the laboratory rat Rattus norvegicus. Ag-gressive Behavior 13:227-242.

Posadas-Andrews A and Roper TJ (1983) Social trans-mission of food preferences in adult rats. Animalbehavior 31:265-271.

Ross S, Smith WI, Wossner BL (1955) Hoarding: Ananalysis of experiments and trends. Journal of Gen-eral Psychology 52:307-326.

Takahashi LK and Lore RK (1980) Foraging and foodhoarding of wild Rattus norvegicus in an urban envi-ronment. Behavioral and Neural Biology 29:527-531.

Whishaw IQ (1988) Food wrenching and dodging: Useof action patterns for the analysis of sensorimotorand social behavior in the rat. Journal of Neuro-science Methods 24:169-178.

Page 242: 0195162854

Chapter 20. Foraging 225

Whishaw IQ (1990) Time estimates contribute to foodhandling decisions by rats: Implications for neuralcontrol of hoarding. Psychobiology 18:460-466.

Whishaw IQ (1991) The defensive strategies of foragingrats: A review and synthesis. The PsychologicalRecord 41:185-205.

Whishaw IQ (1993) Activation, travel distance, and en-vironmental change influence food carrying in ratswith hippocampal, medial thalamic and septal le-sions: Implications for studies on hoarding and the-ories of hippocampal function. Hippocampus 3:373-385.

Whishaw IQ and Oddie SD (1989) Qualitative and quan-titative analyses of hoarding in medial frontal cor-tex rats using a new behavioral paradigm. Behav-ioural Brain Research 33:255-256.

Whishaw, IQ, Oddie SD, McNamara RK, Harris TL,Perry BS (1990) Psychophysical methods for thestudy of sensory-motor behavior using a food-carrying (hoarding) task in rodents. Journal of Neu-roscience Methods 32:123-133.

Whishaw IQ, Dringenberg HC, Comery TA (1992) Rats(Rattus norvegicus) modulate eating speed and vig-ilance to optimize food consumption: Effects ofcover, circadian rhythm, food deprivation, and in-dividual differences. Journal of Comparative Psy-chology 4:411-419.

Whishaw IQ and Gorny BP (1991) Postprandial scan-ning by the rat (Rattus norvegicus): The importanceof eating time and an application of "warm-up"movements. Journal of Comparative Psychology10:39^4.

Whishaw IQ and Gorny BP (1994) Food wrenching anddodging: Eating time estimates influence dodgeprobability and amplitude, Aggressive Behavior20:35-47.

Whishaw IQ and Kornelsen RA (1993) Two types of mo-tivation revealed by ibotenic acid nucleus accum-bens lesions: Dissociation of food carrying andhoarding and the role of primary and incentive mo-tivation. Behavioural Brain Research 55:283-295.

Whishaw IQ and Kolb B (1985) the mating movementsof male decorticate rats: Evidence for subcorticallygenerated movements by the male but regulationof approaches by the female. Behavioural Brain Re-search 17:171-191.

Whishaw IQ and Tome J (1987) Food wresting anddodging: Strategies used by rats (Rattus norvegicus)for obtaining and protecting food from conspecifics.Journal of Comparative Psychology 101:110-123.

Whishaw IQ and Tomie J (1988) Food wrenching anddodging: A neuroethological tests of cortical anddopaminergic contributions to sensorimotor be-havior in the rat, Behavioral Neuroscience 102:110-123.

Whishaw IQ and Tomie J (1989) Food-pellet size mod-ifies the hoarding behavior of foraging rats. Psy-chobiology 17:83-101.

Whishaw IQ and Whishaw GE (1996) Conspecific ag-gression influences food carrying: Studies on a wildpopulation of Rattus norvegicus. Aggressive Be-havior 22:47-66.

Wolfe JB (1939) An exploratory study of food-storing inrats. Journal of Comparative Psychology 28:97-108.

Page 243: 0195162854

Thermoregulation

EVELYN SATINOFF21

To avoid becoming too hot or too cold, ratsbuild dens, nests, and burrows; huddle to-gether; bask in the sun; lie in the shade; swim;sprawl; groom various parts of their body;move from one location to another; sleep; orbecome active. All of these activities come un-der the rubric of behavioral thermoregulation. Ofcourse, rats also mobilize a host of more re-flexive behaviors such as shivering, piloerec-tion, peripheral vasoconstriction, and brownadipose tissue activation to generate and con-serve heat and peripheral vasodilation to loseit. It is apparent that the range of behaviorsthat can be mobilized far outnumbers theavailable reflexive responses. This is whatmakes the study of thermoregulation one ofthe most challenging and interesting areas ofinvestigation in the behavioral sciences.

Thermoregulatory behavior is also rele-vant to a wide range of other research areas,both those concerned with normal regulatoryfunctions and those related to pathologicalconditions. For instance, (1) all metabolicfunctions of the body are affected by bodytemperature, and regulatory mechanisms thatcontrol body fat and its metabolism affectbody temperature (Collins et al., 2001). (2) Inthe process of investigating neural events thatsupposedly related to learning and memory,on a number of occasions scientists havethought that they had discovered a centralcorrelate only to find later that the change wasdue to normal changes in body temperature(Anderson and Moser, 1995). (3) While inves-tigating compounds that might minimize theextent to which brain trauma produces brain

injury, scientists have believed that they dis-covered a therapeutic compound only to laterfind that the therapy resulted secondarilyfrom changes in body temperature (Corbettand Thornhill, 2000). Indeed, most doses ofmost drugs used to investigate the neu-ropharmacology of any behavior also affectbody temperature, and the effect on behaviormay be a secondary effect of the action of thedrug on body temperature (Satinoff, 1979). (4)Postures of thermoregulatory behavior maybe related to symptomology associated withsome pathological conditions (Schallert et al.,1978). (J) Some behaviors, such as lordosis ininfant rat pups, depend on body temperatureand may not be displayed when body tem-perature is too high or too low (Leonard,1987; Satinoff, 1991).

An infant rat is an ectotherm: it does notgenerate heat internally and is largely de-pendent for homeostasis on the environmen-tal temperature in the nest. An adult rat is anendoderm: it regulates its body temperature in-ternally over a broad range of ambient condi-tions. This transition occurs over the first 2months of life as the infant rat gains mobility,body hair, and body mass. Nevertheless, be-cause of its small surface-to-mass ratio, rats arealways threatened by temperature extremes,which make them highly motivated to escapechallenges to normal body temperature.

Over all ages, rats can survive wide fluc-tuations in body temperature, from as low as18° C, a temperature at which bodily func-tions almost cease and at which surgery canbe performed without anesthesia (Arokina et

226

Page 244: 0195162854

Chapter 21. Thermoregulation 227

al., 2002), to as high as 41° C, which is justshort of body temperature at which heatstroke and associated physical damage occur(Lin, 1999) The purposes of this chapter areto sketch behavioral and reflexive thermoreg-ulatory behaviors in adult and immature rats,to describe some of the methods of studyingthermoregulatory behavior, and to outlinesome of the neural mechanisms that mediatetemperature regulation.

THERMONEUTRALITY AND THECONCEPT OF SET POINT

For rats, as for all animals, there is a range ofambient temperatures at which the basal rateof the animal's own heat production equalsthe rate of heat lost to the environment andat which a minimum amount of thermoregu-latory effort is required to maintain a constantbody temperature. The most accurate defini-tions of ambient thermoneutrality are based onboth heat loss and heat production responses.Thus, a "zone of least thermoregulatory ef-fort" can be bounded on the low end by ac-tivity that will raise body temperature and onthe high end by activity that will lower bodytemperature (Satinoff and Hendersen, 1977).

There is some debate over the range ofthermoneutrality for adult rats. It has been de-scribed as being as wide as 18° to 28° C (Pooleand Stephenson, 1977) and as narrow as 29.5°to 30.5° C (Szymusiak and Satinoff, 1981; Ro-manovsky et al., 2002). (Of course, as withanything in science, everything depends onthe measurements used to derive the results.)That is, within this temperature range, ratscontinue with ongoing behavior while not ini-tiating activities to be used specifically to reg-ulate their body temperature. This toleranceis in part related to just what they are doingat the time. While resting in a home cage,their body temperature may fall as low as35° C, whereas when engaged in a strenuousphysical activity, such as solving a problem ina maze or voluntarily running in an activity

wheel, their body temperature may rise ashigh as 41° C. That rats tolerate temperatureswithin this range before actively defendingtheir temperature suggests that body temper-ature within this range defines the rat's com-fort zone. When given a choice, however, ratsmay prefer a much narrower range of bodytemperatures, a range that can be referred toas the thermoneutral zone, or zone of thermalcomfort.

Because rats and other animals regulatetheir body temperature around a relativelyconstant value, this system has been usefullydescribed by control theory. This regulatedbody temperature is referred to as the "setpoint." In engineering terms, set point is thevalue of the input at which the output is zero(Fig. 21-1). Behaviorally defined, set point isthe value of body temperature that an animalwill defend—the reference, or desired, or op-timal body temperature—and it can only beinferred. It is sometimes erroneously assumedto be encoded in the neural structure of ther-moregulatory nuclei in the hypothalamus.But any discussion of the brain structures in-volved in thermoregulation is an anatomicalmodel and set point has no anatomical corre-lates: it is strictly a useful descriptive device.

There are a number of conditions inwhich set point changes, that is, conditions inwhich animals will defend a different bodytemperature.

1. Circadian rhythms. Rats are active during thenight portion of the day-night cycle and sleepmainly during the light portion of the cycle.They allow their body temperature to reachhigh levels before they defend it; at night theyare primarily interested in eating, drinking, andbeing active. Rats mainly sleep in the light pe-riod of a light-dark cycle; their body tempera-ture falls in the light and they allow this to hap-pen (i.e., they do not defend the drop). Thismay be adaptive in that it allows them to in-crease metabolic rate more easily at a timewhen they are searching for and consumingfood and to decrease metabolic rate at a timewhen they are resting.

Page 245: 0195162854

228 REGULATORY SYSTEMS

Figure 21-1. Control diagram for behavioral thermoregulation outlining what is involved when a rat wantsto raise its body temperature. The learned response, such as pressing a lever to turn on a heat lamp, or mov-ing to a warmer environment, occurs when there is a discrepancy (the error signal) between the ideal bodytemperature (the setpoint) and the actual body temperature. The error arises from a disturbance. (This couldbe a cold environment, cooling the brain, or a change in immunological compounds that give rise to fever.)The actual body temperature is monitored and the feedback loops maintain the response at a level appro-priate to the existing discrepancy. The lower feedback path carries information about actual body tempera-ture that is compared with the set temperature at the comparator, or signal mixer. The upper pathway ad-justs the parameters of the response mechanisms in terms of response cost and response effectiveness andoptimizes the effectiveness of the system. Low effectiveness or high response cost might be expected to lowerthe slope of the function relating error to response.

2. Age. When allowed to choose their thermal en-

vironment in a thermally graded alleyway, old

rats prefer a warmer ambient temperature than

do young ones (Florez-Duquet et al., 2001). Of-

ten, the circadian variation in body temperature

is lower in old rats than in young ones (Fig. 21-2).

3. Hypoxia. In rats, resting oxygen uptake changes

about 11% for every 1 ° C change in Tb. There-

fore, when oxygen is in short supply, such as

during hypoxia, a high body temperature could

be injurious. Under hypoxic conditions, rats de-

fend a lower body temperature. This appears to

Figure 21-2. The diurnal fluctuations of body temperature and the temperature the rats selected over 48hours in a thermal gradient. The mean body temperature (measured by telemetry) of the young rats variedfrom about 37.7° to 38.3° C. At the same time they selected ambient temperatures between about 24° and31° C. Note that at night, when the rats were active, and generating more internal heat, they preferred am-bient temperatures lower than in the daytime, when they were mostly sleeping. Body temperatures of oldrats (average 24 months of age) was a little lower than those of the young rats, and they preferred signifi-cantly higher temperatures.

Page 246: 0195162854

Chapter 21. Thermoregulation 229

be adaptive because it lowers metabolic rate,and thus the need for oxygen (Wood and Gon-zalez, 1996).

4. Hormonal state. Regulated temperature varieswith reproductive condition. Female rats in theluteal stage of their estrous cycle appear to havean upward shift in set point and a lowered shiftduring the follicular phase. This shift may be re-lated to changes in levels of progesterone orestrogen, and possibly other nongonadal hor-mone. Because female rats are extremely activeduring the luteal stage, this change may facilitatea higher level of metabolism (Kittrell and Sati-noff, 1988). Because female rats cycle every 4 to5 days, thermoregulatory changes can pose achallenge to experiments in which thermoregu-latory behavior is not a primary concern.

5. Fever. Fever is a condition in which all behav-ioral and reflexive thermoregulatory responseswork in parallel to raise body temperature to anew, higher set level (see Fig. 21-1). It is nor-mally triggered by components of pathogenicorganisms like bacteria, viruses, and fungi,which initiate a cascade of events that ultimatelyresult in the release of brain prostaglandins(Ranels and Griffin, 2003). Correlational studiessuggest that the febrile response has survivalvalue in that it aids in the destruction of thepathological organism (Dantzer, 2001).

6. Other. The set point is apparently changed withmany other agents and conditions; these in-clude the effects of many drugs such as salicy-lates (Satinoff, 1972), stress (Peloso et al., 2002),and brain injury (Satinoff and Prosser, 1988).

THE NEURAL CONTROL OFTEMPERATURE REGULATION

In the late 1800s, investigators discovered thatthermoregulation could be deranged afterbrain damage, and by the 1930s Ranson (1935)used stereotaxic surgery to localize a "heat-loss center" to the preoptic/anterior hypo-thalamus and a "heat-production center" tothe posterior hypothalamus. Subsequent workindicated that thermoregulatory behaviors areinitiated when the preoptic area is heated orcooled or is locally injected with neurotrans-mitter-related compounds. In addition, hypo-

thalamic thermosensitive cells alter their firingrats with changes in their own temperature orchanges in the temperature in other parts of thebody. Thus, substantial early work pointed tothe preoptic/anterior hypothalamus as thebrain "thermostat." However, the concepts ofcenters and thermostats in the brain are muchtoo simplistic. The hypothalamus is impor-tantly involved in a thermal control systemthat involves many parts of the brain, fromthe cortex to the spinal cord. Furthermore, al-though large lesions in this area disrupt ther-moregulation, they do not prevent animalsfrom maintaining their body temperature atnear-normal levels if they have the behavioralresponses to do so. Temperature regulation istoo important to the life of an animal to becontrolled by a simple thermostat (Satinoff,1978, 1983).

THE DEVELOPMENT OFTHERMOREGULATORY BEHAVIOR

Newborn rat pups are blind and hairless andhave reduced immobility at birth. The key prob-lems they face are a lack of thermal insulation,in terms ofboth fur and subcutaneous fat stores,and an unfavorable surface-to-mass ratio. Theyhave a narrow thermoneutral zone rangingfrom 34° to 35° C at birth, which gradually in-creases to the adult range over the next 4 to 6weeks. These ambient temperatures are muchhigher than the usual laboratory temperaturesof 21° to 24° C. Thus, for infant rats, ther-moneutrality consists of temperatures thatwould produce heat stress in adult rats. Al-though brown fat can be activated as early asthe first postnatal day (Kortner et al., 1993) andinfant rats can increase their metabolic ratefrom 10 days of age, or sooner (Nuesslein-Hildesheim and Schmidt, 1993), this capabilityis practically useless because they cannot con-serve heat. Although it might be thought thatinfant rats would be better off drifting into hy-pothermia in the cold, hypeithermic rats de-velop much more slowly that do euthermic rats

Page 247: 0195162854

230 REGULATORY SYSTEMS

(Stone et al., 1976). Therefore, it is desirable forinfant rats to maintain a relatively high bodytemperature. The implications for behavioristsis that any work directed toward studying thebehavior of infant rats requires an ambient tem-perature above that required for the study ofadult rats.

As mentioned earlier, infant rats increasemetabolic rate in response to cold stress,mostly or completely by generating heat intheir brown adipose tissue. However, becauseof their small mass, they quickly become hy-pothermic at ambient temperatures belowtheir thermoneutral zone, and metabolic re-sponses fail. Due largely to difficulties in meas-urement, the extent to which infant rats cancontrol temperature by peripheral vasocon-striction and vasodilation is uncertain. It islikely that the major improvement in tem-perature regulation over the first three weeksof life is due to the development of thermal

insulation (Conklin and Heggeness, 1971).Thus, rat pups are ectothermic (controlled byexternal temperature) and become endother-mic (contribute to the control of their ownbody temperature) only after they developfur. Interestingly, exposing rat pups to a coolenvironment hinders their pelage develop-ment compared with exposing them to awarm environment (Gerrish et al., 1998).

Pups can behaviorally thermoregulate ifthey do not become incapacitated by the cold.Rat pups will seek heat as early as 1 day afterbirth. In thermally graded alleys, rat pups ori-ent and move along the thermal gradient fromcool to warm if they are not immobilized bythe cold (Kleitman and Satinoff, 1982) (Fig.2-3). This result is not surprising because ratpups are born in litters, and litters regulatetheir body temperatures by active group hud-dling. Rat pups in a huddle move from the topof the pile down into it when too much of

Figure 21-3. Positions of 1-day-oldrat pups in a thermal gradient at 10minute intervals for 120 minutes.

Floor Temperature (°C)

9 Day Old Rats

Page 248: 0195162854

Chapter 21. Thermoregulation

their surface area is exposed and they becomecool and they move out to the edges whenthey become too warm (Alberts, 1978). In fact,huddling in rat pups is almost exclusively ther-mally directed during the first week of life.Huddling leads to lower oxygen consumptionin the members of the huddle than if theywere alone, largely because huddled animalsexpose less surface area to the cool surroundand thereby lose less heat.

Thus, newborn rats can only produce andconserve heat behaviorally—by changingtheir position in the huddle or orienting andgoing toward other littermates when they arescattered around the nest side. Huddling orsocial aggregation continues to provide heatcomfort into adulthood, as rats housed ingroups may spend a large portion of their rest-ing time in a pile.

When rat pups are tested at warm ambi-ent temperatures, one can see the earlier ap-pearance of behaviors that seemingly havenothing to do with thermoregulation. For in-stance, there had been reports that in both fe-males and males primed with estradiol ben-zoate, female sexual behaviors—in particular,lordosis and ear wiggling—occurred no ear-lier than 2 to 3 weeks of age. However, whenWilliams (1987) tested pups at 33° to 35° C,both responses were seen in pups of either sexat 4 to 6 days of age. Rat pups also feed in re-sponse to food deprivation and tail pinch,drink in response to angiotensin injections,and respond to odor conditioning, but only ifthey are tested within their thermoneutralzone (Satinoff, 1991). Finally, neural struc-tures have also been found to function muchearlier in rat pups kept within their ther-moneutral zone than in pups that are slightlycool (Horwitz et al., 1982).

BEHAVIORAL METHODS

The measurement of body temperature is notsimple because temperature can vary widelywithin the body. Which temperature to meas-

231

ure depends on what question is being asked.No one really knows what is the "regulatedtemperature." In rats, it could be core or skintemperature or some core-to-skin gradient.

Temperature can be measured on pointson the periphery of the body, such as the tail,which is highly vascularized and thereforeoptimal for losing or conserving body heat(Owens et al., 2002). Temperature can bemeasured within the body cavity, and thistemperature is referred to as core temperature.Core temperature provides a relatively stablevalue that is least affected by environmentaltemperature fluctuations or local metabolicactivity. Temperature can also be measuredwithin the brain. Within the brain, it reflectsboth the core temperature of the body and thetemperature generated by the activity of theneural structure within which the temperature-measuring device is inserted (DeBow and Col-bourne, 2003).

Temperature can be measured with awide range of devices. One commonly usedmethod involves inserting a temperature-measuring probe at least 6 cm into the rectumto measure core temperature. This can be doneacutely, while the rat is restrained in the hand,or chronically, if the animal is placed in a re-straining box so that it cannot chew the probe.Manual insertion of temperature probes has theadvantage of being simple, but the process ofhandling and restraining the rat will quicklyand reliably produce a stress-related increasein body temperature (Eikelboom and Stewart,1982).

There are several ways of sensing tem-perature, but the most common is to use a de-vice that contains a thermistor, which is asemiconductor device made of materialswhose resistance varies as a function of tem-perature. A recorder senses current changesas a function of changes in resistance.

Temperature probes have a number ofadvantages. The size of probes can vary frommillimeters to microns, which make themuseful for receding skin, core, or brain tem-perature. They can also be used acutely or

Page 249: 0195162854

232 REGULATORY SYSTEMS

chronically implanted. They can be quicklyand painlessly inserted into the rectum orplaced on the skin to obtain a temperaturereading, or they can be chronically fixed inplace. They are also relatively inexpensive.The drawbacks of temperature probes are sev-eral: (1) a recording lead must be attached tothe rat, which may limit the animal's freedom;(2) the rat may be able to reach the lead andchew it off; and finally, (3) even though it is amild stressor, it is a stressor, and that may al-ter normal body temperature.

For long-term recordings, a transmittercan be implanted in an animal's body cavityand an AM radio signal from the transmittercan be recorded via computer. A signal can berecorded for periods of weeks or months, aslong as the batteries on the transmitter func-tion. These transmitters come in various sizes,and some can be inserted into the peritonealcavity of rats that are at least 10 days old.

Measures of changes in skin temperature,which can indicate vasoconstriction and va-sodilatation, can be obtained with thermo-couples placed on the skin, such as on the tail,and shielded from air temperature fluctua-tions with insulating tape (Romanovsky et al.,2002). Tail skin temperature can also be meas-ured noninvasively with radiotelemetry (Gor-don et al., 2002). Temperature can also bemeasured by painting a rat, usually on the tail,with thermosensitive liquid crystal paint, al-though this method requires video recordingand calibrating color changes (Romanovsy atal., 2002).

Shivering can be visually observed, re-corded with a movement sensor (Harrod etal., 2002) or recorded using electrical record-ing of muscle activity (Whishaw and Vander-wolf, 1971). Finally, metabolic processes re-lated to thermogensis can be inferred bymeasuring evaporative water loss and/or oxy-gen use in animals placed in calorimeters(Buchanan et al., 2003).

Several behavioral measures have beendeveloped to study temperature regulation inrats. Operant methods present rats with a

lever that they can press to control ambienttemperature. The lever may trigger warm orcold airflow, warm or cold water spray, theonset of radiant heat, and so on. Operantmethods tells an investigator how motivateda rat is to work to control its temperature. Ifall that one wishes to know is the ambienttemperature a rat prefers, it is much easier, onboth the investigator and the rat, to place therat in a thermally graded alleyway, where itcan choose to move to the place it finds ther-mally comfortable. If bedding material isavailable, rats will build nests in a cold envi-ronment and they will modulate food carry-ing and food consumption in response to am-bient temperature changes.

As a cautionary note, some investigatorshave attempted to decrease core tempera-ture by placing rats in a cold environment,but for adult rats, heat-generating mecha-nisms are so effective that they actually raisebody temperature in response to acute coldstress. Core temperature can be reducedmore effectively by lightly spraying a rat's furwith cool water or placing a rat in a coolbath. Because rats have an unfavorable bodysurface to mass ration, their heat generatingmechanisms are unequal to the challenge ofcold-water stress.

STRATEGIES FORTEMPERATURE REGULATION

HEAT STRESS

Rats use four different behaviors when chal-lenged with heat, such as radiant heat orwarm airflow in which environmental tem-perature increases from about 26° C to 41 ° C.These responses are mediated both by activ-ity in peripheral thermoreceptors and by cen-tral cooling of hypothalamic thermoregula-tory regions via blood flow.

1. Rats have a number of reflexive responses, in-cluding vasodilatation, especially of the hairlesstail, to induce body cooling.

Page 250: 0195162854

Chapter 21. Thermoregulation 233

2. Their initial behavioral response is to becomemore active by walking around and rearing.

3. If escape is not possible, rats engage in intensegrooming during which they spread saliva overtheir body to assist evaporation-related cooling.Saliva spreading can be so intense that pro-longed grooming can result in dehydration.

4. Faced with continued heat stress, animals giveup all activity and postural support and lieprone, thus reducing heat production frommetabolic and muscle activity. The extensionposture consists of relaxation and elongation ofthe body in a prone position, with the normallycurved spine straightened horizontally, forelegsplaced under the neck, and hindlimbs foldedoutward and toward the rear. At first the headremains erect but later it rests on the floor, withthe eyes partly closed and retracted (Roberts etal, 1974).

COLD STRESS

Rats have at least five different coping strate-gies when challenged with cold stress, such ascold airflow or body wetting.

1. They display reflexive responses, including pe-ripheral vasoconstriction, especially of the tail,and piloerection of the fur, to increase its insu-lation value.

2. Their initial behavioral response is to becomemore active by walking around and rearing.

3. If escape is not possible, animals increase heatproduction using nonshivering thermogenesis.This consists of energy expenditure and heatproduction obtained by metabolizing brownadipose tissue (brown fat) in response to stim-ulation from the sympathetic nervous system.This response is very effective: total body en-ergy expenditure doubles within minutes of adose of a beta-adrenergic agonist.

4. Under continued cold stress that results in alowering of core temperature, shivering ther-mogenesis is initiated. This consists of the adop-tion of a hunched body posture with posturalsupport in which agonist and antagonist mus-cles contract concurrently, thus generating heatin the absence of locomotion. Shivering can bemild or so violent that an animal can be thrownoff balance. Energy expenditure from brown fatand shivering can be measured indirectly as

oxygen consumption, which requires that ani-mals be tested in a sealed environment in whichairflow is controlled.

5. Fur condition contributes to thermoregulation,and rats have a gland called the Harderian glandthat is adjacent to the eye and secretes Harder-ian liquid (Buzzell, 1996). The liquid is releasedwhen the paws apply pressure around the eyesocket during grooming. The role of Harderianmaterial in control of temperature in rats hasnot been investigated, but it may be similar tothat described for the Mongolian gerbil (Men-ones unguiculatus). For the gerbil, Harderian re-lease decreases at high body temperature (whensaliva production increases) and increases atlow body temperature. This relationship sug-gests that Harderian material improves insula-tion of the pelage (Thiessen, 1989).

ACKNOWLEDGMENTS

Supported by National Institute of Mental Health grant R01-MH41138.

REFERENCES

Alberts JR (1978) Huddling by rat pups: Group behav-ioral mechanisms of temperature regulation and en-ergy consumption. Journal of Comparative andPhysiological Psychology 92:231-235.

Andersen P and Moser El (1995) Brain temperature andhippocampal function. Hippocampus 95:491-498.

Arokina NK, Potekhina IL, Volkova MF (2002) Devel-opment of deep hypothermia in rats with limitedmotor activity. Rossiiskii Fiziologicheskii ZhurnalImeni I.M. Sechenova/Rossiiskaia Akademiia Nauk88:1477-1484.

Buchanan JB, Peloso E, Satinoff E (2003) Thermoregu-latory and metabolic changes during fever in youngand old rats. American Journal of Physiology (Reg-ulatory and Integrative Comparative Physiology)285:R1165-R1169.

Buzzell GR (1996) The Harderian gland: Perspectives.Microscope Research Techniques 34:2-5.

Collins S, Cao W, Daniel KW, Dixon TM, MedvedevAV, Onuma H, Surwit R (2001) Adrenoceptors, un-coupling proteins, and energy expenditure. Experi-mental Biology and Medicine 226:982-990.

Conklin P and Heggeness FW (1971) Maturation of tem-perature homeostasis in the rat. American Journalof Physiology 220:333-336.

Corbett D and Thornhill J (2002) Temperature modu-

Page 251: 0195162854

234 REGULATORY SYSTEMS

lation (hypothermia and hyperthermic conditions)and its influence on histological and behavioral out-comes following cerebral ischemia. Brain Pathology10:145-152.

Dantzer R (2001) Cytokine-induced sickness behavior:Mechanisms and implications. Annals of the NewYork Academy of Sciences 933:222-234.

DeBow S and Colbourne F (2003) Brain temperaturemeasurement and regulation in awake and freelymoving rodents. Methods 2:167-171.

Eikelboom R and Stewart, J (1982) Conditioning of drug-induced physiological responses. Psychological Re-views 89:507-528.

Florez-Duquet M, Peloso E, Satinoff E (2001) Fever andbehavioral thermoregulation in young and old rats.American Journal of Physiology (Regulatory and In-tegrative Comparative Physiology) 280:R1457-R1461.

Gordon CJ, Puckett E, Padnos B (2002) Rat tail skin tem-perature monitored noninvasively by radioteleme-try: Characterization by examination of vasomotorresponses to thermomodulatory agents. Journal ofPharmacological and Toxicological Methods 47:107-114.

Harrod S, Metzger M, Stempowski N, Riccio D (2002)Cold tolerance: Behavioral differences followingsingle or multiple cold exposures. Physiology andBehavior 76:27-39.

Horwitz J, Heller A, Hoffmann PC (1982) The effect ofdevelopment of thermoregulatory function on thebiochemical assessment of the ontogeny of neona-tal dopaminergic neuronal activity. Brain Research235:245-252.

Kittrell EM and Satinoff E (1988) Diurnal rhythms of bodytemperature, drinking and activity over reproductivecycles. Physiology and Behavior 42:477-484.

Kleitman N and Satinoff E (1981) Thermoregulatory be-havior in rat pups from birth to weaning. Physiol-ogy and Behavior 29:537-541.

Kortner G, Schildhauer K, Petrova O, Schmidt I. (1993)Rapid changes in metabolic cold defense and GDPbinding to brown adipose tissue mitochondria of ratpups. American Journal of Physiology (Regulatoryand Integrative Comparative Physiology) 264:R1017-R1023.

Lin MT (1999) Pathogenesis of an experimental heat-stroke model. Clinical Experimental and Pharma-cological Physiology 26:826-837.

Nuesslein-Hildesheim B and Schmidt I (1994) Is the cir-cadian core temperature rhythm of juvenile rats dueto a periodic blockade of thermoregulatory ther-mogenesis? Pflugers Archives 427450-4.

Owens NC, Oootsuka Y, Kanosue K, McAllen RM (2002)Thermoregulatory control of sympathetic fibressupplying the rat's tail. Journal of Physiology (Lon-don) 543:849-858.

Peloso E, Wachulec M, Satinoff E (2002) Stress-inducedhyperthermia depends on both time of day and lightcondition. Journal of Biological Rhythms 17:164-170.

Poole S and Stephenson JD (1977) Body temperatureregulation and thermoneutrality in rats. QuarterlyJournal of Experimental and Cognitive Medical Sci-ences 62:143-149.

Ranels HJ and Griffin JD (2003) The effects ofprostaglandin E2 on the firing rate activity of ther-mosensitive and temperature insensitive neurons inthe ventromedial preoptic area of the rat hypothal-amus. Brain Research 64:42-50.

Ranson SW (1935) The anatomy of the nervous systemfrom the standpoint of development and function.Philadelphia: W.B. Saunders Co.

Roberts WW, Mooney RD, Martin JR (1974) Ther-moregulatory behaviors of laboratory rodents. Jour-nal of Comparative and Physiological Psychology86:693-699.

Romanovsky AA, Ivanov AI, Shimansky YP (2002) Am-bient temperature for experiments in rats: A newmethod for determining the zone of thermal neu-trality. Journal of Applied Physiology 92:1-21.

Satinoff E (1972) Salicylate: Action on normal body tem-perature in rats. Science 176:532-533.

Satinoff E (1978) Neural organization and evolution ofthermal regulation in mammals. Science 201:16-22.

Satinoff E (1979) Drugs and thermoregulatory behavior.In: Body temperature, drug effects and therapeuticimplications (Lomax P and Schonbaum E, eds.), pp.151-181. New York: Marcel Dekker.

Satinoff E (1983) A reevaluation of the concept of thehomeostatic organizatin of temperature regulation.In: Handbood of behavioral neurobiology (SatinoffE and Teitelbaum P, eds.), pp. 443-467. New York:Plenum Press.

Satinoff E (1991) Developmental aspects of behavioraland reflexive thermoregulation. In: Developmentalpsychobiology: New methods and changing con-cepts. (Shair HN, Barr GA, Hofer MA, eds.), pp.169-188. New York: Oxford.

Satinoff E and Hendersen R (1977) Thermoregulatorybehavior. In Handbook of operant behavior (HonigWK and Staddon JER, eds.), pp. 153-173. Engle-wood Cliffs, Colo.: Prentice-Hall.

Satinoff E and Prosser RA (1988) Suprachiasmatic nu-clear lesions eliminate circadian rhythms of drink-ing and activity, but not of body temperature, inmale rats. Journal of Biological Rhythms 3:1-22.

Schallert T, Whishaw IQ, DeRyck, Teitelbaum P (1978).The postures of catecholamine-depletion catalepsy:Their possible adaptive value in thermoregulation.Physiology and Behavior 21:817-820.

Stone EA, Bonnet KA, Hofer MA (1976) Survival and de-

Page 252: 0195162854

Chapter 21. Thermoregulation 235

velopment of maternally deprived rats: Role of bodytemperature. Psychosomatic Medicine 38:242-249.

Szymusiak R and SatinoffE (1981) Maximal oxygen con-sumption defines a narrower thermoneutral zonethan does minimal metabolic rate. Physiology andBehavior 26:689-690.

Thiessen GM (1989) The possible interaction of Hard-erian material and saliva for thermoregulation in theMongolian gerbil, Meriones unguiculatus. Percep-tion Motor Skills 68:3-10.

Whishaw IQ and Vanderwolf CH (1971) Hippocampal

EEG and behavior: Effects of variation in body tem-perature and relation of EEG to vibrissae move-ment, swimming and shivering. Physiology andBehavior 6:391-397.

Williams CL (1987) Estradiol benzoate facilitates lordo-sis and ear wiggling of 4- to 6-day-old rats. Behav-ioral Neuroscience 101:718-723.

Wood SC and Gonzales R (1996) Hypothermia in hypoxicanimals: Mechanisms, mediators, and functional sig-nificance. Comparative Biochemistry and PhysiologyB Biochemistry and Molecular Biology 113:37-43.

Page 253: 0195162854

Stress

JAAP M. KOOLHAAS, SIETSE F. DE BOER,AND BAUKE BUWALDA

22

A considerable part of our current knowledgeof stress and the pathophysiology of stress-re-lated disorders is based on experimental stud-ies in rats. The scientific rationale of thesestudies is that the mechanisms underlyingstress and adaptation have a common biolog-ical basis in animals and humans. Despite thewealth of data and publications on stress re-search in the rat and the wide variety of stressparadigms, one may criticize the validity ofsome of these studies. Studies often seem tofail in particular with respect to their face va-lidity, which means that the model fails to suf-ficiently mimic both the etiology and thesymptomatology of human stress-related dis-orders. For example, many animal studies usestressors that bear little or no relationship tothe biology of the species, that is, to the situ-ations an animal may encounter in its every-day life in a natural habitat. If we want to im-prove and refine our understanding of thecausal mechanisms of stress-related disorders,we need behavioral similarity models that ex-perimentally exploit the common biologicalbasis of animals and humans.

In this chapter, the biology of the rat andits natural defense mechanisms are used as astarting point in the evaluation and descrip-tion of stress models in the rat. In other words,we will focus at tests that explore the capac-ity of rats to cope with ecologically relevantproblems. The individual capacity to deal witheveryday problems in the natural environ-ment is considered to be one of the drivingforces of evolution and speciation. Organismshave become adapted to a dynamic and com-

plex natural environment in which they haveto find their food, deal with conspecifics, orreact to changes in climate. The capacity tocope with environmental challenges largelydetermines the individual survival in the nat-ural habitat. In the course of evolution, ani-mals have developed a wide variety of defensemechanisms to deal with such environmentalchallenges. Central in the biology of the rat isits social nature. Several studies in free rang-ing social groups of animals indicate that thestability of social environment is an importantfactor in health and disease. Unstable socialgroups and failure of social adaptive capacitiesmay lead to serious forms of stress pathology.This is reflected in the relationship betweenposition in the social hierarchy and the inci-dence of certain stress pathologies.

The first studies on the relationship be-tween social position and stress pathologymainly concentrated on cardiovascular disease.For example in mice (Ely, 1981; Lockwood andTurney, 1981; Henry and Stephens-Larson,1985), rats (Henry et al., 1993; Fokkema et al.,1995), and monkeys (Manuck et al., 1983), it hasbeen demonstrated that hypertension and car-diovascular abnormalities are more frequent insocially unstable groups and occur predomi-nantly in the dominant and subdominant malesof the social group. Stomach ulcers are mainlyfound in social outcasts of colonies (Calhoun,1962; Barnett, 1987). Similar observations on therelation between social position and pathologywere reported with respect to immune sys-tem-mediated diseases (Spencer et al., 1996; Ste-fanski et al., 2001).

236

Page 254: 0195162854

Chapter 22. Stress 237

Before embarking on a description of thevarious stress paradigms used in rats, we dis-cuss some issues that are fundamental to allstress models.

STRESS

Central to modern stress research are theterms controllability and predictability. Theseterms date back to a series of experiments byWeiss in the late 1960s (Weiss, 1972). Hedemonstrated that it is not so much the phys-ical nature of an aversive stimulus that in-duces stress pathology but rather the degreein which the stimulus can be predicted andcontrolled. Rats that cannot predict and haveno control over a stressor, such as an electricshock, appeared to have severe damage to thestomach wall, show signs of immunosuppres-sion (Keller et al., 1981; Weiss et al., 1989), andhave the largest rise in plasma corticosterone.Using this type of evidence, the following def-initions of stress can be formulated:

Acute stress is the state of an organism after asudden decrease in the predictability and/orthe controllability of relevant environmentalfactors.

Chronic stress is the state of an organism that oc-curs when relevant environmental aspects havea low predictability and are not, or not very well,controllable over a long period of time.

Although the concept of controllability andpredictability has strongly contributed to thepresent knowledge and insights into the de-velopment of stress pathology, we would liketo make some remarks on the way in whichthese concepts are generally used in an ex-perimental setting.

In most experiments, controllability isoperationally defined as a binary factor, thatis, as full control or complete loss of control.However, in everyday life situations, control-lability is generally graded from absolute con-trol via threat to control in various degrees toloss of control. Few studies consider the im-

portance of a different degree of control in thedevelopment of stress pathology. The impor-tance of such a distinction is demonstrated inexperiments aimed at understanding the de-velopment of hypertension. These experi-ments show that the crucial factor is threat tocontrol rather than loss of control (Koolhaas andBohus, 1989). Apart from the graded natureof controllability and predictability, the fre-quency and duration of stressors are also mat-ters of concern. Usually a distinction is madebetween acute and chronic stress. It is well ac-cepted that the chronic character of stressorsmay indeed result in various forms of stresspathology. However, chronic stress is not avery well defined concept. Many animal mod-els specifically aimed at chronic stress use a se-ries of intermittent acute stressors that maychange daily rather than using the continuouspresence of a stressor. Moreover, chronicstress studies rarely control for the factor timeafter the start of the stress procedure. The hu-man literature indicates that acute stressors ormajor life events may have long-term conse-quences, ultimately leading to a higher inci-dence of disease. Recent studies show that inrats, the experience of a single uncontrollableevent is sufficient to induce changes in be-havior and stress physiological parametersthat range from hours to days, weeks, ormonths (Koolhaas et al., 1997b).

INDIVIDUAL VARIATIONIN COPING STRATEGIES

A wide variety of medical, psychological, andbiological studies in both humans and animalsdemonstrate that individuals can differ con-siderably in their vulnerability for stress-re-lated disorders. Apparently, they differ in theircapacity to cope with environmental de-mands. Factors that have been shown to af-fect the individual coping capacity includegenotype, ontogeny, adult experience, age,social support, and so on. Individual differen-tiation in behavior and physiology is a well-

Page 255: 0195162854

238 REGULATORY SYSTEMS

known phenomenon in many animal species.Several attempts have been made to classifyindividual variation into personalities, tem-peraments, or coping styles that may predictthe response to environmental stressors. Thescarce literature in a number of feral animalspecies suggests that in nature, the dimensionof proactive and reactive coping strategies canbe distinguished (Koolhaas et al., 1999). Au-thors may use different terms to characterizephenotypes, such as shyness and boldness orproactive and reactive or active and passive,but they all seem to share the same basic char-acteristics. Proactive coping is characterizedby an active control of the environment (i.e.,active avoidance, offensive aggression, nestbuilding, etc.). Reactive coping, however, ischaracterized by a more readily acceptance ofthe environment as it is. Detailed analysis ofcoping strategies in rats and mice indicatesthat the most fundamental difference is the de-gree of behavioral flexibility. Reactive copingmales are flexible, whereas the proactive cop-ing is characterized by rigidity and routine for-mation. Recent studies in feral populations in-dicate that this differential flexibility may haveits origin in a differential survival value in na-ture. The challenge for the future is to un-derstand the functional significance of indi-vidual variation in nature by integratingethological, physiological, and ecological ap-proaches in the study of coping strategies.

From a biomedical point of view, theconcept of coping strategies implies that dif-ferent animals have a differential vulnerabil-ity to stress-related disorders. Negative healthconsequences might arise if an animal cannotcope with the stressor or requires very de-manding coping efforts. In view of the differ-ential neuroendocrine reactivity and neuro-biological make-up of proactive and reactivecoping animals, one may expect differenttypes of stress-pathology to develop underconditions in which a particular coping strat-egy fails. Indeed, there are indications thatcoping strategies differ in susceptibility to de-velop cardiovascular pathology, ulcer forma-

tion, stereotypies, depression, and infectiousdisease (Koolhaas et al., 1999).

STRESS MODELS

We discuss some stress models used in adultrats that challenge the defense mechanismsand hence call on the natural adaptive capac-ity of the animal. Most of these models in-volve stressors of the social and physical en-vironment. Few studies use food availabilityas stressor. This is surprising in view of thefact that the controllability and predictabilityof food are crucial to survival in nature andmay strongly challenge the adaptive capaci-ties. Indeed, both food restriction and in-creased caloric intake are reported to affectstress physiology (Rupp, 1999; Seres et al.,2002).

ACUTE STRESS MODELS

Acute Social StressAcute social stress can be studied in the resi-dent-intruder paradigm. This model is basedon the fact that a male rat will defend its ter-ritory against an unfamiliar male intruder.Territorial behavior develops within 1 weekwhen an adult male rat (i.e., >3 months ofage) is housed with a female rat in a large cageof about 0.5 m2. When, after this period, anunfamiliar male conspecific of the same strainand weight is introduced into the home cage,the resident male will attack the intruder anda fight develops. Usually, the resident is thevictor of this social interaction. This paradigmallows analysis of both the winner and theloser of the conflict.

Defeat. When the intruder animal is used asthe experimental animal, one can study theconsequences of social defeat or loss of socialcontrol as stressor. Social defeat by a maleconspecific induces an acute increase in heartrate, blood pressure, and body temperature;strong neuroendocrine responses in plasma

Page 256: 0195162854

Chapter 22. Stress

catecholamines, corticosterone, prolactin, andtestosterone; and changes in central nervousserotonergic neurotransmission (Koolhaas etal., 1997b; Berton et al., 1999; Sgoifo et al.,1999a). These responses, including the be-havioral reaction (flight, immobility), can beconsidered as part of the classic response toan acute stressor. However, a comparison ofa range of different stressors reveals that so-cial defeat may be one of the most severestressors measured in terms of the magnitudeof the corticosterone and catecholamine re-sponse (Koolhaas et al., 1997a). More impor-tant, however, is the time course of thesestress responses.

Recent studies using more chronicrecordings indicate that the various stress pa-rameters have a different time course. The car-diovascular and catecholaminergic response toa 1 hour social defeat diminishes within 1 or 2hours after the defeat, but the corticosteroneresponse lasts for longer than 4 hours. After aninitial rise, plasma testosterone drops belowbaseline levels and remains at extremely lowlevels for at least 2 days. A single social defeatappears to induce a reduction in the circadianvariation in body temperature, growth, sexualinterest, and open field exploration that maylast from 2 to 10 days after the social stress(Koolhaas et al., 1997b). Miczek et al. (1990)found changes in opiate analgesia that last forat least 1 month after the defeat. Althoughmany of these changes can be considered aspart of the symptomatology of human de-pression, we emphasize that social defeat in-duces changes in a variety of physiological andbehavioral parameters, each of which mayhave different temporal dynamics. Hence, anacute stressor may have chronic conse-quences. These lasting changes may be adap-tive but may just as well be considered as theearly signs of stress pathology. The bottomline is that the social defeat model allows fur-ther analysis of the changes in time of factorsknown to be involved in stress and adaptation.By manipulating the frequency, intensity, andtype of previous social experiences, one can

239

obtain insight into the (mal)adaptive nature ofthese changes.

Victory. By using the resident male in theresident-intruder paradigm as experimentalanimal, one can study the consequences ofthreat to control. Although the resident ulti-mately controls its social environment, this ispreceded by a certain degree of unpredictabil-ity and threat to control. This is clearly indi-cated by the fact that the stress response—interms of plasma corticosterone and cate-cholamines, heart rate, and blood pressure—isinitially almost as high as in the defeated in-truder, but these stress parameters rapidly re-turn to baseline levels as soon as the dominancerelationship becomes clear. Typical for the win-ner of the social interaction are the cardiovas-cular abnormalities observed in the electrocar-diogram immediately after the interaction.These abnormalities indicate a strong shift inthe autonomic balance toward high sympa-thetic dominance (Sgoifo et al., 1999b).

Defensive BuryingDefensive burying refers to the natural behaviorof rats of displacing bedding material with typ-ical alternating forward-pushing movements oftheir forepaws (threading or thrusting) andshoveling movements of their heads directedat localized sources of unfamiliar, aversivestimulation/threat. Harmful and noxious ob-jects so buried include electrified prods (Treitet al., 1981), rat chow pellets coated with qui-nine, spouts of bottles containing unpleasanttasting liquids such as pepper sauce and liquidsto which the rats have developed taste aver-sion, flash cubes that discharge near them,tubes that direct airbursts into their faces or de-liver noxious smells, dead conspetifics, andpredators (De Boer and Koolhaas, 2003). Al-though different in form, function, and inten-sity, rats may also bury seemingly harmless ob-jects such as nonelectrified prods, flash cubesthat do not flash, and marbles (Treit, 1985). Byburying unfamiliar and/or harmful objects, in-dividuals can successfully avoid or remove

Page 257: 0195162854

240 REGULATORY SYSTEMS

aversive and possibly life-threatening dangersfrom their habitat. Together with flight, freez-ing, and certain forms of agonistic behavior, de-fensive burying constitutes the unconditioned(innate) species-specific defensive behavioralrepertoire in rats.

The procedure of the shock-prod defen-sive burying test is quite simple and basicallyunchanged since its original description byPinel and Treit (1978). In a test chamber (ei-ther the home cage or a familiarized test cageafter several habituation trials) with sufficientsuitable bedding material on the floor, sub-jects are confronted with a wire-wrappedprod/probe ((ft = 1 cm; 6 to 7 cm long) thatis inserted through a small hole 2 cm abovethe bedding in one of the test chamber walls.The noninsulated wires of the prod are con-nected to a shock source, and whenever thesubject touches the prod with its forepaws orsnout, it receives an electric shock (manuallyoperated or automatically delivered). Theprod either remains electrified during the en-tire test period or is deactivated after the firstcontact. After the first contact with the elec-trified prod, the animars behavior is observedand/or recorded on video for a 10 to 15minute test session. During this observationperiod, a variety of behaviors can be quanti-fied. The repertoire of behavioral reactions iswell delineated and catalogued in rats andmice (Tsuda et al, 1988; De Boer et al, 1991),and methods for its reliable measurementhave become standard equipment in behav-ioral-physiological and pharmacological labo-ratories. It is important to notice that there isa large individual variation in the amount oftime spent burying. Because burying behav-ior is generally negatively correlated with im-mobility behavior, it seems that this differen-tiation is based on a differential expression ofanxiety (De Boer and Koolhaas, 2003).

PredatorRats readily display various forms of defensivebehavior when confronted with a predatorsuch as a cat. Because even rats that never pre-

viously saw a cat show the response and be-cause the response does not habituate, it isconsidered to be an evolutionary ancient andinnate reaction. Although a live cat producesa much stronger response, in a laboratory ex-perimental setting, usually the odor of a cat isused. Cat odor can be presented to the rat byfirst rubbing a cat with cotton wool for a stan-dard period of time and then putting the woolon top of the rat's cage. Dielenberg and Mc-Gregor (2001) developed another way of pre-senting cat odor to a rat. They used a fabriccollar that had been worn by a cat for 3 weeksand presented this in a specific "cat odoravoidance" apparatus. This test cage allowsthe measurement of various avoidance be-haviors and risk assessment behavior. It is im-portant to notice that several studies now in-dicate that the response to cat odor differsfrom the response to fox odor. Fox odor,which can be obtained as the synthetic com-pound trimethylthiazoline (TMT), seems toact more as a generally aversive stimulus(McGregor et al., 2002; Blanchard et al., 2003).

NoveltyNovelty is often used as a minor stressor. Itusually does not induce a maximal stress re-sponse and for that reason is suitable to de-termine stress reactivity. There is a consider-able individual difference in the tendency ofrats to explore a novel environment or novelobjects that is more generally related to cop-ing style and hypothalamus-pituitary-adrenalaxis reactivity (Steimer and Driscoll, 2003).This individual differentiation is considered tobe an animal model for the sensation-seekingtrait in humans (Delhi et al., 1996). There arevarious ways to measure the response to nov-elty. One can introduce a novel object into thehome cage of the experimental animal andmeasure the behavioral and neuroendocrineresponse. However, the response to cleanbedding material can also be used as a stan-dard novelty stress. In this way one takes theadvantage of the fact that cages have to be reg-ularly cleaned anyway.

Page 258: 0195162854

Chapter 22. Stress 241

Forced SwimmingFeral rats, in contrast to mice, often live closeto water and like to swim. They often swimvoluntarily, and the activation of the neu-roendocrine stress response is mainly relatedto the increase in physical activity duringswimming. Porsolt et al. (1977) was the firstto use forced swimming in a tank that had nopossibilities for the animal to escape, as a stres-sor to induce signs of depression. Indeed, theforced swim test strongly activates the neu-roendocrine stress response; when the animalis replaced in the test apparatus 24 hours later,they more readily give up attempts to escape.Without the opportunity to escape, the phys-iological stress response appears to be unre-lated to the physical activity (Abel, 1994a,1994c). The forced swim test was subse-quently somewhat modified and is nowwidely used as a test to measure the efficacyof potential antidepressant drugs. Briefly, thetest uses a cylindrical tank with a diameter ofabout 25 cm and a height of about 50 cm filledwith 30 cm of water at 25° C (Abel, 1994b).Three main types of behavior can be observedwhen the rat is put into the water: swimming,attempts to climb the wall, and floating. Thisfloating behavior increases over the test timeand is generally considered as a form of de-spair in which the animal has given up its at-tempts to escape. However, one may also ar-gue that the animal has two alternative waysto cope with the situation by either activelytrying to escape or quietly floating on the wa-ter surface and saving the energy of useless es-cape attempts. A recent study using proactiveand reactive coping mice confirmed the ideathat the test seems to measure differentialways of coping with an inescapable stressorrather than signs of despair and depression(Veenema et al., 2003). Indeed, treatmentwith an antidepressant drug reduced floatingin the reactive coping male as predicted by awealth of pharmacological literature, but thesame treatment also reduced climbing andswimming in the proactive coping male (un-published observation).

CHRONIC STRESS MODELS

Social StressBecause of its permanent nature, the social en-vironment is frequently used as a way to in-duce naturalistic chronic stressors. The vari-ety of models used to study chronic socialstress may differ in complexity and degree ofexperimental control.

Social Groups. Social stress can be studied inits most complex form using groups orcolonies of rats. This allows an analysis of therelationship between the position of individu-als in the social hierarchy and stress parame-ters. The design of cages used for coloniesmay range from large outdoor enclosures tocages of several square meters with nest boxesor much smaller cages without any further fa-cilities. One of the most extensively studiesparadigms is the visible burrow system de-veloped by Blanchard et al. (1995). The visi-ble burrow system (VSB) consists of a largecentral square and some nest chambers thatare connected to the central square by meansof Plexiglas tunnels. Colonies generally con-sist of four adult males and two adult femalesto facilitate territoriality and aggression. Blan-chard et al. (1995) observed clear signs ofchronic stress in the subordinate males in par-ticular, whereas dominant males developedsymptoms of stress only when the group con-sisted of a relatively large number of highlyaggressive males. The advantage of such a sys-tem is that the model allows the animals toexpress their natural behaviors and defensemechanisms. The disadvantage is the limiteddegree of experimental control and the lim-ited possibilities for continuous monitoring ofphysiological parameters.

Social Instability. The degree of social stressin groups of rats depends mainly on the sta-bility of the social structure. In stable socialgroups, hardly any signs of stress pathologycan be observed. Therefore, researchers tendto increase stress by reducing the social sta-

Page 259: 0195162854

242 REGULATORY SYSTEMS

bility. This can be achieved by forming groupsconsisting of only aggressive males. This leadsto serious and regular dominance fights. How-ever, this procedure might give a considerablebias to the experimental results due to the se-lection procedure. Another way to manipu-late social stability is to mix groups on a reg-ular basis. Lemaire and Mormede (1995)successfully applied this method in rats. Theycould demonstrate that regular mixing of un-familiar groups leads to the development ofhypertension, in particular, in a more sociallyactive strain of rats.

Chronic Subordination. Living as a subordi-nate male in the presence of a dominant maleis generally considered to be a chronic stres-sor. Usually, subordination is studied in largergroups of males such as the visible burrow sys-tem mentioned earlier. Occasionally, groupsof three or even two males are used, living to-gether for a prolonged period of time. Subor-dination is determined by direct observationof the social interactions between the groupmembers.

In species like mice and tree shrews, a sen-sory contact model is frequently used. In thismodel, the subordinate male is housed to-gether with a dominant neighbor, but the twoanimals are separated by a wire mesh screen,allowing visual, auditory, and olfactory con-tact. Dominance relationships are establishedor reconfirmed by removing the wire meshscreen every day or week for a short period oftime (Fuchs et al., 1993; Veenema et al., 2003).Although this is a useful model in these species,we are not aware of its application in rats.

CONCLUDING REMARKS

The stress models described here have in com-mon that they somehow challenge the natu-ral defense mechanisms and hence call on theadaptive capacity of the animal. We wouldlike to emphasize the word natural, becauseits means that, on the basis of the specific evo-

lutionary biology of the species, one might ex-pect the animal to have an adequate answerto a given challenge. This distinguishes the se-lected stress models from models using footshocks, centrifuges, tail suspension, loudnoise, and so on. Although these latter mod-els certainly activate stress physiological sys-tems, one may question the adaptive natureof these responses. By definition, there is amismatch in these models between the chal-lenge or the stressor and the available reper-toire of behavioral, neuroendocrine, and neu-robiological adaptive mechanisms. Becausesuch a mismatch may be fundamental to thedevelopment of stress pathology, this may bethe reason that these models are so popularin stress research. However, we believe thatthe field of stress research has to move towarda more subtle understanding of the factors andprocesses underlying the development ofstress pathology.

Rather than pushing the animal toward astress physiological ceiling, it might be farmore informative to explore the natural fac-tors that determine and modulate the indi-vidual adaptive capacity. These factors in-clude not only perinatal and adult (social)experience but also factors that affect thespeed of recovery after a stressor. One of theselatter factors is social support. It appears thatthe long-term consequences of social defeatare strongly influenced by the social housingconditions after the defeat; that is, defeat com-bined with social isolation has a muchstronger impact in rats than either of the twoalone (Ruis et al., 1999). A second importantquestion that needs to be addressed and thatrequires more naturalistic stress models con-cerns the adaptive nature of the stress re-sponse. Implicit in the interpretation of manystress experiments is that the observedchanges somehow contribute to the develop-ment of pathology. We would argue that oneneeds a more naturalistic setting to obtain ex-perimental evidence for the (mal)adaptiveconsequences of a given stress response. Quitelikely, the stress response of an organism will

Page 260: 0195162854

Chapter 22. Stress 243

be a trade-offbetween the short-term benefitsand the long-term costs in terms of chances topathology and evolutionary fitness. Unfortu-nately, such questions have rarely been ad-dressed in the rat.

Finally, we would like to address a moreethical issue. As in many scientific experi-ments, stress research aims at certain effectsas predicted by specific hypotheses. However,when the predicted effect of a stressor is notfound, there is a strong tendency to blame thestressor rather than the hypothesis. This maylead to the use of extreme stressors that go farbeyond the biological range and hence fallshort in face validity. These include stressorssuch as prolonged restraint (days), high-in-tensity shocks, prolonged conditions of ex-treme crowding, and so on. Again, the biol-ogy and ecology of the species should be theguidelines for contemporary stress research.

REFERENCES

Abel EL (1994a) A further analysis of physiologicalchanges in rats in the forced swim test. Physiologyand Behavior 56:795-800.

Abel EL (1994b) Behavioral and physiological effects ofdifferent water depths in the forced swim test. Phys-iology and Behavior 56:411-414.

Abel EL (1994c) Physical activity does not account forthe physiological response to forced swim testing.Physiology and Behavior 56:677-681.

Barnett SA (1987) The rat: A study in behavior. Chicago:University of Chicago Press.

Berton O, Durand M, Aguerre S, Mormede P, ChaouloffF (1999) Behavioral, neuroendocrine and seroton-ergic consequences of single social defeat and re-peated fluoxetine pretreatment in the Lewis ratstrain. Neuroscience 92:327-341.

Blanchard DC, Markham C, Yang M, Hubbard D,Madarang E, Blanchard RJ (2003) Failure to produceconditioning with low-dose trimethylthiazoline orcat feces as unconditioned stimuli. Behavioral Neu-roscience 117:360-368.

Blanchard DC, Spencer RL, Weiss SM, Blanchard RJ,McEwen B, Sakai RR (1995) Visible burrow systemas a model of chronic social stress: Behavioral andneuroendocrine correlates. Psychoneuroendocrinol-ogy 20:117-134.

Calhoun JB (1962) The ecology and sociology of theNorway rat. Washington, D.C.: GovernementPrinting Office.

De Boer SF and Koolhaas JM (2003) Defensive buryingin rodents: Ethology, neurobiology and psy-chopharmacology. European Journal of Pharma-cology 463:145-161.

De Boer SF, van der Gugten J, Slangen JL (1991) Be-havioral and hormonal indices of anxiolytic and anx-iogenic drug action in the shock-prod defensiveburying paradigm. In: Animal models in psy-chopharmacolgy, pp. 81-96. Basel: Birkhauser Ver-lag.

Delhi F, Piazza PV, Mayo W, Le Moal M, Simon H(1996) Novelty-seeking in rats—biobehavioral char-acteristics and possible relationship with the sensa-tion-seeking trait in man. Neuropsychobiology34:136-145.

Dielenberg RA and McGregor IS (2001) Defensive be-havior in rats towards predatory odors: A review.Neuroscience and Biobehavioral Reviews 25:597-609.

Ely DL (1981) Hypertension, social rank, and aortic ar-teriosclerosis in CBA/J mice. Physiology and Be-havior 26:655-661.

Fokkema DS, Koolhaas JM, van der Gugten J (1995) In-dividual characteristics of behavior, blood pressure,and adrenal hormones in colony rats. Physiologyand Behavior 57:857-862.

Friedman MJ and Schnurr PP (1995) The relationshipbetween trauma, post-traumatic stress disorder, andphysical health. In: Neurobiological and clinical con-sequences of stress: From normal adaptation toPTSD (Friedman MJ, Charney DS, Deutch AY,eds.), pp. 507-524. Philadelphia: Lippincott-RavenPublishers.

Fuchs E, Johren O, Fliigge G (1993) Psychosocial con-flict in the tree shrew: Effects on sympathoadrenalactivity and blood pressure. Psychoneuroen-docrinology 18:557-565.

Henry JP, Liu YY, Nadra WE, Qian CG, Mormede P,Lemaire V, Ely D, Hendley ED (1993) Psychosocialstress can induce chronic hypertension in nor-motensive strains of rats. Hypertension 21:714-723.

Henry JP and Stephens-Larson P (1985) Specific effectsof stress on disease processes. In: Animal stress(Moberg GP, ed.), pp. 161-173. Bethesda: AmericanPhysiological Society.

Keller SE, Weiss JM, Schleifer SJ, Miller NE, Stein M(1981) Suppression of immunity by stress: Effect ofa graded series of stressors on lymphocyte prolifer-ation. Science 213:1397-1400.

Koolhaas JM, De Boer SF, De Ruiter AJ, Meerlo P, SgoifoA (1997a) Social stress in rats and mice. Acta Phys-iologica Scandinavica 161:69-72.

Page 261: 0195162854

244 REGULATORY SYSTEMS

Koolhaas JM and Bohus B (1989) Social control in relationto neuroendocrine and immunological responses. In:Stress, personal control and health (Steptoe A andAppels A, eds.), pp. 295-304. Brussels: John Wiley &Sons Ltd.

Koolhaas JM, Korte SM, De Boer SF, Van Der Vegt BJ,Van Reenen CG, Hopster H, De Jong 1C, Ruis MA,Blokhuis HJ (1999) Coping styles in animals: Cur-rent status in behavior and stress-physiology. Neu-roscience and Biobehavioral Reviews 23:925-935.

Koolhaas JM, Meerlo P, Boer SFd, Strubbe JH, Bohus B(1997b) The temporal dynamics of the stress re-sponse. Neuroscience and Biobehavioral Reviews21:775-782.

Layton B and Krikorian R (2002) Memory mechanismsin posttraumatic stress disorder. Journal of Neu-ropsychiatry and Clinical Neuroscience 14:254-261.

Lemaire V and Mormede P (1995) Telemetered record-ing of blood pressure and heart rate in differentstrains of rats during chronic social stress. Physiol-ogy and Behavior 58:1181-1188.

Lockwood JA and Turney T (1981) Social dominance andstress induced hypertension: Strain differences in in-bred mice. Physiology and Behavior 26:547-549.

ManuckSB, Kaplan JR, ClarksonTB (1983) Behaviorallyinduced heart rate reactivity and atherosclerosis incynomolgous monkeys. Psychosomatic Medicine45:95-108.

McEwen BS (2002) The neurobiology and neuroen-docrinology of stress. Implications for post-trau-matic stress disorder from a basic science perspec-tive. The Psychiatric Clinics North America25:469-494, ix.

McGregor IS, Schrama L, Ambermoon P, DielenbergRA (2002) Not all 'predator odours' are equal: Catodour but not 2,4,5 trimethylthiazoline (TMT; foxodour) elicits specific defensive behaviours in rats.Behavioural Brain Research 129:1-16.

Miczek KA, Thompson ML, Tornatzky W (1990) Shortand long term physiological and neurochemicaladaptations to social conflict. In: NATO ASI SeriesD: Behavioural and social sciences (Puglisi-AllegraS and Oliverio A, eds.), pp. 15-30. Dordrecht:Kluwer.

Porsolt RD, Le Pichon M, Jalfre M (1977) Depression: Anew animal model sensitive to antidepressant treat-ment. Nature 266:730-732.

Ruis MA, te Brake JH, Buwalda B, De Boer SF, Meerlo P,Korte SM, Blokhuis HJ, Koolhaas JM (1999) Housingfamiliar male wildtype rats together reduces the long-term adverse behavioural and physiological effects ofsocial defeat. Psychoneuroendocrinology 24:285-300.

Rupp H (1999) Excess catecholamine syndrome. Patho-physiology and therapy. Annals of the New YorkAcademy of Science 881:430-444.

Seres J, Stancikova M, Svik K, Krsova D, Jurcovicova J(2002) Effects of chronic food restriction stress andchronic psychological stress on the development ofadjuvant arthritis in male Long Evans rats. Annalsof the New York Academy of Science 966:315-319.

Sgoifo A, Koolhaas J, De Boer S, Musso E, Stilli D,Buwalda B, Meerlo P (1999a) Social stress, autonomicneural activation, and cardiac activity in rats. Neuro-science and Biobehavioral Reviews 23:915-923.

Sgoifo A, Koolhaas JM, Musso E, De Boer SF (1999b)Different sympathovagal modulation of heart rateduring social and nonsocial stress episodes in wild-type rats. Physiology and Behavior 67:733-738.

Spencer RL, Miller AH, Moday H, McEwen BS, Blan-chard RJ, Blanchard DC, Sakai RR (1996) Chronicsocial stress produces reductions in available splenictype II corticosteroid receptor binding and plasmacorticosteroid binding globulin levels. Psychoneu-roendocrinology 21:95-109.

Stefanski V, Knopf G, Schulz S (2001) Long-term colonyhousing in Long Evans rats: Immunological, hor-monal, and behavioral consequences. Journal ofNeuroimmunology 114:122-130.

Steimer T and Driscoll P (2003) Divergent stress responsesand coping styles in psychogenetically selected Ro-man high-(RHA) and low-(RLA) avoidance rats: Be-havioural, neuroendocrine and developmental as-pects. Stress 6:87-100.

Treit D (1985) Animal models for the study of anti-anx-iety agents: A review. Neuroscience and Biobehav-ioral Reviews 9:203-222.

Treit D, Pinel JPJ, Fibiger HC (1981) Conditioned de-fensive burying: A new paradigm for the study ofanxiolytic agents. Pharmacology, Biochemistry, andBehavior 15:619-626.

Tsuda A, Yoshishige I, Tanaka M (1988) Behavioral fieldanalysis in two strains of rats in a conditioned de-fensive burying paradigm. Animal Learning Behav-ior 16:354-358.

Veenema AH, Meijer OC, de Kloet ER, Koolhaas JM(2003) Genetic selection for coping style predictsstressor susceptibility. Journal of Neuroendocrinol-ogy 15:256-267.

Weiss JM (1972) Influence of psychological variables onstress-induced pathology. In: Physiology, emotion andpsychosomatic illness. CIBA Foundation Symposium(Porter R and Knight J, eds.). Amsterdam: Elsevier.

Weiss JM, Sundar SK, Becker KJ, Cierpal MA (1989) Be-havioral and neural influences on cellular immuneresponses: Effects of stress and interleukin-1. Jour-nal of Clinical Psychiatry 50:43-55.

Yehuda R, McFarlane AC, Shalev AY (1998) Predictingthe development of posttraumatic stress disorderfrom the acute response to a traumatic event. Bio-logical Psychiatry 44:1305-1313.

Page 262: 0195162854

Immune System

HYMIE ANISMAN ANDALEXANDER W. KUSNECOV

23

The major function of the immune system isto monitor the organism's internal environ-ment for signs of tissue damage and microbialinfiltration (e.g., bacteria and viruses). In con-sidering the factors that govern immune func-tioning, it seems that in the main, immuneprocesses are similar across species (or strainswithin a species), although characteristic dif-ferences occur within certain immune com-ponents. In this report we provide a briefoverview of the functioning of the immunesystem, followed by a description of how im-mune alterations can affect central nervoussystem (CNS) processes and behavior in therat. We also describe how factors that have animpact on psychological processes, most no-tably stressors, may come to affect immunefunctioning. In so doing, we introduce nu-merous caveats concerning the conditions andlimitations that determine the nature of theeffects observed, making it clear that disen-tangling the impact of various manipulationson immune activity is complex.

In addition to experimental stressor ma-nipulations that affect immune functioning,laboratory conditions or experimental para-digms may act as stressors that can affect im-mune competence (or activate products of theimmune system, such as cytokines, that couldaffect CNS activity). In general, stressors canbe subdivided into those of a processive na-ture (i.e., those that involve appraisal of a sit-uation or higher order sensory cortical pro-cessing, including events such as exposure toa novel environment, change of social condi-tions, predator exposure, restraint, cold swim,

foot or tail shock, or conditioned fear cues),which are either psychogenic or neurogenic(being purely psychological in nature or in-volving physical or painful stimuli, respec-tively). In addition, stressors may include sys-temic (e.g. metabolic) insults, such as bacterialor viral infection (Herman and Cullinan,1997), which in rats induce many (althoughnot all) of the central neurochemical alter-ations provoked by processive stressors, andthus might also be expected to provoke be-havioral changes that often are elicited byaversive events (Anisman et al, 2002). The na-ture of the immune and cytokine changes in-duced is dependent on several stressor and or-ganismic characteristics. Of particular note inthis regard is that even mild stressors mayhave an impact on immune functioning of therat, including standard laboratory manipula-tions, such as handling and injection proce-dures, and transport from the breeder. Ofcourse, intense insults such as surgery, espe-cially that involving direct brain manipula-tions (e.g., lesions, stimulation, cannulation),may have particularly potent actions on cen-tral cytokine functioning (Fassbender et al.,2001) and hence may influence neuronal ac-tivity and behavioral output.

A BRIEF PRIMER ONIMMUNITY IN THE RAT

Stimulation of immune cells, as depicted inFigure 23-1, results in an exquisitely orches-trated immune response that exhibits antigen

245

Page 263: 0195162854

246 REGULATORY SYSTEMS

Figure 23—1. Simplified diagram ofcellular interactions in the immunesystem. An immune response is ini-tiated on the presentation of antigen(Ag) by antigen presenting cells(APC) to T and B lymphocytes. Onstimulation with Ag, lymphocytesenter a proliferative phase, out ofwhich they emerge as effector cellscapable of producing Ab, if B cells,or aiding the latter, if T cells. Regu-lation of lymphocyte proliferativeand effector functions is mediated bycytokines that are produced by mostcells of the immune system and canaffect central nervous system func-tion.

specificity, learning, and a high degree of au-tocrine and paracrine regulation. Immuno-logical memory in the immune system,demonstrated by the rapid and robust re-sponse to an antigen (i.e., foreign molecule)that the immune system has previously en-countered is mediated by T and B lympho-cytes (henceforth called T and B cells), themain cells of the immune system responsiblefor its adaptive or acquired characteristics.1

Other cells of the immune system serveaccessory functions in that they can presentantigen to lymphocytes, regulate the activityof lymphocytes, and enzymatically digest for-eign cells (e.g., bacteria) and necrotic tissuethrough a process called phagocytosis. Acces-sory cells include dendritic cells, which spe-cialize as antigen-presenting cells (APCs), andpolymorphonuclear phagocytes (monocytesand macrophages). It should be noted thatmacrophages are actually monocytes in a ma-ture, more differentiated state that includesphagocytic function—for example, in theCNS, microglial cells assume the role of resi-dent macrophages.

In any given immune response, there arethree identifiable processes: induction, activa-tion, and effector function. The first twoprocesses represent stages in which antigen ispresented to and stimulates lymphocytes, cul-minating in their activation. Once activated,

lymphocytes can differentiate into blast cellsthat mediate effector functions, namely anti-body and cytokine production, and cytotoxi-city. Antibody production is the primary func-tion of B cells, which can also serve as APCs,whereas T cells perform regulatory and cyto-toxic functions.2 Specifically, there are twosubtypes of T cells: T helper (aka CD4-posi-tive cells) and T cytotoxic cells (aka CD8-pos-itive cells). The helper CD4+ T cells (Th cells)were originally defined by their ability to helpB cells produce antibody against antigenicstimuli, although it is now known that "help"has a broader definition in that immune re-sponses can be suppressed or downregulatedby Th cells, an important self-regulatory func-tion of the immune system that can preventexcessive inflammation and reduce the risk ofautoimmunity. This inhibitory or anti-inflam-matory function is mediated by the Th2 sub-set of Th cells, with the Thl subset principallyidentified with proinflammatory immuneresponses.

Activation of B cells leads to productionof soluble immunoglobulin (Ig) moleculesmeasurable in bodily fluids (e.g., bloodplasma/serum). There are five classes of Igmolecules (IgA, IgD, IgE, IgM, and IgG),which represent the antibodies induced byantigen. Most classes of antibody are pro-duced during the primary immune response,

Page 264: 0195162854

Chapter 23. Immune System 247

although IgM predominates. Reactivation ofB cells by antigen results in a secondary hu-moral response of greater affinity and avidityfor the antigen and is mostly IgG.

All T cells express a T-cell receptor(TCR), which recognizes specific antigenicpeptides and initiates intracellular signals thatresult in proliferation and differentiation. Ad-ditional surface molecules, CD4 and CDS, alsomediate helper and cytotoxic T-cell functions.Thus, CD4+ Th cells are regulatory, whereasCD8+ cytotoxic T cells directly lyse infectedand neoplastic cells. Lytic functions are alsocarried out by large granular lymphocytescalled natural killer (NK) cells.

Induction of a T-cell response requires arecognition step in which antigen specificallybinds to the TCR. The recognition step be-tween antigen and TCR involves a physical in-teraction between the T cell and an APC. Forexample, macrophages, after digesting a largeforeign substance by phagocytosis, load the di-gested peptide fragments onto intracellularclass II molecules of the major histocompati-bility complex (MHC) that can then presentthe antigenic peptide to a T cell that expressesa TCR with specific complementarity to thepeptide sequence. This results in activationand differentiation of the T cell.

Much of the effector function of the im-mune response involves the production of sol-uble substances that promote the eliminationof antigen (e.g., antibody molecules). Cy-tokines, which play a fundamental role in thisrespect, are protein molecules synthesizedand secreted by cells to serve as autocrine/paracrine signaling and growth and differen-tiation factors. Fibroblasts and endothelialcells secrete cytokines, but so do cells of theimmune system (and as is now well known,the CNS), including T cells, monocytes, andmacrophages. The major characteristics of cy-tokines are that (1} they are rarely constitu-tively expressed, (2) cellular effects are re-ceptor mediated, and (3) they are generallypleiotropic (Thompson, 1998). The lattercharacteristic is particularly germane to neu-

roscience, because cytokines both affect andare produced in the CNS. Indeed, one of thefastest growing areas in neuroscience researchis the role of cytokines in CNS pathology, aswell as in normal behavior.

Disturbances of various arms of the im-mune system are associated with different dis-ease states. Diminished activity of T and Bcells has been associated with acquired im-munodeficiency syndrome, whereas as re-duced NK-cell activity may be related to viralillness and neoplastic disease. Furthermore, itis possible that the recognition abilities of theimmune cell may become compromised, orsuppressor cells (or inhibitory cytokines) maynot operate adequately, leading to the im-mune system attacking the self, hence favor-ing the development of autoimmune disor-ders, such as lupus erythematosus andrheumatoid arthritis (Abbas et al., 2000). Fur-thermore, dysfunction within any given sys-tem can be effected through different pro-cesses. Lymphocytes in circulation may bealtered, proliferation of these lymphocytesmay be compromised, or the killing potentialof the lymphocytes may be impaired. Suchchanges may develop owing to dysfunctionswithin the immune system itself, or they mayreflect the influence of processes (e.g., neu-roendocrine) that ordinarily regulate immuneactivity (Besedovsky and Del Rey, 2001).

In considering analyses of immune func-tioning, a variety of in vitro and in vivo pro-cedures are available in the rat, but the ulti-mate test of whether a treatment has positiveor negative effects concerns its effects on theorganism's well-being in the face of challenge;that is, has vulnerability to pathology been al-tered? This can be examined in terms of al-tered vulnerability to viral or bacterial insults,recovery from illnesses or rate of wound heal-ing, and development or exacerbation ofpathological symptoms, both in randomlybred animals or in those with specific vulner-abilities (e.g., Fisher 344 versus Lewis rats). Inaddition, one can examine the effects of thetreatment on specific attributes or arms of the

Page 265: 0195162854

248 REGULATORY SYSTEMS

immune system; this includes simply count-ing the number of specific types of immunecells (per fixed volume of blood). Althoughthis does not characterize functional changesthat may have occurred, it is certainly impor-tant to know whether a given treatment, suchas a stressor, alters the cell trafficking patternsof immune cells. Functional attributes, in con-trast, include the level of T- or B-cell prolifer-ation in response to specific stimulants (mito-gens) that are present in secondary immuneorgans (e.g., spleen) or in circulation, antibodyresponses to antigenic stimuli, or the killingability of specific immune cells on being ap-propriately challenged (e.g., NK cell toxicity).

INTERACTIONS BETWEEN THEIMMUNE SYSTEM AND THE BRAIN

Multidirectional communication, involvingconcurrent and sequential interactions, isknown to occur among the immune, en-docrine, autonomic, and central nervous sys-tems, likely involving several processes. For in-stance, cytokines released from activatedimmune cells may influence CNS activity with-out actually entering the brain parenchyma butby stimulating afferent vagal fibers (Dantzer,2001) or receptors at circumventricular andother vascular regions (Nadeau and Rivest,1999). Cytokines may also gain access to thebrain at circumventricular organs (Nadeau andRivest, 1999) that lack an efficient blood-brainbarrier or through saturable carrier-mediatedtransport mechanisms (Banks, 2001), eventu-ally reaching distal sites through volume diffu-sion (Konsman et al., 2000). Finally, cytokinesmay nonselectively stimulate cells of largeblood vessels and small capillaries, thereby per-mitting greater passage of cytokines into thebrain (Rivest et al., 2000).

Once in the brain, cytokines can bind tospecific cytokine receptors (Cunningham andDe Souza, 1993), thus influencing neuronalfunctioning (Anisman and Merali, 1999). Theroutes by which these cytokines come to af-

fect limbic processes (e.g., stimulation of thecentral amygdala) remain to be determined,although stimulation of the parabrachial nu-cleus and paraventricular hypothalamus maybe particularly important in this regard (Bullerand Day, 2002). Ultimately, cytokines affectneuroendocrine and central neurotransmitterprocesses, and vice versa (Dunn, 1995; Anis-man and Merali, 1999), and thus it is to be ex-pected that dysfunctions of one system mayhave repercussions on other systems.

Signaling of the CNS by peripheral cy-tokines is complemented by cytokine synthe-sis within the brain, most likely within astro-cytes and microglia but possibly withinneuronal tissues (Rivest and Laflamme, 1995).Indeed, in rats, central cytokine bioactivity in-creases appreciably in response to a variety ofphysical and chemical insults (e.g., brain in-jury, cerebral ischemia, kainic acid or 6-hy-droxydopamine lesions, seizure) (Rothwelland Luheshi, 2000), as well as systemic or cen-tral challenge with bacterial endotoxin orviruses (e.g., Nadeau and Rivest, 1999). More-over, stressors may influence central cytokineexpression or protein levels (Nguyen et al.,1998). These various challenges may affect theregulation of ligand-receptor interactions andmay affect acute-phase reactions (Black, 2002;Nguyen et al., 2002), thus influencing a widerange of neurological disease states (Nguyenet al., 2002). However, it is still unclear underwhat conditions cytokines act in a reparatorycapacity or when they promote neuronaldamage.

Aside from contributions to neurologicaldisorders, inflammatory immune processesmay contribute to psychological disturbances,such as depression (Maes, 1999). To be sure,depressive illness has been thought to stemfrom neurochemical alterations, such as vari-ations of norepinephrine and serotonin, aswell as neuropeptides such as corticotropin-releasing hormone (CRH), which may be pro-voked by stressful encounters. Interestingly,when administered systemically, cytokines in-fluence hypothalamus-pituitary-adrenal (HPA)

Page 266: 0195162854

Chapter 23. Immune System 249

activity and promote hypothalamic and ex-trahypothalamic neurotransmitter changesreminiscent of those elicited by stressors(Anisman and Merali, 1999, 2002). These neu-rochemical changes may in turn promote be-havioral changes characteristic of depressivestates (e.g., induction of anhedonia [Anismanet al., 2002]), which may be antagonized byantidepressant treatments (Merali et al., 2003).Essentially, the view was adopted that the im-mune system may be part of a regulatory loopthat, by virtue of its effects on neurochemicalprocesses, might contribute to the symptomsof mood and anxiety-related disorders (Anis-man and Merali, 2002). Indeed, as in rats, inhumans receiving cytokine immunotherapy(e.g., interleukin-2 or interferon-a adminis-tration in the treatment of certain cancers andhepatitis C), profound depressive symptomsmay evolve that may be attenuated by treat-ment with the selective serotonin reuptake in-hibitor paroxetine (Musselman et al., 2001).Of course, it is possible that the cytokinesthemselves are insufficient to engender de-pression but rather reflect an interaction be-tween the cytokine treatment and the distressbeing experienced by the patient with canceror hepatitis C, especially as the effects of animmunological challenge may be appreciablyincreased if administered on the backdrop ofa sustained stressor regimen (Tannenbaum etal., 2002).

STRESS, CENTRAL PROCESSES, ANDIMMUNOLOGICAL ALTERATIONS

Just as immune activation can affect CNSfunctioning, it appears that stressors, via theirimpact on neuroendocrine, autonomic, andcentral nervous system processes, may affectimmune activity across a range of differentspecies, including laboratory (e.g., mice, rats,monkeys) and agricultural animals (e.g., fowl,pigs, cattle). It is generally thought that acutestressors provoke several biological defen-sive responses of adaptive significance. Acute

stressors of modest severity increase HPA ac-tivity, neuropeptide changes at hypothalamicand extrahypothalamic sites, and monoamine(norepinephrine, dopamine, and serotonin)synthesis and utilization (Anisman and Mer-ali, 1999; Anisman et al., 2002). With morechronic stressors, further compensatory neu-rochemical changes evolve, which may alsoserve to maintain the organism's well-being(Lopez et al., 1999). Yet, if the stressor persists,the wear and tear on biological systems maybecome excessive (allostatic overload), andthe availability of resources sufficiently di-minished, culminating in increased vulnera-bility to pathology (McEwen, 2000).

Individual differences in response to ad-verse events have long been appreciated, andsome rats appear to be relatively vulnerableto stressor-provoked neurotransmitter andneuroendocrine changes, and one can imag-ine that vulnerability to allostatic load wouldvary in a similar fashion. Furthermore, stres-sor encounters, as well as cytokine challenges,may result in the sensitization of neuronalfunctioning, so that later stressor experiencesmay elicit more profound neurochemicalchanges (Tilders and Schmidt, 1999). It hassimilarly been observed that rat pups sub-jected to immunological challenges within afew days postpartum may, as adults, displayincreased stressor reactivity (Shanks et al.,2000), just as early life traumatic events maypromote such outcomes in humans (Heimand Nemeroff, 2001). These early life sensiti-zation effects are not limited to neuroen-docrine changes but are also evident with re-spect to immune responses mounted afteradult stressor experiences (Shanks et al., 2000).

Just as stressors affect neuroendocrineand central neurotransmitter functioning,both psychogenic and neurogenic insults af-fect various aspects of immune functioning inthe rat, such as suppression of splenic NK cellactivity, mitogen-stimulated cell proliferation,the plaque-forming cell response followingantigenic stimulation, and macrophage activ-ity (Kusnecov et al., 2001). The specific

Page 267: 0195162854

250 REGULATORY SYSTEMS

changes observed are dependent on factorssuch as stressor severity and history, experi-ential, organismic, genetic and ontogeneticfactors, and the specific immune parametersand compartments (e.g., spleen versus blood)being examined. Moreover, many of the im-mune alterations are independent of tissuedamage associated with stressors, becausethey are provoked by psychogenic stressors(e.g., stressor-related odors, cues associatedwith stressors, and psychosocial stressors)(Kusnecov and Rabin, 1994; Kusnecov et al,2001; Moynihan and Stevens, 2001).

It is important to distinguish between theimmune changes exerted as a function ofstressor severity and chronicity. For instance,in rats, mild stressors may augment immu-nity, whereas more protracted stressors mayhave the opposite effect. From an adaptiveperspective, this seems to make sense. Stress-ful events, as clear threats to the organism'swell-being, should encourage increased celltrafficking, cell proliferation, and cytotoxicity.However, as the stressor continues, and theallostatic load increases, the reduced resourceavailability may render the organism morevulnerable to immunological disturbances(Dhabbar and McEwen, 1999). For instance,acute stressors enhanced the delayed-type hy-persensitivity response (DTH), an in vivomeasure of T-cell-mediated immunity (Dhab-bar and McEwen, 1999; Dhabhar, 2000). Thismeasure requires initial sensitization withantigen, followed days to weeks later by chal-lenge with the sensitizing antigen. This pro-motes an inflammatory response character-ized by increased redness and swelling of thechallenged part of the body (typically the foot-pad or pinnae of the ear). When rats receiveda single session of restraint immediately be-fore challenge with the sensitizing chemicalDNFB, the DTH response of challenged ratswas increased, and this effect was sensitive tothe severity of the stressor (Dhabhar andMcEwen, 1999; Dhabhar, 2000).

It should be underscored that in these stud-

ies, the immune-enhancing effects of stressorswere examined in relation to elicitation of theDTH response in rats that had previouslybeen exposed to the sensitizing antigen. Thus,the findings may be related to how acutestressors affect the memory T-cell response,rather than how stressors influence the im-mune response in otherwise naive T cells. Inthis regard, acute stressors impaired the DTHresponse following introduction of sheep ery-throcytes into the lung (Blecha et al., 1982),raising the possibility that the nature of theoutcomes observed may be related to eitherdifferences in the specific antigens used or thelocalization of antigens to different immuno-logical compartments (dermal tissue versusupper respiratory mucosal surfaces). Further-more, stressor effects on the DTH response inrats (Flint and Tinkle, 2001) may differ fromthose seen in mice (Wood et al., 1993), al-though it is difficult to equate stressor severi-ties across species.

In considering the factors that influencestressor-elicited immune changes, both Kus-necov et al. (2001) and Moynihan and Stevens(2001) made the point that the effects ofchronic stressors on immune functioning maybe related to the specific immune response be-ing examined. Using in vitro assessment ofcell-mediated immunity, such as lymphocyteproliferation to T-cell mitogens, concanavalinA (ConA) or phytohemagglutinin (PHA,) ithas commonly been reported that acute stres-sors suppress rat splenocyte proliferation. Yet,the antigen-specific proliferative response tocholera toxin-sensitized spleen cells can be en-hanced by the same stressor (Kusnecov andRabin, 1993).

As mentioned, acute stressors cause re-ductions in splenic and blood lymphocyte cel-lular proliferation to the nonspecific T-cell mi-togens (Kusnecov and Rabin, 1994). Such anoutcome, however, is not consistently ob-served after more prolonged stressors (e.g.,foot shock, immobilization, isolation). Shortperiods of social isolation or swim stress in-

Page 268: 0195162854

Chapter 23. Immune System

hibited lymphocyte proliferation in blood andspleen lymphocytes, whereas more prolongedstressor exposure reversed this effect (Kus-necov et al., 2001). However, studies using im-mobilization/restraint or electric shock indi-cated that the immunosuppressive effects ofacute stressors were retained after a chronicstressor regimen (Lysle et al., 1987; Batumanet al., 1990). In explaining these divergent find-ings, the nature and severity of the stressorsused may prove to be relevant. Once again,however, important species differences mayalso exist in this respects, as acute stressors en-hanced the activity of splenic B cells or acces-sory cells in mice (Lu et al., 1998; Shanks andKusnecov, 1998).

Of the studies that have examined the ef-fects of stressors on immune functioning, themajority involved neurogenic stressors,whereas some used psychogenic insults. Be-cause these stressors may activate differentneural circuits (Lopez et al., 2001; Anisman etal., 2002) and may have different effects on pe-ripheral processes, the fact that both types ofstressors generally elicited similar effects isconsistent with the notion that the psycho-logical ramifications of stressors were respon-sible for the observed outcomes. However, itis important to note that not all psychogenicstressors exert similar effects, because neuralcircuitry activated by predatory stressors maynot be the same as that associated with otherpsychogenic stressors. Indeed, one of the mostpotent stressors for rats (and humans) is socialdisruption, a perturbation that may have par-ticularly marked effects on immune function-ing and susceptibility to infectious disease(Sheridan, 1998). Interestingly, social defeat inrats followed by lipopolysaccharide adminis-tration 1 week later resulted in a diminishedcorticosterone response, increased circulatinginterleukin-1 levels, and increased mortality inresponse to the lipopolysaccharide (Carborezet al., 2002). Similarly, 1 week of social disrup-tion provoked resistance to the immunosup-pressive effects of glucocorticoids (Stark et al.,

251

2001), a situation that can result in excessiveproduction of cytokines in response tolipopolysaccharide and ultimately increasedmortality (Quan et al., 2001). Thus, protractedstress does not necessarily vanquish host de-fense mechanisms against environmentalpathogens but may actually give rise tochanges that may result in unexpectedly seri-ous health complications.

CONCLUSION

The original conception of the immune sys-tem as an autonomously functioning systemhas been invalidated by research demonstrat-ing neurohormonal regulation of immunecells. This is supported further by the "hard-wired" infrastructure of noradrenergic andpeptidergic innervation of lymphoid organs.Together with the hormones of the neu-roendcrine system, these sympathetic andparasympathetic connections to the immunesystem are subject to influences from up-stream neural processes that respond to envi-ronmental stressors, such that any given stres-sor is capable of communicating to andaltering the functional state of the immunesystem.

As discussed, a number of in vitro and invivo parameters of immune function can bealtered by stressor exposure, although the spe-cific immunological readout depends on arange of stressor and organismic variables thatawait full characterization. It should be notedthat regardless of the immunological impactof the CNS response to a psychogenic or neu-rogenic stressor, there may be important im-plications for CNS function in terms of com-pensatory changes in the immune system. Aswe discussed, cytokines are the chief regulatoryproducts of the immune system, and it is wellestablished that a number of cytokines influ-ence various motivational and cognitive be-haviors, which includes the recruitment of dis-tinct stress-related pathways in the brain,

Page 269: 0195162854

252 REGULATORY SYSTEMS

including the locus coeruleus, amygdala, hip-pocampus, and hypothalamus. Although acutestressors can increase immune responding, thishas largely been demonstrated using benignimmunological stimuli. However, in principle,among stressed organisms the immune re-sponse to an infectious agent could result in ex-cess production of proinflammatory cytokinesthat serve to differentially affect the CNS rela-tive to nonstressed organisms. Moreover, thiswould have an impact on a CNS already en-gaged in adjusting to the psychogenic stressorthat initially altered the immune response andthus might affect CNS pathology.

ACKNOWLEDGMENTS

This work was supported by the Canadian Institutes of HealthResearch (H.A.), U.S. Public Health Service grants DA14186(A.W.K.) and MH60706 (A.W.K.), and National Institute forEnvironmental and Health Science Rutgers University and theUniversity of Medicine and Dentistry of New Jersey CenterGrant P30 ES05022. The authors are indebted to Zul Meraliand Shawn Hayley for their comments. H.A. holds a CanadaResearch Chair in Neuroscience.

NOTES

1. More detailed information on the immune sys-tem can be found in Abbas et al. (2000) and Janeway andTravers (2001).

2. All immune cells, including lymphocytes, orig-inate in the bone marrow, although T cells require fur-ther maturation in the thymus gland; B cells maturelargely in the bone marrow and fetal liver.

REFERENCES

Abbas AK, Lichtman AH, Pober JS (2000) Cellular andmolecular immunology, 4th ed. Philadelphia: WBSaunders Company.

Anisman H, Kokkinidis L, Merali Z (2002) Further evi-dence for the depressive effects of cytokines: Anhe-donia and neurochemical changes. Behavior andImmunity 16:544-556.

Anisman H and Merali Z (2002) Cytokines, stress, anddepressive illness. Brain, Behavior and Immunity16:513-524.

Anisman H and Merali Z (1999) Anhedonic and anxio-genic effects of cytokine exposure. Advances in Ex-perimental and Medica Biology 461:199-233.

Banks WA (2001) Cytokines, CVSs, and the blood-brain-barrier. In: Psychoneuroimmunology, 3rd ed. (AderR, Felten DL, Cohen N, eds.), pp. 483-498. NewYork: Academic Press.

Batuman OA, Sajewski D, Ottenweller JE, Pitman DL,Natelson BH (1990) Effects of repeated stress on Tcell numbers and function in rats. Behavior and Im-munity 4:105-117.

Besedovsky HO and Del Rey A (2001) Cytokines as me-diators of central and peripheral immune-neuroen-docrine interactions. In Psychoneuroimmunology,3rd ed. (Ader R, Felten DL, Cohen N, eds.), pp.483-498. New York: Academic Press.

Black PH (2002) Stress and the inflammatory response:A review of neurogenic inflammation. Behavior andImmunity 16:622-653.

Blecha F, Barry RA, Kelley KW (1982) Stress-induced al-terations in delayed-type hypersensitivity to SRBCand contact sensitivity to DNFB in mice. Proceed-ings of the Society of Experimental and BiologicalMedicine 169:239-246.

Buller KM and Day TA (2002) Systemic administrationof interleukin-lbeta activates select populations ofcentral amygdala afferents. Journal of ComparativeNeurology 452:288-296.

Carborez SG, Gasparotto OC, Buwalda B. Bohus B(2002) Long-term consequences of social stress oncorticosterone and IL-lbeta levels in endotoxin-challenged rats. Physiology and Behavior 76:99-105.

Cunningham ET and De Souza EB (1993) Interleukin 1receptors in the brain and endocrine tissues. Im-munology Today 14:171-176.

Dantzer R (2001) Cytokine-induced sickness behavior:Mechanisms and implications. Annals of the NewYork Academy of Science 933:222-234.

de Groot J, Ruis MA, Scholten JW, Koolhaas JM,Boersma WJ (2001) Long-term effects of social stresson antiviral immunity in pigs. Physiology and Be-havior 73:145-158.

Dhabhar FS (2000) Acute stress enhances while chronicstress suppresses skin immunity. The role of stresshormones and leukocyte trafficking. Annals of theNew York Academy of Science 917:876-893

Dhabhar FS and McEwen BS (1999) Enhancing versussuppressive effects of stress hormones on skin im-mune function. Proceedings of the National Acad-emy of Sciences 96:1059-1064.

Fassbender K, Schneider S, Bertsch T, Schlueter D, Fa-tar M, Ragoschke A, Kuhl S, Kischka U, HennericiM (2001) Temporal profile of release of interleukin-1/3 in neurotrauma. Neuroscience Letters 284:135-138.

Page 270: 0195162854

Chapter 23. Immune System 253

Flint MS and Tinkle SS (2001) C57BL/6 mice are resist-ant to acute restraint modulation of cutaneous hy-persensitivity. lexicological Science 62:250-256.

Heim C and NemeroffCB (2001) The role of childhoodtrauma in the neurobiology of mood and anxietydisorders: Preclinical and clinical studies. BiologicalPsychiatry 49:1023-1039.

Herman JP and Cullinan WE (1997) Neurocircuitry ofstress: Central control of hypothalamo-pituitary-adrenocortical axis. Trends in Neurostience 20:78-84.

Janeway C and Travers P (2001) Immunobiology, 5thed. New York: Garland Publishing.

Konsman JP, Parnet P, Dantzer R (2002) Cytokine-in-duced sickness behaviour: mechanisms and impli-cations. Trends in Neuroscience 25:154-159.

Kusnecov AW and Rabin BS (1994) Stressor-induced al-terations of immune function: mechanisms and is-sues. International Archives of Allergy and Im-munology 105:107-121.

Kusnecov AW and Rabin BS (1993) Inescapable foot-shock exposure differentially alters antigen- and mi-togen-stimulated spleen cell proliferation in rats.Journal of Neuroimmunology 44:33-42.

Kusnecov AW, Sved A, Rabin B (2001) Immunologic ef-fects of acute versus chronic stress in animals. In:Psychoneuroimmunology, 3rd ed. (Ader R, FeltenDL, Cohen N, eds.), pp. 265-278. New York: Acad-emic Press.

Lopez JF, Akil H, Watson SJ (1999) Neural circuits me-diating stress. Biological Psychiatry 146:461-471.

Lu ZW, Song C, Ravindran AV, Merali Z, Anisman H(1998) Influence of a psychogenic and a neurogenicstressor on several indices of immune functioningin different strains of mice. Brain, Behavior and Im-munity 12:7-22.

Lysle DT, Lyte M, Fowler H, Rabin BS (1987) Shock-in-duced modulation of lymphocyte reactivity: Sup-pression, habituation, and recovery. Life Science41:1805-1814.

Maes M (1999) Major depression and activation of theinflammatory response system. Advances in Exper-imental and Medical Biology 461:25-45.

McEwen BS (2000) Allostasis and allostatic load: Impli-cations for neuropsychopharmacology. Neuropsy-chopharmacology 22:108-124.

Merali Z, Brennan K, Brau P, Anisman H (2003) Disso-ciating anorexia and anhedonia elicited by inter-leukin-1/3: Antidepressant and gender effects on re-sponding for "free chow" and "earned" sucroseintake. Psychopharmacology 165:413-418.

Moynihan JA and Stevens SY (2001) Mechanisms ofstress-induced modulation of immunity in animals.In: Psychoneuroimmunology, 3rd ed. (Ader R, Fel-ten DL, Cohen N, eds.), pp. 227-249. New York:Academic Press.

Musselman DL, Lawson DH, GumnickJF, ManatungaA, Penna S, Goodkin R, Greiner K, Nemeroff C,Miller AH (2001) Paroxetine for the prevention ofthe depression and neurotoxicity induced by highdose interferon alpha. The New England Journal ofMedicine 344:961-966.

Nadeau S and Rivest S (1999) Regulation of the gene en-coding tumor necrosis factor alpha (TNF-alpha) inthe rat brain and pituitary in response in differentmodels of systemic immune challenge. Journal ofNeuropathology and Experimental Neurology58:61-77.

Nguyen MD, Julien J-P, Rivest S (2002) Innate immu-nity: The missing link in neuroprotection and neu-rodegeneration. Nature Reviews 3:216-227.

Nguyen KT, DeakT, Owens SM, Kohno T, Fleshner M,Watkins LR, Maier S (1998) Exposure to acute stressinduces brain interleukin-lj8 protein in the rat. Jour-nal of Neuroscience 19:2799-2805.

Quan N, Avitsur R, Stark JL, He L, Shah M, CaligiuriM, Padgett DA, Marucha PT, Sheridan JF (2001) So-cial stress increases the susceptibility to endotoxicshock. Journal of Neuroimmunology 115:36-45.

Rivest S and Laflamme N (1995) Neuronal activity andneuropeptide gene transcription in the brains of im-mune-challenged rats. Journal of Neuroendocrinol-ogy 7:501-525.

Rivest S, Lacroix S, Vallieres L, Nadeau S, Zhang J,Laflamme N (2000) How the blood talks to the brainparenchyma and the paraventricular nucleus of thehypothalamus during systemic inflammatory andinfectious stimuli. Proceedings of the Society forExperimental Biology and Medicine 223:22-38.

Rothwell NJ and Luheshi G (2000) Interleukin 1 in thebrain: biology, pathology and therapeutic target.Trends in Neuroscience 23:618-625.

Shanks N and Kusnecov AW (1998) Differential immunereactivity to stress in BALB/cByJ and C57BL/6Jmice: In vivo dependence on macrophages. Physi-ology and Behavior 65:95-103.

Shanks N, Windle RJ, Perks PA, Harbuz MS, Jessop DS,Ingram CD, Lightman SL (2000) Early-life exposureto endotoxin alters hypothalamic-pituitary-adrenalfunction and predisposition to inflammation. Pro-ceedings of the National Academy of Sciences97:5645-5650.

Sheridan JF (1998) Norman Cousins Memorial Lecture1997. Stress-induced modulation of anti-viral im-munity. Brain Behavior and Immunity 12:1-6.

Song C, Merali Z, Anisman H (1999) Variations of nu-cleus accumbens dopamine and serotonin followingsystemic interleukin-1, interleukin-2 or interleukin-6 treatment. Neuroscience 88:823-836.

Stark JL, Avitsur R, Hunzeker J, Padgett DA SheridanJF (2002) Interleukin-6 and the development of so-

Page 271: 0195162854

254 REGULATORY SYSTEMS

cial disruption-induced glucocorticoids resistance.Journal of Neuroimmunology 124:9-15.

Tannenbaum B, Tannebaum G, Anisman H (2002) Neu-rochemical and behavioral alterations elicited by achronic intermittent stressor regimen: Implicationsfor allostatic load. Brain Research 953:82-92.

Thomson A (1998) The cytokine handbook, 3rd ed. SanDiego: Academic Press.

Tilders FJH and Schmidt ED (1999) Cross-sensitizationbetween immune and non-immune stressors. A rolein the etiology of depression? Advances in Experi-mental Medicine and Biology 461:179-197.

Wood PG, Karol MH, Kusnecov AW, Rabin BS (1993)Enhancement of antigen-specific humoral and cell-mediated immunity by electric footshock stress inrats. Brain Behavior and Immunity 7:121-134.

Page 272: 0195162854

Development \

Page 273: 0195162854

This page intentionally left blank

Page 274: 0195162854

Prenatal Behavior

SCOTT R. ROBINSON ANDMICHELE R. BRUMLEY

24

For much of the history of comparative, phys-iological, and developmental psychology, theexperimental demonstration that a particularbehavioral capacity could be expressed soon af-ter birth was sufficient to conclude that the be-havior developed without benefit of experi-ence—in other words, that the behavior was"innate." Development before birth wasmostly viewed as a process of maturation, in-volving cellular and tissue-level interactionsthat were guided by regulatory genes. A newperspective on fetal development has emerged,however, from research that has demonstratedcomplex behavioral organization, sensory re-sponsiveness, and capacity for learning duringthe prenatal period. In large part, this researchhas been advanced by the use of animal mod-els of fetal behavioral development, promi-nently including studies of the rat fetus(Robinson and Smotherman, 1992a, 1995;Smotherman and Robinson, 1997).

Like other placental mammals, rats areborn after a period of physiological dependencyon the mother. Because the fetus's life supportderives from the placental connection to themother's uterus, researchers face a significantchallenge to gain experimental access to fetalsubjects for behavioral study. This challenge hasbeen overcome by methods that involve block-ade of the spinal cord of the pregnant rat, per-mitting surgical exteriorization of the uterusand fetuses while avoiding the activity-suppressing effects of general anesthesia(Smotherman and Robinson, 1991). When im-mersed in a bath containing physiological saline

maintained at maternal body temperature(37.5° C), this preparation provides visual andexperimental access to individual fetal subjects.Fetal rats can be observed without manipula-tion through the semitransparent wall of theuterus (in utero). Clearer visualization andmore direct experimental access are providedby delivering individual fetuses from the uterusinto the saline bath, either with the amniotic sacintact (in amnion) or after the embryonic mem-branes have been removed (ex utero). Thesemethods permit experiments involving videorecording of fetal motor behavior, presentationof chemical and tactile stimuli, administrationof drugs, or surgical manipulation of the centralnervous system (CNS) in test sessions lasting upto 2 hours. Developmental changes in fetal be-havior are measured by cross-sectional experi-mental designs, with fetal subjects provided bypregnancies at different gestational ages fromthe inception of fetal movement (El6, or 16days postconception) through term (E21-22).The advent of these methods for studying thebehavior of the rat fetus has provided a windowon prenatal development through which the fe-tus may be studied in naturalistic environments.

ECOLOGY OF FETAL DEVELOPMENT

Adult behavior represents a continuous andchanging interaction between the animal and itsenvironment. Ethologists have long maintainedthat it is imperative to recognize the relevantfeatures of an animal's environment if its be-

257

Page 275: 0195162854

258 DEVELOPMENT

havior is to be fully understood. Although somehave argued that the same perspective must beadopted in studying the behavior of the fetus(Smotherman and Robinson, 1988; Ronca et al.,1993), too often prenatal behavior is treated asthough the developing embryo existed in an en-vironmental vacuum. The prenatal environ-ment is buffered from some perturbations aris-ing in the world outside of the mother. But thefetal milieu also constitutes a rich and dynami-cally changing environment, in which the fetusdevelops in relation to the mother and to sib-lings that share the same needs for space andlife support.

The key maternal element of the fetal en-vironment is the uterus, comprising the highlyvascularized endometrium that provides a siteof attachment for the placenta and a muscu-lar myometrium that elastically constrains thefetus and provides periodic physical stimula-tion during nonlabor contractions (Jenkin andNathanielsz, 1994). Within the uterus, eachfetus is surrounded by its own embryonicmembranes—the chorion and amnion—which maintain a volume of amniotic fluidthat bathes the fetus throughout gestation.The concentric envelopes of the maternal ab-domen, uterus, chorion, amnion, and amni-otic fluid create a series of barriers that blockor attenuate sensory stimuli originating in theoutside world (particularly visual stimuli, butto a lesser extent, acoustic, mechanical, andchemical stimuli). At the same time, the innerworld provides a rich source of chemicalstimuli in amniotic fluid, which contains hun-dreds of chemical compounds (Wirtschafterand Williams, 1957), and mechanical stimuliarising from uterine contractions and loco-motion, changes of posture, grooming, andother active movements of the mother (Roncaet al., 1993). Siblings in utero provide anothersource of both chemical and mechanical stim-ulation that is likely to influence fetal behav-ior and development. Androgens producedby male fetuses masculinize the behaviorand morphology of female siblings in utero(Meisel and Ward, 1981), and movements of

Figure 24-1. Changes in fetal body mass (left axis) and am-niotic fluid volume (right axis) during the last 6 days of ges-tation in the rat (E16-E21). Points show means; bars depictSEM.

adjacent fetuses can affect the amount andpattern of fetal activity (Brumley and Robin-son, 2002).

Sensory stimuli arising from the internalenvironment are likely to exert differential ef-fects at different times during gestation, ow-ing to pronounced changes in fetal growth inrelation to the physical environment in utero.For example, in late gestation the fetal rat dou-bles in mass every 1.4 days, and this growthrate occurs at the same time that amnioticfluid peaks and declines in volume, producinga marked reduction in free space available forfetal movement (Fig. 24-1). Fetuses are sensi-tive to differences in the physical environ-ment, as evidenced by different rates of spon-taneous movement observed in utero, inamnion, and ex utero (Smotherman andRobinson, 1986). For these reasons, studies in-volving fetal subjects should carefully selectand control environmental conditions at thetime of behavioral testing.

MOVEMENT AND SENSATIONDURING PRENATAL DEVELOPMENT

Perhaps the most prominent aspect of fetal be-havior is the fact that all fetuses move spon-taneously before birth. Early researchers were

Page 276: 0195162854

Chapter 24. Prenatal Behavior 259

in agreement that fetal motility appears to in-volve a random and purposeless collection ofkicks, jerks, and twitches (Hamburger, 1973).Application of quantitative methods for char-acterizing fetal movements and video tech-nology for analyzing movement sequences indetail have forced a revision in the conclusionthat fetal activity lacks organization. Rat fe-tuses begin to express movement on El6,when small-amplitude movements of theforelimbs, head, and body trunk can be seen.Over the next few days, the rate of movementincreases markedly and then remains stablefrom about El8 through term (Smothermanand Robinson, 1986). Not only does theamount of activity increase, but also move-ments become more organized, showingquantitative patterning in the form of cyclic-ity and synchrony. Cyclic motor activity,which involves waxing and waning periods ofactivity of about one cycle per minute, is char-acteristic of the spontaneous movements of avariety of species, including fetal and infantrats (Smotherman et al., 1988). Because it con-tinues to be expressed in both forelimbs andhindlimbs after midthoracic spinal cord tran-section, cyclicity appears to be generated fromnumerous sources in the CNS (Robertson andSmotherman, 1990).

Synchrony is another form of temporalpatterning that involves the nearly simultane-ous movement of two or more body parts(Robinson and Smotherman, 1988). Interlimbsynchrony is expressed among all pairwiselimb combinations, with movements becom-ing more tightly coupled (intermovement in-tervals <0.2 second) as gestation proceeds(Kleven et al., in press). High levels of inter-limb synchrony are evident between fore-limbs as early as El8 of gestation, with syn-chrony becoming prominent in hindlimbs andbetween girdles 1 day later (El9). Like cyclic-ity, interlimb synchrony does not depend onneural control from rostral sources in theCNS; fetuses continue to exhibit synchronousbouts of limb activity after cervical transectionof the spinal cord (Robinson et al., 2000).

At the same time that rat fetuses areshowing pronounced developmental changesin the organization of spontaneous move-ment, they also are capable of expressing re-sponsiveness to sensory stimulation in tactile,chemical, and proprioceptive modalities. Fe-tal rats exhibit simple motor responses to cu-taneous tactile stimulation on El6. Tactilestimuli can be presented in a controlled man-ner through the use of von Frey filaments,which are calibrated to bend when a specifiedforce is applied at a point. Fetuses show sim-ple withdrawal reflexes to punctate stimuli atyounger ages but express coordinated wipingor scratching responses directed at the site ofstimulation by E20 (Smotherman and Robin-son, 1992). General tactile stimulation can beprovided by stroking an area of the skin witha soft paintbrush. Fetal responses to such stim-ulation vary by age and locus of application;gentle stroking has been shown to facilitatecoordinated body and hindlimb movementsin fetuses near term (Robinson and Smother-man, 1994), including expression of a stereo-typic leg extension response (Robinson andSmotherman, 1992a). This response consistsof elevation of hindquarters and extension ofthe hindlegs after brush strokes to the anogen-ital region, which mimics the licking behaviorof the mother directed toward newborn rats(Moore and Chadwick-Dias, 1986).

Responsiveness to chemical stimuli hasbeen extensively studied by infusing smallvolumes (20 /al) of chemosensory fluids intothe mouth of the fetus through a cannula in-serted through the lower jaw, with a flangedtip resting on the midline of the tongue(Smotherman and Robinson, 1991). Fetusesshow motor responses to chemosensory infu-sion as early as El8, and by E20 they expresscoordinated action patterns (Fig. 24-2). Intra-oral infusion of lemon extract, or other fluidswith strong olfactory components, reliablyevokes facial wiping responses in the E20 andE21 rat fetus (Robinson and Smotherman,1991a). Similarly, infusion of milk can elicitcomponents of postnatal suckling behavior

Page 277: 0195162854

260 DEVELOPMENT

Saline

Milk

Dimethyl Disulfide

Infant Formula

30% Sucrose

10% Lactose

Corn Oil

95% Ethanol

Mint (liquid)

Lemon (liquid)

Citral (liquid)

Citral (gas)

Cyclohexanone (gas)

0 20 40 60 80 100

% Fetuses Responding

Figure 24-2. Percentage of rat fetuses responding to intra-oral infusions of various chemosensory stimuli (in liquid orgas phase) with facial wiping or stretch responses.

(Robinson and Smotherman, 1992b). It islikely that fetal responses to such compoundchemical stimuli are mediated by multiplechemosensory systems. Stimuli that are pre-dominantly perceived in one modality, suchas sucrose or citral (an artificial lemon odor),can promote changes in motor activity in theE20 fetus, suggesting that gustatory and ol-factory systems are functional by this time.Further, fetuses that are prepared by surgicaltransection of the olfactory bulbs show a re-duction, but not elimination, of responsive-ness to complex chemosensory stimuli suchas lemon, suggesting that other orosensorysystems, including the trigeminal system, playa role in modulating fetal chemosensory re-sponsiveness (Smotherman and Robinson,1992).

FETAL ACTION PATTERNS

Highly sequenced and coordinated forms ofbehavior in the rat fetus, commonly referredto as fetal action patterns, can be evoked by ex-

perimental presentation of appropriate stim-uli. A prominent example of a fetal action pat-tern is facial wiping. Facial wiping involvesplacement of the forepaw onto the face fol-lowed by a downward stroke of the pawacross the face, generally from ear to nose(Robinson and Smotherman, 199la). This re-sponse, which resembles grooming and aver-sion responses in adult rats (Richmond andSachs, 1980), can be reliably evoked in rat fe-tuses by intraoral infusion of novel chemosen-sory fluids on the last 2 days of gestation (E20and E21). On El9, however, fetuses show fa-cial wiping when they are tested inside theamniotic sac but not after the membranes arestripped away (Robinson and Smotherman,1991a). Thus, E19 fetuses appear to require ad-ditional biomechanical support to stabilize thehead for paw-face contact to be established(Fig. 24-3, left).

Punctate tactile stimulation applied tothe perioral region also can evoke facial wip-ing in the fetus. Facial wiping in response tochemosensory or tactile stimulation has beenused as a bioassay to assess neurochemical-and stimulus-induced changes in fetal sensorythresholds (Smotherman and Robinson,1992). For example, administration of opioid-ergic drugs, such as the opioid agonist mor-phine, reduces fetal responsiveness to sensorystimuli. Endogenous opioid activity can beevoked by presenting biologically relevant flu-ids, such as milk or amniotic fluid, into themouth of the fetus. When challenged with anintraoral lemon infusion 60 seconds after ex-posure to milk or amniotic fluid, rat fetusesshow diminished facial wiping responses, butwiping is reinstated if subjects are pretreatedwith an opioid receptor antagonist, such asnaloxone (Smotherman and Robinson, 1992;Korthank and Robinson, 1998).

Facial wiping is just one of a collection oforganized action patterns expressed by the ratfetus. These behaviors have proved to be use-ful in the study of the ontogeny and neuralsubstrates of behavior in the prenatal rat(Robinson and Smotherman, 1992a). Experi-

Page 278: 0195162854

Chapter 24. Prenatal Behavior 261

Figure 24-3. (Left) Percentage of El9 and E20 rat fetuses exhibiting facial wiping to lemon when tested in amnion (IN) orex utero (EX). (Right) Percentage of mouthing responses that resulted in oral grasping of an artificial nipple by E19, E20,and E21 rat fetuses.

mental manipulations that mimic patterns ofpostnatal stimulation have revealed that manyaction patterns have their neurobehavioralorigins in the fetus.

ORAL GRASP RESPONSE

Fetuses that are presented with an artificialnipple near the mouth respond with lateralhead movements, oral behavior, and activeseizing of the nipple (Robinson et al., 1992).Continuous improvement in the ability tograsp the nipple occurs between E19 and E21(Fig. 24-3, right), just days before functionalsuckling at the lactating dam's nipple is re-quired of the newborn pup. Oral grasping ofan artificial nipple provides an alternativemethod for infusing fluids into the mouth ofthe fetus or neonate and has been used as aconditioned stimulus in classical conditioningexperiments with fetal rats (see later).

STRETCH RESPONSE

Shortly after milk is delivered into the mouthof the newborn rat during suckling, the pupexhibits dorsiflexion of the back, elevation ofthe head, and coordinated hindleg extension.Stretch responses can be experimentallyelicited in E20 and E21 rat fetuses by the in-traoral infusion of bovine light cream, whichis similar in fat, water, and other constitutentsto mature rat milk. In contrast to the promptstretch of neonates, the fetal stretch responseoccurs with an average latency of 180 secondsafter infusion of milk (Robinson and Smoth-erman, 1992b).

ORAL ACTIVITY

Various forms of oral activity in the fetus areevoked after chemosensory and perioral tac-tile stimulation. These include mouthing in

Page 279: 0195162854

262 DEVELOPMENT

response to oral presentation of milk andmouthing, licking, sucking, and biting in re-sponse to perioral presentation of an artificialnipple (Robinson et al., 1992). Oral activityalso occurs during spontaneous activity whenfetuses remain within the amniotic sac, sur-rounded by their own amniotic fluid.

ALTERNATED STEPPING

Administration of various neurochemicals tothe fetus has been shown to produce bouts ofalternated stepping behavior. L-DOPA (L-beta-3,4-dihydroxyphenylalanine) elicits fore-limb stepping (Robinson and Kleven, in press)that is similar to the air-stepping of neonatalrats (van Hartesveldt et al., 1990), and sero-tonin agonists elicit hindlimb stepping (Brum-ley et al., 2003). It is presumed that these neu-roactive drugs engage neural substrates thatgive rise to functional locomotion in the post-natal animal.

RESPONSE TO UMBILICAL

CORD COMPRESSION

Rat fetuses show a stereotyped behavioral andphysiological response to acute hypoxia pro-duced by compression of the umbilical cord(Robinson and Smotherman, 1992a). After oc-clusion of the umbilical cord with a mi-crovascular clamp, fetuses show an initial de-crease and then pronounced but transientincrease in spontaneous activity, accompaniedby heart rate deceleration. Neonatal rats donot show a comparable response to hypoxia,suggesting that the fetal response may be anadaptation unique to the prenatal period.

BRADYCARDIA

Heart rate is a physiological measure that isused as an index of sensory responsiveness inmany animal models. Rat fetuses exhibit atransient bradycardia after oral infusion of anovel chemosensory stimulus such as lemon,immediately before expression of the milk-

evoked stretch response, and during hypoxiainduced by umbilical cord occlusion. Episodesof tachycardia in response to stimulation havenot been reported in the rat fetus but can beexpressed by rat pups within 5 to 7 days afterbirth (Smotherman et al., 1991).

PROPRIOCEPTIVE STIMULIAND MOTOR LEARNING

The expression of species-typical action pat-terns demonstrates that fetal rats are respon-sive to exteroceptive stimuli, such as cuta-neous, olfactory, and gustatory cues. However,fetuses also are responsive to stimuli arisingfrom their own movements, as demonstratedin a paradigm developed to probe the sensori-motor dynamics of limb movement in the ratfetus (Robinson and Kleven, in press). This mo-tor learning paradigm is similar to experimentsconducted with human infants (Thelen, 1994),in which subjects are trained to learn a new pat-tern of interlimb coordination by restricting themovement of two limbs with an interlimbyoke. To effect yoke motor learning, two legsare fitted with a length of thread that providesa physical connection between the limbs for a30 minute training period. During training,movement of one leg causes the yoked leg alsoto move. Thus, proprioceptive feedback is gen-erated by both active and passive movement.Conjugate movements occur when the twolegs initiate movement at the same time andfollow parallel spatial trajectories. Accordingly,E20 rat fetuses show a significant, gradual in-crease in conjugate movements betweenyoked limbs during the training period, sug-gesting that they can use kinesthetic feedback(in the absence of explicit reinforcement) to al-ter patterns of interlimb coordination duringspontaneous motor activity (Robinson andKleven, in press). After the yoke is removed andthe legs are no longer physically coupled, fe-tuses continue to show conjugate leg move-ments for the next 15 to 25 minutes. Such per-sistence makes it unlikely that the yoke is

Page 280: 0195162854

Chapter 24. Prenatal Behavior 263

eliciting only reflexive movements. The elas-ticity and degree to which the yoke effectivelycouples the two legs influence the number ofconjugate movements during the training pe-riod (Fig. 24-4), suggesting that even minorchanges in the quality of movement-relatedfeedback can alter the expression of coordi-nated motor behavior in the rat fetus.

EXPOSURE LEARNING

Considerable evidence indicates that fetusescan modify their behavior as a consequenceof prenatal sensory experience. For instance,the rat fetus can habituate to repeated expo-sures of a sensory stimulus (Robinson andSmotherman, 1995). On E20 and E21, fetusesinitially exhibit an increase in motor activityand bradycardia in response to oral infusionof lemon, but responsiveness wanes after 5 to10 infusions. Dishabituation by infusion of anovel mint solution reinstates fetal responsesto lemon, confirming that the response decre-ment is due to central responsiveness and not

to peripheral effects such as receptor adapta-tion or effector fatigue.

Evidence also suggests that exposurelearning in the fetus provides a scaffold forlater behavior. Familiar tastants experiencedin utero can be distinguished from novel tas-tants as measured by overall fetal motor ac-tivity a few days later (Robinson and Smoth-erman, 1991b) and by mouthing and lickingbehavior in later-term fetuses and neonates(Mickley et al., 2000). Cues present in amni-otic fluid (Hepper, 1987) or maternal diet(Hepper, 1988) are preferred during taste orodor tests conducted at postnatal ages. In ad-dition, prenatal exposure to a particular odorcue can influence subsequent associativelearning with that odor in newborn pups(Chotro et al., 1991). Because many con-stituents of maternal diet can cross the pla-centa to enter fetal circulation and amnioticfluid, simple exposure learning may play animportant role in the establishment of dietarypreferences during prenatal development(Robinson and Smotherman, 1991b).

Unyoke

- Elastic Yoke

- Thread Yoke

- Rigid Yoke

Time (min)

Figure 24-4. Conjugate movements expressed as a per-centage of hindlimb activity during yoke motor learning inthe E20 rat fetus. During training, the hindlimbs are fittedwith an interlimb yoke made of elastic thread (elastic), flex-ible suture (thread), or thread stiffened with cyanoacrylate(rigid). Following the 30 minute training period, the yokeis removed (dotted vertical line). Control subjects are fittedwith a yoke that is cut at the beginning of training(Unyoke). Points show mean percent conjugate move-ments in successive 5 minute intervals; bars depict SEM.

ASSOCIATIVE LEARNING

Associative learning also can be acquired andexpressed before birth. In the classical condi-tioning of activity in the fetal rat (Robinsonand Smotherman, 1991b), a neutral condi-tioned stimulus (sucrose CS) is paired with anunconditioned stimulus that activates fetal be-havior (lemon US). After four CS-US pairings,E20 fetuses respond to the CS alone with aconditioned increase in motor activity. Phys-iological responses also can be conditioned inthe fetal rat. For example, intraoral infusionsof milk reduce fetal sensory responsiveness bypromoting activity in the endogenous opioidsystem (Smotherman and Robinson, 1992).After paired presentations of an artificial nip-ple (CS) and milk (US), fetuses exhibit a re-duction in cutaneous responsiveness when re-exposed to the nipple CS alone, an effect thatis mediated by a conditioned increase in opi-

Page 281: 0195162854

264 DEVELOPMENT

old activity (Robinson and Smotherman,1995).

The effects of associative learning in thefetus can have lasting effects on fetal behav-ior and response to stimulation. In ataste/odor aversion paradigm of perinatallearning, intraperitoneal (IP) injection of LiClis used as an aversive US (Robinson andSmotherman, 1991b). On El 7, injection ofmint odor (a neutral CS) into the amniotic sacfollowed by IP injection of LiCl has an im-mediate effect of suppressing fetal activity andevoking body curls. When reexposed to themint CS 2 days later, E19 fetuses exhibit a con-ditioned motor response that resembles theunconditioned response of El7 fetuses. Like-wise, newborns that have been exposed as fe-tuses to apple juice paired with LiCl injectionare less likely to suckle at nipples painted withapple juice and to exhibit longer latencies tomove across a short runway suffused with ap-ple odor to suckle from a lactating dam.Therefore, associative learning, like exposurelearning, can be expressed by rat fetuses be-fore birth and is capable of influencing post-natal behavioral development.

CONCLUSIONS

Study of the rat fetus in vivo provides a sim-ple mammalian system that permits investi-gation of early neurobehavioral development.Unlike other model systems that are promi-nently used in basic neuroscience research,such as surgically reduced preparations orphylogenetically simpler organisms, study ofthe rat fetus provides an alternative animalmodel that offers the unique advantage ofdevelopmental continuity with behavioralfunctions in the adult mammal. Research hasconfirmed that behavioral organization, coor-dinated movement, sensory responsiveness,and learning all can be expressed in the pre-natal rat. Study of the rat fetus thus is well sit-uated to complement existing research ap-proaches with adult rats as well as simpler

systems to understand basic problems in be-havioral neuroscience and neurobehavioraldevelopment.

ACKNOWLEDGMENT

S.R.R. is supported by National Institute of Child Health andHuman Development grant HD-33862.

REFERENCES

Brumley MR, Fleenor RA, Simmons LL, Robinson SR(2003) Serotonin agonists alter motor activity andpromote hindlimb stepping in the intact and mid-thoracic transected E20 rat fetus (abstract). Devel-opmental Psychobiology 43:249.

Brumley MR and Robinson SR (2002) Responsiveness ofrat fetuses to sibling motor activity: Communica-tion in utero? (abstract). Developmental Psychobi-ology 41:73.

Chotro MG, Cordoba NE, Molina JC (1991) Acute pre-natal experience with alcohol in the amniotic fluid:Interactions with aversive and appetitive alcoholorosensory learning in the rat pup. DevelopmentalPsychobiology 24:431-451.

Hamburger V (1973) Anatomical and physiological ba-sis of embryonic motility in birds and mammals. In:Behavioral embryology (Gottlieb G, ed.), pp. 51-76.New York: Academic Press.

Hepper PG (1987) The amniotic fluid: An importantpriming role in kin recognition. Animal Behaviour35:1343-1346.

Hepper PG (1988) Adaptive fetal learning: prenatal ex-posure to garlic affects postnatal preferences. Ani-mal Behaviour 36:935-936.

Jenkin G and Nathanielsz PW (1994) Myometrial activ-ity during pregnancy and parturition. In: Textbookof fetal physiology (Thorburn GD and Harding R,eds.), pp. 405-414. Oxford: Oxford University Press.

Kleven GA, Lane MS, Robinson SR (in press) Develop-ment of interlimb movement synchrony in the ratfetus. Behavioral Neuroscience.

Korthank AJ and Robinson SR (1998) Effects of amnioticfluid on opioid activity and fetal responses tochemosensory stimuli. Developmental Psychobiol-ogy 33:235-248.

Meisel RL and Ward IL (1981) Fetal female rats are mas-culinized by male littermates located caudally in theuterus. Science 220:437-438.

Mickley GA, Remmers-Roeber DR, Grouse C, WalkerC, Dengler C (2000) Detection of novelty by peri-natal rats. Physiology and Behavior 70:217-225.

Page 282: 0195162854

Chapter 24. Prenatal Behavior 265

Moore CL and Chadwick-Dias AM (1986) Behavioral re-sponses of infant rats to maternal licking: Variationswith age and sex. Developmental Psychobiology19:427-438.

Richmond G and Sachs BD (1980) Grooming in norwayrats: the development and adult expression of acomplex motor pattern. Behaviour 75:82-96.

Robertson SS and Smotherman WP (1990) The neuralcontrol of cyclic activity in the fetal rat. Physiologyand Behavior 47:121-126.

Robinson SR and Kleven GA (in press) Learning to movebefore birth. In: Prenatal development of postnatalfunctions (Hopkins B and Johnson S, eds.). West-port, CT: Greenwood Publishing Group.

Robinson SR and Smotherman WP (1988) Chance andchunks in the ontogeny of fetal behavior. In: Behav-ior of the Fetus (Smotherman WP and Robinson SR,eds.), pp. 95-115. Caldwell, NJ: Telford Press.

Robinson SR and Smotherman WP (199la) The amni-otic sac as scaffolding: Prenatal ontogeny of anaction pattern. Developmental Psychobiology 24:463-485.

Robinson SR and Smotherman WP (199lb) Fetal learn-ing: Implications for the development of kin recog-nition. In: Kin recognition (Hepper PG, ed.), pp.308-334. Cambridge: Cambridge University Press.

Robinson SR and Smotherman WP (1992a) Fundamen-tal motor patterns of the mammalian fetus. Journalof Neurobiology 23:1574-1600.

Robinson SR and Smotherman WP (1992b) Organiza-tion of the stretch response to milk in the rat fetus.Developmental Psychobiology 25:33-49.

Robinson SR and Smotherman WP (1994) Behavioral ef-fects of milk in the rat fetus. Behavioral Neuro-science 108:1139-1149.

Robinson SR and Smotherman WP (1995) Habituationand classical conditioning in the rat fetus: Opioid in-volvements. In: Fetal development: A psychobio-logical perspective (Lecanuet JP, Krasnegor NA,Fifer WP, Smotherman WP, eds.), pp. 295-314.New York: Lawrence Erlbaum & Associates.

Robinson SR, Blumberg MS, Lane MS, Kreber L (2000)Spontaneous motor activity in fetal and infant ratsis organized into discrete multilimb bouts. Behav-ioral Neuroscience 114:328-336.

Robinson SR, Hoeltzel TCM, Cooke KM, Umphress SM,Murrish DE, Smotherman WP (1992) Oral capture

and grasping of an artificial nipple by rat fetuses. De-velopmental Psychobiology 25:543-555.

Ronca AE, Lamkin CA, Alberts JR (1993) Maternal con-tributions to sensory experience in the fetal andnewborn rat (Rattus norvegicus}. Journal of Com-parative Psychology 107:61-74.

Smotherman WP and Robinson SR (1986) Environ-mental determinants of behaviour in the rat fetus.Animal Behaviour 34:1859-1873.

Smotherman WP and Robinson SR (1988a) The uterusas environment: The ecology of fetal experience. In:Handbook of behavioral neurobiology, vol 9, De-velopmental psychobiology and behavioral ecology(Blass EM, ed.), pp. 149-196). New York: PlenumPress.

Smotherman WP and Robinson SR (1991) Accessibilityof the rat fetus for psychobiological investigation.In: Developmental psychobiology: New methodsand changing concepts (Shair HN, Hofer MA, BarrG, eds.), pp. 148-164. New York: Oxford UniversityPress.

Smotherman WP and Robinson SR (1992) Prenatal ex-perience with milk: Fetal behavior and endogenousopioid systems. Neuroscience and BiobehavioralReviews 16:351-364.

Smotherman WP and Robinson SR (1997) Prenatal on-togeny of sensory responsiveness and learning. In:Comparative psychology: A handbook (GreenbergG and Haraway MM, eds.), pp. 586-601. New York:Garland Press.

Smotherman WP, Robinson SR, Hepper PG, Ronca AE,Alberts JR (1991) Heart rate response of the rat fe-tus and neonate to a chemosensory stimulus. Phys-iology and Behavior 50:47-52.

Smotherman WP, Robinson SR, Robertson SS (1988)Cyclic motor activity in the fetal rat (Rattus norvegi-cus). Journal of Comparative Psychology 102:78-82.

Thelen E (1994) Three-month-old infants can learn task-specific patterns of interlimb coordination. Psycho-logical Science 5:280-285.

van Hartesveldt C, Sickles AE, Porter JD, Stehouwer DJ(1990) L-DOPA-induced air-stepping in developingrats. Developmental Brain Research 58:251-255.

Wirtschafter ZT and Williams DW (1957) The dynam-ics of protein changes in the amniotic fluid of nor-mal and abnormal rat embryos. American Journalof Obstetrics and Gynecology 74:1022-1028.

Page 283: 0195162854

Infancy

JEFFREY R. ALBERTS25

Born after a mere 22 days of gestation, theNorway rat pup is understandably immature.Over the next 3 to 4 weeks, development israpid and dramatic. The newborn (day 0 to 4),infant (day 5 to 10), juvenile (days 11 to 17),and weanling (days 17 to 28 and beyond)stages are distinct and dynamic. The presentchapter offers brief overviews of pup growthand differentiation of behavioral systems.There follows a kind of "ethogram" of earlypostnatal behavior in Rattus norvegicus, in-tended to put into a more natural contextsome of the processes of sensory and motordevelopment, with an emphasis on informa-tion that can be used for designing and inter-preting tests with immature rats.

SOME DEVELOPMENTAL SEQUENCES

There are a variety of descriptive studies ofthe young Norway rat's behavior. WillardSmall's diary-like daily observations is repletewith rich descriptions and some simple but re-vealing tests (Small, 1899). The account byHolies and Woods (1964) is more objectivelyobservational, based mostly on daily time-samples of a mother and litter in a laboratorycage. Altman and Sudarshan (1971) describedpostnatal rat development by reporting the re-sults of a comprehensive battery of motortests. There are, in addition, accounts andanalyses of different features of develop-ment—sensory and perceptual ontogenesis(Alberts, 1984), physiological and regulatorydevelopment (Adolph, 1971), and different

topics in behavioral maturation. Integrativestudies of maternal behavior and pup devel-opment are also available (e.g., Rosenblatt,1965; see also Chapter 27).

With such resources at hand, it is onlypractical here to recognize some of the strik-ing sequences that comprise the Norway rat'srapid and dramatic behavioral ontogeny frombirth to postnatal week 3 or 4, after which itweans to independence.

GROWTH AND SEQUENTIALCHANGES IN APPEARANCE

Domesticated rat pups weigh about 7 gramsat birth (day 0). Once they are cleaned by thedam and begin stable breathing, they developa rich, red coloration, visible over the furlessbody. The glabrous pups' skin is so thin thatthe mother's milk can be seen in their stom-ach through the abdominal wall. The new-born has mere bulges for eyes and folds of skinwhere the external ears (pinnae) will form.

Close inspection reveals arrays of finewhiskers on the mystacial pads. The nares(nostrils) are open, because breathing occursthrough the nose. There are milk teeth, butthe incisors have yet to appear. Paws andclaws are nicely formed but the pups can notstand, grip, or ambulate.

Postnatal growth is rapid (20 grams byday 10, 30 grams on day 15). Each day, thepup's general appearance changes. The steadyincrease in body mass includes skeletalgrowth and differentiation accompanied bymuscle growth by which antigravitational

266

Page 284: 0195162854

Chapter 25. Infancy 267

support and movements are sustained. Sub-cutaneous fat adds some insulation. Beginningon about day 5, there are signs of fur growthand coat coloration is visible. By day 10, thepup is covered with fur, although a coat con-taining both guard hairs and underhairs is at-tained only after at least 3 weeks of age, if notlater (Gerrish and Alberts, 1995). The pinna ofthe external ear separates from the head byabout day 10, and the external auditory mea-tus opens on day 12. Eyelids unseal by day 15.

SEQUENCES OF SENSORY DEVELOPMENT

At birth, some, but not all, sensory systemsare functional. Moreover, those that are func-tional in the neonate are not functionally com-plete (i.e., within each modality, the pups con-tinue to develop responsiveness to a greaterrange of stimuli, increased sensitivity to lowerlevels of stimulation, and improved acuity,discrimination, and recognition). Thus, it isimportant to distinguish between onset offunction and the subsequent development offunction (Alberts, 1984).

Onset of sensory function is thought to oc-cur via an immutable sequence, possibly uni-versal for all vertebrates (Gottlieb, 1971; Alberts,1984), with tactile, vestibular, auditory, and vi-sual function beginning in this order. Othermodalities have not been sequenced this way,but at birth, a Norway rat pup also displays rudi-mentary function in its tactile, vestibular, ther-mal, and chemosensitive systems.

Tactile function is present only on somebody regions. Pups respond to punctate prob-ing with a von Prey hair in the perioral area, onthe forepaws, and around the anogenital region.There is a general, rostral-to-caudal topographicspread of tactile sensitivity. The vibrissae(whiskers) may be functional as "tactile hairs,"but this has not been studied behaviorally.Rather, it has been found that on the day ofbirth, after 30 minutes of stroking of a vibrissalpad, there was an increase in 2-deoxyglucose up-take in the trigeminal sensory nucleus (Wu andGonzalez, 1997) ipsilateral to the stimulation.

Prenatal vestibular responses to tiltingstimulation (angular acceleration) has beendemonstrated by movement reactions and bytachycardia reflexes (Ronca and Alberts, 1994,2000). Newborns demonstrate righting re-sponses (although these become more reliableand robust with age). Indeed, the strength ofmost vestibular reflexes appears to increaseontogenetically (see Altman and Sudarshan,1974). Geotaxis has long been considered oneof the infant rats' characteristic responses (e.g.,Crozier and Pincus, 1929), but this behavior hasbeen reconsidered (Krieder and Blumberg,1999) and reinterpreted (Alberts et al., in press).

The chemical senses are also operative atbirth but undergo extensive development. Be-cause the rat is an obligate nose breather, thedevelopment of nasal sampling (e.g., sniffing)is significant (Welker, 1964; Alberts and May,1980a). In one of the only olfactometric assaysof chemosensitivity in rat pups, it was foundthat sensitivity to natural and chemical odorsincreases gradually, until at least 17 days ofage (Alberts and May, 1980b). Trigeminal,vomeronasal, and main (cranial nerve I)chemosensitive receptors transmit olfactory in-formation, and each of these has an ontoge-netic timetable. Experience and level of stimu-lation are components to the timing of thesedevelopments (Alberts, 1981; Brunjes, 1994).

Taste function has been demonstrated byintroducing fluid stimuli into the mouth andmeasuring behavioral responses (e.g., Hall andBryan, 1981). There also are electrophysiolog-ical data that support the presence of gustatoryfunction in the rat pup (e.g., Hill, 1987). Again,increased range and levels of sensitivity follow.The pups' impressive abilities to integrate de-velopmentally into their behavior complex in-formation from the chemical senses is exem-plified by the findings that cues in amnioticfluid and mother's milk can capture control ofolfactory and taste-guided behaviors, such assuckling (Pedersen and Blass, 1982) and foodrecognition (Galef and Sherry, 1973).

Temperature sensitivity is seen in a new-bom's thermogenic response to cooling. Al-

Page 285: 0195162854

268 DEVELOPMENT

though there are no quantitative assessmentsof the temperature sense, simply moving apup from one incubator to another, within a23° C room, is sufficient to trigger an episodeof nonshivering thermogenesis (Efimova, etal., 1992). Temperature sensitivity is also seenin pups' movements on a thermal gradient(Kleitman and Satinoff, 1982). A pup's sensi-tivity to temperature gradients, indicated bythe ability to spatially discriminate tempera-ture differences, probably refelects discrimi-nation of temperature differences across bodyareas. Related to this, perhaps, is the findingthat conductive warmth to the ventrum is apotent reinforcer for newborn rats in an op-erant task (Flory et al., 1997).

Pups can respond to acoustic stimulationbefore their ears open, but there is a dramaticincrease in sensitivity when, around day 12,the auditory meatus opens and drains. Audi-tory sensitivity increases earlier for low fre-quencies, with sensitivity to high frequenciesdeveloping later (Brunjes and Alberts, 1981).

Rat pups are sensitive to light before theeyelids unseal: Pups tend to move away fromsources of light (negative geotaxis). When theeyes open, the pups can resolve displays of atleast l°2l' visual angle, as evidenced by anoptokinetic head nystagmus to a moving ar-ray (Brunjes and Alberts, 1981). Depth per-ception also develops and is experience de-pendent (Tees, 1976).

Eye opening does not necessarily signalthe onset of pattern sensitivity. For example,hyperthyroidism accelerates eye opening inrats, but the precocially unsealed lids exposea visual system that is not similarly advancedin its development. The milestone of eyeopening cannot be used as a marker for visualdevelopment (Brunjes and Alberts, 1981).

SEQUENCES OF MOVEMENT ANDPOSTURAL DEVELOPMENT

Respiratory movements are prominent in thenewborn's behavioral repertoire. The onsetand establishment of regular breathing move-

ments occur within the first hour after partu-rition. Respiration rate in the calm, 1-day-oldresting pup is about 2 cycles per second (cps).This basal rate increases to about 4 cps by day7 to 9 (Alberts and May, 1980a).

The breathing pup also behaves by curl-ing and extending its body along the sagittalplane. When actively curling, it forms a Cshape. Conversely, the extension can be socomplete that the pup's spine forms the shapeof lordosis. Stretching and extending by anewborn within the first couple of hours ofpostnatal existence create the squirming ap-pearance of newborn rats.

On day 1, the ventrum-down orientationis a pup's primary orientation (c.f., Fraenkeland Gunn, 1940). This is seen when a pup isplaced on its back. They actively reorient andright themselves and resume their primaryorientation. Pellis et al. (1991) traced ontoge-netic changes in the pups' righting (turningfrom supine to prone). There are develop-mental sequences of righting strategies, whichthey analyzed in terms of a sequence of forms,each triggered by tactile and vestibular stim-uli. Their detailed account of ontogenetic se-quence is striking and provides recognizablebehavioral modules that appear in othersettings.

When a newborn or infant is prone, itslegs are often splayed outward. It may moveits head from side to side. Such scanningmovements may be a form of olfactory, tac-tile, or thermal sampling. By day 4 to 5, scan-ning movements recruit additional spinal seg-ments and the forelimbs. Pups frequently turntheir head to one side and extend the con-tralateral forelimb, effectively pushing, orpunting, their body into a partial rotation. Be-cause the hindlimbs tend to be inert at theseages, the pups' "punting behavior" movesthem in rough circles.

Pups begin using a quadraped stancearound day 10 and then begin to crawl (day10 to 11) and then walk (day 12 to 13) and run(day 15). The developmental kinematics oflimb and interlimb movments involved in lo-comotion have been described (e.g., Bekoff

Page 286: 0195162854

Chapter 25. Infancy 269

and Trainer, 1979; Stehouwer and VanHartesveldt, 2000).

Limb movements during face- and body-washing sequences have been analyzed de-velopmentally (Richmond and Sachs, 1980).Ventral grooming can occur when the pup as-sumes a stereotyped vertical pose, balancedon hindlimbs and tail. In this position, it canlean forward and groom its ventral surface,even reaching the genital area. Self-groomingincreases and then differentiates sexually, withmales engaging in more self-stimulation of theanogenital area during the prepubertal period(Moore and Rogers, 1984).

Locomotor activity has been measureddevelopmentally with a variety of "stabilime-ter" devices. Under several different, well-con-trolled conditions, isolated pups gradually in-crease level of activity from day 1 to at leastday 12. Then there is a 10-fold increase in ac-tivity, followed by a dramatic dimunition tolower, more stable levels (Campbell et al.,1969). Interestingly, when the ontogeny ofgeneral activity is measured in the presence ofan anesthetized rat dam, the 15-day activitypeak is not seen; under such conditions, thereis a generally linear increase in activity fromdays 5 to 30 (Randall and Campbell, 1976).With this observation of different resultswhen developmental measures are made un-der more natural conditions, we turn to a re-view of the developing pups' "species-typical"behavior.

The fetal rat is expelled from themother's body during a 6 hour labor, withuterine and abdominal contractions thatsqueeze and push the fetus through the birthcanal to the outside world (Ronca et al., 1993).Each newborn is encased within an amnioticsac that the mother orally removes and con-sumes, along with amniotic fluid and pla-centa. After the umbilical connection is sev-ered, the newborn rat begins postnatalbehavior. Interestingly, the ontogeny of post-natal breathing is facilitated not by the physio-logical stimulus of hypoxia, but rather by the sen-sory stimulus of the birth process, which iscomposed of specific tactile, proprioceptive,vestibular, and thermal events (Ronca and Al-berts, 1995a, 1995b). The operational statusof the perinates' sensory-perceptual systems,however immature it may be, is sufficientlysensitive across modalities and adequately"tuned" within each modality that the new-born can respond adaptively to its new world(Alberts and Ronca, 1993).

When the rat dam finishes parturition andhas consumed the placentas and cleaned thepups, she gathers the pups and assembles thembeneath her body, usually within a constructednest. This form of contact behavior, or brood-ing, allows for conductive heat exchange fromthe dam's body to those of the pups withwhich she is in contact. Under these condi-tions, augmented by an insulative nest, thepups' body temperatures rise to about 35° C.

AN ETHOGRAM OF ANORWAY RAT'S EARLY LIFE

The newborn rat is a stunningly immatureand incomplete creature, when its appear-ance, morphology, sensory capacities, motorabilities, coordination, and behavioral reper-toire are compared with those of the adult.But the same features that are "incomplete"in relation to the adult phenotype can simul-taneously be seen as "complete," well articu-lated, and serving as adaptation when evalu-ated within the context of newborn's world.

SUCKLING

If breathing is a pup's initial adaptive postnatalbehavior, suckling is probably the next. Thefirst nipple attachment, like those that followthroughout the 3 to 4 weeks of suckling, is un-der olfactory control. Anosmic pups do notsuckle (see Alberts, 1976). A maternal cue forsuckling was demonstrated by Teicher andBlass (1976; see also Hofer et al., 1976). Wash-ing the dam's ventrum eliminates suckling andpainting the nipples with a distillate of the washreinstates the pups' behavior (Teicher andBlass, 1978). Nevertheless, the olfactory cue or

Page 287: 0195162854

270 DEVELOPMENT

cues that stimulate suckling do not attract pupsto the nipples. Rather, the odor activates thepups, and their increased activity (extending,scanning, squirming, probing) brings them intocontact with the dam and a nipple. Oral re-flexes, triggered by perioral tactile cues, are re-sponsible for nipple apprehension.

The specific odors that can activate thepup to suckle appear to be learned. Normally,amniotic fluid and saliva are potent stimuli(Teicher and Blass, 1978). Pederson and Blass(1976) showed that a pup's initial nipple at-tachment can be activated by an odor intro-duced into the amniotic fluid on El7, about 5days before term. The underlying mecha-nisms of this experience-dependent processand the extent to which this represents a nat-uralistic form of classical conditioning remaintopics of study, but much has been learnedabout the initiation and expression of suckling(Blass and Teicher, 1980). One implication ofthese findings is that the pups' normal re-sponses to amniotic fluid are similarly learned.

The typical sequence for mother-initiatednursing bouts begins when the dam investi-gates the pups, sniffing and licking them andhandling them with her forepaws (Rosenblatt,1965). This stimulation activates them. Thepups' stretching and probing induce thekyphosis posture in the dam (Stern, 1988), andthe pups attach to nipples. The litter's sucklingbehavior creates a series of neuroendocrineevents in the dam that culminate in simulta-neous release of milk to all 12 nipples. Thepups in turn display such a dramatic, reflexive,whole-body response "stretch response" to thereceipt of milk that the pups' stretch responsecan be used as a bioassay of intramammarypressure and milk letdown (Lincoln et al.,1973).

After a milk letdown and stretch response,pups often release the nipple and begin tosquirm and root into the dam's ventrum. Be-cause a common event in the dam simultane-ously triggers the behavior of all of the pups,after a letdown the litter often becomes a scath-ing mass that settles as the pups reattach to a

nipple. This sequence of letdowns in the damand the explosive behavior of the pups occursevery 6 to 9 minutes during a nursing bout.Viewed in the broader context of develop-mental time, the greatest frequency and dura-tion of nursing bouts are early in the postnatalperiod. Letdowns begin to decrease around day18 (Cramer et al., 1990).

Pups live in close association with themother's body for much of each day from day0 to day 15, after which the mother begins towithdraw behaviorally. Leon et al. (1978) sug-gested that thermal factors limit the mother'sability to remain in prolonged contact with thelitter. Stern and Azzara (2000) dispute this in-terpretation. The issue may involve the ther-mal parameters that are normally experiencedby the dam and the litter in the nest; body heatand metabolic demands of lactation and devel-opment are, under any circumstances, centralfactors in ontogenetic regulations.

Until about day 15, the major limiting fac-tor in the pups' milk intake is the availabilityof mother's milk (Friedman, 1975; Hall andRosenblatt, 1977). Satiety mechanisms in thepups do not determine when the pups cease in-gesting milk. At early ages, under natural con-ditions with the dam, the pups are essentiallysuckling machines that ingest all of the avail-able milk. Nevertheless, regulatory mecha-nisms for ingestion can be revealed under ex-perimental conditions (Hall and Williams,1983).

HUDDLING

Contact behavior, or huddling, is also preva-lent in rat repertoire. Huddling begins soonafter birth and, under most circumstances, ismaintained throughout adult life. Figure 25-1illustrates some typical huddles of pups at twoages. Few behaviors are so fundamental andenduring. As the mother makes longer andmore frequent excursions from the nest(Cramer et al., 1990), the huddle becomes thepups' immediate environment (Alberts andCramer, 1988).

Page 288: 0195162854

Chapter 25. Infancy 271

Figure 25-1. Huddles of rat pups at5 and 20 days of age. Despite the dra-matic sensory and motor develop-ments, huddling is a continuousfeature of the pups' behavioral reper-

Much is known about huddles of pups. Ahuddle is far more than a pile of bodies. Thebehavior of individual pups actively forms andmaintains the group. Through the behavior ofindividuals, a group behavior emerges (Alberts,1978a; Schank and Alberts, 1997a, 1997b). Thisgroup displays a form of group-regulatory behav-ior in which its total exposed surface area varieswith ambient temperature. By huddling thisway, body heat can be retained and metabolicenergy conserved (Alberts, 1978a).

The pups' huddling behavior is undermultisensory control (Alberts, 1978b) at allages, with age-related hierarchies of sensorydominance. For instance, thermal stimuli aremore salient than olfactory cues for elicitinghuddling by 5- and 10-day-old rats, whereasolfactory cues are dominant in the 15- and 20-day-olds (Alberts and Brunjes, 1978). More re-cent analyses indicate that 7-day-old pups areindifferent to the activity state of adjacentpups, but by day 10 their huddling is affectedby the movements and activity state of theirlittermates (Schank and Alberts, 1997b).

The cutaneous contact derived from hud-dling is a form of thermotactile stimulationthat serves as a reinforcer for the associativelearning of odors. Although suckling per se

and the ingestion of milk can both function asreinforcers of behavior, the formation of odorpreferences for huddling is not affected bysuckling or milk but is apparently inducedby thermotactile stimulation (Alberts andMay,1984). Under different conditions, tactilestimulation alone (stroking) can also have re-inforcing properties (Sullivan and Hall, 1988).

VOCAL EMISSIONS AND SOME

REFLEX-LIKE REACTIONS

When infant and juvenile rat pups are isolatedand cool, they emit vocalizations in the 40 to50 kHz range that are anthropomorphicallytermed "ultrasounds," although the emissionsare easily detectable by adult rats. A burst ofhigh-frequency or ultrasonic vocalization(USV) appears to attract the attention of amother and can lead to retrieval behaviors(Allin and Banks, 1972).

Two types of analysis prevail in the con-trols of USV production by rat pups. One isthermal and regulatory; the other emphasizesisolation and inferred internal states, such asdistress. Although these views are not mutu-ally exclusive, the literature appears dividedand contradictory.

toire

Page 289: 0195162854

272 DEVELOPMENT

Without doubt, the loss of body heat anddecreased body temperature provide a potentstimulus for the onset of USV production.Heating the pup reduces such USVs. A regu-latory explanation of infant rat USV posits thatcooling creates a homeostatic perturbation inthe pup, leading to responses that involvechanges in respiration. The pup's altered res-piratory maneuvers can produce as a byprod-uct laryngeal noises in the high-frequencyrange (due simply to the small size of the pup'sanatomy). The oxygen demands of increasedmetabolic heat production have been hy-pothesized as a basis of USV; more recently,abdominal maneuvers that enhance bloodflow to the heart under cool conditions thatalter blood viscosity have also been identified(see Chapter 35).

The alternate view emphasizes the re-moval and replacement of social or maternalcues (e.g., Hofer and Shair, 1978, 1987). Manyof the studies that were intended to demon-strate USV responses to isolation failed to con-trol for the rapid heat loss sustained by thesmall, thermally fragile rat pups, and it has be-come clear that these pups are behaviorallyand physiologically sensitive to small and briefepisodes of cooling (Blumberg et al., 1992).There do appear, however, to be instances inwhich nonthermal factors can stimulate oraugment USV production, requiring more

systematic and integrative studies to attain asolid understanding of this aspect of the pup'sbehavior.

USVs appear to play a role in attractingthe dam to an isolated pup and eliciting re-trieval or transport. Brewster and Leon (1980)described the appearance and ontogenetic dis-appearance of the transport reflex in rat pups,whereby stimulation to the skin on the backof young pups elicits a reflexive vetroflexionof the tail, relative immobility, and raising ofthe hindlimbs. Such responses augment themother's efficiency in carrying the pup. Fig-ure 25-2 illustrates some of their findings, in-cluding how the transport reflex wanes de-velopmentally and diminishes greatly afterday 24, when pups are independently mobileand presumably no longer require maternaltransport.

Other "disappearing reflexes" presentcompelling cases of developmentally coordi-nated timing and function in relation to thecycle of maternal behavior. At birth, pups aregenerally incapable of urinating or defecatingspontaneously. Nevertheless, in response toanogenital stimulation, they display a mic-turition reflex. The mother's licking releasesurination from the pup, and this behavior notonly serves the immediate needs of the pupsbut it maintains an interactive bond betweenmother and pups and in fact serves the hy-

Figure 25-2. The transport reflex(upper left) facilitates the mother'sability to carry pups because theirlimbs are retracted and held close tothe body. The reflex is triggered bythe tactile stimulation involved ingrasping and lifting them; dorsal re-gions 1 and 2 (upper right) were themost sensitive areas for evoking thereflex. The graph shows the strength(probability) of the pups' transportreflex, emerging as they increase insize and waning as they approachindependence.

Page 290: 0195162854

Chapter 25. Infancy

drational needs of the dam (Friedman et al.,1981; Alberts and Gubernick, 1983). Lactatingrats have a robust sodium appetite (Richterand Barelare, 1938; Alberts and Gubernick,1983; Gubernick and Alberts, 1983) and avidlyseek and consume the pups' hypotonic urine.As the pups' kidneys develop, solute concen-tration increases, and as urine becomes hy-pertonic, spontaneous urination develops andmaternal licking disappears.

Moore and Chadwick-Diaz (1986) de-scribe another reflex that supports mother-offspring interactions during licking and thatdisappears with development. The dam re-quires access to a pup's anogenital area to ef-ficiently stimulate the micturition reflex. Thepup's anogenital region is most accessiblewhen the pup is supine, but, as we have noted,the righting reflex is strong and reliable. A dra-matic inhibition of the righting response canbe produced by tactile pressure to the pup'sventrum. This is the sort of stimulation thatthe mother's snout and paws provide duringanogenital licking of a supine pup. This inhi-bition of righting wanes developmentally,with the kind of coordinate timing that sug-gests ontogenetic adaptation.

WEANING

Weaning is a developmental process uniqueto and universally represented in mammalianyoung. Formally, weaning is the shift from in-gestion of mother's milk to independent in-gestion of solid food and water. In the Nor-way rat, weaning occurs naturally acrosspostnatal days 14 to 34, or so. It is commonlaboratory practice "to wean" a litter (i.e., toseparate offspring from the dam) around day21. This can be done, but here we discuss amore gradual, naturally occurring process. Ingeneral, weaning represents the achievementof independence from reliance on maternal orparental resources.

In the domain of ingestive behavior,weaning is composed of two processes thatproceed simultaneously but largely indepen-

273

dently. One is the abandonment of suckling,and the other is the onset of independent feed-ing and drinking. Hall and Williams (1983)synthesized a large body of data and arguethat indeed suckling and feeding are separablebehaviors, controlled by different stimuli andmediated by different neural pathways. Thedevelopmental sequences during weaningseem consistent with such an analysis.

The mother's milk production peaksaround day 15. A litter of eight pups may take60 ml of milk per day around that time (Fried-man et al., 1981). Time spent nursing is about10 hours per day (Cramer et al., 1990). Con-trary to some hypotheses, weaning in rats isnot caused by a deficit in the dam's ability toproduce sufficient milk relative to the nutri-tive requirements of the developing litter. At15 days of age, litters consume all of the milkavailable from a dam, whereas 20-day-old lit-ters do not (Thiels et al., 1988). The diminu-tion in suckling derives from the behavioralinteractions with the dam. As the dam's pat-tern of maternal behavior creates more fre-quent and longer departures from the nest andlitter, pups suckle less. The less they suck, theless likely they are to resume suckling (Thielset al., 1988). Conversely, the act of sucklingmaintains itself. Thus, it appears that the de-crease in a pup's suckling response to a damdiminishes as a consequence of decreased ac-cess. In addition, there is a separate but coor-dinate set of experiences that lead to the on-set of feeding, food site selection, and tastepreferences. Homeostatically regulated as-pects of intake also develop, in part on the ba-sis of experience.

Onset of feeding begins with sampling ofsolid foods. Depending on the configurationand characteristics of the pups' environment,such sampling begins between days 14 and 16.Galef and associates provided a rich accountof the cues and behavioral processes that con-tribute to the sampling and selection of dietby weanling rats. After 2 weeks of postnatallife, a rat pup has experienced a variety ofchemical cues in its environment, including

Page 291: 0195162854

274 DEVELOPMENT

flavors in mother's milk; odors associatedwith maternal and sibling urine, feces, saliva,and breath; and secretions of the sebaceousand preputial glands. Thus, when the youngrat ventures from the nest, there are numer-ous cues that it can and does recognize as fa-miliar. For example, pups readily detect"residual cues" left by conspecifics at a feed-ing site (Galef and Heiber, 1976).

Wild and domesticated rat pups reared inseminatural environments in a laboratorymake their first departures from the natal nestbetween days 16 and 19 (Galef and Clark,1971). The mother is usually outside the nestentrance when pups egress, but experimentalanalyses indicate that the pups are not seek-ing the dam. Rather, they respond to visual,acoustic, and/or olfactory cues (Leon andMoltz, 1972; Alberts and Leimbach, 1980;Galef, 1983). The cues that trigger egressionfrom the nest and approach to feeding sitesare remarkably general. There are no prefer-ences for the mother in relation to anotheradult (Galef and Clark, 1971), and it seems thatmere stimulus strength determines probabil-ity of approach (Gerrish and Alberts, 1995).

Adult rats tend to congregate at feedingsites. When pups approach conspecifics, it islikely that they approach an area where adultsare feeding. It appears that the onset of solidfood intake may involve the power of a gen-eral approach mechanism bringing the pupinto the vicinity of safe, ingesta that has beenor is being eaten by adult conspecifics. Oncethe pups are in the vicinity of food that adultsare eating or an area where rats have eatenand where there remain some residual cues,they may make recognize flavors that havebeen present in mother's milk (Galef and Hen-derson, 1972) and sample such foods if the ratthey have suckled from has eaten there. Theyalso learn associations between food odorsborne by rats, based on chemicals in thebreath of the conspecifics (Galef et al., 1988).By the time pups egress from the nest, theyare capable of rapidly recognizing and learn-ing which substances are associated with

postingestional signals of nutritive content. Inaddition, they are equipped with basic homeo-static controls for regulated intake of calories,electrolytes, and water. With such capabili-ties, the pups are capable of sustaining them-selves after only a few weeks of postnataldevelopment.

DESIGNING AND INTERPRETINGTESTS WITH RAT PUPS

A full discussion of the strategies for testingyoung rats is unfortunately beyond the scopeof the present chapter. It can be inferred fromthe material presented briefly herein that thespecial features of the immature rat and theirrapid developmental transformations presentboth challenges and opportunities for re-searchers who want to probe them for sys-tematic data and insights into developmentalstatus.

Context, both global and proximal, canbe a powerful factor in testing the immaturerat, more so than for the adult. Small changescan make big differences in outcome. Air andsurface temperatures are vitally importantand may require careful control and moni-toring for reliable results. The behavioral andphysiological effects and the hedonic value oftemperatures vary with age, so valid cross-agecomparisons may require testing at differenttemperatures if behavioral state or metabolicrate is to be equated. Newborns displayunique behavioral capabilities when testing isconducted at warm air temperatures (Hall,1979; Johanson and Hall, 1980).

Olfactory context also can be important.Testing in the presence of familiar odors canenhance learning and memory performance.

Knowledge of the pups' behavioral reper-toire during development can be used to craftuseful and robust measures. Similarly, an ap-preciation of the components and timing ofmaternal activities can be very important incalibrating schedules of deprivation or re-ward. The pups' physiological characteristics

Page 292: 0195162854

Chapter 25. Infancy 275

can require careful calibration for the effectsof privation, including dehydration, stomachclearing and stomach fill, across ages.

Knowledge of the developmental se-quences in relation to developmental niches(Alberts and Cramer, 1988) can be used tocraft useful and robust stimuli for eliciting andrewarding behavior. It is always important tobe aware of the task demands with pups ateach point in their development so that per-formance deficits are not misattributed.

This information, although detailed andmultileveled, is not cause for despair. Aware-ness of the pups' natural environment andtheir species- and age-typical repertoires canusually explain, if not predict, relations andpatterns that are coherent and manageableand that can be made tractable for testing andunderstanding.

References

Alberts JR (1976) Olfactory contributions to behavioraldevelopment in rodents. In: Mammalian olfaction,reproductive processes, and behavior, pp. 67-93.New York: Academic Press.

Alberts JR (1978a) Huddling by rat pups: Group behav-ioral mechanisms of temperature regulation and en-ergy conservation. Journal of Comparative and Psy-siological Psychology 92:231-240.

Alberts JR (1978b) Huddling by rat pups: Multisensorycontrol of contact behavior. Journal of Comparativeand Physiological Psychology 92:220-230.

Alberts JR (1981) Ontogeny of olfaction: Reciprocal rolesof sensation and behavior in the development ofperception. In: Development of perception: Psy-chobiological perspectives, Vol 1 (Aslin RN, AlbertsJR, Petersen MR, eds.), pp. 321-357. New York: Aca-demic Press.

Alberts JR (1984) Sensory perceptual development in theNorway rat: A view toward comparative studies. In:Comparative perspectives on memory develop-ment (R Kail and N Spear, eds.), pp. 65-101. NewYork: Plenum.

Alberts JR and Brunjes PC (1978) Ontogeny of thermaland olfactory determinants of huddling in the rat.Journal of Comparative and Physiological Psychol-ogy 92:897-906.

Alberts JR and Cramer CP (1988) Ecology and experi-ence: Sources of means and meaning of develop-

mental change. In: Handbook of behavioral neuro-biology, Vol 9 (Blass EM, ed.), pp. 1-39. New York:Plenum Publishing Corp.

Alberts JR and Gubernick DJ (1983) Reciprocity and re-source exchange: A model of parent-offspring rela-tions. In: Symbiosis in parent-offspring interactions(Rosenblum LA and Moltz H, eds.). New York:Plenum Press.

Alberts JR and Leimbach MP (1980) The first foray: Ma-ternal influences in nest egression in the weanlingrat. Developmental Psychobiology 13:417-429.

Alberts JR and May B (1980a) Development of nasal res-piration and sniffing in the rat. Physiology and Be-havior 24:957-963.

Alberts JR and May B (1980b) Ontogeny of olfaction:Development of the rat's sensitivity to urine andamyl acetate. Physiology and Behavior 24:965-970.

Alberts JR and May B (1984) Nonnutritive, thermotac-tile induction of filial huddling in rat pups. Devel-opmental Psychobiology 17:161-181.

Alberts JR, Motz B, and SchankJC (2004) Positive geotaxisin rats: A natural behavior and an historical correction.Journal of Comparative Psychology 118:123-132.

Alberts JR and Ronca AE (1993) Fetal experience re-vealed by rats: Psychobiological insights. Early Hu-man Development 35:153-166.

Allin JT and Banks EM (1972) Functional aspects of ul-trasound production by infant albino rats (Rattusnorvegicus). Animal Behavior 20:175-185.

Altman J and Sudarshan K (1975) Postnatal developmentof locomotion in the laboratory rat. Animal Behav-iour 23:896-920.

BekofTA and Trainer W (1979) The development of in-terlimb co-ordination during swimming in postna-tal rats. Journal of Experimental Biology 83:1-11.

Blass EM and Teicher MH (1980) Suckling. Science210:15-22.

Blumberg MS, Efimova IV, Alberts JR (1992) Thermo-genesis during ultrasonic vocalization by rat pupsisolated in a warm environment: A thermographicanalysis. Developmental Psychobiology 25:497-510.

Bolles RC and Woods PJ (1964) The ontogeny of behav-iour in the albino rat. Animal Behaviour 12:427-441.

Brewster J and Leon M (1980) Facilitation of maternaltransport by Norway rat pups. Journal of Compar-ative and Physiological Psychology 94:80-88.

Brunjes PC (1994) Unilateral nans closure and olfactorysystem development. Brain Research Reviews19:146-160.

Brunjes PD and Alberts JR (1981) Early auditory and vi-sual function in normal and hyperthyroid rats. Be-havioral and Neural Biology 31:393-412.

Campbell BA, Lytle LD, Fibiger HC (1969) Ontogeny ofadrenergic arousal and cholinergic inhibitory mech-anisms in the rat. Science 166:637-638.

Page 293: 0195162854

276 DEVELOPMENT

Cramer CP, Thiels E, Alberts] (1990) Weaning in rats:I. Maternal behavior. Developmental Psychobiol-ogy 23:479-493.

Crozier WJ and Pincus G (1929a) Analysis of the geo-tropic orientation of young rats. I. Journal of Gen-eral Physiology 13:59-80.

Fleming A (1965) Maternal behavior. In: Determinantsof infant behavior, Vol 3 (Foss BM, ed.). London:Methuen & Co Ltd.

Flory GS, Langley CM, Pfister JF, Alberts JR (1997) In-strumental learning for a thermal reinforcer in1-day-old rats. Developmental Psychobiology30:41^47.

Fraenkel GS and Gunn DL (1940) The orientation of an-imals. London: Oxford University Press.

Friedman MI (1975) Some determinants of milk inges-tion in suckling rats. Journal of Comparative andPhysiological Psychology 89:636-647.

Friedman MI, Bruno JP, Alberts JR (1981) Physiologicaland behavioral consequences in rats of water recy-cling during lactation. Journal of Comparative andPhysiological Psychology 95:26-35.

Galef BG (1983) Utilization by Norway rats (R. norvegi-cus) of multiple messages concerning distant foods.Journal of Comparative Psychology 97:364-371.

Galef BG and Clark MM (1971) Social factors in the poi-son avoidance and feeding behavior of wild anddomesticated rat pups. Journal of Comparative andPhysiological Psychololgy 75:341-357.

Galef BG and Heiber L (1976) Role of residual olfactorycues in the determination of feeding site selectionand exploration patterns of domestic rats. Journal ofComparative and Physiological Psychology 90:727-739.

Galef BG and Henderson PW (1972) Mother's milk: Adeterminant of the feeding preferences of weaningrat pups. Journal of Comparative and PhysiologicalPsychology 78:213-219.

Galef BG, Mason JR, Preti G, Bean NJ (1988) Carbondisulfide: A semiochemical mediating socially-in-duced diet choice in rats. Physiology and Behavior42:119-124.

Galef BG and Sherry DF (1973) Mother's milk: Amedium for transmission of cues reflecting the fla-vor of the mother's diet. Journal of Comparativeand Physiological Psychology 83:374-378.

Gerrish CJ and Alberts JR (1995) Differential influenceof adult and juvenile conspecifics on feeding byweanling rats (Rattus norvegicus): A size-related ex-planation. Journal of Comparative Psychology109:61-67.

Gottlieb G (1971) Ontogenesis of sensory function inbirds and mammals. In: The biopsychology of de-velopment (Tobach E, Aronson LR, Shaw E, eds.).New York: Academic Press.

Gubernick DJ and Alberts JR (1983) Maternal licking of

young: Resource exchange and proximate controls.Physiology and Behavior 31:593-601.

Hall WG (1979a) Feeding and behavioral activation ininfant rats. Science 205:206-209.

Hall WG (1979b) The ontogeny of feeding in rats: I. In-gestive and behavioral responses to oral infusions.Journal of Comparative and Physiological Psychol-ogy 93:977-1000.

Hall WG and Bryan TE (1980) The ontogeny of feedingin rats: II. Independent ingestive behavior. Journalof Comparative and Psychological Psychology94:746-756.

Hall WG and Rosenblatt JS (1977) Suckling behavior andintake control in the developing rat pup. Journal ofComparative and Physiological Psychology 91:1232-1247.

Hall WG and Williams CL (1983) Suckling isn't feeding,or is it? A search for developmental continuities.Advances in the Study of Behavior 13:219-254.

Hill DL (1987) Development of taste responses in therat prabrachial nucleus. Journal of Neurophysiology57:481-495.

Hofer MA, Brunelli SA, and Shair H (1993) Ultrasonicvocalization responses of rat pups to acute separa-tion and contact comfort do not depend on mater-nal cues. Developmental Psychobiology 26:81-95.

Hofer MA and Shair H (1978) Ultrasonic vocalizationduring social interaction and isolation in 2-week-oldrats. Developmental Psychobiology 11:495-504.

Johanson IB and Hall WG (1980) The ontogeny of feed-ing in rats: III. Thermal determinants of early in-gestive responding. Journal of Comparative andPhysiological Psychology 94:977-992.

Kleitman N and SatinofFE (1982) Thermoregulatory be-havior in rat pups from birth to weaning. Physiol-ogy and Behavior 29:537-541.

Kreider JC and Blumberg MS (1999) Geotaxis in 2-week-old Norway rats (Rattus norvegicus): A reevalua-tion. Developmental Psychobiology 35:35-42.

Leon M, Croskerry PG, Smith GK (1978) Thermal con-trol of mother-young contact in rats. Physiology &Behavior 21:793-811.

Leon M and Moltz H (1972) Maternal pheromone: Dis-crimination by pre-weanling albino rats. Physiologyand Behavior 7:265-267.

Lincoln DW, Hill A, Wakerly JB (1973) The milk-injec-tion reflex of the rat; An intermittent function notabolished by surgical levels of anesthesia. Journal ofEndocrinology 57:459-476.

Moore CL and Chadwick-Diaz AM (1986) Behavioral re-sponses of infant rats to maternal licking: Variationswith age and sex. Developmental Psychobiology19:427-438.

Moore CL and Rogers SA (1984) Contribution of self-grooming to onset of puberty in male rats. Devel-opmental Psychobiology 17:243-253.

Page 294: 0195162854

Chapter 25. Infancy 277

Pedersen PE and Blass EM (1982) Prenatal and postnataldeterminants of the 1st suckling episode in albinorats. Developmental Psychobiology 15:349-355.

Pellis VC, Pellis SM, Teitelbaum P (1991) A descriptiveanalysis of the postnatal development of contact-righting in rats (Rattus norvegicus). DevelopmentalPsychobiology 24:237-267.

Randall PK and Campbell BA (1976) Ontogeny of be-havioral arousal: Effect of maternal and sibling pres-ence. Journal of Comparative and Physiological Psy-chology 90:453-459.

Richmond G and Sachs BD (1980) Grooming in Norwayrats: The development and adult expression of acomplex motor pattern. Behaviour 75:82-96.

Richter CP and Barelare B (1938) Nutritional require-ments of pregnant and lactating rats studies by theself-selection method. Endocrinology 23:15-24.

Ronca AE and Alberts JR (1994) Sensory stimuli associ-ated with gestation and parturition evoke cardiacand behavioral responses in fetal rats. Psychobiol-ogy 22:270-282.

Ronca AE and Alberts JR (2000) Effects of prenatal space-flight on vestibular responses in neonatal rats. Jour-nal of Applied Physiology 89:2318-2324.

Ronca AE and Alberts JR (1995) Simulated uterine con-tractions facilitate fetal and newborn respiratory be-havior in rats. Physiology and Behavior 58:1035-1041.

Ronca AE and Alberts JR (1995) Maternal contributionsto fetal experience and the transition from prenatalto postnatal life. In: Fetal behavior: A psychobio-logical perspective (Lecanuet J-P, Krasnegor N,Smotherman WP, eds.), pp. 331-351. Hillsdale, NJ:Lawrence Erlbaum Associates.

Ronca AE, Lamkin CA, Alberts JR (1993) Maternal con-tributions to sensory experience in the fetal andnewborn rat. Journal of Comparative Psychology107:61-74.

Rosenblatt JS (1965) The basis of synchrony in the be-havioral interaction between the mother and heroffspring in the laboratory rat. In: Determinants ofinfant behavior, Vol 3 (Foss BM, ed.), pp. 3—44. Lon-don: Methuen & Co Ltd.

SchankJ and Alberts J (1997) Self-organized huddles ofrat pups modeled by simple rules of individual be-havior. Journal of Theoretical Biology 189:11-25.

Schank JC and Alberts JR (1997) Aggregation and theemergence of social behavior in rat pups modeledby simple rules of individual behavior. In: Interna-tional Conference on Complex Systems (Bar-YamY, ed.), pp. 1-8. Nashua, NH: New England Com-plex Systems Institute.

Small WS (1899) Notes on the psychic development ofthe young white rat. American Journal of Psychol-ogy 11:80-100.

Stehouwer DJ and Van Hartesveldt C (2000) Kinematicanalyses of air-stepping in normal and decerebratepreweanling rats. Developmental Psychobiology36:1-8.

Stern JM (1988) A revised view of the multisensory con-trol of maternal behaviour in rats: Critical role oftactile inputs. In: Ethoexperimental analysis of be-haviour (Blanchard RJ, Brain PF, Blanchard DC,Parmigiani S, eds.). II Ciocco: Martinus Nijhoff.

Stern JM and Azzara AV (2000) Thermal control ofmother-young contact revisited: Hyperthermic ratsnurse normally. Physiology and Behavior 77:11-18.

Stone EA, Bonnet KA, Hofer MA (1976) Survival and de-velopment of maternally deprived rats: Role of bodytemperature. Psychosomatic Medicine 38:242-249.

Sullivan RM and Hall WG (1988) Reinforcers in infancy:Classical conditioning using stroking or intraoral in-fusions of milk as USC. Developmental Psychobi-ology 21:215-224.

Tees R (1976) Perceptual development in mammals. In:Studies on the development of behavior and thenervous system, Vol 3 (Gottlieb G, ed.), pp.282-326. New York: Academic Press.

Teicher MH and Blass EM (1976) Suckling in newbornrats: Eliminated by nipple lavage, reinstated by pupsaliva. Science 193:422-425.

Thiels E, Cramer CP, Alberts JR (1988) Behavioral in-teractions rather than milk availability determinedecline in milk intake of weanling rats. Physiologyand Behavior 42:507-515.

Welker WI (1964) Analysis of sniffing of the albino rat.Behaviour 22:223-244.

Wu CC and Gonzalez MF (1997) Functional develop-ment of the vibrissae somatosensory system of therat: (14c)2-Deoxyglucose metabolic mapping study.Journal of Comparative Neurology 384:323-336.

Page 295: 0195162854

Adolescence

RUSSELL W. BROWN

The rapid development of cognitive and mo-tor behavior in the laboratory rat is theessence of survival for the animal. Rats areweaned from the mother at approximately 21days of age. Therefore, within 21 days, the an-imal must have the ability to avoid predators,locate food caches, and consume food and ispreparing to establish a home terrirory. Basedon this rapid development, the rat provides anexcellent model for the development of be-havior and the brain structures that mediatethese behaviors. The following sections de-scribe the appearance of motor abilities, socialand play behavior, sexual maturity, and sen-sory function, followed by a section on cogni-tive function, including a brief discussion of thedevelopment of simple stimulus associationsmade in conditioned taste aversion followedby complex associations necessitated in spatialmemory. Finally, a brief discussion is includedon the development of brain structures andtheir possible role in these behaviors.

APPEARANCE OF BEHAVIORS

AMBULATORY ABILITIES

The newborn rat on the first day of life (post-natal day 1 [Pi]) is essentially motionless butvery quickly develops behaviors that can helpit to secure and locate food later in life. By P8,the animal is able to crawl but does not havemuch use of its hindlimbs. By P12 and Pi 3, theraised posture necessary for walking develops,although the animal does not yet move swiftlyand the hindlimbs often slip and are dragged

behind the body. The important act of rearingon the hindlimbs presupposes functional mat-uration of the hindlimbs and does not emergewith appreciable frequency in the open field un-til Pi 8. Rearing on the hindlimbs is typically ob-served in aroused adult rats and may representan acute form of investigatory response, whichmay reflect a preparatory move for climbing.By P21, most animals are able to traverse a 0.5cm wide path, ascend and descend a wire meshor ladder surface, and ascend or descend a ropewith few errors (Altman and Sudarshan, 1975).Therefore, the rat develops many ambulatoryskills that are at least approaching adult-likeabilities by the time it is weaned from themother at P21.

TESTING

One of the simplest tasks used to test grossambulatory ability in the rat is the locomotorbox, or locomotor arena. The arena used totest locomotor behavior is typically coveredwith a network grid of infrared beams. Eachtime an infrared beam is broken by the rat, anactivity count is scored, either manually bythe experimenter or a computer program at-tached to the arena. This task has been usedprevalently in behavioral pharmacology re-search to model addiction behavior (for areview, see Kalivas and Pierce, 1997), and au-tomated computer programs have been de-veloped to provide a variety of measurementsfor locomotor behavior, including horizontaland vertical activity, bouts of movement, andtotal distance traveled to name a few.

278

26

Page 296: 0195162854

Chapter 26. Adolescence

ABILITY TO EAT AND OBTAIN FOOD

Handling and Food-Eating PostureAs adults, rats are able to use skilled move-ments to reach, manipulate, and eat food.Handling food with the forepaws requires bi-manual coordination in the rat, which is theability of an organism to use both forelimbswhile carrying out two different behaviors tosolve a task. A recent systematic study of thedevelopment of food-obtaining behavior wasperformed by Brenda Coles and Ian Whishaw(Brenda Coles, unpublished master's thesis).The typical eating posture is characterized bya rat sitting back on its haunches using thehindlimbs as the base of support for the bodyin a slightly splayed manner. The forelimbsare held underneath the snout and are used tohold and manipulate food (Fig. 26-1). There-fore, for a rat to be able to eat food in a typi-cal adult manner, forelimbs must be suffi-ciently developed to manipulate the food toeat, and the hindlimbs must be sufficientlymature to support the animal's body weight.It is not until about PI8 that pups begin to situp on their haunches when eating small foodpellets; they are still close to the ground withthe hindlegs splayed outward in an exagger-ated manner, providing a wide base of sup-port. By P21, eating posture becomes sexuallydimorphic, in that male rats are able to sit un-aided in an adult-like posture to eat all sizes

Figure 26-1. Adult-like eating posture in the rat. When arat is eating a piece of food, the hindlimbs provide supportfor the body, whereas the forelimbs are free to manipulatethe piece of food. (Adapted from Coles BLK and WhishawIQ, unpublished data.)

279

of food pellets, but females are still using lit-termates or objects for support to attain thesame goal. By P24, this sexual dimorphism hasdisappeared, and both female and male ratsare able to eat with an adult-like posture.

ReachingAnother skilled motor behavior that developsin the rat is reaching for food. A reach con-sists of lifting the forelimbs from the ground,positioning the elbows inward, so that thepaws are adjacent to the mouth and claspingthe food with the digits. These movementsare executed mainly with the upper forelimb.As the limb is positioned for grasping, theaperture of the digits is adjusted to anticipatethe size of the food and the food is graspedand manipulated with the tips of the digits.The food is then supinated approximately 90°as the paw is retracted in toward the body andthen supinated 90° again for the rat to retrievethe food with its mouth from the paw(Whishaw and Tomie, 1989). Results haveshown that rats do not attempt to reach forfood until PI9, but at this point reaches arenot successful. This may be due to the factthat rats are unable to aim their body ordemonstrate pronation. Reaching attemptsbecome more frequent by P21. By P23, ratsattempt to use different types of strategies tobecome somewhat succesful at reaching forthe food, but the precise reaching does not de-velop until about P26.

Reaching can be studied in at least twodifferent ways. In the tray-reaching test, ratscan be trained to reach for chicken feed placedin a tray mounted approximately 1 cm in frontof a cage equipped on one side with metal barsspaced approximately 0.5 cm apart. A morecomplex version of the skilled reaching task isthe single-pellet precision reaching task. In thistask, the rat must reach through a narrow slotto obtain a single pellet of food placed on ashelf. Both of these tasks have proved usefulfor testing of skilled motor behavior in rats,with animals learning the tray-reaching taskmore rapidly than the single-pellet precision

Page 297: 0195162854

280 DEVELOPMENT

reaching task, indicating that precision skilledreaching may be the more complex motor be-havior to perform for the rat.

Securing of the Food from ConspecificsThere are two behaviors that have been stud-ied in laboratory rats to secure food from con-specifics: robbing and dodging. Robbing be-havior is characterized by a rat walking upalongside the other rat's body from the rearand attempting to grasp the food either withthe mouth or with a reach of one of the fore-limbs. Dodging is a defensive tactic in whichthe feeding rat uses forequarter rotation andhindlimb stepping movements to escape froma conspecific attempting to steal the food(Whishaw and Tomie, 1989). Successful dodg-ing requires a full 180° turn initiated by a con-traversive rotation of the front half of the bodyaway from the robber, with steps being takenipsilateral to the turn direction by both theforelimb and hindlimb. Robbing behavior ap-pears prevalent beginning at Pi7, and rats aresuccessful at stealing food from conspecifics.One reason that robbers are so successful atthis age is that pups do not seem to be able tododge the perpetrator, likely due to the factthat the hindlimbs have not yet reached ma-turity to aid in dodging the robber. Animalsbegin to attempt to dodge the robber at PI9but are not very successful, and dodging doesnot become effective or near maturity untilapproximately P25 (B. L. K. Coles and I. Q.Whishaw, unpublished data).

Social/Play BehaviorRats engage in various forms of play behav-ior, with social play in the form of playingfighting being among the most commonlyreported. The frequency of play fightingreaches its peak during adolescence and de-clines after puberty. In rats, play fighting in-volves attack and defense of the nape of theneck, which, if contacted, is gently nuzzled.As juveniles (beginning around P30), themost frequently used defensive tactic is torotate to the supine position when the nape

is contacted. This results in an on-top/on-bottom orientation referred to as pinning.Different play behaviors develop at differentrates, and play fighting becomes rougher, es-pecially in males, as the animal reaches adultmaturity (approximately P60). One of themore frequent play behaviors observed im-mediately after weaning is wrestling, in whichtwo animals roll and tumble with one an-other. Another frequent play behavior thatoften initiates play is pouncing; animals ex-hibiting this behavior invariably engage in aplay bout that is long enough for the recipi-ent animal responds. This typically precedesany other behavior in the play sequence. Asanimals reach late adolescence (around P50),boxing becomes prevalent, in which two ani-mals stand upright facing one another andmake pawing movements toward each other(Meaney and Stewart, 1981).

Sensory FunctionWhen the rat is born, both its eyes and ears areclosed. The only sensory functions that arefunctional on the first day of life are taste, odor,and touch. This is known because the rat canlearn associations involving taste, odor, ortouch stimuli on the first day of life and evenhas the abilities before birth (Smotherman,1982). Obviously, the animal is not yet able toforage for its own food, and essentially the an-imal is completely dependent on the motherfor food and protection. The ears do not openuntil approximately P8 to P9, and the eyes donot open until approximately P15 to P16.Therefore, gradually through the first 3 weeksof life, all sensory abilities appear to allow theanimal to obtain and forage food for itself.

Mating BehaviorIn male rats, the onset of mounting typicallyoccurs between P41 and P45. Frequency ofmale mounting of the female increases be-tween P46 and P50 and decreases between P51and P55. The onset of mounting is associatedwith anogenital sniffing and chase behaviors.Lordosis is the behavior in which a female indi-

Page 298: 0195162854

Chapter 26. Adolescence 281

cates to the male that she is viable for mating.This behavior is defined as the female rat ele-vating its anogenital region accompanied by adownward arching of its back and twitching ofthe ears. Lordosis begins to appear in femalerats at about P42 and is demonstrated with in-creasing frequency through P55.

In conclusion, the development and ap-pearance of behaviors to obtain and securefood seem to follow the rostrocaudal gradientof maturation, with the forelimbs leadinghindlimb maturation by a few days. A time-line of behavioral development is presented inFigure 26-2. A discussion of the correlation ofthe development of the central nervous sys-tem and behavior is included at the end of thechapter.

COGNITION

One of the most important and complex as-pects of a rat's behavioral repertoire is the abil-ity to locate food in the natural environmentand avoid predators. For the rat to performthis ability, it must learn associations betweenstimuli. Cognitive abilities develop rapidly

Eyes Open

Food Sniffing

Eat Hard Food

Robbing

Reaching (chicken feed)

Reaching (single pellet)

Dodging

Upright Posture

Age: Number indicates day of age

Figure 26-2. Timeline of behavioral development.(Adapted from Coles BLK and Whishaw IQ, unpublishedmaster's thesis.)

during the first 3 weeks of life, and for manycognitive associations, animals are able to per-form at near-adult levels by the time they areweaned. Based on the fact that cognitive abil-ties in rats covers a vast area of research, thissection focuses primarily on the developmentof the ability to avoid an aversive stimulus (ataste) and the ability to find a goal location,defined as spatial memory. Avoiding an aver-sive stimulus involves relatively simple stim-uli assocations, whereas spatial memory in-volves more complex stimulus associations.This section concludes with a discussion of theinfantile amnesia phenomenon, which is rapidforgetting that appears to occur during earlydevelopment.

CONDITIONED TASTE AVERSION

A unique type of learning that has beendemonstrated in the laboratory rat as well ashumans is conditioned taste aversion (CTA).One of the basic abilities that the rat must ac-quire early in development is the ability toavoid ingesting poisonous substances. In theCTA paradigm, rats are trained with a taste(such as saccharin or sucrose) that is tempo-rally paired with an injection of lithiumcholoride (LiCl) that produces illness. As inhumans, CTA appears to be a unique form oflearning is that adult rats have shown the abil-ity to acquire an aversion to a particular tastein one trial (a single pairing of the taste andLiCl injection), and the memory for this as-sociation is persistent for weeks.

Interestingly, rats have demonstrated theability to acquire taste aversion as early as be-fore birth in utero (Smotherman, 1982) andhas demonstrated the ability to avoid a par-ticular taste as early as PI (Schweitzer andGreen, 1984). Training adult animals on thistask is relatively simple. Based on the fact thatrats prefer saccharin and sucrose solutions,these substances can be mixed in with thedrinking water. The animal is then presentedwith this solution, and verification of imbibingthe solution can be done with a specialized

Page 299: 0195162854

282 DEVELOPMENT

drinking tube that allows for the measuring ofvolume. Immediately after presentation of thesweetened water, the animal is given an in-traperitoneal injection of LiCl (20 mg/kg). Ona retention test the next day, the animal shoulddemonstrate a drastic reduction in consump-tion of the sweetened water.

In preweanling animals, CTA is a bitmore difficult to perform. At very early ages,before about 10 days of age, a gavage is madethrough insertion of a tube through the cheekof the young animal and injection of the so-lution through the gavage into the animal'smouth. Another way to administer the solu-tion is through an injection directly into theanimal's mouth. As in adults, immediately af-ter presentation of the substance, an in-traperitoneal injection of LiCl is given. There-fore, CTA can be acquired at a very early ageand is a fairly simple learning paradigm to usein the laboratory rat.

INFANTILE AMNESIA

Although rats can learn many different typesof complex associations, early in developmentthese learned associations appear to be quitesusceptible to rapid forgetting in infant rats,referred to in the literature as infantile amne-sia (Spear and Riccio, 1994). Infantile amnesiaoccurs for all altricial mammals in which it hasbeen tested.

There is an interesting contradiction inthe literature concerning infantile amnesia.Smotherman (1982) demonstrated in an ele-gant series of studies that rats can learn toavoid a taste associated with illness as early asembryonic day 19 and that this association ap-pears to be well retained in memory, as ratstested 2 to 3 weeks after birth still demonstratelearning of the aversion. On the other hand,studies have shown that infant rats demon-strate rapid forgetting of the CTA association,as animals trained on the CTA paradigm atPI, PlO, orPlS demonstrate rapid forgettingof this association tested just a few days later.

However, animals slightly older at P20demonstrate retention of the taste aversionwhen tested 25 days after the conditionedstimulus-unconditioned stimulus pairing(Schweitzer and Green, 1982). What this lit-erature seems to suggest is that brain struc-tures that mediate CTA are still developing upto 18 days of age but reach maturity at aboutweanling age. This seems to make sense evo-lutionarily, as rats must begin searching fortheir own food around this time point indevelopment.

How is this contradiction in the literatureresolved? Is it possible that an animal can learnand remember an association better in uterothan it can after birth? In both studies, the tastewas presented through injection by the ex-perimenter, and the animal was testedthrough imbibing the water themselves. Itcertainly seems that acquiring such an aver-sion is hard-wired, as animals can learn this as-sociation before birth. However, retaining itin memory seems to be the larger problem.The resolution to the contradiction likely liesin experimental methodology. In the one case,the taste is injected into the amniotic fluid,whereas in the other case, the taste is pre-sented through the mouth. Therefore, it maybe that rats can learn certain associations toavoid particular stimuli in utero, however, theaction of drinking of the fluid and its associa-tion with illness may not develop until muchlater. This suggests that brain mechanismsthat underlie CTA do not develop untilaround weaning age, but other biologicallyimportant stimuli that the animal can learn toavoid may actually be acquired in utero. Thisalso demonstrates that the ability to learn as-sociations and retain them in memory de-pends on the nature of the stimulus presenta-tion and the type of stimuli that must beassociated. Depending on the association tobe learned and the age of the animal, brainmechanisms may not yet have developed tomediate remembering of certain types ofassociations.

Page 300: 0195162854

Chapter 26. Adolescence

SPATIAL MEMORY

Memory for a spatial location in the rat is anextremely important survival skill, as ratsmust remember the location of food cachesand the availability of these resources. Thereare two primary skills that have been testedin the young rat. One skill is the use of distalcues to locate a hidden spatial location, oth-erwise known as "place" navigation, and theother is the use of a visible cue to locate a vis-ible spatial location, otherwise known as"cued" navigation. Clearly, finding a hiddenlocation through the use of distal cues appearsto be a much more difficult task than simplyapproaching a visible cue to find a spatial lo-cation. Indeed, adult animals demonstrate amore rapid learning curve in cued navigationthan in place navigation, and it appears thecued task is simpler than the place task.

A very effective spatial memory task isthe Morris water maze (MWM). The MWMis named for its creator, Richard G. M. Mor-ris, who published a now seminal paper onmethodologies used to train animals on thistask (1981). This task involves training rats tolocate a "hidden" or "cued" platform in a poolof water. In both versions of the MWM, thewater is colored with powdered paint or pow-dered milk so that the animal must use extra-maze cues to locate the platform, which islocated approximately 1 to 2 cm below the wa-ter surface. In the cued version, the platformis cued through placement of a wooden blockon top of the platform so that its location caneasily be visualized by the swimming rat.

There are several advantages of this taskover other spatial tasks, such as the radial armmaze, including the fact that animals do nothave to be placed on food restriction to be mo-tivated to locate the platform and training canbe completed within 1 day. One disadvantageof this task, according to some investigators,is that rats are being removed from their "nat-ural" environment (land) to locate a safe lo-cation. The argument against this notion is

283

that rats are amphibious animals and are ableto swim quite well even on the first day thatthe eyes are open on PI5 to PI6. Another dis-advantage is that rats can experience hy-pothermia while swimming, especially whenthey are young and not able to thermoregu-late as well as do adults. This problem can typ-ically be handled by drying the animals be-tween trials and keeping the water at anambient temperature for this age of rat(around 23° to 25° C for young rats insteadof 19° to 20° C for adult rats) (Brown andWhishaw, 2000).

Finally, proper training methodologymust be used with young rats. Very short in-tertrial intervals that are 1 minute or less canoften cause fatigue in the young animal, pro-ducing poor performance not necessarily re-lated to cognitive ability. One training tech-nique that has been shown to be effective andrelatively efficient is to begin training on PI7,and animals are trained for 3 consecutive days.Two trial blocks of four trials are given oneach day, however, these trial blocks arespaced apart by approximately 2 to 3 hours,and individual trials are spaced apart by atleast 5 minutes. This allows the animal propertime to rest between trials, and the animalshould demonstrate an asymptote of learningequivalent to adult levels by the last day oftraining (Pi9) (Kraemer and Randall, 1996).One final note is to make certain that thereare proper extramaze cues around the pool,including posters, desks, the experimenter,and other visual cues that are relatively closeto the edge of the pool (within 1 meter). Us-ing proper extramaze cues helps to facilitatethe animal's performance, because rats usethese cues to locate the platform.

There has been a considerable amount ofdebate as to the spatial abilities in the prewean-ling and early postweanling rat. Rudy and col-leagues (1987) demonstrated that rats cannotbegin to learn to rely on distal cue navigationuntil approximately 20 days of age and are notadept at using distal cues to locate a spatial lo-

Page 301: 0195162854

284 DEVELOPMENT

cation until about 23 days of age. However,several other studies have reported that rats canlearn a spatial discrimination based on distalcues at an earlier age. Brown and Whishaw(2001) reported that both proximal and distalcue navigation appear to develop at the samerate, and both are functional by 19 days of agebut are not mature at 18 days of age. This sug-gests that similar mechanisms may mediateboth place and cue navigation, contradictingthe idea that these two types of navigation maydevelop at different rates.

Although the 19-day-old rat can find a spa-tial location, it appears the ability to retain amemory for that location does not develop un-til much later. In fact, animals trained on a spa-tial task at 19 to 21 days of age demonstratecomplete forgetting of that location just 3 daysafter training (Brown and Kraemer, 1997). Incontrast, adult rats have demonstrated memoryfor a spatial location as long as 3 months aftertesting (Sutherland and Dyck, 1984). Therefore,it appears that that the ability to learn a spatiallocation develops at an early age, even beforeweaning, but the ability to remember a spatiallocation may not develop until adulthood.

WHAT DOES THIS TELL US ABOUTDEVELOPMENT OF THE BRAIN

AND BRAIN FUNCTION?

Throughout the rat's first 3 weeks of life, thenervous system of the rat undergoes a rapidprocess of development. Surprisingly, therehave been few attempts to correlate motor be-havior with the details of anatomical changesin the brain. In general, the appearance of mo-tor behaviors seems to follow the rostrocaudalgradient of maturation of the cerebral cortex;therefore, the mouth and forelimbs mature be-fore the more caudally located hindlimbs re-ceive their spinal cord connections.

MOTOR DEVELOPMENT

Regarding the motor cortex in the rat, thisbrain area has a representation of the body

that is composed of a large forelimb area lo-cated in the anterior portion of the motor cor-tex and the hindlimb area that is located moreposterior. Following the rostrocaudal gradientof development, the more anterior forelimbareas are formed first, and thus mature firstbefore the hindlimbs. The corticospinal tractis the projection from the motor cortex to thespinal cord; it also is thought to contribute toskilled limb movements. Growth of the corti-cospinal tract follows the same neurogeneticprinciple of growth: the axons from cortexterminate in an anterior (older) to posterior(younger) pattern. Thus, the anatomicalconnections for the forelimbs mature beforethose for the hindlimbs (see Fig. 26-3). Thispattern of development is in accord with thebehavioral literature, which has demonstratedthat forelimb use matures before the hind-limbs (Bayer and Altman, 1991).

Figure 26-3. Diagram of the formation of the developmentof anatomical connections between layer V output neuronsof the corticospinal tract and terminations in the spinal cord.(Adapted from Bayer SA and Altman J, 1991.)

Page 302: 0195162854

Chapter 26. Adolescence

CONDITIONED TASTE AVERSION

There have been three primary brain struc-tures implicated in CTA: the parabrachial nu-cleus (PEN), amygdala, and insular cortex. Allthree of these areas are known to be impor-tant in modulating gustatory responses. ThePEN is located in the brain stem. It sends di-rect major projections to the amygdala and tothe insular cortex via the ventromedial nucleiof the thalamus and is known to play a majorrole in taste. The amygdala is the central brainstructure known to be involved in emotion,is part of the limbic system, and is located inthe temporal lobe. Ablations of specific nucleiwithin the amygdala have been shown to pro-duce deficits in aversive conditioning, demon-strating the importance of this brain area inthis type of learning. The insular cortex is thearea of the cerebral cortex that surrounds therhinal sulcus. The insular cortex has been re-ferred to as the visceral cortex because it re-ceives taste and visceral information from thethalamus. It has been postulated that the in-sular cortex receives convergences of primarysensory inputs not seen within any other sen-sory areas of the cortex. It appears that thesethree brain areas—the PEN, amygdala, and in-sular cortex—form a circuit in the brain thatis responsible for the acquisition of condi-tioned taste aversions (Bermudez-Rattoni,1995). There have not been any attempts tomap out the development of the brain cir-cuitry between these three areas and to cor-relate it with CTA. However, a connection be-tween the amygdala and medial prefrontalcortex develops at P19, suggesting that theconnectivity of the amygdala to the cortex ismaturing at this age (Cunningham et al.,2003).

SPATIAL MEMORY

It has been shown that an important brainstructure to mediate cognition, especially spa-tial memory, is the hippocampus. Althoughall areas of the hippocampus proper and hip-pocampal formaton have been implicated in

285

spatial memory function, the maturity of themossy fiber system at Pi8 may be especiallyimportant. The mossy fibers are axons of thedentate granule cells projecting onto the py-ramidal cells of the dentate hilus (CA4) andCA3 pyramidal cells, and this mossy fiber sys-tem appears to be mature by approximatelyP18 (Zimmer, 1978). Based on past results thathave shown animals develop spatial abiltiesnear adult levels by Pi9, it appears that ma-turity of the mossy fiber connections of thedentate gyrus-CA4 and dentate gyrus-CA3areas at this age may be especially importantin mediating spatial memory performance atthis age.

In summary, this chapter just touches onthe vast literature describing the developmentof cognitive and motor abilities in the labora-tory rat. There are several important issues tokeep in mind when training preweanling orearly postweanling laboratory rats. The firstissue is training methodology. Rats at an earlyage have an ever-expanding behavioral reper-toire, but to comprehend their abilities, theanimal must be trained properly. This meansthat longer intertrial intervals may need to beused, and special provisions in terms of envi-ronment may be needed to be taken into con-sideration. Another important issue involvesthe complexity and rapid development of cog-nitive and motor abilities in the rat. A rat thatdoes not have an ability on a particular day ofdevelopment may have that ability at near-adult levels on the next day. With this rapiddevelopment in abilities and brain structures,the developing rat provides an ideal but com-plex subject to study the ontogeny of learn-ing, memory, and motor abilities. Finally, it isalways important to keep in mind the stressincurred on the young animal. Taking apreweanling away from the mother may pro-duce undue stress, especially if the animal isaway from the mother for an extended periodof time. Attempting to understand a younganimal's abilities is a daunting task, and onein which every aspect of the training envi-ronment and methodology must be kept inmind.

Page 303: 0195162854

286 DEVELOPMENT

REFERENCES

Altman J and Sudarshan K (1975) Postnatal developmentof locomotion in the laboratory rat. Animal Behav-ior 23:8096-8920.

Bayer SA and Altman J (1991) Neocortical development.New York: Raven Press.

Bermudez-Rattioni F (1995) The role of the insular cor-tex in the acquisition and long lasting memory foraversively motivated behavior. In: Plasticity in thecentral nervous system (McGaugh JL, Bemudez-Rattoni F, Prado-Alcala RA, eds.). Mahwah, NJ: Erl-baum.

Brown RW and Kraemer PJ (1997) Ontogenetic differ-ences in retention of spatial learning tested with theMorris water maze. Developmental Psychobiology30:329-341.

Brown RW and Whishaw IQ (2000) Similarities in thedevelopment of place and cue navigation by rats ina swimming pool. Developmental Psychobiology37:238-245.

Coles BLK and Whishaw IQ (1996) Neural changes inforelimb cortex and behavioural development. Un-published master's thesis, Lethbridge, Alberta,Canada: University of Lethbridge.

Cunningham MG, Bhattacharya S, Benes FM (2002)Amygdalo-cortical sprouting continues into earlyadulthood: Implications for the development of nor-

mal and abnormal function during adolescence.Journal of Comparative Neurology 453:116-130.

Kraemer PJ and Randall CK (1995) Spatial learning inpreweanling rats trained in a Morris water maze.Psychobiology 23:144-152.

Pierce RC and Kalivas PW (1997) A circuitry model ofthe expression of behavioral sensitization to am-phetamine-like psychostimulants. Brain ResearchBrain Research Reviews 25:192-216.

Rudy JW, Stadler-Morris S, Albert P (1987) Ontogeny ofspatial navigation behaviors in the rat: Dissociationof "proximal"- and "distaT-cue-based behaviors.Behavioural Neuroscience 101:62-73.

Schweitzer L and Green L (1982) Acquisition and ex-tended retention of a conditioned taste aversion inpreweanling rats. Journal of Comparative and Phys-iological Psychology 96:791-806.

Smotherman WP (1982) Odor aversion learning by therat fetus. Physiology and Behavior 29:769-771.

Spear NE and Riccio DC (1994) Memory: Phenomena andprinciples. Needham Heights, Mass.: Allyn & Bacon.

Whishaw IQ and Tomie J-A (1989) Food-pellet size mod-ifies the hoarding behavior of foraging rats. Psy-chobiology 17:93-101.

Zimmer J (1978) Development of the hippocampus andfascia dentata: Morphological and histochemical as-pects. Maturation of the Nerv Sys, Progress in BrainResearch, Vol. 48. MA Corner, Ed. Elsevier/NorthHolland Press: Amsterdam.

Page 304: 0195162854

Maternal Behavior

STEPHANIE L. REES, VEDRAN LOVIC,AND ALISON S. FLEMING

27

Parental or maternal behavior is an ultimate testof individual fitness, as the ability of individualsto raise healthy offspring is necessary both forthe individuals' capacity to pass on their genesand for species survival. Although the ultimatefunction of raising healthy offspring is commonamong different species, there is considerablevariability in the expression of maternal behav-ior across mammalian species, with the mostmarked distinctions occurring between altricialand precocial species. In the altricial species,such as rodent species, the young are born inan immature state, often in litters and usuallyinto a stable nest or home environment wherethe young remain for a considerable period be-fore weaning (Weisner and Sheard, 1933; Flem-ing and Li, 2002). In contrast, in some precocialspecies, the young have fully developed sensoryand motor abilities within hours of birth(Gonzalez-Mariscal and Poindron, 2003). In thischapter, only the behavior of laboratory rats,Rattus norvegicus, is discussed, a familiar, well-studied example of an altricial species.

Laboratory rats have proved to be a goodmodel for the study of hormonal (Rosenblatt,2002), sensory (Stern, 1996), neural (Numanand Sheehan, 1997), experiential (Li and Flem-ing, 2003), and developmental (Fleming et al.,2002) factors that control maternal behavior.The study of maternal behavior itself has usebecause it is a highly organized behavior thatcan be used as a model of social behavior. Thestudy of maternal behavior can be used in theanalysis of the effects of exposure to drugs ofabuse (cocaine: Mattson et al., 2003), to ther-

apeutic agents such as antipsychotics (Li et al.,in press), or to prenatal and/or postnatalstress, alcohol, maternal separation, or "en-richment" (Kuhn and Schanberg, 1998). Ma-ternal behavior in and of itself is now knownto have marked effects on the neurologicaland behavioral development of offspring,with the quantity of licking by the mother al-tering the development of the pups stress andendocrine systems (Liu et al., 1997), brain de-velopment, and cognitive, affective, and socialbehaviors (Hofer, 1994; Fleming et al., 2002).

This chapter describes maternal behaviorof the laboratory rat and outlines various meth-ods of observing and quantifying this behavior.Although in some rodent biparental speciesmales also show parental behavior, this is notthe case for most rodents, including R. norvegi-cus. However, under certain experimental con-ditions (see later), males also show many of thecomponents of behavior normally shown bythe mother rat (Rosenblatt et al., 1996; Rosen-blatt and Ceus, 1998). We describe the generaland specific methods for the testing of maternalbehavior. Also, several environmental and situ-ational factors that affect the expression of ma-ternal behavior must be considered.

DESCRIPTION OFMATERNAL BEHAVIOR

Based on a long history of research in the area(Weisner and Sheard, 1933), there is a rela-tively complete picture of the phenomenol-

287

Page 305: 0195162854

288 DEVELOPMENT

ogy of rat maternal behavior. Under the in-fluence of maternal hormones, the newmother rat is maternally responsive to new-born pups as soon as they emerge from thebirth canal (Rosenblatt and Lehrman, 1963).Typically rats give birth to litters of 8 to 16pups, and these litters tend to consist of ap-proximately an equal number of males and fe-males. At parturition (birth), the mother pullsoff the amniotic sac, eats the placentas, andcleans the pups (Hudson et al., 1999). Withinthe first 30 minutes after parturition, she re-trieves all of the pups to a nest site, mouthesand licks them, and adopts a nursing postureover them. No prior experience with pups isneeded for this immediate maternal respon-siveness (Fleming and Rosenblatt, 1974). Pupsremain with the mother rat until weaningwhich occurs normally between postnataldays 22 and 30. As the pups develop, they movetoward independence, first eating crumbsfrom around the mother's mouth and fur andeventually eating rat chow at a distance fromthe nest and drinking from the provided wa-ter bottle. Over time, pups spend less time inthe nest and with the mother rat and moth-ers often actively move away from pups whenthey approach.

Once maternal behavior has been exhib-ited at parturition under hormonal influences,the behaviour is sustained through experi-ences with the pups acquired during the post-partum period. This effect of experience isalso influenced by input from the pups them-selves (Li and Fleming, 2003).

Described here are some of the moreimportant maternal behaviors displayed byrat dams, especially during the first 10 dayspostpartum.

RETRIEVAL

Retrieval of pups consists of the mother ratcarrying a pup to the present nest or to a newlocation where a new nest will be constructed.This behavior is almost always observed afterparturition, when pups tend to be scattered

around the nest. Pup retrieval, however, isalso evident during the retrieval test, and thisbehavior is thought to be a measure of a rat's"motivation" to be maternal. A very mater-nally responsive rat retrieves the pups quicklyand efficiently to the nest site, whereas a non-maternal rat does not retrieve the pups. Amother rat will frequently pick up pups whileshe and the pups are in the nest and reposi-tion the pups within the nest. This type of be-havior is not considered retrieval; rather, it isreferred to as pup pick-ups or mouthing.

PUP LICKING

Pup licking is an important source of stimu-lation for newborn pups. In the past severaldecades, a number of studies have demon-strated the important effect of pup licking onoffspring's emotionality (Francis and Meaney,1999), cognition (Liu et al., 2000; Lovic andFleming, 2003), and physiology (Kuhn andSchanberg, 1998), as well as propagation ofmaternal behavior (Fleming et al., 2002).There are two types of licking: pup body lick-ing and pup anogenital licking. Body lickingcan be observed during various circumstancesin the maternal cage (e.g., before retrieval, be-tween retrievals, during nursing), whereasanogenital licking tends to be observed whilepups are nursing and are on their backs. Dur-ing anogenital licking, pups show reflexivehind extremities extensions in response toanogenital stimulation. This type of stimula-tion is important to ensure pup urination anddefecation and plays an important role in sex-ual development of the male pup (Moore,1984). The ability to observe pup licking, andto differentiate between two types of licking,is initially difficult but can be acquired withina few hours of observing maternal behavior.

NURSING POSTURES

The purpose of crouching postures is to allowthe pups access to teats and milk, to regulatetheir temperature, and to protect them from

Page 306: 0195162854

Chapter 27. Maternal Behavior 289

environmental elements. The definition ofcrouching differs greatly across studies and lab-oratories. There are two general types of nurs-ing postures: hovering and crouching. Hover-ing is a posture in which a rat is positionedover some or all of the pups in the nest, butthe female is not quiescent. She is actively lick-ing pups, moving the nest material, self-grooming, or moving pups within the litterwhile hovering. Despite the mother rat beingactive, at least some pups have access to herteats (Fig. 27-1 A). Crouching is considered tobe a quiescent posture, and it usually occursin response to sufficient stimulation by pups.A mother rat tends to stop other activities (al-though anogenital licking is sometimes ob-served) and develops a characteristic posturewith her extremities spread out and backarched. Crouching is sometimes divided into

low crouching and high crouching postures,depending on the degree of the arch of thespinal column (Figs. 27-1B and 27-lC). Athird nursing posture that is rarely observedduring testing, especially with younger pups,is a supine posture. Here the mother rat lieson her side, giving the pups access to her nip-ples. This posture is observed during longerperiods of undisturbed maternal behavior inthe nest with older pups (>10 days old; seeFig. 27-1D).

NEST BUILDING AND OTHERMATERNAL BEHAVIORS

Other maternal behaviors that are typicallyobserved and recorded during a maternal testare nest building (in which the mother ratgathers nesting material to a nest site) and

Figure 27-1. Illustrations of different nursing postures. (A) Hovering. The mother rat is over all or most ofthe pups while engaging in other behaviors such as mouthing, repositioning, licking pups, fixing the nest, orself-grooming. (B) Low crouch. Soon after the mother rat settles in the nest, the frequency of behaviors otherthan low crouching decreases. She extends her limbs with her back slightly arched allowing the pups accessto her ventrum. (Q High crouch. This posture is characterized by high arching of the back and significant limbextensions by the mother rat that allow the rooting pups access to her ventrum. Typically, mother rats ceaseall other activity during high crouch nursing. (D) Supine. This posture is characterized by the mother ratlying/sleeping on her side with pups attached to her teats. This posture is observed with older pups and af-ter longer periods of undisturbed nest environment.

Page 307: 0195162854

290 DEVELOPMENT

sniffing pups. For nest building, often a nestrating can be used (in which 1 indicates nonest; 2, some nest [some nest materialarranged in a single location]; 3, moderatenest [low walls of a nest are evident]; 4, goodnest [walls of nest are distinct and surround acavity]; and 5, excellent nest [walls of nest arehigh and can conceal pups]). The time that themother rat spends in the nest can also berecorded.

NONMATERNAL BEHAVIORSIN MATERNAL CONTEXT

Mother rats engage in a number of nonma-ternal behaviors in a maternal testing context,and it is a good practice to record these be-haviors while testing for maternal behavior,to assess general activity level and what themother is doing when she is not respondingto the pups. Mother rats often groom them-selves, usually after retrieving the pups.Mother rats also engage in a number of ex-plorative behaviors such as sniffing air and dig-ging of the bedding. They eat, occasionallysettle in a cage corner, and sleep away frompups.

CONDITIONS OF OBSERVINGAND QUANTIFYING

MATERNAL BEHAVIOR

To assess maternal behavior in rats, a numberof standardized procedures have been devel-oped that vary somewhat from one laboratoryto another, usually in the duration of the ob-servations. The following described observa-tion strategies are derived from work in sev-eral laboratories.

TEST CAGE AND ENVIRONMENT

Testing should be conducted in a large, clearPlexiglas cage. Rats should be transferred tothe test cage either during late pregnancy, orat least 1 day before testing begins, for habit-

uation and to reduce effects of novelty. Typ-ical dimensions of test cages are approxi-mately 45 X 40 X 20 cm, although cages ofdifferent dimensions can be used. It is impor-tant that the test cage is transparent, to allowa clear view of behaviors, and that the type ofcage is consistent across testing, because thesize and condition of the cage can have a sig-nificant impact on maternal behaviors. Woodshavings bedding should be spread over thebottom of the cage at an approximate depthof 1.5 cm, and nesting material should also beprovided (e.g., two paper towels shreddedinto 2 to 3 cm pieces). In addition to the cagesetting, laboratory setting should be stan-dardized, with ambient temperature main-tained at about 22° C, 40% to 50% humidity,and 12 hour light-dark photo period. Testingis usually done during the light phase or in-active phase of the cycle. Observations can bedone during the dark phase as well, but ratstend to be more active during this phase. Forall tests, litters should be culled to an equalnumber of pups; eight pups is standard (fourmales and four females).

METHODS OF RECORDING BEHAVIOR

There are two general methods of quantify-ing maternal behavior in the laboratory set-ting. One method involves periodic samplingof ongoing maternal behavior, sometimes re-ferred to as spot checks (e.g., recording whatthe mother does at 100 different 5 secondpoints during the day) (Francis et al., 2002).This type of testing is almost always donewith rats whose nests have not been dis-turbed. The other type of testing involves acontinuous observation period (10 to 120minutes) with either a nondisrupted or dis-rupted maternal nest. Several continuous ob-servations can be made over several days toget a good idea of changes in maternal be-havior over time. The method of choiceshould be based on the nature and needs ofthe experiment, as well as the resources andconstraints at hand.

Page 308: 0195162854

Chapter 27. Maternal Behavior 291

Paper and Pencil Technique(One-Zero Time Sampling)A table can be created, with columns repre-senting time and rows representing behaviors.This creates boxes that represent each behav-ior at every time point (typically a 5 secondtime point). By checking each behavior that isseen during a time point, the frequency ofeach behavior over the test can be determinedwhen the number of checks over each row areadded. This method of recording is conven-ient but not as precise as the use of an eventrecorder.

Event Recording of Ongoing BehaviorThis method of recording consists of using acomputer with an event-recording program(e.g., Behavioral Evaluation Strategy and Tax-onomies [BEST] Program) that can be tailoredfor each user in terms of length of the test andnumber of behaviors recorded. Each behaviorto be recorded is represented by a key on thekeyboard. When the mother rat performs abehavior, the designated key is pressed untilthe mother rat ceases to perform this behav-ior. This allows the tester to record both fre-quency (each time a key is pressed) and dura-tion (how long a key is pressed) of eachbehavior over the test.

MATERNAL TEST PROTOCOLS

The following are not the only types of ma-ternal tests used, but they are the simplest andmost typical. The test of choice should be basedon the nature and needs of the experiment.

Continuous Observation ofNondisturbed Maternal BehaviorA nondisturbed maternal test is a basic obser-vation of natural, nondisturbed maternal be-haviors. At no point during this test are themother rat and her litter disturbed and thetester can choose to start observations at anypoint during the day by simply beginning torecord ongoing behaviors. These tests canrange in length from 10 minutes to several

hours. Rats can be tested on one particular dayor over several continuous or alternate days.Unless processes of weaning are of interest,testing is usually undertaken over the first2 weeks postpartum.

This test protocol allows the tester to de-termine whether experimental manipulationsaffect the natural display of maternal behav-ior—that is, how much time the mother ratspends with the pups and if her pattern of be-haviors is similar to that shown by a non-manipulated mother rat.

Maternal Behavior Test with RetrievalMaternal behavior test with retrieval is simi-lar to an undisturbed maternal test, exceptthat the mother and pups are separated for abrief period before the beginning of the test.The tester removes the pups from the nest,leaving the nest relatively intact. Pups are re-moved for about 5 minutes and must be main-tained at room temperature. After this briefseparation period, pups are returned to thetest cage in the corner diagonally opposite thenest. The recordings should start immediatelyas maternally responsive rats will start re-trieving the pups without a delay.

This test is often performed in combina-tion with subsequent spot checks across theday. These spot checks are done at approxi-mately 2 hour intervals after the retrieval test.The retrieval test mainly assesses retrieval be-havior that is an indication of maternal moti-vation. Retrieval behavior varies greatly be-tween postpartum rats, sensitized virgin rats,and juvenile rats, so the degree of maternalmotivation can easily be assessed with thistest. Less motivated rats, such as virgin rats,show little, if any, retrieval behavior, whereasmore motivated rats, such as newly parturi-ent rats that are under the influence of ma-ternal hormones, show efficient retrieval of allpups in the cage.

Maternal Memory TestThe role of maternal hormones subsides overthe first few days after parturition. Then ma-

Page 309: 0195162854

292 DEVELOPMENT

ternal behavior is maintained through experi-ence, sometimes referred to as maternal mem-ory or maternal experience effect (Li and Flem-ing, 2003). The latency of maternal behaviorafter a period of separation from pups isshorter if rats have had an earlier maternal in-teraction with pups than if they have not.

The maternal memory test is a variationof the retrieval test except that it measureshow well rats "remember" their experiencewith pups during the postpartum period orearlier induction period. In general, much likeother tests of memory, maternal memory testhas two phases: the experience or exposurephase and the test phase.

After parturition, pups are taken awayfrom the mother rat at 15 minute intervals,not allowing her to have the experience of be-ing maternal. After a certain period, from sev-eral hours later to the next day, the motherrat is exposed to foster pups obtained from adonor mother rat. Most likely, this mother ratwill readily display maternal behavior. Themother rat is left with the pups for at least 1hour and then separated from any pup-relatedcues for several days (e.g., 10). After this pe-riod of isolation, the mother rat is tested againwith new foster pups.

Mother rats with intact "maternal mem-ory" display maternal behavior within 1 or 2days, whereas mother rats treated with somekind of manipulation (e.g., treatment withprotein synthesis inhibitors or lesions of thenucleus accumbens [Li and Fleming, 2003])will not show maternal behavior. Alterna-tively, mother rats with poor maternal mem-ories have long latencies to show maternal be-havior, similar to virgin rats.

Maternal Motivation and Preferencefor Pup-Related Stimuli TestsAlthough rat maternal behavior is organized,the mechanisms of its control are quite nu-merous and dependent on each other, as mo-tor, motivational, experiential, attentional,and other factors all play a role in the expres-sion of maternal behavior (e.g., Mattson et al.,

2003). Maternal behavior tests in the maternalcontext are useful for assessing actual mater-nal responsiveness and performance (i.e., ap-petitive component of maternal behavior),however, they are not necessarily the besttools for the assessment of individual mecha-nisms (e.g., motivation) that contribute to ma-ternal behavior. A description follows of atesting procedure that can be used to assesspup hedonic values and avoidance of pup-associated stimuli.

One test that can assess motivation is theconditioned place preference test, and it hasbeen used to assess how rewarding pups are.The test is useful because it can discriminatebetween rats that show similar maternal be-havior during a typical maternal test, as out-lined earlier, yet do not find pups equally re-warding (Morgan et al., 1999; Mattson et al.,2003). Testing is done in an apparatus thatconsists of two white Plexiglas boxes (22 X40 X 30 cm), with each box differing in wallpatterns (horizontal black bars versus verticalblack bars) and texture of the floor (smoothversus perforated). The test consists of twophases: the exposure phase and the test phase.During the exposure phase, rats are randomlyassigned to have one specific box associatedwith pups. On two alternate days, rats are ex-posed to pups in the assigned box, and on theother two alternate days, rats are exposed tothe other box (Mattson et al., 2003).

During the test phase, on test day, the ratsare placed in the test apparatus without bar-riers between the boxes, allowing them tofreely spend time in either of the boxes. En-tries and time spent in each box is recordedduring a 10 minute testing period. Rats thatfind pups rewarding will spend significantlymore time in the box that was previously as-sociated with pups than in the other box. Forexample, multiparous and primiparous ratsfind pups more rewarding than virgin rats andthus spend more time in the box associatedwith pup-related cues.

Another way to determine how mater-nally "motivated" rats are, without actually

Page 310: 0195162854

Chapter 27. Maternal Behavior 293

testing maternal behavior, is to test rats in apreference or choice test. Two stimuli (pupurine odors versus diestrous female urineodors or cool vocalizing pups versus warmnonvocalizing pups) are placed simultane-ously at two ends of a Y maze, or single stim-uli are presented consecutively on successivetrials, and the relative times spent in proxim-ity to each stimulus during a 5 minute test iscomputed (Fleming et al., 1989; see also Far-rell and Alberts, 2002).

Testing of Maternal Behaviorin Nonpostpartum RatsAlthough initially "fearful" and avoidant ofpups, virgin female rats can be induced to be-come maternal, without showing lactation,through continuous exposure to pup stimula-tion. We describe two procedures that areused to induce maternal behavior in virginrats.

Sensitization. Not having the advantage ofhormonal priming associated with pregnancyand parturition, virgin females and males arenot maternally responsive when first pre-sented with newborn foster pups (Rosenblatt,1967; Fleming and Luebke, 1981; Rosenblattand Ceus, 1998). After 5 to 10 days of contin-uous exposure to pups, virgin rats eventuallybegin to respond maternally (Rosenblatt,1967), showing a pattern of behavior thatclosely resembles, but does not perfectly par-allel, the behavior of a new mother rat (Lon-stein et al., 1999). To be considered maternalduring the sensitization procedure, virgin ratsare continuously exposed to foster pups untilthe maternal criterion is reached—that is, thevirgin rat retrieves all foster pups to the nestsite over 2 consecutive days during retrievaltests. Maternal tests of sensitized virgin ratsshould follow the protocols outlined earlier.Foster pups need to be replaced daily becausevirgin rats do not lactate and pups need milkto survive (donor mothers should not be leftwith fewer than six pups because lactationmust be maintained in these mothers). One

problem that may occur during the sensitiza-tion procedure is cannibalization. Becausevirgin rats find pups aversive, virgin rats oc-casionally kill the pups. If this occurs duringthe test, the test should be stopped, injuredpups should be euthanized, and healthy pupsshould be returned to their mothers. A virginthat cannibalizes 2 days in succession shouldbe removed from the experiment.

Hormonal Priming. Virgin rats can also be in-duced to become maternal through the ad-ministration of maternal hormones, namelyestrogen and progesterone (Bridges, 1984).Virgin rats are ovariectomized and subcuta-neously implanted with a 2 mm Silastic cap-sule of estrogen in the dorsal region of theneck. Three days after these initial proce-dures, three 30 mm Silastic capsules of pro-gesterone are subcutaneously implanted inthe same area. Ten days after this procedure,the progesterone capsules are removed andthe estrogen capsule is left. One day after theremoval of the progesterone capsules, the vir-gin rat can be exposed to pups for maternaltesting. As in the sensitization procedure, pupsmust be replaced daily because hormonallyprimed virgin rats do not lactate. Althoughcloser to paralleling maternal behavior of thepostpartum rat than the behavior of the sen-sitized virgin, hormone priming of virgin ratsdoes not account for the actions of other hor-mones, such as prolactin, oxytocin, and corti-costerone, which are also altered in the lacta-tional/postpartum state.

To assess relative sensitivities to hor-mones in rats under different conditions, thehormone regimen can be either lengthened orshortened (Bridges, 1984).

Tests of juvenile Maternal BehaviorImmediately after weaning, male and femalejuvenile rats spontaneously show maternalbehavior (Rees and Fleming, 2001), whichsubsides around the onset of puberty. In nat-ural environments, mother rats are typicallygiving birth to a second litter when their first

Page 311: 0195162854

294 DEVELOPMENT

litter is being weaned (Gilbert et al., 1983), andhence, juvenile rats are exposed to pupsaround weaning.

Juvenile maternal testing is similar to thevirgin sensitization procedure, in that overseveral days, juvenile rats are exposed to pupsand juvenile behavior is recorded during a ma-ternal test. There are, however, some differ-ences between adult and juvenile maternal be-havior because juvenile rats are much smallerthan adult rats and tend to drag rather thancarry pups to the nest.

Also due to their size, juvenile rats can-not adopt a rigid arched back posture overpups and instead tend to lie on top of the pups.Juvenile rats may show all components of ma-ternal behavior before they show retrieving,which is the opposite pattern observed inadult maternal rats. Juvenile maternal behav-ior can be of interest because it allows thetester to investigate the development of ma-ternal behavior and, in a more general sense,of early social behavior.

ISSUES TO BE CONSIDEREDWHEN TESTING

There are some factors to consider when test-ing maternal behavior. If possible, these fac-tors should remain constant (or should becontrolled for) across experimental groups be-cause confounds may arise.

Separation of Mother Rat from PupsOne of the most well-documented experi-mental manipulations of maternal care is earlymaternal separation (Lehmann and Feldon,2000). It is well known that any type of sepa-ration can alter pup development (Hofer,1994) and the mother's behavior toward pups(Pryce et al., 2001).

Very short periods of separation (<15minutes), termed "early handling," increasesthe mother's behavior toward pups over sev-eral time points after the separation period.On the other hand, longer periods of separa-tion (>1 hour) do not increase maternal re-

sponsiveness (Pryce et al., 2001). If some typeof manipulation that separates the mother ratfrom pups, for example drug administration,is being investigated in conjunction with ma-ternal behavior, the disruptive effects of sep-aration should be considered in the behavioralanalysis.

Cage SizeWhen testing adult males or virgin females orjuveniles of either gender, in the inductionparadigm, the size of the cage affects the speedwith which rats begin to show retrieval be-havior. Very small cages, where the rat andthe pups are forced into close proximity, re-sults in faster latencies to initiate retrieving(across days), than in larger cages, where ratscan maintain a distance from the pups.

Orcadian RhythmsMaternal behavior, like other behaviors, has acircadian rhythm. Typically, mother rats in-teract more with their pups during the day orlight part of the cycle (Leon et al., 1984).

Size of the LitterOne important factor that should be and caneasily be controlled in studies investigatingmaternal behavior is the size of the litter. Ithas been demonstrated that the larger the lit-ter, the less licking is received by the litter, perpup and per litter (Fleming et al., 2002).Mother rats, however, spend more time withlarger litters than with smaller litters (De-viterne et al., 1990). To be able to compare re-sults across studies, the litter size must remainconstant, because changes in the litter size alsochange pup growth rate and development(Agnish and Keller, 1997).

Age of LitterMaternal behavior changes as the pups age(Stern and MacKinnon, 1978). There is morematernal behavior, especially retrieval, re-ceived by 1-week-old pups than by older pups,perhaps because pups are more active and in-dependent as they age. Hence, if donor pups

Page 312: 0195162854

Chapter 27. Maternal Behavior 295

are used during testing, it is recommendedthat they be young (1 to 5 days old).

Gender Ratio of the LitterAnother issue to consider is the gender ratioof the litter. Before testing, a litter should beculled so that there is an equal number of fe-male and male pups. In a litter with an equalgender ratio, there are gender differences inlicking received by pups. Male pups receivemore anogenital licking from the mother thando female pups (Moore, 1985). This prefer-ence is based on many factors, one beingtestosterone and another being pheromones(Moore, 1986).

After the first postpartum week, motherrats will also begin to retrieve male pups be-fore female pups during a retrieval test (De-viterne and Desor, 1990).

Also, one study shows that female ratsraised in unisexual litters have lower frequen-cies of pregnancy, but when pregnancy andbirth occur, litter sizes are much larger thanthose of rats raised in bisexual litters (Sharpe,1975). If gender is an important issue being in-vestigated, the difference in maternal behav-ior received between genders can easily be as-sessed by marking female and male pups withdifferent colors and tracking the behavior ofthe mother rat toward each gender and color.

Strain DifferencesThere are differences in maternal behavioracross strains of rats. For example, Long-Evans mother rats show more licking, espe-cially anogential licking, toward their pupsthan do Fisher 344 mother rats, although gen-der differences in licking are retained (Mooreet al., 1997). Long-Evans mother rats alsoshow a more ordered litter, more physicalcontact with their pups, and more frequentnursing postures than both Wistar andSprague-Dawley mother rats (Maclver andJeffrey, 1967). These differences in maternalbehavior across strains remain constant re-gardless of the strain of foster pup (Maclverand Jeffrey, 1967). Finally, in different labora-

tories, there are marked differences in the la-tency with which rats begin to show mater-nal behavior during the induction procedure(Terkel and Rosenblatt, 1971). These differ-ences may well be related to differences be-tween the strains in "emotionality" or "neo-phobia" (Fleming and Luebke, 1981). Hence,when comparing studies, strain of rat shouldbe taken into account due to the fact that thereare strain differences in maternal behavior.

Use of Foster PupsAlthough maternal rats accept any pups pre-sented to them, evidence suggests that theycan differentiate between their own and fos-ter pups and that they treat them differently.For example, foster pups from split litters, inwhich half of the pups are foster pups and halfare pups from the mother being tested, weighmuch less by postnatal day 30 and show a re-duced capacity to survive food deprivationthan biological pups (Ackerman et al., 1977),suggesting that foster pups are treated differ-ently than the mother's own offspring.

Parity of the MotherMaternal behavior changes with parity, ornumber of births, because primiparous (gavebirth to one litter) and multiparous (gave birthto more than one litter) rats will react differ-ently to pups (Wright et al., 1977). There arealso differences in the neural circuit of mater-nal rats as a function of parity with increasesin GFAP (Featherstone et al., 2000) and opi-ate receptors (Bridges and Hammer, 1992) inthe medial preoptic area. This demonstratesthat unless parity is being compared, all ratsmust be of the same reproductive condition(nulliparous, primiparous, or multiparous) tocontrol for the potential confound of parity.

CONCLUSION

Maternal behavior can easily be recordedthrough careful observations and well-plannedprocedures. Many factors should be consid-

Page 313: 0195162854

296 DEVELOPMENT

ered while planning a maternal behaviorstudy because maternal behavior is a complexbehavior, influenced by many factors and reg-ulated by multiple mechanisms. Being a reli-able and robust behavior, it can also be usedin laboratory demonstrations to illustrate, forexample, the role of hormones or of olfactorycues in the regulation of behavior, in the ex-pression of learning, or in the principles of re-inforcement within a species-specific charac-teristic context.

ACKNOWLEDGMENTS

We give many thanks to Alison Diaz for production of the il-lustrations used in this chapter.

REFERENCES

Ackerman SH, Hofer MA, Weiner H (1977) Some ef-fects of a split litter cross foster design applied to 15day old rat pups. Physiology and Behavior19:433-436.

Agnish ND and Keller KA (1997) The rationale forculling of rodent litters. Fundamental and AppliedToxicology 38:2-6.

Bridges RS (1984) A quantitative analysis of the roles ofdosage, sequence, and duration of estradiol andprogesterone exposure in the regulation of mater-nal behavior in the rat. Endocrinology 114:930-940.

Bridges RS and Hammer RP Jr (1992) Parity-associatedalterations of medial preoptic opiate receptors infemale rats. Brain Research 578:269-274.

Ceger P and Kuhn CM (1998) Responses to maternalseparation: Mechanisms and mediators. Interna-tional Journal of Developmental Neuroscience16:261-270.

Deviterne D and Desor D (1990) Selective pup retriev-ing by mother rats: Sex and early development char-acteristics as discrimination factors. DevelopmentalPsychobiology 23:361-368.

Deviterne D, Desor D, Krafft B (1990) Maternal behav-ior variations and adaptations, and pup develop-ment within litters of various sizes in Wistar rats.Developmental Psychobiology 23:349-360.

Farrell WJ and Alberts JR (2002) Maternal responsive-ness to infant Norway rat (Rattus norvegicus) ultra-sonic vocalizations during the maternal behavior cy-cle and after steroid and experiential inductionregimens. Journal of Comparative Psychology116:286-296.

Featherstone RE, Fleming AS, Ivy GO (2000) Platicityin the maternal circuit: Effects of experience andpartum condition on brain astrocyte number in fe-male rats. Behavioral Neuroscience 114:158-172.

Fleming AS, Kraemer GW, Gonzalez A, Lovic V, ReesS, Melo A (2002) Mothering begets mothering: Thetransmission of behavior and its neurobiologyacross generations. Pharmacology, Biochemistry,and Behavior 73:61-75.

Fleming AS and Li M (2002) Psychobiology of maternalbehavior and its early determinants in nonhumanmammals. In Handbook of parenting: Biology andecology, Vol 2 (Borenstein MH, ed.). Mahwah, NJ:Lawrence Erlbaum Associates.

Fleming AS and Luebke C (1981) Timidity prevents thevirgin female rat from being a good-mother: Emo-tionality differences between nulliparous and par-turient females. Physiology and Behavior 27:863-868.

Fleming AS and Rosenblatt JS (1974) Maternal behaviorin the virgin and lactating rat. Journal of Compara-tive and Physiological Psychology 86:957-972.

Francis DD and Meaney MJ (1999) Maternal care andthe development of stress responses. Current Opin-ion in Neurobiology 9:128-134.

Francis DD, Young LJ, Meaney MJ, Insel TR (2002) Nat-urally occurring differences in maternal care are as-sociated with the expression of oxytocin and vaso-pressin (Via) receptors: Gender differences. JNeuroendocrinol 14:349-353.

Gilbert AN, Burgoon DA, Sullivan KA, Adler NT (1983)Mother-weanling interactions in Norway rats in thepresence of a successive litter produced by postpar-tum mating. Physiology and Behavior 30:267-271.

Gonzalez-Mariscal G and Poindron P (2002) Parentalcare in mammals: immediate internal and sensoryfactors of control. In: Hormones, brain and behav-ior, Vol. 1 (Eds, Pfaff DW, Arnold AP, Etgen AM,Fahrbach SE, Rubin RT). Elsevier Science: SanDiego.

Hofer MA (1994) Early relationships as regulators ofinfant physiology and behavior. Acta Pediatrica397:S9-S18.

Hudson R, Cruz Y, Lucio A, Ninomiya J, Martinez-Gomez M (1999) Temporal and behavioral pattern-ing of parturition in rabbits and rats. Physiology andBehavior 66:599-604.

Lehmann J and Feldon J (2000) Long-term biobehavioraleffects of maternal separation in the rat: Consistentor confusing? Reviews in Neuroscience 11:383-408.

Leon M, Adels L, Coopersmith R, Woodside B (1984)Diurnal cycle of mother-young contact in Norwayrats. Physiology and Behavior 32:999-1003.

Li M and Fleming AS (2003) Differential involvement ofnucleus accumbens shell and core subregions in ma-

Page 314: 0195162854

Chapter 27. Maternal Behavior 297

ternal memory in postpartum female rats. Behav-ioral Neuroscience 117:426-445.

Liu D, Diorio J, Day JC, Francis DD, Meaney MJ (2000)Maternal care, hippocampal synaptogenesis andcognitive development in rats. Nature Neuro-science 3:799-806.

Liu D, Diorio J, Tannenbaum B, Caldji C, Francis D,Freedman A, Sharma S, Pearson D, Plotsky PM,Meaney MJ (1997) Maternal care, hippocampal glu-cocorticoids receptors, and hypothalamic-pituitary-adrenal responses to stress. Science 277:1659-1662.

Lonstein JS, Wagner CK, De Vries GJ (1999) Compari-son of the "nursing" and other parental behaviorsof nulliparous and lactating female rats. Hormonesand Behavior 36:242-251.

Lovic V and Fleming AS (2004) Artificially reared femalerats show reduced prepulse inhibition and deficitsin the attentional set-shifting task—reversal of ef-fects with maternal-like licking stimulation. Behav-ioural Brain Research, 148(1-2):209-219.

Mattson BJ, Williams SE, Rosenblatt JS, Morrell JL(2003) Preferences for cocaine- or pup-associatedchambers differentiates otherwise behaviorallyidentical postpartum maternal rats. Psychopharma-cology (Berlin) 167:1-8.

Mclver AH and Jeffrey WE (1967) Strain differences inmaternal behavior in rats. Behaviour 28:210-216.

Morgan HD, Watchus JA, Milgram MW, Fleming AS(1999) The long lasting effects of electrical simula-tion of the medial preoptic area and medial amyg-dala on maternal behavior in female rats. BehavioralBrain Research 99:61-73.

Moore CL (1984) Maternal contributions to the devel-opment of masculine sexual behavior in laboratoryrats. Developmental Psychobiology 17:347-356.

Moore CL (1985) Sex differences in urinary odors pro-duced by young laboratory rats (Rattus norvegicus).Journal of Comparative Psychology 99:336-341.

Moore CL (1986) A hormonal basis for sex differencesin the self-grooming of rats. Hormones and Behav-ior 20:155-165.

Moore CL, Wong L, Daum MC, Leclair OU (1997)Mother-infant interactions in two strains of rats: Im-plications for dissociating mechanism and functionof a maternal pattern. Developmental Psychobiol-ogy 30:301-312.

Numan M and Sheehan TP (1997) Neuroanatomical cir-cuitry of mammalian maternal behavior. Annals ofNew York Academy of Sciences 807:101-125.

Pryce CR, Bettschen D, Feldon J (2001) Comparison ofthe effects of early handling and early deprivationon maternal care in the rat. Developmental Psy-chobiology 38:239-251.

Rees SL and Fleming AS (2001) How early maternal sep-aration and juvenile experience with pups affect ma-ternal behavior and emotionality in adult postpar-tum rats. Animal Learning and Behavior 29:221-233.

Rosenblatt JS (1967) Nonhormonal basis of maternal be-havior in the rat. Science 156:1512-1514.

Rosenblatt JS (2002) Hormonal basis of parenting inmammals. In: Handbook of parenting, Vol 2 (Bor-nenstein MH, ed.). Mahwah, NJ: Lawrence ErlbaumAssociates.

Rosenblatt JS and Ceus K (1998) Estrogen implants inthe medial preoptic area stimulate maternal behav-ior in male rats. Hormones and Behavior 33:23-30.

Rosenblatt JS, Hazelwood S, Poole J (1996) Maternal be-havior in male rats: Effects of medial preoptic arealesions and presence of maternal aggression. Hor-mones and Behavior 30:201-215.

Rosenblatt JS and Lehrman DS (1963) Maternal behav-ior in the laboratory rat. In: Maternal behavior inmammals (Rheingold HL, ed.). New York: Wiley.

Sharpe RM (1975) The influence of the sex of litter-mateson subsequent maternal behavior in Rattus norvegi-cus. Animal Behavior 23:551-559.

Smotherman WP, Wiener SG, Mendoza SP, Levine S(1976) Pituitary-adrenal responsiveness of rat moth-ers to noxious stimuli and stimuli produced by pups.CIBA Foundation Symposium 45:5-25.

Stern JM (1996) Somatosensation and maternal care inNorway rats. Advances in the Study of Behavior25:243-293.

Stern JM and MacKinnon DA (1978) Sensory regulationof maternal behavior in rats: Effects of pup age. De-velopmental Psychobiology 11:579-586.

Trekel J and Rosenblatt JS (1971) Aspects of nonhor-monal maternal behavior in the rat. Hormones andBehavior 2:161-171.

Wiesner BP and Sheard NM (1933) Maternal behaviorin the rat. Edinburgh: Oliver.

Page 315: 0195162854

Play and Fighting

SERGIO M. PELLIS AND VIVIEN C. PELLIS28

Fighting involves attack by one animal anddefense by another. After a successful deflec-tion of an opponent's attack, the defendermay then launch its own attack; this leads toa defense by the other animal. The relativeskills in attack and defense determineswhether an attacker succeeds (Geist, 1971).Rats are no exception to these general prin-ciples and have been a favorite subject for thestudy of aggression (Blanchard and Blan-chard, 1994). A conceptual and descriptivedifficulty has arisen because rats engage inboth agonistic and playful forms of fighting.The problem has been in identifying whetherthey differ and how and by what criteria anobserver can determine when one grades intothe other.

Serious fighting is most often seen inadults, and play fighting is most often seenin juveniles (Pellis and Pellis, 1987). To es-tablish the basic parameters of these be-haviors, the discussion begins with the seri-ous fighting of adults; this is then contrastedwith the playful fighting of juveniles. Weshow that serious and playful fighting aredistinct behaviors in rats, with play derivedfrom sex, not aggression. Further, we showthat rats have modified the sexual contentof play fighting and have co-opted its use inadulthood for quasi-aggressive purposes.To understand the subtle, but important,differences between serious and playfulfighting, an understanding of the targetsthat rats compete for during these contestsis essential.

TARGETS OF AGONISTIC ATTACK

In launching an attack, an attacker seeks to gainsome advantage over their opponent; this mayinvolve throwing the opponent off balance orstriking or biting the opponent (Geist, 1978). Of-ten, blows with feet or specialized structures—such as horns and antlers—or bites are directedto particular body targets. If body targets havebeen subjected to a long history of such attacks,thickening of the skin (to limit penetration ofteeth or horns) in that region or of the under-lying skeleton (which absorbs the impact of ablow) occurs (Pellis, 1997). Rats direct their bitesto specific areas of the opponent's body.

To facilitate the analysis of serious fight-ing, and so identify the targets being attackedand the tactics being used for attack and de-fense of those targets, a resident-intruder par-adigm is useful. In this paradigm, an unfamil-iar male rat is placed into the home enclosureof a resident male; in this situation, the resi-dent does most of the attacking and the in-truder does most of the defending (Blanchardand Blanchard, 1990; Kemble, 1993).

Using the resident-intruder paradigm, ithas been shown repeatedly that the residentdirects most of its bites to the lower dorsumand flanks and that the intruder directs retal-iatory bites to the resident's face. This patternof bite delivery for attackers and defenders hasbeen shown to be true for a variety of do-mestic strains as well as for wild-caught rats(Blanchard and Blanchard, 1990). Further-more, a pattern of scarring consistent with

298

Page 316: 0195162854

Chapter 28. Play and Fighting 299

these targets has been found in free-living rats(Blanchard et al., 1985).

Analyses of the fighting movementsadopted by contestants of several species ofmurid rodents (the family to which rats be-long; see Chapter 1) show that the face is atarget of both attack and defense (Pellis, 1997).The defender launches a bite at the attacker'sface either as retaliation for a bite to theirlower flanks or dorsum, or when approachedand hemmed in by an attacker (Pellis andPellis, 1987; Blanchard and Blanchard, 1994).That is, a defensive bite to the face occurs inresponse to an attack by an opponent.

In contrast, it is the attacker that maneu-vers itself into a position from which to launchan attack at the defender's face (Pellis and Pel-lis, 1992). Although for most encounters attack-ers are more likely to launch bites to the flanksand lower dorsum, the possibility of a shift toan attack to the face poses functional challengesto the defender. Similarly, the possible retalia-tion by the defender in delivering bites to its op-ponent's face poses a problem to the attacker.

EFFECTS OF TARGET LOCATION ONTACTICS OF ATTACK AND DEFENSE

In the typical encounter between resident andintruder male rats, the resident approaches ina lateral orientation and moves toward its op-ponent's rump. In response, the intruder rearsonto its hind feet, fans out its vibrissae, andfaces the opponent. From this position, the in-truder can continually turn to face the residentas it attempts to move to the defender's rear.

A biting lunge to the opponent's rump canbe countered with a biting lunge to the at-tacker's face by the defender. The attacker maythen stand on all fours and orient laterally infront of the upright defender for a protractedperiod of time, seemingly motionless. The lackof movement, however, is only apparent;frame-by-frame analyses of videorecordings re-veal that small movements by one animal are

countered by small movements by the other(Blanchard and Blanchard, 1994). To break thisstalemate, the attacker has to risk retaliation.

The most typical maneuver used by theattacker to break this stalemate is for it topress forward against the intruder while main-taining the lateral orientation (Fig. 28-1). In

Figure 28-1. The lateral attack tactic in serious fighting. Theattacker, usually standing on all fours, orients laterally in frontof a defender (a). Then, as the defender orients to the at-tacker's face, the attacker presses, with its lower flank, againstthe defender's exposed ventrum (b). If the attacker managesto unbalance the defender, the attacker may then seize thisopportunity to lunge and bite at the defender's exposed lowerdorsum (c). (Adapted from Pellis and Pellis [1987]).

Page 317: 0195162854

300 DEVELOPMENT

this way, the attacker can press its lower flankagainst the defender's ventrum, while keep-ing its own head out of reach, as the defenderwould have to relinquish its postural stabilityby lunging obliquely toward the attacker'shead. To do so would then put the defenderin a vulnerable position for a counterattack bythe attacker. From this position, the attackercan lunge at the defender's lower flank. Evenif unsuccessful, this tactic of attack can lead tothe defender losing its footing. The attackercan then lunge and bite the defender's lowerflank or rump, with a reduced chance of be-ing bitten on the face by the defender. The lat-eral attack maneuver can be seen as incorpo-rating a defensive component that protectsthe attacker's head from a retaliatory strike bythe defender (Pellis, 1997).

From the defender's perspective, there isa potential tradeoff between standing itsground versus turning to flee as soon as theattacker's last attack is checked. Althoughstanding its ground may lead the attacker toswitch to a head attack, fleeing exposes thetarget on the defender most likely to be bit-ten, the rump. The same applies to anothercommonly used defensive tactic, that of turn-ing to supine (Adams, 1980). As is the case forthe upright tactic, lying supine protects thedorsum from attack, and facing the attackerwith vibrissae fanned out provides a barrieragainst the attacker's continued attempts togain access. Unfortunately for the defender, ahighly motivated attacker may forgo furtherattempts to gain access to the dorsum andflanks and switch its attack to the face. Fromthe supine position, the defender is even lesswell positioned than when facing its opponentfrom a standing position to block such an at-tack. Again, there is a tradeoffby the defenderin either remaining supine or fleeing the mo-ment that turning to supine has deflected anattack to its rump.

The tradeoff between different tactics isinfluenced by several factors, including differ-ences in the defender's prior experience witha particular attacker as well as the physical

constraints of the test enclosure (Pellis et al.,1989; Pellis and Pellis, 1992). For example, ratsare more likely to use the supine defense insmaller enclosures (Boice and Adams, 1983).Presumably, in confined conditions, the op-tion to turn and flee is more dangerous.

This brief overview of the tactics of at-tack and defense shows that such tactics areused to compete for access to particular bodytargets. It is clear that the movements per-formed during aggressive encounters need tobe analyzed functionally in relationship tothose body targets (Pellis, 1997). Unfortu-nately, much of the comparative and experi-mental literature continues to provide mostlynumerical frequencies of the occurrence ofprespecified tactics, such as the lateral, up-right, and supine positions (Alleva, 1993). Theproblem is that a change in the frequency ofany particular tactic cannot be interpretedgiven that the change could be due to the ac-tion of the attacker, defender, or both (Cools,1985).

Several approaches can be used to ana-lyze interactions so as to discern how themovements of the interactions are related tothe attack and defense of particular targets.One such approach is to use choreographicnotation techniques to monitor the spa-tiotemporal pattern of movements by thebody parts of both animals simultaneously(Pellis, 1989; Foroud and Pellis, 2003).

Another approach is to identify the mo-ments in an encounter—referred to as "deci-sion points"—where one animal, in commit-ting itself to an action, can no longer influencethe subsequent action made by his opponent(Pellis, 1989). For example, consider an at-tacker approaching from the rear and lungingat the opponent's rump. If the attacker leapsinto the air as it lunges, then what the oppo-nent chooses to do to defend its rump can nolonger be influenced by the attacker, who isnow in the air and has limited its own ma-neuverability. In this decision point, the typi-cal murid rodent has three tactics of defensefrom which to choose: (1) to leap or run away,

Page 318: 0195162854

Chapter 28. Play and Fighting

(2) to rotate to supine and face the attacker,or (3) to rotate while standing to an uprightposition and face the attacker (Pellis et al.,1992).

These techniques for evaluating the tac-tics of attack and defense are contingent onknowing the body targets over which the an-imals compete. Naturally, the simplest way toidentify such targets is to score the opponent'swounds (Blanchard and Blanchard, 1990).Such a direct approach can be helpful, but ithas a limitation. If both opponents are skilledat combat, an attack by one may be success-fully countered by the other; this may resultin a situation where neither animal succeedsin delivering a bite (Geist, 1971). Also, suc-cessful parrying of an attack by a defendermay lead the attacker to deliver a bite oppor-tunistically on some other body location (Pel-lis and Pellis, 1988). Therefore, wound marksmay not necessarily reveal the body targetsaround which tactics of attack and defense areorganized (Pellis, 1997).

Choreographic descriptive methods canbe used to determine the targets by identify-ing those body parts that are approached byone combatant and withdrawn by the other(Pellis, 1989). In turn, the targets so revealedcan be substantiated by techniques that di-rectly assess the targeting of the attacker, suchas by placing an anesthetized intruder into thehome cage of a resident, and then recordingthe body targets that are bitten. Given that theintruder cannot move, the choice of target bythe resident is not constrained by the in-truder's defensive actions (e.g., Blanchard etal., 1977; Pellis and Pellis, 1992). Similarly,models of intruders, such as euthanized ratsthat are frozen into a specific posture, can beplaced into the resident's cage (Kruk et al.,1979).

Regardless of the combination of tech-niques used, knowing the body parts targetedby combatants facilitates the analysis of howthe tactics that are used are organized to at-tack and defend those targets. The frameworkprovided by such an analysis of targets and

301

tactics provides a basis from which to evalu-ate the causal mechanisms accounting forspecies, sex, and age differences and experi-mental effects on overt fighting behavior(Pellis, 1997).

TARGETS IN THE PLAYFIGHTING OF RATS

When the content of play fighting is scored interms of the behavior patterns used, rats arefound to rotate to supine, stand upright, facethe partner in a lateral orientation, and lungeand flee (Poole and Fish, 1975). That is, thecontent of play involves the same behaviorpatterns that are seen in serious fighting(Grant and Macintosh, 1966). Some re-searchers have argued that on the basis of therelative frequency and probabilistic relation-ship between these behavior patterns in playfighting compared with serious fighting, thatthe two types of fighting are causally gener-ated by different motivational systems (Pooleand Fish, 1975; Panksepp, 1981). Others haveclaimed that the small differences between theplay fighting of juveniles and the serious fight-ing of adult rats reveal that play fighting is animmature form of serious fighting (Silverman,1978; Taylor, 1980). Both playful and seriousfighting involve competitive interactions thatuse the same species-typical behavior patternsof aggression. Scoring the relative frequencyof these behavior patterns and their sequen-tial organization leads to a debate of "Just howdifferent is different?"

A descriptive approach using choreo-graphic methods revealed that during playfighting, rats compete for access to the part-ner's nape. This nape contact rarely involvesbiting; rather, the attacker nuzzles its partner'snape with its snout (Pellis, 1988). That is, theattacking partner uses tactics of attack to gainaccess to the nape; the defending partner thenuses tactics of defense to avoid such contactand to launch its own attacks on the originalattacker's nape, which then defends and so on

Page 319: 0195162854

302 DEVELOPMENT

Figure 28-2. In play fighting, the at-tacker uses tactics of attack, such aslunging, to gain access to its part-ner's nape (a). In turn, the defenderuses tactics of defense, such as ro-tating to supine, to avoid this con-tact (b). From this supine position,the defender may then launch itsown attack on the original attacker'snape (c). The original attacker maythen use defensive tactics to blocksuch an attack (d). (Adapted fromPellis and Pellis [1987]).

(Fig. 28-2). Thus, although the targets in se-rious fighting are the lower dorsum and flanksand, to a lesser extent, the face, during playfighting the target is the nape area, which isnuzzled rather than bitten (Pellis and Pellis,1987).

The difference between the two forms offighting in rats therefore is not a quantitativedifference, as was debated earlier, but ratheris a qualitative one. Indeed, although a playfight may occasionally escalate to seriousfighting, when it does, the partners stop com-peting for nuzzling contact of the nape andswitch to biting attacks to their partner's rump(Takahashi and Lore, 1983; Pellis and Pellis,1991). The similarities in postures betweenplay fighting and serious fighting reflect thatthere is a standard species-typical set of tacticsfor attack and defense, which may be used incompetitive interactions not only in conspe-cific aggression and play but also in sexual andpredatory encounters (Pellis, 1988).

Analysis of those species-typical posturesor behavior patterns reveals that they aremodified in play fighting compared with seri-ous fighting to accommodate the differencesin targets (Pellis and Pellis, 1987). These find-ings support the view that play fighting is adistinct behavioral system, and not merely im-mature aggression. However, the findings for

rats need to be reconciled with those from abroader comparative literature that show thatfor a wide range of mammals and for somebirds, play fighting in juveniles involves thesame targets of attack and defense as in adultserious fighting (Aldis, 1975). Why are ratsseemingly aberrant in this regard?

THE ORIGINS OF PLAYFIGHTING IN RATS

In rats, nuzzling of the nape can occur duringadult sexual encounters, and partners may usea variety of defensive tactics to avoid suchcontact (Pellis, 1988). Therefore, one possibil-ity is that the proponents of play fighting asimmature serious fighting may have beenright, except that they have used the wrongadult behavioral system for their comparison.That is, play fighting in rats may simply be im-mature sexual behavior. Indeed, comparingthe play behavior of rats with that of othermurid rodents confirms this possibility (Pellis,1993). During play fighting, murid rodents donot compete for access to agonistic targets butrather compete for those typically contactedduring the precopulatory phase of sexual en-counters. The targets competed for duringplay and sexual behavior can be very different

Page 320: 0195162854

Chapter 28. Play and Fighting

from those attacked during serious fighting.For example, in Djungarian hamsters, seriousfighting involves bites directed at the rumpand sides of the face, whereas duringcourtship and play fighting, it is the front ofthe mouth that is nuzzled (Pellis and Pellis,1989).

Earlier researchers likely missed the con-nection of play fighting to sexual behavior inrats and other murid rodents because they fo-cused on scoring behavior patterns, ratherthan on the targets around which those be-havior patterns are organized (Fagen, 1981;Hole and Einon, 1984). Furthermore, studentsof animal play have tended to label play as sex-ual only when the behavior patterns used bythe participants are those obviously derivedfrom sexual contexts, such as mounting(Mitchell, 1979; Fagen, 1981). Because juve-nile play fighting in murid rodents mimics theadult precopulatory phase of sexual behavior(Pellis, 1993), mounting is rare, and so thecommonality in targets would have had to beidentified to make the connection (Pellis,1988).

TACTICS USED IN PLAY FIGHTINGAND THEIR DEVELOPMENT

Play fighting in rats, from its earliest onset—in the days before weaning—involves napecontact (Pellis and Pellis, 1997). It is only inthe latter stages of the juvenile phase, as ratsapproach puberty, that they occasionallymount one another after nape contact (Pellisand Pellis, 1990). This gradual onset, frommore appetitive components to more con-summatory ones, supports the possibility thatjuvenile play fighting is an immature stage ofthe sexual motivational system. Comparativedata support this hypothesis.

Both montane and prairie voles competefor access to the nape during play fighting andduring precopulatory behavior (Pellis et al.,1989; Pierce et al., 1991). During sexual en-counters, female montane voles are more

303

likely to use upright defensive tactics to blocktheir partner's access to their own napes thanare female prairie voles; instead, females ofthis species are more likely to rotate to supineto avoid nape contact than are montane voles(Pierce et al., 1991). Correspondingly, duringjuvenile play fighting, prairie voles use thesupine defense more often than do montanevoles, whereas montane voles use the uprightdefense more often than do prairie voles (Pel-lis et al., 1989). That is, in the organization ofdefense in both of these vole species, playfighting resembles the species-specific patternof their adult sexual behavior (Pellis and Pel-lis, 1998). Rats, however, differ.

During adult precopulatory encounters,the most likely defensive response of a femalerat to a male's contact with her nape is for herto evade by leaping or running away or by lat-erally dodging away. Rotating around the lon-gitudinal axis of the body, thus moving thenape away while turning to face the male, isthe least likely response (<10% of cases) (Pel-lis and Iwaniuk, 2004). In marked contrast, injuvenile play fighting, only around 20% to30% of defense involves evasion, whereas 60%to 80% of cases involve rotating around thelongitudinal axis to supine or nearly supine(Pellis and Pellis, 1990). That is, in rats, playfighting inverts the pattern typical of adult sex-ual behavior (Pellis and Pellis, 1998). There-fore, unlike what may be the case for someother species of murid rodents, in rats, playfighting has been modified in that it is notsimply an immature version of adult sexualbehavior. That play fighting in juvenile ratsrepresents a novel behavioral system is sup-ported by two lines of evidence.

Rats, but not other species of murid ro-dents that do not have the rat-like modifiedpatterns of play fighting, experience emo-tional and cognitive deficits if deprived of playfighting in the juvenile phase (Einon et al.,1978). Furthermore, rats, but not other muridrodents, use play fighting in adulthood in non-sexual contexts (Pellis, 1993; 2002). The novelfunctions of play fighting in adult rats show

Page 321: 0195162854

304 DEVELOPMENT

that a pattern of play derived from sexualbehavior can be co-opted for use in quasi-aggressive situations. In turn, this shows thatthe link between play fighting and aggressionin rats that was identified by some earlier re-searchers may not have been entirely illusory(Sirverman, 1978; Taylor, 1978).

PLAY FIGHTING ASQUASI-AGGRESSION IN RATS

Play fighting in rats, as defined by competi-tion for snout-to-nape contact, persists wellbeyond puberty, into adulthood (Pellis andPellis, 1990). In the juvenile phase, the mostcommon response when contacted on thenape is for the rat to rotate around its longi-tudinal axis to a fully supine position. For fe-male rats, even as the frequency of play fight-ing wanes after puberty, the rotation to supineremains the most likely defensive response.The situation for male rats differs.

With the onset of puberty, males are morelikely to rotate only partially around the longi-tudinal axis of the body—thus keeping at leastone hindpaw in contact with the ground—thanthey are to rotate fully to supine. From thispartially rotated position, the defending malecan use its flank to push laterally against theattacker, or it can rear upright and face the at-tacker. All male rats undergo this change intheir preferred defensive strategy, but after pu-berty, they also modify their pattern of defensedepending on the identity of their play partner(Pellis and Pellis, 1992).

When attacked playfully by a female ora subordinate male, a male rat most likely usesthe partial rotation tactic. However, when at-tacked by a dominant male, the subordinatemale most likely uses the complete rotationtactic. Indeed, two subordinate males behavesymmetrically toward one another, with bothlikely to use the partial rotation tactic whenattacked by the other. However, if one be-comes dominant after the removal of the res-ident dominant male, they then behave asym-

metrically toward one another, in that thesubordinate male now rotates fully to supine(Pellis and Pellis, 1990,1991; Pellis et al., 1993).

From an analysis of this modulation ofplay fighting in rats, and with the absence ofsuch nonsexual uses of play in adulthood byother species of murid rodents, it appears thatamong adult male rats, play fighting is usedfor social assessment and manipulation (Pel-lis, 2002). Among familiars, subordinate malesinitiate more play with the dominant maleand, when contacted playfully by him, re-spond in a more juvenile fashion. That is, sub-ordinate males appear to use play for main-taining affiliative bonds with the dominantmale. In contrast, a subordinate male mayplay more roughly so as to challenge the sta-tus of the dominant male. Similarly, bothdominants and subordinates can use playfighting to challenge the status of an unfamil-iar male encountered in a neutral arena (Pel-lis and Pellis, 1991, 1992; Pellis et al., 1993;Smith et al., 1999).

CONCLUSION

Play fighting and serious fighting superficiallyresemble each other because both use manyof the same tactics with which to compete foraccess to bodily targets. In rats, serious andplayful fighting differ qualitatively as they in-volve competition over different targets, withthe playful targets having been derived fromsexual behavior. These findings suggest thatserious and playful fighting have differentcausal and functional properties. Surprisingly,the two may overlap in ways that do not seemto be the case for other murid rodents. Ratshave evolved novel control mechanisms overplay fighting, and this emancipation from itsregulatory origins has allowed play fighting tobe co-opted for nonsexual uses, one of whichis aggressive competition for enhanced socialstatus (Pellis, 2002; Pellis and Iwaniuk, 2004).

Interestingly, the physiological profile ofrats, in terms of their endocrinological re-

Page 322: 0195162854

Chapter 28. Play and Fighting 305

sponses when engaged in play fighting, is sim-ilar to that which is present when they areengaged in serious aggression, with the phys-iological distinction between the two becom-ing increasingly blurred with age (Hurst et al.,1996). Nonetheless, neonatal treatment withtestosterone propionate leads to an elevationof play fighting, but the changes in play do notaffect which rats become dominant, suggest-ing that different developmental processesregulate playful and serious fighting (Pellis etal., 1992). Therefore, although at some levelsof analysis playful and serious fighting are dis-tinct, at other levels, the two are similar. Itappears that in rats, then, there has been acomplex evolutionary relationship betweenserious and playful fighting. These two formsof fighting may have had separate origins, butsome of their current functions overlap, as dosome of the neurobehavioral mechanismsthat regulate them.

REFERENCES

Adams DB (1980) Motivational systems of agonistic be-havior in muroid rodents: A comparative reviewand neural model. Aggressive Behavior 6:295-346.

Aldis O (1975) Play fighting. New York: Academic Press.Alleva E (1993) Assessment of aggressive behavior in ro-

dents. In: Methods in neurosciences. Paradigms forthe study of behavior, Vol 14 (Conn PM, ed.), pp.111-137. New York: Academic Press.

Blanchard DC and Blanchard RJ (1990) The colonymodel of aggression and defense. In: Contemporaryissues in comparative psychology (Dewsbury DA,ed.), pp. 410-430. Sunderland, Mass.: Sinauer Asso-ciates, Inc.

Blanchard RJ and Blanchard DC (1994) Environmentaltargets and sensorimotor systems in aggression anddefence. In: Ethology and psychopharmacology(Cooper SJ and Hendrie CA, eds.), pp. 133-157).New York: John Wiley & Sons.

Blanchard RJ, Blanchard DC, Pank L, Fellows D (1985)Conspecific wounding in free ranging Rattus norvegi-cus. The Psychological Record 35:329-335.

Blanchard RJ, Blanchard DC, Takahashi T, Kelly MJ(1977) Attack and defensive behaviour in the albinorat. Animal Behaviour 5:622-634.

Boice R and Adams N (1983). Degrees of captivity and

aggressive behavior in domestic Norway rats. Bul-letin of the Psychonomic Society 21:149-152.

Cools AR (1985) Brain and behavior: hierarchy of feed-back systems and control of input. In: Perspectivesin ethology. Mechanisms, Vol 6 (Bateson PPG andKlopfer PH, eds.), pp. 109-168. New York: PlenumPress.

Einon DF, Morgan MJ, Kibbler CC (1978) Brief periodsof socialization and later behavior in the rat. De-velopmental Psychobiology 11:213-225.

Fagen R (1981) Animal play behavior. New York: Ox-ford University Press.

Foroud A and Pellis SM (2003) The development of"roughness" in the play fighting of rats: A Labanmovement analysis perspective. DevelopmentalPsychobiology 42:35-43.

Geist V (1971) Mountain sheep. Chicago: University ofChicago Press.

Geist V (1978) On weapons, combat and ecology. In:Advances in the study of communication and affect,Vol 4 (Krames LP, Pliner P, Aloway T, eds.), pp.1-30. New York: Plenum Press.

Grant EC and Macintosh JM (1966) A comparison ofsome of the social postures of some common labo-ratory rodents. Behaviour 21:246-259.

Hole GT and Einon DF (1984) Play in rodents. In: Playin animals and man (Smith PK, ed.), pp. 95-117. Ox-ford: Basil Blackwell.

Hurst JL, Barnard CJ, Hare R, Wheeldon EB, West CD(1996) Housing and welfare in laboratory rats:Time-budgeting and pathophysiology in single sexgroups. Animal Behaviour 52:335-360.

Kemble ED (1993) Resident-intruder paradigms for thestudy of rodent aggression. In: Methods in neuro-sciences. Paradigms for the study of behavior, Vol14 (Conn PM, ed.), pp. 138-150. New York: Aca-demic Press.

Kruk MR, van der Poel AM, de Vos-Frerichs TP (1979)The induction of aggressive behavior by electricalstimulation in the hypothalamus of male rats. Be-haviour 70:292-322.

Mitchell G (1979) Behavioral sex differences in nonhumanprimates. New York: Van Nostrand Reinhold Co.

Panksepp J (1981) The ontogeny of play in rats. Devel-opmental Psychobiology 14:327-332.

Pellis SM (1988) Agonistic versus amicable targets of at-tack and defense: Consequences for the origin, func-tion and descriptive classification of play-fighting.Aggressive Behavior 14:85-104.

Pellis SM (1989) Fighting: The problem of selecting ap-propriate behavior patterns. In: Ethoexperimentalapproaches to the study of behavior (Blanchard RJ,Brain PF, Blanchard DC, Parmigiani S, eds.), pp.361-374. Dordrecht, the Netherlands: Kluwer Aca-demic Publishers.

Page 323: 0195162854

306 DEVELOPMENT

Pellis SM (1993) Sex and the evolution of play fighting:A review and model based on the behavior ofmuroid rodents. Play Theory and Research 1:55-75.

Pellis SM (1997) Targets and tactics: The analysis of mo-ment-to-moment decision making in animal com-bat. Aggressive Behavior 23:107-129.

Pellis SM (2002) Keeping in touch: Play fighting and so-cial knowledge. In: The cognitive animal: empiricaland theoretical perspectives on animal cognition(Bekoff M, Allen C, Burghardt GM, eds.), pp.421^27. Cambridge, Mass.: MIT Press.

Pellis SM and Iwaniuk AN (2004) Evolving a playfulbrain: A levels of control approach. InternationalJournal of Comparative Psychology 17:90-116.

Pellis SM and Pellis VC (1987) Play fighting differs fromserious fighting in both the target of attack and tac-tics of fighting in the laboratory rat Rattus norvegi-cus. Aggressive Behavior 13:227-242.

Pellis SM and Pellis VC (1988) Play-fighting in the Syr-ian golden hamster Mesocricetus auratus Waterhouseand its relationship to serious fighting during post-weaning development. Developmental Psychobiol-ogy 21:323-337.

Pellis SM and Pellis VC (1989) Targets of attack and de-fense in the play fighting by the Djungarian ham-ster Phodopus campbelli: Links to fighting and sex. Ag-gressive Behavior 15:217-234.

Pellis SM and Pellis VC (1990) Differential rates of at-tack, defense and counterattack during the devel-opmental decrease in play fighting by male and fe-male rats. Developmental Psychobiology 23:215-231.

Pellis SM and Pellis VC (1991) Role reversal changes dur-ing the ontogeny of play fighting in male rats: At-tack versus defense. Aggressive Behavior 17:179-189.

Pellis SM and Pellis VC (1992) Juvenilized play fightingin subordinate male rats. Aggressive Behavior 18:449-457.

Pellis SM and Pellis VC (1997) The pre-juvenile onset ofplay fighting in rats (Rattus norvegicus). Develop-mental Psychobiology 31:193-205.

Pellis SM and Pellis VC (1998) The play fighting of ratsin comparative perspective: A schema for neurobe-havioral analyses. Neuroscience and BiobehavioralReviews 23:87-101.

Pellis SM, Pellis VC, Dewsbury DA (1989) Different lev-els of complexity in the play fighting by muroid ro-dents appear to result from different levels of in-tensity of attack and defense. Aggressive Behavior15:297-310.

Pellis SM, Pellis VC, Kolb B (1992) Neonatal testosteroneaugmentation increases juvenile play fighting butdoes not influence the adult dominance relation-ships of male rats. Aggressive Behavior 18:437^447.

Pellis SM, Pellis VC, McKenna MM (1993) Some subor-dinates are more equal than others: Play fightingamongst adult subordinate male rats. Aggressive Be-havior 19:385-393.

Pellis SM, Pellis VC, Pierce JD Jr, Dewsbury DA (1992)Disentangling the contribution of the attacker fromthat of the defender in the differences in the in-traspecific fighting of two species of voles. Aggres-sive Behavior 18:425-435.

Pierce JD Jr, Pellis VC, Dewsbury DA, Pellis SM (1991) Tar-gets and tactics of agonistic and precopulatory behav-ior in montane and prairie voles: Their relationship tojuvenile play fighting. Aggressive Behavior 17:337-349.

Poole TB and Fish J (1975) An investigation of playfulbehaviour in Rattus norvegicus and Mus musculus(Mammalia). Journal of the Zoological Society, Lon-don 175:61-71.

Silverman P (1978) Animal behaviour in the laboratory.New York: Pica Press.

Smith LK, Fantella S-L, Pellis SM (1999) Playful defen-sive responses in adult male rats depend on the sta-tus of the unfamiliar opponent. Aggressive Behav-ior 25:141-152.

Takahashi LK and Lore RK (1983) Play fighting and thedevelopment of agonistic behavior in male and fe-male rats. Aggressive Behavior 9:217-227.

Taylor GT (1980) Fighting in juvenile rats and the on-togeny of agonistic behavior. Journal of Compara-tive and Physiological Psychology 94:953-961.

Page 324: 0195162854

Sex

WILLIAM J. JENKINS AND JILL B. BECKER29

The genes that an animal inherits set off a cas-cade of events that result in the sexual differ-entiation of the external genitalia and centralnervous system. Simply, genetic sex deter-mines gonadal sex and gonadal sex determinesphenotypic sex.1 During the late prenatal andearly postnatal development of the rat, expo-sure to testosterone produced by the matur-ing testes produces "organizational effects" onneuronal differentiation, growth, and survivalthat result in masculinization of the brain. Inthe absence of testosterone, the brain is fem-inized2 (Breedlove et al, 2002).

As an adult, the animal responds to itsown gonadal hormones (testosterone for themale, estradiol and progesterone for the fe-male) with the initiation of sex-typical behav-iors, including reproductive behaviors. Sexu-ally differentiated brain structures generatesexually dimorphic copulatory behaviors as afunction of "activational effects" of circulatinghormone levels of the adult. For example,male rats that are castrated neonatally displaylimited masculine copulatory behaviors andexhibit lordosis if treated with estradiol andprogesterone in adulthood. Furthermore, fe-male rats that are treated with testosterone orestradiol during early development show in-creased male-typical and decreased female-typical sexual responses (Breedlove et al.,2002).

To study sexual behavior in the rat, it isnecessary to understand the endocrinology ofthe reproductive system of both females andmales. Hormones produced by the gonadsfeed back to the brain to stimulate reproduc-

tive behavior. In this chapter, we provide anoverview of the reproductive systems and theneural systems that mediate sexual behaviorand discuss how to study sexual behavior inrats. We address these issues in females firstand in males next.

SEXUAL BEHAVIOR INTHE FEMALE RAT

Sexual behavior in the female rat is a functionof a complex interplay of hormones and en-vironmental circumstances. Female rats arenonseasonal, spontaneous ovulators that dis-play a 4 to 5 day estrous cycle (Fig. 29-1) com-posed of four basic stages: diestrus 1 (ormetestrus), diestrus 2 (or diestrus), proestrus,and estrus (when the female rat is sexually re-ceptive). This is one of the most rapid ovar-ian cycles among mammals, and it is madepossible by truncating the cycle immediatelyafter ovulation.

THE ESTROUS CYCLE

Follicular PhaseThe ovarian cycle begins with the develop-ment of follicles from oocytes in the ovary.During the follicular phase, the hypothalamusproduces a pulsatile release of gonadotropin-releasing hormone (GnRH) that stimulatesthe release of two primary gonadotropinsfrom the anterior pituitary: follicle-stimulatinghormone (FSH) and luteinizing hormone (LH).These gonadotropins stimulate the growth of

307

Page 325: 0195162854

308 DEVELOPMENT

Figure 29—1. Plasma concentrations of hormones acrossthe female rat estrous cycle. LH, luteinizing hormone; FSH,follicle-stimulating hormone. (Adapted from McCarthy andBecker, 2002.)

follicles and estradiol synthesis in the ovariangranulosa cells. There also is increased steroido-genesis in the ovary, and estradiol secretionincreases gradually during this phase. In therat, this stage is 2 days long. The first day iscalled diestrus 1 or metestrus, and the secondday is called diestrus 2 or just diestms.

Periovulatory PeriodThe time just before and after ovulation is dy-namic. Estradiol reaches its peak concentra-tions of 100 to 150 pg/ml around noon on theday of proestrus approximately 6 to 12 hoursbefore ovulation. The peak of estradiol trig-gers a surge of GnRH from the hypothalamusand the release of prolactin from the anteriorpituitary that induces LH and FSH release toreach their surge levels. Eight to 10 hourslater, the mature follicle releases the ovumand transforms into the corpus luteum. Pro-gesterone also increases a few hours beforeovulation about 4 to 6 hours after the estra-diol surge, during the afternoon of proestrus(Freeman, 1994).

Estrus is the period of sexual receptivityand the day of ovulation. Onset of sexual re-ceptivity occurs shortly after the start of thedark phase of the light-dark cycle and pre-cedes ovulation by a few hours in most rats.Ovulation, induced by the LH surge onproestrus, occurs 4 to 6 hours after nightfall,

and sexual receptivity persists for 12 to 20hours (depending on whether the femalemates). Note that behavioral receptivity oc-curs 36 to 48 hours after estradiol begins to in-crease and 4 to 6 hours after the increase inprogesterone. Baseline serum concentrationsof estradiol at "vaginal estrus" or behavioralestrus are approximately 3 to 12 pg/ml (Mc-Carthy et al., 2002).

In many female mammals, ovulation isfollowed by an additional phase called theluteal phase, which is maintained by hormonesproduced by the corpus luteum. In rats, thecorpus luteum becomes functional only if thefemale engages in sexual behaviors that acti-vate a progestational reflex (twice-daily surgesof prolactin). These surges of prolactin main-tain the corpus luteum so that progesteroneis released and implantation can occur if theeggs are fertilized. The corpus luteum is main-tained for approximately 12 days, if the rat ispregnant, the placenta then assumes respon-sibility for the maintenance of progesteronesecretion. In the event of an unfertile mating,the female rat may exhibit "pseudopreg-nancy," a 12-day period of anestrus inducedby the secretion of progesterone from the cor-pus luteum (Smith et al., 1975; Gunnet et al.,1983).

Stimulation of the vaginocervical area bythe male or by the experimenter can inducepseudopregnancy. This is mentioned, becauseto determine where a female rat is in her ovar-ian cycle the investigator must obtain cellsfrom the vaginal epithelium. If one is tooforceful during this procedure, pseudopreg-nancy can be the result.

Determining Estrous CyclePhase by Vaginal LavageThe stages of the estrous cycle can be deter-mined by examining morphological changesin vaginal epithelial cells under light mi-croscopy (Figs. 29-2 and 29-3). This is usuallydone by flushing the vagina with saline usingan eyedropper and 0.9% saline. The tip of theeyedropper is filled with a small amount (1 to

Page 326: 0195162854

Chapter 29. Sex 309

Figure 29-2. Vaginal cytology varies with the female ratestrous cycle. See Table 29-1 for description of cell types.

2 drops) of saline and then inserted into thevagina of the female rat. The vagina is flushedtwo or three times with the saline, or until thesaline becomes cloudy, then the fluid is placedonto a slide. The eyedropper is thoroughlyrinsed with distilled water, and the next rat issampled. For obtaining data from a largenumber of rats at a time, we use a piece ofPlexiglas about 4 X 8 cm onto which Silasticadhesive has been affixed to make barriersabout 1 X 1 cm. With this plate, samples from32 rats can be obtained at one time. The re-sulting samples (vaginal smear) are examined

Figure 29-3. Vaginal cytology as assessed by saline lavageunder light microscopy. A, Diestrus I; B, diestrus II; C,proestrus; D, estrus. See Table 29-1 for description of celltypes.

under a light microscope while the sample isstill wet (see later).

How to Hold the Rat for Vaginal SmearsThere are two ways to hold the rat to obtainthe sample. The ventral approach is used mostcommonly in rats that have been handled.The female is picked up and held in one handwith the belly exposed, using the little fingerto hold out one back leg. The experimenter'sother hand is used to insert the eyedropperinto the vagina (it is a good idea to hold thetail with the lower fingers of the hand, hold-ing the eyedropper to steady the hand andkeep the tail from flailing). The other ap-proach is a dorsal approach in which the fe-male is standing facing away from the exper-imenter and then her hindquarters are liftedslightly by grasping the base of the tail; theeyedropper is then inserted into the vagina un-der the tail.

The greatest concern about collectingvaginal lavage samples is to not be overly ag-gressive so that the animal does not becomepseudopregnant. It may seem that the dorsalapproach is easier for those not experiencedat handling rats, but the rats tend to strugglequite a lot with this approach, and becausethey have two feet on a surface, they canachieve quite a lot offeree if so motivated. Itis also not as easy to see how far the eye-dropper has been inserted. We generally havebetter success in obtaining good samples andat not inducing pseudopregnancy with theventral approach, even if the animal handlersare inexperienced.

How to Read the SlidesStaining the cells is not necessary. Examinethe liquid obtained under a XlO to X20 ob-jective with a blue or green filter. Interpretingthe stage of estrous cycle from the cell mor-phology requires practice and experience witha particular rat (Table 29-1; see Figs. 29-2 and29-3). Exactly what the cells will look like ata particular stage of the cycle varies with thetime of day (especially on proestrus) and the

Page 327: 0195162854

310 DEVELOPMENT

Table 29-1. Description of Estrous Cycle Changes in Vaginal Cell Morphology

Day of Cycle Cell Morphology in Smear

Diestrus 1(metaestrus)

Diestrus 2 (diestrus)

Proestrus

Estrus

Pseudopregnant

Small, round cells called leukocytes (occasionally during early metaestrus, cells can beseen that look like the nucleated large cells of proestrus. The only way to tell the differ-ence is in retrospect—the next day the animal has leukocytes and is not in estrus).

Small, round cells called leukocytes plus some large, round cells without nuclei are seen.

Proestrus is characterized by a predominance of nucleated epithelial cells, which are largeand round and have an easily visible nucleus. These appear individually and in dustersand can be interspersed with cornified squamous epithelial cells. Leukocytes are not pres-ent in the smear. To see the nuclei, the examiner must focus up and down; the cells looklike donuts. Some spindly cells may be seen as well.

Large, irregularly shaped cells are referred to as cornified cells because occasionally theyclump together and resemble an ear of corn in the husk. Cornified squamous epithelial celbare large and irregular, have no visible nucleus, and contain a granular cytoplasm. Theseappear individually and in very large clusters. Nucleated epithelial cells and leukocytesare not present on the smear.

Usually a combination of all of the above cell types is seen, or a smear looks like onefrom diestrus with irregularly shaped cells. This persists for about 12 days.

rat. Rather than trying to guess whether a ratis in a particular stage of the cycle, it is betterto describe the cell types that appear in thesmear. After 7 to 8 days, it is usually possibleto determine the stage of the cycle of a femaleand what the cytology looks like for eachstage. We collect data for at least two com-plete cycles before using a female rat in an es-trous cycle-timed experiment, because it cantake this long to be confident, based on thecytology, of the stage of the cycle of a rat. Inaddition, when we begin collecting estrous cy-cle data, we group together house rats that arein the same phase of the estrous cycle to takeadvantage of the effects of pheromones of es-trous cycle synchrony (Schank et al., 1992).

SEXUAL BEHAVIOR OF THE FEMALE RAT

Mating behavior occurs in the female rat dur-ing the period of sexual receptivity that beginsshortly after lights-out (for rats on a 12 hourlight-dark cycle) on the evening of proestrusthrough the early morning of estrus (Barnett,1975). Copulation in rats consists of three pri-mary interactions between the male and fe-male. These are defined by the behavior of themale, which consists of mounting, intromission,

and ejaculation. These events can be reliablydistinguished under experimental conditionsas a function of various characteristics andqualities associated with each (see later). Themale mounts the female from the rear, clasp-ing and palpating her flanks with his forepaws.When the female is sexually receptive, thismounting results in the female assuming a re-flexive posture known as lordosis. When in lor-dosis, the female arches her back and dorsi-flexes her tail. If the mount is accompanied byinsertion of the penis into the vagina, it is re-ferred to as an intromission. The mating se-quence is characterized by repeated intromis-sions before the male ejaculates.

Lordosis and Preceptive BehaviorsLordosis is usually quantified by determiningthe lordosis quotient (LQ), which is the per-centage of time that the female exhibits lor-dosis when the male mounts. The LQ is cal-culated by dividing the number of times thefemale displays lordosis by the total number ofmale contacts, multiplied by 100%. The LQ isthe most common measure of sexual recep-tivity used for female rats. Some investigatorsgive a qualitative rating to the intensity of lor-dosis, but it is more common to use the LQ.

Page 328: 0195162854

Chapter 29. Sex 311

To determine the LQ, it is necessary tocarefully observe the female rat's posture.This is most easily accomplished by video-recording the behavioral interaction, as this al-lows for a given sequence to be observed mul-tiple times and at slow motion if necessary.When a female rat assumes the lordotic pos-ture, she arches her back and often elevatesher head. Lordosis is accompanied by tempo-rary immobility that sometimes persists evenafter the mount, intromission, or ejaculationis over. Lordosis can also be elicited by an ex-perimenter who grasps the female lightly withthe index finger and thumb along the lateralregion of the hindquarters. Sexually receptivefemales will have an LQ that approaches100%; however, researchers often opera-tionally define sexual receptivity as an LQ thatis greater than 50%.

Lordosis may be reflexive in nature, butthe female is not a passive participant in thecopulatory bout. Female rats engage in a se-ries of behaviors that include approaching themale, presenting, hopping, darting, and earwiggling to solicit sexual behavior from themale. These proceptive, or solicitation, be-haviors may represent an index of female sex-ual initiative or motivation, and during thesebehaviors the female emits ultrasonic vocal-izations that are thought to attract the male(Beach, 1976; White et al., 1993). These be-haviors are observed under laboratory testingconditions, although their display depends onthe context in which copulation occurs.Specifically, the behaviors described earlieroccur when the rats mate in pairs in relativelysmall test cages (Erskine, 1989).

Under more natural circumstances, inwhich rats mate in groups in larger, morecomplex arenas, the female displays a differ-ent set of solicitational behaviors, includingapproach to, orientation to, and run awayfrom the male. Because of the nature of thesebehaviors, they are observed under conditionsin which there is sufficient physical space forthem to occur. To quantify these behaviors,the incidence of each behavior (ear wiggling,

hopping, and darting) is recorded from visualobservation or videorecording of a sexual in-teraction (Erskine, 1989).

Pacing BehaviorUnder seminatural conditions, the female isable to use solicitational behaviors sequen-tially; under these circumstances, the femalecontrols or "paces" the rate of copulation.This is important for the female rat becausethe interval between intromissions is criticalfor triggering the neuroendocrine reflex thatresults in the progestational state necessaryfor pregnancy. Males and females differ intheir optimal latencies between intromissionsduring copulation. Males prefer regular, rapidintromissions that lead to a fairly quick ejacu-lation. Females, on the other hand, requirelonger intervals between intromissions to op-timize vaginocervical stimulation receivedduring the copulatory bout (Erskine, 1989).

In the wild, both males and femalesachieve their preferred rate of copulation byengaging in group mating. In the laboratory,females can also pace the rate of copulation ifsex occurs in an environment where the fe-male rat both approaches and avoids the male.In our laboratory, the female rat paces in atesting arena that is divided into two cham-bers with an opaque barrier (Fig. 29-4). Themale is physically tethered to the larger por-tion of the arena so that he has complete mo-bility up to the barrier but he cannot cross thebarrier. The female, on the other hand, hasfull access to both sides of the chamber. Itshould be noted, however, that if a tetheredmale is used during a test of sexual behavior,it is important that the male is trained withthe tether before experimental data are col-lected. Often, when a male is placed in a tetherfor the first time, he is more interested in try-ing to get out of the tether than he is in theestrous female!

Observations of paced mating behaviorhave led researchers to focus on two compo-nents of paced copulatory behavior that mayreflect two different aspects of the interaction:

Page 329: 0195162854

312 DEVELOPMENT

number of exits after a mount (for example)is divided by the number of mounts (X100%)to obtain percent exits after mounts. After afemale exits the male side following a givencoital stimulus, we also monitor the numberof seconds that the female stays away fromthe male before crossing over to the male.Add the total time that the female avoids themale and divide by the total number of exitsfor a given coital stimulus to obtain the meanreturn latency for mounts, intromissions, andejaculations. Again, when a female rat is pac-ing, both percent exits and return latenciesincrease with the intensity of the coitalstimuli.

Figure 29-4. Schematic of pacing apparatus. The appara-tus is made of clear Plexiglas that measures 61 cm long X30 cm wide X 46 cm high. The barrier is placed approxi-mately one third of the length of the chamber away fromthe wall. It is made of an opaque Plexiglas and measures 20cm long X 1 cm wide X 25 cm high. The female rat hasfree access to either side of the chamber, while the male isresticted through the use of a flexible tether and harness tothe larger side of the chamber. For additional details seeXiao and Becker (1997).

(1) percent exits (number of times a femaleleaves the male after a given coital stimu-lus/total number of coital stimuli X 100%)and (2) return latency (amount of time femaleavoids male after a given coital stimuli). Bothof these measures increase as sexual stimulibecome more intense (mount < intromis-sion < ejaculation). Therefore, pacing de-pends on the ability of the female to discrim-inate between sexual stimuli and to performappropriate motoric behavioral responses as aresult.

Two measures of pacing are determinedfrom behavioral observations. Percent exitsare calculated from the number of times thatthe female rat exits the male side of thechamber after receiving a mount, intromis-sion, or ejaculation. If using an apparatus likethat shown in Figure 29-4, an exit is deter-mined by the female completely crossing theopaque barrier to avoid the male rat. The

IMPORTANCE OF PACING BEHAVIOR

As mentioned, the female rat requiresvaginocervical stimulation to initiate the neu-roendocrine reflex that maintains the imma-ture corpora lutea. It is not simply a matter ofa minimum amount of vaginocervical stimu-lation to trigger this event, however; rather,the pattern in which the vaginocervical stim-ulation is received is critical for the female rat.Female rats exhibit high rates of pseudopreg-nancy with as few as five intromissions whenthey were allowed to pace the rate of copula-tion compared with females that receive fivenonpaced intromissions. This discrepancy de-creases as the number of intromissions re-ceived increases, nevertheless, females pacingthe rate of intromissions always show higherrates of pseudopregnancy (Gilman et al.,1979). Furthermore, when female rats pacethe rate of intromissions, behavioral estrus istruncated compared with those rats that havenonpaced sex (Erskine et al., 1982). These re-sults indicate that although pacing the rate ofcopulation is not necessary for the inductionof pregnancy or pseudopregnancy, the femalecan optimize the vaginocervical stimulationreceived under circumstances in which she ac-tively controls the rate of copulation and thatdoing so results in physiologically relevantconsequences.

Page 330: 0195162854

Chapter 29. Sex 313

NEURAL SUBSTRATES OF

RECEPTIVE, PROCEPTIVE, AND

PACED SEXUAL BEHAVIOR

Female sexual behavior in the rat is contin-gent on the presence of the ovarian hormonesestradiol and progesterone. Mating behaviorin the female is disrupted by ovariectomy andis reestablished by estradiol and progesteroneacting sequentially, These hormones activatethe neural systems necessary for these behav-iors by binding to intracellular receptors thatactivate genomic processes that result in theinduction of specific proteins (Pfaff et al.,1994).

The importance of estradiol and proges-terone for sexual receptivity has been demon-strated in experiments with ovariectomizedrats to remove the endogenous source ofovarian hormones. If no hormones are given,an ovariectomized female will not be recep-tive and will reject attempts by a male rat tomate with her. If an ovariectomized female istreated with estradiol alone, after about 6 to10 days she will exhibit sexual receptivity, butthe behavior does not appear normal. If pro-gesterone alone is given, sexual behavior isnever induced even after weeks of proges-terone treatment.

On the other hand, when estradiol is ad-ministered about 48 hours before proges-terone, as occurs during the estrous cycle, nor-mal-appearing sexual receptivity results. Thisis because estradiol primes the brain to be sen-sitive to progesterone by inducing the syn-thesis of progesterone receptors in the hypo-thalamus. Progesterone then acts on thesereceptors to induce sexual behavior (Mc-Carthy et al., 2002).

The hypothalamus is where neuroen-docrine information is integrated to initiate fe-male receptive behavior. The ventromedialhypothalamus (VMH) and the medial preop-tic area of the hypothalamus (MPOA) are twosites believed to be critical to the display ofsexual behavior in the female rat. Copulatorybehavior induces consistent and reliable in-

creases in the immediate-early gene c-fos inthese two structures (Erskine, 1993; Pfaus etal., 1993; Coolen et al., 1996).

The VMH is crucial to the display of lor-dosis in the female rat. Lesions to the VMHdisrupt sexual receptivity. Electrical stimula-tion of the VMH facilitates the lordosis re-sponse as a function of hormone treatment.Direct estradiol implants into the VMH in-duce sexual receptivity and facilitate sexual re-sponses in ovariectomized female rats in con-junction with systemic administration ofprogesterone. Sequential implants of estradioland progesterone into VMH induce sexual be-havior in female rats that is comparable to lev-els observed by systemic treatment with estra-diol and progesterone (Pfaff et al., 1994;McCarthy et al., 2002).

Implants of an estradiol antagonist intothe VMH block the induction of sexual re-ceptivity by systemically administered hor-mones, and damage to the VMH that blockssexual receptivity also reduces the amount oftime that a female spends near a tetheredmale. Similarly, progesterone implants intothe VMH elicit hopping, darting, and ear wig-gling from female rats. The efferents project-ing from the VMH to various midbrain sitesappear to be especially important for the dis-play of lordosis, however, although lesions tothese areas have no consistent effect on solic-itational behaviors displayed by the female(Erskine, 1989; PfafF et al., 1994; McCarthy etal., 2002).

If the VMH is involved in the facilitationof lordosis, it appears that the MPOA nor-mally inhibits this reflex. Lesions to theMPOA reduce the amount of estradiol re-quired to induce lordosis and lead to higherlordosis quotients as a function of hormonalpriming. In contrast, electrical stimulation ofthe MPOA disrupts lordosis. This role of theMPOA is not without controversy, however,as different effects of MPOA lesions have beenobserved as a function of testing conditions(Erskine, 1989; Pfaff et al., 1994; McCarthy etal., 2002).

Page 331: 0195162854

314 DEVELOPMENT

There is also evidence that the MPOA isinvolved in the display of preceptive behav-iors. Both nonspecific and axon-sparing le-sions of the MPOA diminish preceptive be-haviors (Whitney, 1986; Hoshina et al., 1994).When the female rat is tested under condi-tions that allow the female to escape the male,lesions to the MPOA result in the femalespending less time with the male and receiv-ing fewer intromissions. Lesions to the MPOAare also associated with an increase in the rateat which females leave a male after a givensexual interaction, as well as the amount oftime they stay away from the male after thisinteraction. These behavioral effects are ob-served without apparent effect on the LQs ofthese females (Yang et al., 2000). The MPOAmay also be involved in encoding vaginocer-vical stimulation, and lesions of the MPOAcause pseudopregnancy in the rat. Togetherthese studies suggest that lordosis is under thecontrol of both the VMH and MPOA, whichact in opposition to one another, whereas pro-ceptive behaviors are associated with both(Gunnet et al., 1983; Haskins et al., 1983).

The neural substrates of pacing behaviorhave received less attention; however, the dis-play of this behavior depends on the ability ofthe female rat to interpret the vaginocervicalstimulation received during the copulatorybout. Vaginocervical stimulation is relayedthrough the pelvic and pudenal nerves, andtranssection of these nerves disrupts the female'sability to pace copulation (Erskine, 1992) with-out affecting the display of sexual receptivity orproceptivity (Rowe et al., 1993). Vaginocervicalstimulation is associated with increased meta-bolic activity and Fos-immunoreactivity in theMPOA, mesencephalic reticular formation, bednucleus of the stria terminalis dorsal raphe, andmedial amygdala (Erskine, 1993; Pfaus et al.,1993; Rowe et al., 1993; Wersinger et al., 1993;Polston et al., 1995). The medial amygdala ap-pears to be critical for processing vaginocervicalstimulation and relaying information to hypo-thalamic centers that mediate neuroendocrinefunction (Polston et al., 2001a, 2001b).

Pacing behavior relies on the female'sability to interpret coital stimuli, engage in anappropriate motoric response, and temporallysequences contacts with the male. Given thiscomplex interplay among sensorimotor func-tion and motivation, the striatum and nucleusaccumbens (NAcc) have been investigated fortheir role in pacing behavior (Becker, 1999).

Extracellular dopamine concentrationsincrease in both striatum and NAcc in femalerats that are pacing the rate of copulation com-pared with those having nonpaced sex (Mer-melstein et al., 1995). In females that are in-duced into sexual receptivity with sequentialimplants of estradiol and progesterone intothe VMH, estradiol in the striatum enhancespercent exits and estradiol in the NAcc in-creases return latencies (Xiao et al., 1997). Le-sions of the striatum affect the percent exitmeasure after ejaculations (Jenkins et al.,2001). Lesions of the NAcc increase the femalerat's rate of rejection of a male attempting tomount, without affecting LQs exhibited bythese females and in some instances result ina complete avoidance of the male (Rivas et al.,1990, 1991; Jenkins et al., 2001). Previous sex-ual experience results in an enhanced increasein dopamine concentrations in the NAcc of fe-male hamsters (Kohlert et al., 1999). Thesestudies provide evidence that the striatum andNAcc serve important roles as neuroanatom-ical substrates of paced copulatory behavior inthe female rat.

SEXUAL MOTIVATION IN THE FEMALE RAT

Female sexual behavior has traditionally beenthought of in terms of receptive versus pro-ceptive behaviors. We have stressed that thefemale is not a passive participant in copula-tory behavior. It is only recently, however,that researchers have begun to look at sexualmotivation in females (despite the fact that theword preceptive implies "female initiative" toengage in sexual behavior [Beach, 1976]). Thisdiscrepancy is probably due in large part tothe fact that female rats who have received

Page 332: 0195162854

Chapter 29. Sex 315

hormone priming require stimulation from amale to exhibit lordosis. As such, they haveoften been viewed as victims of their hor-monal status and environmental circum-stance, rather than as initiators of the sexualinteraction.

Nevertheless, female rats in estrus willcross an electric grid to gain access to an in-tact male (Meyerson et al., 1973). They alsoshow a preferences for odors of sexually ac-tive males to those of estrous females (Bakkeret al., 1996), and they prefer to spend timewith intact, sexually active males (Pfeifle et al.,1983). Female rats even perform operant re-sponses to gain access to sexually active males,develop conditioned place preferences tochambers where they have had sex versuschambers where they were alone, and de-velop conditioned place preferences for pacedsex versus nonpaced sex (French et al., 1972;Oldenburger et al., 1992; Matthews et al.,1997; Paredes et al., 1999; Jenkins et al., 2003).Under traditional testing conditions, how-ever, the female's full display of preceptive be-havior is not observed. In fact, the view of thefemale as a passive participant in the copula-tory bout may be more a function of testingcontext than reflective of female sexual be-havior. Obviously, experiments that allow thefemale to pace the rate of copulation have pro-vided new insight into female sexual behaviorand motivation.

SEXUAL BEHAVIOR INTHE MALE RAT

Although the female shows sexual receptivityonly during behavioral estrus, the male willmate at any time that he comes into contactwith a receptive female. Typically, a male willperform 8 to 10 intromissions before achiev-ing an ejaculation, which involves the depo-sition of a sperm plug into the female's vagina.During the mating sequence, the male emitsultrasonic vocalizations. There are two calls:(1) with the initial contacts, the male emits the

mating call, and (2) just before ejaculation, heemits the preejaadatory call. These calls pro-mote immobility of the female during lordo-sis and in the wild may serve to attract femalesto the male (McClintock et al., 1982; White etal., 1990). After ejaculation, there is a post-ejaculatory call that is thought to serve to keepthe female away from the male for a while(Anisko et al., 1978). This is important becauseafter an ejaculation, the male enters a quies-cent period of inactivity known as the post-ejaculatory refractory period. This can last from2 to 5 minutes, at which point the male usu-ally reinitiates sexual behavior with the fe-male. The pattern repeats itself until the malebecomes sexually sated (Bermant, 1967; Adler,1969). As a side note, sexually sated males of-ten reinitiate copulatory behavior if presentedwith a novel receptive female, a phenomenonhumorously referred to as the "Coolidge ef-fect" (Bermant et al., 1968).

HOW TO MEASURE MALESEXUAL BEHAVIOR

Having to differentiate among mounts, intro-missions, and ejaculations can seem like aguessing game. After closely observing thesexual interactions between the male and fe-male rat, however, researchers can quicklyand reliably distinguish each. Again, it is agood idea to videorecord sessions to score sex-ual behavior. The key to determining whatthe male has just done is to observe his be-havior during the dismount. After a mount,male rats typically dismount the female gen-tly, simply moving off of the female. They of-ten groom themselves and remount the fe-male quickly. Intromissions, on the otherhand, are a little flashier. After intromitting,the male generally springs off of the female.Just before the dismount, there is a sharppelvic thrust, and the male often flails hisforepaws out to the side during the thrust andjust before the dismount. Grooming of thegenital region often follows intromissions,and sometimes the erect penis can still be seen

Page 333: 0195162854

316 DEVELOPMENT

outside the body cavity immediately after theintromission. To the novice, differentiatingbetween mounts and intromissions is oftenthe most difficult part of scoring sexual be-havior in the male rat.

The ejaculation is often the easiest sexualinteraction to determine early on. The ejacu-lation involves several pronounced pelvicthrusts. Instead of dismounting right away,males typically continue to grasp the female af-ter achieving an ejaculation and the female of-ten has to kick him away. There is marked de-crease in muscle tension in the male rat afteran ejaculation, which can be described as a"melting away" of tension to people new to ob-serving sexual behavior in the rat. Ejaculationsare generally followed by prolonged groomingbouts; the male rat often avoids the female andsometimes takes a little nap. Once the male re-covers from this "refractory period," he oftenreinitiates sexual behavior; however, the num-ber of intromissions that precede an ejaculationcan be decreased dramatically as the durationof the copulatory bout increases.

In addition to the type of interaction (i.e.,mount, intromission, ejaculation) in whichthe male engages, researchers often collect in-formation on the latency to first mount, in-tromission, and ejaculation. Information iscollected on the number of intromissions pre-ceding an ejaculation, and noncontact erec-tions are often used as an index of the male'smotivation to engage in sexual behavior.

NEURAL SUBSTRATES OF SEXUALBEHAVIOR IN THE MALE:MOTIVATION VERSUS ABILITY

Many of the same structures implicated in fe-male sexual behavior are also important for thesexual behavior in males. For example, theMPOA is critical for male sexual behavior. Le-sions to the MPOA disrupt the male's ability toperform copulatory behaviors (Heimer et al.,1966, 1982), although there is some evidencethat such disruption is most likely in those ratsthat are sexually naive (Dejonge et al., 1989).

Furthermore, the neuroanatomical sub-strates of the ability to engage in sexual be-havior are dissociable from those involved inthe motivation to engage in sexual behavior.This dissociation was demonstrated by an in-genious experiment conducted by Everitt andcolleagues (1990). In this experiment, malerats were trained to bar press for a sexually re-ceptive female. Once the behavior was firmlyestablished, males received lesions to eitherthe MPOA or basolateral amygdala, and themales were placed into the operant chamber.Males receiving lesions to the MPOA readilybar pressed for receptive female rats. Once thefemale was presented, however, the malefailed to copulate with her. On the other hand,males receiving lesions to the amygdala didnot bar press as much for the female, but ifthe female was presented to them, they wouldcopulate. Clearly, the male's motivation to en-gage in sexual behavior with the female is dis-sociable from his ability to engage in sexualbehavior with the female (Everitt, 1990).

As in the female, it appears that the MPOAserves more than a consummately role in sex-ual behavior. Work from Elaine Hull's labora-tory suggests that the dopamine systems of theMPOA are also associated with anticipatory as-pects of sexual behavior (Hull et al., 1999). Adopamine agonist infused into the MPOA fa-cilitates male sexual behavior (Hull et al., 1986),and a dopamine antagonist infused into MPOAreduces the number of ejaculations achieved bythe male (Pehek et al., 1988). Treatment withdopamine antagonists also reduces the numberof intromissions that precede ejaculation anddelays the initiation of copulation (Pehek et al.,1988; Pfaus et al., 1989). MPOA lesions in themale rat can also alter partner preferences, cer-tainly suggesting a role for MPOA in the mo-tivation to engage in sexual behavior (Paredeset al., 1998).

SEXUAL MOTIVATION IN THE MALE RAT

The striatum and NAcc are also involved inmale sexual behavior. As in the female rat, the

Page 334: 0195162854

Chapter 29. Sex 317

NAcc is implicated in the motivation to engagein sexual behavior. In experiments from theEveritt laboratory, animals with basolateralamygdala lesions that had ceased bar pressingfor access to females reinitiated bar pressingwhen amphetamine was infused into the NAcc,indicating that increased dopamine in the NAccis important for sexual motivation in the maleas well as the female (Everitt, 1990). Extracel-lular concentrations of dopamine increase inthe NAcc when male rats are presented with areceptive female rats behind a wire screen andthen again during copulation itself. Increases inextracellular concentrations of dopamine in thestriatum, however, are observed only duringcopulatory behavior (Pfaus et al., 1990), and thereinitiation of sexual behavior when a new fe-male is presented (i.e., the Coolidge effect) isassociated with increases in NAcc dopamine(Fiorino et al., 1997). Certainly, other neuralstructures are involved in the sexual behaviorof the male rat, but the MPOA, amygdala, andNAcc have received an enormous amount ofattention from researchers interested in under-standing the neural substrates of male sexualbehavior, and motivation (for a discussion ofother neural structures and neurotransmittersinvolved in male sexual behavior, see Bitran etal., 1987; Coolen et al., 1998; Hull et al., 1999;Pfaus, 1999).

The male rat's motivation to engage insexual behavior has never been doubted. Var-ious measures have been used to demonstratethe male rat's motivation to engage in sexualbehavior. Males will perform operant behav-iors to gain access to sexually receptive fe-males (Everitt, 1990), exhibit noncontact erec-tions as a function of estrous female rat cues(e.g., Liu et al., 1998), and develop condi-tioned place preferences for areas in whichthey have engaged in sexual behavior (e.g.,Mehara et al., 1990). Researchers also use levelsearching (Mendelson et al., 1989) and latencyto initiate sexual contact as an index of themale rat's motivation to engage in sexual be-havior (Fiorino et al., 1999), although corre-lational and factor analyses indicate that an-

ticipation of and initiation of sexual behaviormay represent different conceptual mecha-nisms (Pfaus et al., 1990).

CONCLUDING REMARKS

Because the female assumes a reflexive pos-ture during copulation, researchers have as-sumed that the male rat not only initiates sex-ual contact but also controls the rate at whichthe copulation occurs (see Bermant, 1967).We now know that that the male is not theonly active participant in the copulatory boutprovided the testing arena allows for the fe-male to pace, but there remain some long-held conceptions of male sexual behavior andmotivation that have been challenged. For in-stance, although it is clear that female rats findonly sexual behavior that occurs at their pre-ferred interval rewarding (Jenkins et al., 2003),it has traditionally been thought that malesfind sexual behavior rewarding under any cir-cumstance provided they achieve an ejacula-tion. Recently, however, Martinez and Pare-des (2001) demonstrated that males, likefemales, develop conditioned place prefer-ences only for sex that occurs at their pre-ferred rate (Martinez et al., 2001). Therefore,the idea that males are motivated to engagein sex under any circumstances may have tobe modified. Further research and investiga-tion should clarify this issue.

We have seen that the male and femalerat share many commonalities in terms of theanatomical substrates of sexual behavior. An-other common feature in the sexual behaviorof males and females is the role that the stria-tum and NAcc play in mediating sexual be-havior and motivation.

NOTES

1. This concept has been the dominant theory inneuroendocrinology for more than 40 years. Recent ev-idence, however, indicates that there may also be con-

Page 335: 0195162854

318 DEVELOPMENT

tributions from the Y and/or X chromosomes that de-termine if there are sex differences in the brain (inde-pendent of the ~Sry gene that determines if an animalhas testes; Carruth LL, Reisert I, et al. (2002) Sex chro-mosome genes directly affect brain sexual differentia-tion. Nature Neuroscience 5:933-934; and Xu J, Bur-goyne P, et al. (2002) Sex differences in sex chromosomegene expression in mouse brain. Human MolecularGenetics 11:1409-1419.

2. There is also evidence that complete feminiza-tion requires exposure to low levels of estradiol duringdevelopment, so the story becomes more complicatedall the time (Fitch RH and Denenberg VH [1998] A rolefor ovarian hormones in sexual differentiation of thebrain. Behavioral and Brain Sciences 21:311-352).

REFERENCES

Adler NT (1969) Effects of the male's copulatory be-havior on successful pregnancy of the female rat.Journal of Comparative Physiology and Psychology69:613-622.

Anisko JJ, Suer SF, et al (1978) Relation between 22-kHzultrasonic signals and sociosexual behavior in rats.Journal of Comparative Physiology and Psychology92:821-829.

Bakker J, Van Ophemert J, et al (1996) Sexual differen-tiation of odor and partner preference in the rat.Physiology and Behavior 60:489-494.

Barnett SA (1975) Reproductive behavior. In: The rat: Astudy in behavior, p. 138. Chicago: The Universityof Chicago Press.

Beach FA (1976) Sexual attractivity, proceptivity, and re-ceptivity in female mammals. Hormones and Be-havior 7:105-138.

Becker JB (1999) Gender differences in and influences ofreproductive hormones on dopaminergic functionin striatum and nucleus accumbens. Pharmacology,Biochemistry, and Behavior 64:803-812.

Bermant G (1967) Copulation in rats. Psychology To-day. 1:52-60.

Bermant G, Lott D, et al (1968) Temporal characteris-tics of the Coolidge effect in male rat copulatory be-havior. Journal of Comparative and PhysiologicalPsychology 65:447-452.

Bitran D and Hull EM (1987) Pharmacological analysisof male rat sexual behavior. Neuroscience andBiobehavior Reviews 11:365-389.

Breedlove SM and Hampson E (2002) Sexual differenti-ation of the brain and behavior. In: Behavioral en-docrinology (Becker JB, Breedlove SM, Crews D,McCarthy MM, eds.), pp. 75-115. Cambridge,Mass.: MIT Press.

Carruth LL, Reisert I, et al (2002) Sex chromosomegenes directly affect brain sexual differentiation. Na-ture Neuroscience 5:933-934.

Coolen LM, Peters HJ, et al (1998) Anatomical interre-lationships of the medial preoptic area and otherbrain regions activated following male sexual be-havior: A combined for and tract-tracing study.Journal of Comparative Neurology 397:421-435.

Coolen LM, Peters HJPW, et al (1996) Fos immunore-activity in the rat brain following consummatory el-ements of sexual behavior: A sex comparison. BrainResearch 738:67-82.

Dejonge FH, Louwerse AL, et al (1989) Lesions of theSDN-POA inhibit sexual behavior of male Wistarrats. Brain Research Bulletin 23:483-492.

Erskine MS (1989) Solicitation behavior in the estrousfemale rat: A review. Hormones and Behavior23:473-502.

Erskine MS (1992) Pelvic and pudendal nerves influencethe display of paced mating behavior in response toestrogen and progesterone in the female rat. Be-havioral Neuroscience 106:690-697.

Erskine MS (1993) Mating-induced increases in FOS pro-tein in preoptic area and medial amygdala of cyclingfemale rats. Brain Research Bulletin 32:447-^51.

Erskine MS and Baum MJ (1982) Effects of paced coitalstimulation on termination of estrus and brain in-doleamine levels in female rats. Pharmacology, Bio-chemistry, and Behavior 17:857-861.

Everitt BJ (1990) Sexual motivation: A neural and be-havioural analysis of the mechanisms underlying ap-petitive and copulatory responses of male rats. Neu-roscience and Biobehavioral Reviews 14:217-232.

Fiorino DF, Coury A, et al (1997) Dynamic changes inthe nucleus accumbens dopamine efflux during thecoolidge effect in male rats. Journal of Neuroscience17:4849-^855.

Fiorino DF and Phillips AG (1999) Facilitation of sexualbehavior and enhanced dopamine efflux in the nu-cleus accumbens of male rats after D-amphetamine-induced behavioral sensitization. Journal of Neuro-science 19:456-463.

Fitch RH and Denenberg VH (1998) A role for ovarianhormones in sexual differentiation of the brain. Be-havioral and Brain Sciences 21:311.

Freeman ME (1994) The neuroendocrine control of theovarian cycle of the rat. In: The physiology of re-production, 2nd ed (Knobil E and Neill JD, eds.).New York: Raven Press, Ltd.

French D, Fitzpatrick D, et al (1972) Operant investiga-tions of mating preference in female rats. Journal ofComparative Physiology and Psychology 81:226-232.

Gilman DP, Mercer LF, et al (1979) Influence of femalecopulatory behavior on the induction of pseudo-

Page 336: 0195162854

Chapter 29. Sex 319

pregnancy in the female rat. Physiology and Be-havior 22:675-678.

Gunnet JW and Freeman ME (1983) The mating-induced release of prolactin: A unique neuroen-docrine response. Endocrinology Reviews 4:44-61.

Hansen S, Kohler C, et al (1982) Effects of ibotenic acid-induced neuronal degeneration in the medialpreoptic area and the lateral hypothalamic area onsexual behavior in the male rat. Brain Research239:213-232.

Haskins JT and Moss RL (1983) Action of estrogen andmechanical vaginocervical stimulation on the mem-brane excitability of hypothalamic and midbrainneurons. Brain Research Bulletin 10:489-496.

Heirner L and Larsson K (1966) Impairment of matingbehavior in male rats foloowing lesions in the pre-optic-anterior hypothalamic continuum. Brain Re-search 3:248-263.

Hoshina Y, Takeo U, et al (1994) Axon-spring lesion ofthe preoptic area enhances receptivity and dimin-ishes proceptivity among components of female ratsexual behavior. Behavioural Brain Research 61:197-204.

Hull EM, Bitran D, et al (1986) Dopaminergic controlof male sex behavior in rats: Effects of an intra-cerebrally-infused agonist. Brain Research 370:73-81.

Hull EM, Lorrain DS, et al (1999) Hormone-neuro-transmitter interactions in the control of sexual be-havior. Behavioural Brain Research 105:105-116.

Jenkins WJ and Becker JB (2001) Role of the striatumand nucleus accumbens in paced copulatory behav-ior in the female rat. Behavioural Brain Research121:119-128.

Jenkins WJ and Becker JB (2003) Females devlop condi-tioned place preference for sex at their preferred in-terval. Hormones and Behavior 43:503-507.

Kohlert JG and Meisel RL (1999) Sexual experience sen-sitizes mating-related nucleus accumbens dopamineresponses of female Syrian hamsters. BehaviouralBrain Research 99:45-52.

Liu Y, Sachs BD, et al (1998) Sexual behavior in malerats after radiofrequency or dopamine-depleting le-sions in the nucleus accumbens. Pharmacology, Bio-chemistry, and Behavior 60:585-592.

Martinez I and Paredes RG (2001) Only self-paced mat-ing is rewarding in rats of both sexes. Hormonesand Behavior 40:510-517.

Matthews TJ, Grigore M, et al (1997) Sexual reinforce-ment in the female rat. Journal of ExperimentalAnalysis of Behavior 68:399-410.

McCarthy MM and Becker JB (2002) Neuroendocrinol-ogy of sexual behavior in the female. In: Behavioralendocrinology (Becker JB, Breedlove SM, Crews D,McCarthy MM, eds.), pp. 117-151. Cambridge,Mass.: MIT Press/Bradford Books.

McClintock MK, Anisko JJ, et al (1982) Group matingamong Norway rats. II. The social dynamics ofcopulation: Competition, cooperation, and matechoice. Animal Behavior 30:410-425.

Mehara BJ and Baum MJ (1990) Nalozone disrupts theexpression but not the acquisition by male rats of aconditioned place preference response for an oes-trous female. Psychopharmacology 101:118-125.

Mendelson SD and Pfaus JG (1989) Level searching: Anew assay of sexual motivation in the male rat.Physiology and Behavior 45:337-341.

Mermelstein PG and Becker JB (1995) Increased extra-cellular dopamine in the nucleus accumbens andstriatum of the female rat during paced copulatorybehavior. Behavioral Neuroscience 109:354-365.

Meyerson BJ and Lindstrom L (1973) Sexual motivationof in the female rat. Acta Physiologica Scandinavica(Supplement) 389:1-80.

Oldenburger WP, Everitt BJ, et al (1992) Conditionedplace preference induced by sexual interaction in fe-male rats. Hormones and Behavior 26:214-228.

Paredes RG, Tzschentke T, et al (1998) Lesions of themedial preoptic area anterior hypothalamus(MPOA/AH) modify partner preference in malerats. Brain Research 813:1-8.

Paredes RG and Vazquez B (1999) What do female ratslike about sex? Paced mating. Behavioural Brain Re-search 105:117-127.

Pehek EA, Warner RK, et al (1988) Microinjection of cis-flupenthixol, a dopamine antagonist, into the me-dial preoptic area impairs sexual behavior of malerats. Brain Research 443:70-76.

Pfaff DW, Schwartz-Giblin S, et al (1994) Cellular andmolecular mechanisms of female reproductive be-haviors. In: The physiology of reproduction, 2nd ed(E Knobil and JD Neill, eds.), pp. 107-220. NewYork: Raven Press.

Pfaus JG (1999) Neurobiology of sexual behavior. Cur-rent Opinions in Neurobiology 9:751-758.

Pfaus JG, Damsma G, et al (1990) Sexual behavior en-hances central dopamine transmission in the malerat. Brain Research 530:345-348.

Pfaus JG, Kleopoulos SP, et al (1993) Sexual stimulationactivates c-fos within estrogen concentrating re-gions of the female rat forebrain. Brain Research624:253-267.

Pfaus JG, Mendelson SD, et al (1990) A correlational andfactor analysis of anticipatory and consummatorymeasures of sexual behavior in the male rat. Psy-choneuroendocrinology 15:329-340.

Pfaus JG and Phillips AG (1989) Differential effects ofdopamine receptor antagonists on the sexual be-havior of male rats. Psychopharmacology 98:363-368.

Pfeifle JK and Edwards DA (1983) Midbrain lesions elim-

Page 337: 0195162854

320 DEVELOPMENT

inate sexual receptivity but spare sexual motivationin female rats. Physiology and Behavior 31:385-389.

Polston EK and Erskine MS (1995) Patterns of inductionof the immediate-early genes c-fos and egr-1 in thefemale rat brain following differential amounts ofmating stimulation. Neuroendocrinology 62:370-384.

Polston EK and Erskine MS (200 la) Excito toxic lesionsof the medial amygdala differentially disrupt pro-lactin secretory responses in cycling and mated fe-male rats. Journal of Neuroendocrinology 13:13-21.

Polston EK, Heitz M, et al (200lb) NMDA-mediated ac-tivation of the medial amygdala initiates a down-stream neuroendocrine memory responsible forpseudopregnancy in the female rat. Journal of Neu-roscience 21:4104-4110.

Rivas FJ and Mir D (1990) Effects of nucleus accumbenslesion on female rat sexual receptivity and procep-tivity in a partner preference paradigm. BehaviouralBrain Research 41:239-249.

Rivas FJ and Mir D (1991) Accumbens lesion in femalerats increases mount rejection without modifyinglordosis. Revista Espanola de Fisiologia 47:1-6.

Rowe DW and Erskine MS (1993) c-Fos proto-oncogeneactivity induced by mating in the preoptic area, hy-pothalamus and amygdala in the female rat: Role ofafferent input via the pelvic nerve. Brain Research621:25-34.

Schank JC and McClintock MK (1992) A coupled-oscilla-tor model of ovarian-cycle synchrony among femalerats. Journal of Theoretical Biology 157:317-362.

Smith MS, Freeman ME, et al (1975) The control ofprogesterone secretion during the estrous cycle andearly pseudopregnancy in the rat: Prolactin, gona-tropin and steroid levels associated with rescue ofthe corpus luteum of pseudopregnancy. Endocrinol-ogy 96:219-226.

Wersinger SR, Baum MJ, et al (1993) Mating-inducedFOS-like immunoreactivity in the rat forebrain: Asex comparison and a dimorphic effect of pelvicnerve transection. Journal of Neuroendocrinology5:557-568.

White NR, Cagiano R, et al (1990) Changes in matingvocalizations over the ejaculatory series in rats (Rat-tus norvegicus). Journal of Comparative Psychology104:255-262.

White NR, Gonzales RN, et al (1993) Do vocalizationsof the male rat elicit calling from the female? Be-havior and Neural Biology 59:76-78.

Whitney JF (1986) Effect of medial preoptic lesions onsexual behavior of female rats is determined by testsituation. Behavioral Neuroscience 100:230-235.

Xiao L and Becker JB (1997) Hormonal activation of thestriatum and the nucleus accumbens modulatespaced mating behavior in the female rat. Hormonesand Behavior 32:114-124.

Xu J, Burgoyne P, et al (2002) Sex differences in sexchromosome gene expression in mouse brain. Hu-man Molecular Genetics 11:1409-1419.

Yang LY and Clemens LG (2000) MPOA lesions affectfemale pacing of copulation in rats. Behavioral Neu-roscience 114:1191-1202.

Page 338: 0195162854

Environment

ROBBIN L GIBB

30

Desirable qualities in rats have been selectedfor through captive breeding. These qualitiesinclude tameness, curiosity, and reduced fearand aggression in response to handling (Bar-nett, 1975). Genetic selection may have madelaboratory rats well suited to the laboratory,but the question arises of what exactly arelaboratory conditions. For example, within acolony, standard housing may range from sin-gle housing in Plexiglas shoeboxes to grouphousing. Across laboratories, variation may beeven greater. I discuss how environmental pa-rameters influence the physiology and behav-ior of rats and thus the experimental outcome.This review considers lighting, humidity, air-flow, noise, cage construction, diet, and socialopportunity. Maternal influences, age at wean-ing, animal care, exercise regimens, and en-richment are also discussed. These factors caninduce changes in brain weight, cerebral vas-cularization, adrenal size, and body weight.

HOUSING CONSIDERATIONS

The environment of the rat can be consideredat the macroenvironment and microenviron-ment levels. Macroenvironment refers to theambient conditions in the animal colony;these include lighting, temperature, humidity,airflow, and noise. Microenvironment refers tothe conditions within the cage. The materialused to construct the cage, the design and size,the type of lid and bedding, access to food,and social opportunity are variables that affectmicroenvironment. Cage design also influ-

ences the ventilation, lighting, temperature,and noise level that its occupants perceive. Forexample, hanging wire cages are well venti-lated and reduce animal contact with theirexcreta but do not allow the animal to mod-ify its microenvironment. Macroenvironmen-tal conditions have a greater impact on ani-mals housed in this type of cage. Shoeboxcages with bedding are less well ventilated butallow the animal to make alterations in its en-vironment. In addition, bedding affords theanimal opportunity to dig, a natural behaviorthat laboratory rats share with their wild pre-decessors (Canadian Council on Animal Care[CCAC], 1984). A study by Krohn et al. (2003)used telemetry to monitor heart rate, bodytemperature, and blood pressure of rats kepton three types of flooring: grid floor, plasticfloor, or bedding. Grid flooring caused eleva-tions in blood pressure and heart rate and thuswas rated most undesirable. Bedding wasmost acceptable. Anzaldo and colleagues(1994) allowed rats to inhabit cages that wereequipped with L-shaped partitions (high-perimeter housing), cages that allowed in-creased floor space with a three-dimensionaldesign, or standard cages. The three-dimen-sional cage was least preferred; the high-perimeter cage was most preferred. The pref-erence of the rats for the "high-perimeter"caging may have reflected their thigmotactic(edge-using) tendencies. As the populationdensity within the cages increased, the pref-erence for the high-perimeter caging overstandard caging was reduced. This studyshowed that rats preferred social interaction

321

Page 339: 0195162854

322 DEVELOPMENT

and security over increased floor space andthat population density within a cage can al-ter the choice of spatial design.

LIGHTING

Three characteristics of light that affect hous-ing conditions are intensity, quality (wave-length), and photoperiod (CCAC, 1993). Mon-itoring the intensity of light in the animalcolony ensures adequate illumination for pro-viding animal care without causing blindnessin the rats. Lighting that is considered normalfor humans can cause retinal damage in rats,especially albinos (Bellhorn, 1980). A study byWasowicz et al. (2002) showed that retinas ofpigmented animals such as Long-Evans ratsare affected by prolonged exposure to a mod-erate light source. Light intensity varies be-tween cages near the light source and thosecloser to the floor. In some animal rooms, asmuch as an 80-fold difference in illuminationcan occur in the vertical dimension (Schofieldand Brown, 2003). Animals housed in the up-permost levels of racks are more susceptibleto blindness than are those housed lowerdown. Eye problems could interfere with ex-periments that depend on the animal detect-ing visual cues.

Few studies have been conducted on theeffects of light quality on rats. Spalding et al.(1969) showed the wavelength of ambientlight influenced wheel running in mice, andthe degree of the effect was dependent on thestrain of mouse. A study of the effects of dif-ferent types of fluorescent lighting (full spec-trum, cool white, black, etc.) on organ andbody weights in mice showed that the type oflighting affected both organ and body weightin males but not in females (Saltarelli and Cop-pola, 1979). The CCAC (1993) recommendsthat light used in animal housing be as closeto natural sunlight as possible.

Light synchronizes circadian rhythmwith environmental time through photo-transduction by retinal ganglion cells (Bersonet al., 2002). Circadian rhythms control the

sleep-wake cycle of an animal and can influ-ence its performance in an experiment, espe-cially in older animals (Winocur and Hasher,1999; Poulos and Borlongan, 2000). Most an-imal housing revolves on a 12 hour light-darkcycle. If lights are switched on during the darkcycle or left on for a 24 hour period, retinalganglion cells respond by altering the circa-dian cycle. Breeding cycles of rats can be af-fected by circadian timing. Hoffman (1973)reported that a 12 hour light-dark cycle pro-duces a 4 day estrous period in Sprague-Daw-ley rats, whereas a 16-8 hour light-dark cycleincreases the estrous period to 5 or more days.Circadian rhythms influence the physiology ofrats. Changes in body temperature, corticos-terone levels, neurotransmitter receptor bind-ing, drug sensitivity, size of experimentally in-duced cortical infarct, and motor activity havebeen associated with the circadian cycle (Ixartet al., 1977; Vinall et al., 2000; Benstaali et al.,2001; Rebuelto et al., 2002). As such, sched-ules for behavioral testing and surgical proce-dures should be consistent to reduce variationin experimental outcome.

TEMPERATURE

Although rats have fur coats, they are sensi-tive to fluctuations in ambient temperature.The normal temperature for a rat room is inthe range of 20° to 24° C. This range allowsoptimal growth of rats and seems most com-patible with their behavioral preferences (All-mann-Iselin, 2000). Temperatures outside thisrange induce activity and metabolic changesthat can affect experimental design. Dose-response curves for drugs can be shifted bychanges in ambient temperature. A 4° C vari-ation in temperature can cause a 10-fold vari-ation in drug toxicity (Clough, 1987). Shifts inambient temperature also affect the amountof food and water that an animal consumes.Changes in ingestive behaviors can alter theeffective dose of an administered drug.

The number of inhabitants in the cage in-fluences the temperature of the microenvi-

Page 340: 0195162854

Chapter 30. Environment 323

ronment. Body temperature can be influencedby changes in animal care personnel, stormyweather, and handling (Clough, 1987).

HUMIDITY

Relative humidity in an animal facility is rec-ommended to be approximately 50%, althougha range from 40% to 70% can be tolerated(CCAC, 1993). Airborne microorganisms areless viable at a relative humidity of 50%. Lowhumidity can cause health problems such as dryskin and ringtail, whereas high humidity can in-crease ammonia production from the cages(Clough, 1987), thereby increasing the inci-dence of respiratory distress.

VENTILATION

Ventilation within the animal room influencestemperature, humidity, and air quality. Cagedesign and placement also affect the airflowat the microenvironment level. Draft-freeventilation that allows 15 to 20 air exchangesper hour is recommended. Rats housed inshoebox cages with filter tops require moni-toring to ensure that ammonia from soiledbedding does not reach toxic levels. High lev-els of ammonia in the environment are asso-ciated with respiratory distress or disease(Broderson et al., 1976). The human thresh-old level for detection of ammonia (8 parts permillion [ppm]) is above the concentration ca-pable of inducing pathology (Schofield andBrown, 2003; CCAC, 1993). Bedding shouldalso be free from aromatic carcinogens andpesticides. Both contaminants are associatedwith sawdust bedding (Clough, 1987).

NOISE

Whether a sound has damaging effects de-pends on its loudness, frequency, and dura-tion. Sounds of 160 decibels will cause dam-age to hearing in rats and in humans. It isrecommended that animal room noise doesnot exceed 85 decibels, although auditory

damage in rats has been found after intermit-tent exposure to sounds at 83 decibels (CCAC,1993). Rats hear sounds that range in fre-quency from 1000 Hz to 100,000 Hz, depend-ing on the strain (Gamble, 1982). Thus, theyare insensitive to lower-frequency tones thatfall within the human auditory range but theirupper threshold is well beyond the humanrange. This makes monitoring noise in animalfacilities more difficult. Sounds that we are in-capable of detecting in addition to those thatwe hear may cause changes in plasma corti-costerone levels, immune system function, re-productive fitness, and body weight in rats(Clough, 1987). Certain sudden, loud soundscan induce a startle response or audiogenicseizure in rats and mice. Nursing dams havebeen known to cannibalize their young afterexposure to sudden, loud noise. Sounds canalso induce aggression or changes in toleranceto electric shock (Gamble, 1982).

Rodents produce ultrasonic vocalizationsto communicate during mating, aggressivebehaviors, and maternal care (Harding andMcGinnis, 2003; CCAC, 1984; Von Fritag etal., 2002; Smotherman et al., 1974). Excessivenoise can reduce the effectiveness of thismeans of communication.

Noise in the environment can influencethe development of audition in young ro-dents. Chang and Merzenich (2003) demon-strated that rats reared in the presence of con-tinuous moderate noise showed delayedauditory cortical maturation. This effect wasreversed by returning the animals to a normalacoustic environment. Thus, animal holdingfacilities should be well away from sources ofmechanical noise, because constant exposureto sounds can alter the timing of normal au-ditory development in young rats.

DIETARY CONSIDERATIONS

The nutritional requirements of an animal canbe influenced by many factors. Genetic strain,sex, age, physiological status, and environ-

Page 341: 0195162854

324 DEVELOPMENT

ment contribute to the nutrient requirementsof the rat. Rats of different genetic strainsgrow at variable rates and thus have specificnutritive requirements. Because male ratsgrow faster and have a higher proportion ofbody protein than do females, they require ahigher proportion of protein in their diet. Sim-ilarly, growing, lactating, and postoperativeanimals require a higher percentage of dietaryprotein than do adult animals that are simplymaintaining their body weight. Rats livingin cooler conditions increase food intake tomaintain a constant body temperature, whereasrats living in warmer conditions reduce foodintake and may require higher nutritional den-sity in their food.

DIETARY RESTRICTION OR OPTIMIZATION

The negative consequences of ad libitum feed-ing has now been established in every out-bred, inbred, and hybrid cross strain of rat ex-amined (Keenan et al., 2000). Rat dietformulations are based on the nutritional re-quirements of weanling rats and lactatingdams and contain between 18% and 23% pro-tein. Animals in the growing or nursing phasesof life require approximately 15% protein intheir diet, whereas adult animals in a mainte-nance phase require 5% to 12% protein(Keenan et al., 2000). Animals recoveringfrom surgical procedures also require moreprotein in their diet, to ensure rapid healing.Because it is simpler to give all rats in a colonythe same food, most research facilitiesovernourish their adult inhabitants. Unre-stricted access to food is "unnatural" and com-promises the health of the animal. Among lab-oratory animals, only rodents are commonlygiven ad libitum access to food. Other specieshave their food intake restricted in accordancewith good scientific and veterinary practices(Keenan et al., 2002).

Dietary restriction has a positive impacton the health of rats. Feeding rats ad libitumhighly nutritious rat chow causes obesity, di-abetes, and tumors; shortens the life span; and

tends to reduce cognitive performance partic-ularly as the animal ages (Means et al., 1993).Formation of free radicals and/or gly cationreactions of sugars with proteins may be re-sponsible for the aging effects associated withad libitum feeding. Dietary restriction is asso-ciated with increased production of proteinsknown to enhance neuroplasticity and conferresistance to metabolic insult such as brain-derived neurotrophic factor (Mattson et al.,2002; Mattson et al., 2003). Anson and col-leagues (2003) have shown that the pattern offeeding dietary restricted animals affects thedegree of benefit derived from the procedure.Mice fed every other day ate the same amountas unrestricted animals and maintained theirbody weight but showed an increased resist-ance of neurons in the brain to the effects ofexcitotoxic stress. The number of dendriticspines found on neurons in the rat neocortexdecline with aging, but 24-month-old rats thatwere restricted to every other day feeding hadthe same number of dendritic spines as did6-month-old rats fed ad libitum (Moroi-Fetterset al., 1989).

A study by Markowska and Savonenko(2002) showed that the effectiveness of dietaryrestriction varies with the genetic strain of therat. For example, Fisher 344 rats failed to showsignificant benefit from dietary restriction butoffspring of a Fisher 344 and Brown Norwaycross showed improvement on tests of bothcognitive and sensorimotor behaviors.

CONTROL OF ENVIRONMENT

Joffe et al. (1973) raised rats in an environmentin which the animals were able to controlfood, light, and water availability by bar press-ing. These rats, compared with animals raisedin standard housing under identical food,light, and water availability conditions, weremore exploratory, less emotional, and moreconfident in open-field testing. This study sup-ports the notion that animals prefer to exertcontrol on their environment and that having

Page 342: 0195162854

Chapter 30. Environment 325

such control reduces their affective responseto stress.

SOCIAL OPPORTUNITY

Rats are social creatures, and they benefitfrom opportunities for social interaction. So-cial isolation is known to cause changes in be-havior (i.e., alcohol consumption) and tem-perament, as well as physiology, includingchanges in size of the adrenal and thyroidglands (Baker et al., 1979) in adult rats. A studyby Hurst and colleagues (1997) examined theeffects of housing male rats as singles or ingroups of three in two joined but dividedcages. The type of barrier that was used to di-vide the cages was used to vary the degree ofsocial contact between the two cages. Ratshoused in isolation engaged more frequentlyin behaviors related to escape or seeking so-cial information. Singly housed male ratswere more aggressive if not exposed to neigh-bors or other cagemates, yet they showed re-duced corticosterone concentration and or-gan pathology compared with group-housedrats. Although single housing may reduce so-cial stress, animals thus housed are motivatedto seek social interaction. Sharp and col-leagues (2003) examined the effect of grouphousing on stress responses to witnessingcommon experimental procedures and hus-bandry. They found that group housingreduced the stress response to witnessing tailinjections, restraint, cage changes, and de-capitation. Group housing thus is preferableto solitary housing; if housing rats singly ispart of the experimental design, it should bejustified to and approved by the local animalcare committee (CCAC, 1993). Short-term so-cial isolation of rats (4 to 7 days) has beenshown to increase the frequency of socialinteraction when the opportunity arises(Niesink and van Ree, 1982), suggesting thatrats find social interaction rewarding. How-ever, overcrowding leads to stress and in-creased aggression.

MATERNAL INFLUENCES

Natural variations in maternal care can influ-ence the cognitive development of offspring.Meaney and his group have shown that moth-ers that spend more time in an arched-backnursing posture and licking and groomingtheir pups have offspring that show enhancedspatial learning in adulthood (Liu et al., 2000).These animals have elevated levels of gluta-mate receptors and growth factors in the hip-pocampus. Levine (1967) demonstrated thatremoving the pups from the nest for brief timeintervals during the early postnatal period re-sulted in reduced response to stress in adult-hood. This procedure was called "handling."It is now known that "handling" reduces basallevels of corticosterone (Levine, 1967; Beaneet al., 2002) and alters expression of gluco-corticoid receptors in the hippocampus andfrontal cortex (Diorio et al., 1993; Bodnoff etal., 1995; Liu et al., 2000). It has been shownthat mothers who experienced "handling" ininfancy had offspring that showed reducedplasma steroids in response to novel stimuli(Denenberg and Whimbey, 1963; Levine,1967). This finding indicates that early experi-ences of a mother can affect stress responsesin subsequent generations.

We have shown that complex housing ofa pregnant dam throughout the duration ofher pregnancy ameliorates the behavioral dev-astation normally associated with perinatalcortical lesion. Both normal and frontal lesionoffspring showed enhanced spatial cognitionafter prenatal condominium experience (Gibbet al., 2001). Similar results were found withprenatal tactile stimulation. Pregnant damswere "petted" with a soft hairbrush threetimes a day for 15 minutes throughout the du-ration of their pregnancy. Offspring that weregiven postnatal day 3 lesions of frontal cortexshowed marked improvement on behavioraltests. Both prenatal complex housing and tac-tile stimulation altered neuronal morphologyin sham-operated and lesioned animals. Thus,experiences of the mother rat can have an im-

Page 343: 0195162854

326 DEVELOPMENT

pact on the behavior and physiology of heroffspring. Social isolation of rats during thepreweaning period of life (3 to 6 hours per dayfor 5 days) alters the behavior of both the pupsand the mother rat by increasing their activ-ity and the number of mother-pup interac-tions (Zimmerberg et al., 2003). Taken to-gether, these studies show that although thereis some natural variation in maternal care thatcan influence the behavior of offspring, it isprudent to attempt to control the early expe-riences of rat pups to reduce the confoundingeffects of variations in experience.

AGE AT WEANING

Preweanling rats are sensitive to a maternalpheromone present in the mother's feces andrespond by eating maternal feces to promotedevelopment of myelin in the brain. Rats re-spond to this pheromone until 27 days of age,but rats deficient in myelin continue to re-spond to the pheromone beyond this age(Schumacher and Moltz, 1985). Weaning ratpups too early has been shown to have seri-ous physiological and behavioral conse-quences. Increased susceptibility to gastricpathology and delayed maturation of re-sponses to restraint stress were noted in pupsweaned at 15 days rather than at 22 days ofage (LaBarba and White, 1971; Ackerman etal., 1975; Milkovic et al., 1975).

HUSBANDRY VARIABLES

Many environmental factors can act as un-controlled variables in an experiment: musicin the colony, strong smells, different careconditions, animal transportation, and evenfrequency of cage cleaning.

Using music in the colony as a controlledsource of sound is thought to be helpful in re-ducing the disruptiveness of uncontrollednoise. Sounds associated with normal hus-bandry procedures have a greater impact on

the magnitude of the subsequent stress re-sponse in animals accustomed to silence thanin those used to constant sounds. Music canalso provide a form of enriching experiencefor rats. Rauscher, Robinson, and Jens (1998)showed that rats exposed in utero and 60 dayspostpartum to Mozart compositions wereable to complete a maze more quickly andwith fewer errors than were rats exposed towhite noise or silence.

Rats possess a highly specialized sense ofsmell. Just as much of our behavior is guidedby sight, rats use their keen sense of smell tofamiliarize themselves with their environ-ment. Pheromones are smells that help ratsidentify the presence of neighbors and cangive signals that affect development, repro-ductive fitness, and some behaviors of nearbyrats. Strong cleaning odors, ammonia buildup,and the use of perfumes by laboratory per-sonnel can all interfere with the acquisition ofodor information by rats.

Another important variable to consider isthe handling of the animal provided by bothexperimenter and animal care personnel.Some animal health care technicians treat theanimals they care for like pets and handlethem a great deal, whereas others treat themas though they are wild and handle them withreticence. Likewise, some technicians talk toand handle animals in the cage they are clean-ing, whereas others avoid contact. These dis-parate methods of handling can cause the an-imals to mount varying degrees of a stressresponse, which could affect their perfor-mance when subjected to testing procedures.

Transportation of animals via airplane ortruck affects corticosterone levels and immunefunction (CCAC, 1993). It is recommended thata minimum period of adjustment of 2 days beallowed to ensure stabilization of physiologicalparameters. Timed-pregnant females that aresubjected to transportation stress may have off-spring that are very different behaviorally andneuroanatomically from offspring from moth-ers that do not experience this stress duringpregnancy (Stewart and Kolb, 1988).

Page 344: 0195162854

Chapter 30. Environment 327

Frequency of cage cleaning affects the fit-ness and number of useable rats at weaning.Litters that had their cages changed twice perweek had more healthy survivors than did lit-ters exposed to once-a-week bedding changes(Cisar and Jayson, 1967). This may have re-sulted from increased exposure to ammoniain the once-a-week litters or from increasedhandling in the twice-a-week litters, but thefrequency of cage cleaning may affect the ex-perimental outcome.

EXERCISE

A rat's natural inclination is to explore its en-vironment for food and mating opportunities.Access to a running wheel provides laboratoryanimals with a means for exploration beyondthe limits of their caged environment. Al-though exercise is not normal behavior of an-imals in the wild, there is mounting evidencethat exercise is beneficial to the health of rats.Exercise has been shown to increase the pro-duction of neurotrophic factors in the centralnervous system (Gomez-Pinilla et al., 2001)and neurons in the hippocampus and motorcortex (van Praag et al., 1999; Galvez et al.,2002). There also is evidence that exercise re-duces an animal's response to stress (Green-wood et al., 2002). In addition, exercise hasbeen shown to be therapeutic but not pro-phylactic for rats that have sustained corticalinjury (Gentile et al., 1987).

ENRICHMENT

In the 1940s, Donald Hebb raised a group oflaboratory rats in his home. When he testedthese animals as adults in a maze, he foundthat they made fewer errors than did animalsraised under standard laboratory conditions.This was the first demonstration that an en-riching environment can influence the behav-ioral performance of rats. Rosenzweig and hiscolleagues (1971) extended this finding by

showing that the brains of "enriched" animalswere heavier and showed an increase in cor-tical thickness, acetylcholinesterase activity,synaptic contacts, and dendritic arborization.Rats housed in a complex environment alsoundergo brain changes that include increasesin glial density and vasculature (Black et al.,1987). It is interesting to note that environ-mental enrichment has a greater effect onopen-field behavior and body weight of wildrats than on their domesticated counterparts(Huck and Price, 1975). It appears that geneticchanges accompanying domestication havemade laboratory rats more resistant to the in-fluences of experience.

Greenough and Black (1992) proposedthat environment can influence brain mor-phology in one of two ways: experience-de-pendent and experience-expectant changes inthe brain. Experience-expectant changes oc-cur during development and require properinput for a system like the visual system to de-velop normally. This involves stabilizing use-ful synapses and deleting redundant ones. Ex-perience-dependent changes are those thatallow experiences to alter the animal through-out its life span. Such experiences includemaze learning and motor learning.

We have determined that the impact ofcomplex housing on neuroanatomical changesin the cerebral cortex varies with age and sexin rats (Kolb et al., 2003). Animals placed inenriched environments as adults showed anincrease in spine density, whereas animals en-riched at weaning showed a decrease in spinedensity. Male rats at all ages showed increasesin dendritic length, yet only adult femalesshowed similar increases.

Enriching experience is not only derivedfrom complex housing. Sensory stimulationand behavioral testing can also be consideredforms of enrichment. Brief periods of tactile orolfactory stimulation after brain injury in ratscan improve their behavioral outcome (Gibband Kolb, 2000; C. L. R. Gonzalez, unpublishedobservations) not only during development butalso in adulthood. Participation in an experi-

Page 345: 0195162854

328

ment exposes an animal to a variety of experi-ences to which animals in the colony are notsubjected. These experiences have the poten-tial to alter the subsequent behavior and neu-roanatomy of test subjects. Kolb and colleagues(2003) conducted an experiment to determinethe relative effects of solving the place versionof the Morris water task (learning condition),swimming in the pool without a platform(yoked condition: animals were allowed toswim for the same length of time that it tookthe animals in the learning condition to find theplatform), and no behavioral testing on spinedensity and dendritic arbor in the occipital cor-tex. Rats that solved the problem had the great-est dendritic arbor and spine density, yet ratsthat swam for an equal length of time showeda significant elevation of these measures abovethe baseline found for nontested animals. Thisresult shows that simple participation in theexperiment was sufficient to change corticalcircuitry.

DEVELOPMENT

CONCLUSION

Enrichment can take many forms: access torunning wheels, handling, sensory stimula-tion, group housing, or complex housing (Fig.30-1). There is a debate as to whether stan-dard housing for rats should include someform of enrichment. Proponents of this viewbelieve that standard housing produces "im-poverished" animals that have underdevel-oped brains and a limited behavioral reper-toire. But despite the minimal stimulationprovided by standard housing, very littlepathological behavior has been ascribed torats raised in these conditions. Although en-vironmental stimulation produces a smarterrat, the value of "intelligence" to animals thatdo not need to compete for food, housing, ormating opportunities is difficult to assess. Theuse of laboratory rats that have been rearedin standard housing as a baseline group tostudy the effects of the environment on brain

Figure 30-1. Various forms of housing for laboratory rats, a, Group housing, b, Running wheel access.c, Complex housing.

Page 346: 0195162854

Chapter 30. Environment 329

development and function have yielded manyvaluable insights over the past 60 years. Al-though it is important to consider optimalhousing conditions for the experimental sub-jects, one should be mindful of the impact ofadding enriching devices or protocols to stan-dard laboratory rat care. Minimal changes inenvironmental conditions can have huge ef-fects on the behavior and physiology of labanimals.

REFERENCES

Ackerman SH, Hofer MA, Weiner H (1975) Age at ma-ternal separation and gastric erosion susceptibilityin the rat. Psychosomatic Medicine 37:180-184.

Allmann-Iselin I (2000) Husbandry. In: The LaboratoryRat (Krinke GJ, ed.), pp. 45-72. London: AcademicPress.

Anson RM, Guo Z, De Cabo R, lyun T, Rios M,Hagepanos A, Ingram DK, Lane MA, Mattson MP(2003) Intermittent fasting dissociates beneficial ef-fects of dietary restriction on glucose metabolismand neuronal resistance to injury from calorie in-take. Proceedings of the National Academy of Sci-ence U S A 100:6216-6220.

Anzaldo AJ, Harrison PC, Maghirang R-G, Gonyou HW(1994) Increasing welfare of laboratory rats with thehelp of spatially enhanced cages. Animal Welfare In-formation Center Newsletter 5(3): 1-2.

Baker HJ, Lindsey JR, Wiesbroth SH (1979) The labo-ratory rat. London: Academic Press.

Barnett SA (1975) The rat: A study in behavior. Chicago:The University of Chicago Press.

Beane ML, Cole MA, Spencer RL, Rudy JW (2002)Neonatal handling enhances contextual fear condi-tioning and alters corticosterone stress responses inyoung rats. Hormones and Behavior 41:33-40.

Bellhorn RW (1980) Lighting in the animal environ-ment. Laboratory Animal Science 20:440—450.

Benstaali C, Mailloux A, Bogdan A, Auzeby A, Touitou Y(2001) Circadian rhythms of body temperature andmotor activity in rodents their relationships with thelight-dark cycle. Life Sciences 68:2645-2656.

Berson DM, Dunn FA, Takao M (2002) Phototransduc-tion by retinal ganglion cells that set the circadiandock. Science 295:1070-1073.

Black JE, Sirevaag AM, Greenough WT (1987) Complexexperience promotes capilliary formation in youngrat visual cortex. Neuroscience Letters 83:351-355.

Bodnoff SR, Humphreys AG, Lehman JC, DiamondDM, Rose GM, Meaney MJ (1995) Enduring effects

of chronic corticosterone treatment on spatial learn-ing, synaptic plasticity, and hippocampal neu-ropathology in young and mid-aged rats. Journal ofNeuroscience 15:61-69.

BrodersonJR, Lindsey JR, Crawford JE (1976) The roleof environmental ammonia in respiratory my-coplamosis of rats. American Journal of Pathology85:115-130.

Canadian Council on Animal Care (1984) CCAC guide.Ottawa: CCAC.

Canadian Council on Animal Care (1993) CCAC guide.Ottawa: CCAC.

Chang EF and Merzenich MM (2003) Environmentalnoise retards auditory cortical development. Sci-ence 300:498-502.

Cisar CF and Jayson G (1967) Effects of frequency ofcage cleaning on rat litters prior to weaning. Labo-ratory Animal Care 17:215-218.

Clough C (1987) Quality in laboratory animals. In: Lab-oratory animals: An introduction for new experi-menters (Tuffery AA, ed.), pp. 79-97. Chichester:Wiley-Interscience.

Denenberg VH and Whimbey AE (1963) Behavior ofadult rats ils modified by the experiences their moth-ers had as infants. Science 142:1192-1193.

Diorio D, Viau V, Meaney MJ (1993) The role of the me-dial prefrontal cortex (cingulate gyrus) in the regu-lation of hypothalamic-pituitary-adrenal responsesto stress. Journal of Neuroscience 13:3839-3847.

Galvez R, Soskin PN, Cho JH, Grossman AW, Greee(2002) Voluntary exercise increases the number ofnew neurons in the adult rat motor cortex in a timedependent fashion. In: Society for Neuroscience,program No. 662.610. Orlando, FL: Society forNeuroscience.

Gamble MR (1982) Sound and its significance for labo-ratory animals. Biological Reviews 57:395-421.

Gentile AM, Beheshti 2, Held JM (1987) Enrichmentversus exercise effects on motor impairments fol-lowing cortical removals in rats. Behavioral andNeural Biology 47:321-332.

Gibb R and Kolb B (2000) Comparison of the effects ofpre- and postnatal tactile stimulation on functionalrecovery following early frontal cortex lesions. So-ciety for Neurosciences Abstracts. 26 Programnumber 366.9.

Gibb R, Gonzalez CLR, Kolb B (2001) Prenatal enrichmentleads to improved functionaol recovery followingperinatal frontal cortex injury: Effects of maternalcomplex housing. In: Society for Neurosciences. Ab-stracts. Vol 27 Program number 476.4.

Gomez-Pinilla F, Ying Z, Opazo P, Roy RR, Edgerton VR(2001) Differential regulation by exercise of BDNF andNT-3 in rat spinal cord and skeletal muscle. EuropeanJournal of Neuroscience 13:1078-1084.

Page 347: 0195162854

330 DEVELOPMENT

Greenough WT and Black JE (1992) Induction of brainstructure by experience. Substrates for cognitive de-velopment. In: Minnesota Symposium on Child De-velopment (Nelson CA, ed.), pp. 155-200. Hillsdale:Lawrence Erlbaum.

Greenwood BN, Hinde JL, Nickerson M, Thompson K,Fleshner M (2002) Freewheel running changes thebrain's response to acute, uncontrollable stress. So-ciety for Neurosciences Abstracts. Vol. 28 Programnumber 843.10.

Harding SM and McGinnis MY (2003) Effects of testos-terone in the VMN on copulation, partner prefer-ence, and vocalizations in male rats. Hormones andBehavior 43:327-335.

Hoffman JC (1973) The influence of photoperiods on re-productive functions in female mammals. In: En-docrinology. II. Handbook of Physiology. Section 7,pp. 57-77. Washington, D.C.: American Psycho-logical Society.

Huck UW and Price EO (1975) Differential effects of en-vironmental enrichment of the open-field behaviorof wild and domestic Norway rats. Journal of Com-parative Physiology and Psychology 89:892-898.

Hurst JL, Barnard CJ, Nevison CM, West CD (1997)Housing and welfare in laboratory rats: Welfare im-plications of isolation and social contact amongcaged males. Animal Welfare 6:329-347.

Ixart G, Szafarczyk A, Belugou JL, Assenmacher I (1977)Temporal relationships between the diurnal rhythmof hypothalamic corticotrophin releasing factor, pi-tuitary corticotrophin and plasma corticosterone inthe rat. Journal of Endocrinology 72:113-120.

Joffe JM, Rawson RA, MuclikJA (1973) Control of theirenvironment reduces emotionality in rats. Science180:1383-1384.

Keenan KP, Ballam GC, Haught DG, Laroque P (2000)Nutrition. In: The laboratory rat (Krinke GJ, ed.),pp. 57-75. London: Academic Press.

Kolb B, Gibb R, Gorny G (2003) Experience-dependentchanges in dendritic arbor and spine density in neo-cortex vary with age and sex. Neurobiology ofLearning and Memory 79:1-10.

Krohn TC, Hansen AK, Dragsted N (2003) Telemetry asa method for measuring the impact of housing con-ditions on rats' welfare. Animal Welfare 12:53-62.

LaBarba RC and White JL (1971) Litter size variationsand emotional reactivity in BALB/c mice. Journalof Comparative and Physiological Psychology75:254-257.

Levine S (1967) Maternal and environmental influenceson the adrenocortical response to stress in weanlingrats. Science 156:258-260.

Liu D, Caldji C, Sharma S, Plotsky PM, Meaney MJ(2000) Influence of neonatal rearing conditions onstress-induced adrenocorticotropin responses and

norepinepherine release in the hypothalamic para-ventricular nucleus. Journal of Neuroendocrinology12:5-12.

Markowska AL and Savonenko A (2002) Retardation ofcognitive aging by life-long diet restriction: Impli-cation for genetic variance. Neurobiology of Aging23:75-86.

Mattson MP, Duan W, Guo Z (2003) Meal size and fre-quency affect neuronal plasticity and vulnerabilityto disease: Cellular and molecular mechanisms.Journal of Neurochemistry 84:417-431.

Mattson MP, Duan W, Chan SL, Cheng A, Haughey N,Gary DS, Guo Z, Lee J, Furukawa K (2002) Neuro-protective and neurorestorative signal transductionmechanisms in brain aging: modification by genes,diet and behavior. Neurobiology and Aging 23:695-705.

Means LW, HigginsJL, Fernandez TJ (1993) Mid-life on-set of dietary restriction extends life and prolongscognitive functioning. Physiology and Behavior54:503-508.

Milkovic K, Paunovic JA, Joffe JM (1975) Effects of pre-and postnatal litter size on development and be-havior of rat offspring. Developmental Psychobiol-ogy 9:365-375.

Moroi-Fetters SE, Mervis RF, London ED, Ingram DK(1989) Dietary restriction suppresses age-relatedchanges in dendritic spines. Neurobiology and Ag-ing 10:317-322.

Niesink RJ and van Ree JM (1982) Short-term isolationincreases social interactions of male rats: A para-metric analysis. Physiology and Behavior 29:819-825.

Poulos S and Borlongan C (2000) Artificial lighting con-ditions and melatonin alter motor performance inadult rats. Neuroscience Letters 280:33-36.

Rauscher FH, Robinson D, Jens JJ (1998) Improved mazelearning through early music exposure in rats. Neu-rological Research 20:427-432.

Rebuelto M, Ambros L, Montoya L, Bonafine R (2002)Treatment-time-dependent difference of ketaminepharmacological response and toxicity in rats.Chronobiology International 19:937-945.

Rosenzweig MR (1971) Effects of environment on de-velopment of brain and behavior. In: The bio-psychology of development (Tobach E and Shaw E,eds.), pp. 303-342. New York: Academic Press.

Saltarelli CG and Coppola CP (1979) Influence of visiblelight on organ weights of mice. Laboratory AnimalScience 29:319-322.

Schofield JC and Brown MJ (2003) Animal care and use: Anonexperimental variable. In: Essentials for animal re-search: A primer for research personnel (Bennett B,Brown M, and Schofield J, eds) Second Edition. [Bookavailable online]. Retrieved February 10,2004 from the

Page 348: 0195162854

Chapter 30. Environment 331

World Wide Web: http://www.nal.iosda.gov/awic/pubs /noawicpubs / essentia-htm.

Schumacher SK and Moltz H (1985) Prolonged respon-siveness to the maternal pheromone in the post-weanling rat. Physiology and Behavior 34:471-473.

Sharp J, Zammit T, Azar T, Lawson D (2003) Stress-likeresponses to common procedures in male ratshoused alone or with other rats. ContemporaryTopics In Laboratory Animal Science 41:8-14.

Smotherman WP, Bell RW, Starzec J, Elias J, ZachmanTA (1974) Maternal responses to infant vocaliza-tions and olfactory cues in rats and mice. Behaviorand Biology 12:55-66.

SpaldingJF, Archuleta RF, Holland LM (1969) Influenceof the visible colour spectrum on activity in mice.Laboratory Animal Care 19:50-54.

Stewart] and Kolb B (1988) The effects of neonatal go-nadectomy and prenatal stress on cortical thicknessand asymmetry in rats. Behavioral and Neural Bi-ology 49:344-360.

van Praag H, Kempermann G, Gage FH (1999) Runningincreases cell proliferation and neurogenesis in the

adult mouse dentate gyrus. Nature Neuroscience2:266-270.

Vinall PE, Kramer MS, Heinel LA, Rosenwasser RH(2000) Temporal changes in sensitivity of rats tocerebral ischemic insult. Journal of Neurosurgery93:82-89.

Von Fritag JC, Schot M, van den Bos R, Spruijt BM(2002) Individual housing during the play period re-sults in changed responses to and consequences ofa psychosocial stress situation in rats. Developmen-tal Psychobiology 41:58-69.

Wasowicz M, Morice C, Ferrari P, Callebert J, Versaux-Botteri C (2002) Long-term effects of light damageon the retina of albino and pigmented rats. Investi-gations in Ophthalmology and Visual Science43:813-820.

Winocur G and Hasher L (1999) Aging and time-of-dayeffects on cognition in rats. Behavioural Neuroscience113:991-997.

Zimmerberg B, Rosenthal AJ, Stark AC (2003) Neonatalsocial isolation alters both maternal and pup behav-iours in rats. Developmental Psychobiology 42:52-63.

Page 349: 0195162854

This page intentionally left blank

Page 350: 0195162854

Defense and %/•Social Behavior

Page 351: 0195162854

This page intentionally left blank

Page 352: 0195162854

Antipredator Defense

D. CAROLINE BLANCHARD ANDROBERT J. BLANCHARD

31

NATURAL BEHAVIORS

A focal aspect of the class of defensive behav-iors of the rat is that they are "natural." This im-plies, first, that they occur in nature, that is, inthe real world, outside a laboratory setting. Insuch real-world settings, behaviors are subjectto their natural consequences; some have goodoutcomes for the animals that have performedthem, whereas others may result in a range ofwoes from inconvenience to disaster. In linewith the tenets of Darwinian evolution, the con-ditions under which such behaviors occur havean important effect on the success of their out-come, such that particular behaviors are adap-tive if and only if they occur in situations thatare conducive to their success. Thus, the con-cept of "natural behavior" suggests a strong re-lationship between both environmental (physi-cal and social) stimuli and stimuli arising fromthe animal itself, and particular behaviors. Athird aspect of "natural" to many biologists andpsychologists is that evolution and some degreeof genetic determination are involved in shap-ing such a behavior; that is, it is not solely theproduct of specific learning.

DEFENSE IN RATS:SPECIFIC BEHAVIORS

In both wild and laboratory rats (Rattusnorvegicus), the specific defensive behaviorsthat have been described (e.g., Blanchard

and Blanchard, 1969, 1989; Pinel and Treit,1978; Blanchard RJ et al., 1980, 1989, 1990,1991; Blanchard DC et al., 1981, 1991; Die-lenberg et al., 1999, 2001; McGregor et al.,2002) include the following.

Flight: rapid movement (typically running)away from a threat source

Hiding or sheltering: entering and remaining ina place where the animal is less visible or thatthe threat cannot enter

Freezing: immobility, also called crouchingwhen a specific posture is maintained

Alarm cries: cries of about 22 kHz that are emit-ted when familiar conspecifics are present, andto which they may respond defensively

Defensive threat: defensive upright or standingposture facing the oncoming threat, typicallyaccompanied by tooth exposure and screams

Defensive attack: biting at the oncoming threat,often after a jump toward it

Risk assessment: a pattern of investigation of thethreat source, including scanning it from a dis-tance while freezing; and a "stretch attend" or"stretch approach" behavior in which the ani-mal adopts a stretched-out, low-back posturewhile oriented toward the threat source andmay show short bursts of movement inter-spersed with periods of immobility. Closer ap-proaches and even contact may occur, againtypically involving the low-back stretchedposture.

Defensive burying: discrete threat stimuli are of-ten covered with substrate or other materials.It is not certain if this is specifically a defensive

335

Page 353: 0195162854

336 DEFENSE AND SOCIAL BEHAVIOR

response or a more general response to dis-agreeable objects, because rats also bury deadconspecifics, unpalatable food, and novel ob-jects (Wilkie et al., 1979; Blanchard RJ et al.,1989). "Burying" may also involve an elementof risk assessment, in that tossing objects ontoanimals or other stimuli with ambiguousthreat characteristics may cause these stimulito respond in such a fashion as to make theirthreat potential clearer.

This list is likely to be incomplete, withnew categories of defense awaiting descrip-tion and analysis. In addition, subtle variationsand combinations of the above may occur un-der particular circumstances. For example"tunnel guarding" behavior of colony subor-dinate males appears to combine sheltering(in the tunnels of a visible burrow system)with defensive threat and attack toward thedominant, should the latter attempt to en-ter the tunnel in which the subordinate islodged (Blanchard RJ et al., 2001 A). Thesebehaviors make obvious a point that weelaborate later, that features of the situationhave a very strong influence on the formof the defense seen: No tunnel, no tunnelguarding.

STIMULI THAT ELICIT DEFENSE

TYPES OF THREAT STIMULI

Defensive behaviors occur in response tothreats to an animal's life or body. Thesethreats can be divided into four categories:(1) predators, (2) attacking conspecifics, (3)threatening features of the environment(lightning, fire, high places, and water), and(4) nonconspecific but dangerous resourcecompetitors. The list suggests several impor-tant points. First, it is active danger to life orlimb that elicits defensive behavior, not a pas-sive threat such as hunger, nor a threat or chal-lenge to some resource that does not involveimmediate danger to the animal. Second, thedistinction between animate threats and inan-

imate threats will turn out to be important forboth the type of defensive behavior offeredand for laboratory models attempting tomeasure these defenses.

An additional point concerning types ofthreat stimuli is that some of these are clear,immediate, salient, and embodied in a specificobject, whereas others are ambiguous. Cer-tain sounds or odors may serve as distinct cuesthat danger is or has been near but provide lit-tle information on the exact location or iden-tity of the danger or whether it is still present.As Table 31-1 indicates, these features are as-sociated with very different types of defense.A localizable threat is important for sometypes of defensive behavior, such as flight, tobe effective. Simply running around withoutreference to the location of the threat sourcemay attract attention without removing thesubject from the presence of the threat. Sim-ilarly, an embodied threat is necessary for de-fensive attack to be effective; biting at soundsor odors does not reduce the threat they maysignal.

Threat Ambiguity and Risk AssessmentThe major type of defense to ambiguousthreat is risk assessment, an active investi-gation of the threat stimulus or situationthat typically involves a stretched, low-backposture that appears to minimize the chanceof the animal being detected as it goes aboutthe business of investigating the threat. Thestretched posture is very similar to the pos-ture adopted by a stalking animal; anothersituation in which the joint goals of ap-proach to another animal while escaping de-tection by that animal might be inferred. Inaddition, the movement parameters of thestretch approach are such as to minimize de-tection: the animal's forward movement isbrief and interspersed with periods of im-mobility, during which detection by thethreat is less likely. What is clear is that riskassessment is associated with threat-relatedinformation gathering. A very clever set ofstudies by Pinel et al. (1989) indicated that

Page 354: 0195162854

Chapter 31. Antipredator Defense 337

Table 31-1. Stimuli, Behaviors, Outcomes, and Affect Associated WithThreat Detection or with Defense to Specific Threats

Aspect Threat detection Defense

Stimuli

Behaviors

Outcome

Affect

Animate movementPredator odorsAlarm criesConspecific odorsRisk assessment

Stretch attendStretch approach

Hiding-avoidance

Danger detectedorSafety detected

Anxiety

PredatorAttacking conspecific

FlightFreezingDefensive threat and attackHiding-avoidanceConspecific back defenseAlarm vocalizationsEscapeAvoidance of detectionFrightening off threatProtection of selfProtection of body partsWarning conspecifics (respectively)Fear

proximity to a stimulus while in this stretchattend/approach mode was associated withlearning of defensiveness to that stimulus,whereas equal or greater proximity withoutrisk assessment was not.

Ambiguity of threat does not necessarilyimply that the stimulus is itself amorphous ordisembodied. The attack potential of a con-specific or even a predator may be ambigu-ous, even if the animal itself is clearly present.In such cases, the threatened animal's first re-sponse may be risk assessment. When a dom-inant rat is tethered such that it cannot ap-proach a subordinate, the subordinate reactsby stretch-attending or stretch-approachingthe dominant, presumably to check out thethreat features of the latter (unpublished ob-servations). Similarly, a predator at a distanceand not approaching the subject may elicit riskassessment rather than flight. When ap-proached from a distance by the human ex-perimenter, a wild rat first freezes while ori-ented to the predator, with an abrupttransition to flight when the latter is about 3meters away, presumably when its threat po-tential becomes clear (Blanchard DC et al.,1981).

Risk assessment behavior is adaptive be-cause defensive behaviors are costly in termsof time and energy. Displaying them or con-tinuing to display them when there is nothreat may be very wasteful. Risk assessmentis part of the process of deciding when tostop being defensive. Not being defensivewhen there is a real threat can be disaster-ous, however, and risk assessment may helpthe animal to determine that a danger is gen-uine, leading to the expression of more spe-cific defenses. Finally, displaying the wrongdefensive behavior for a particular threat andsituation may be as dangerous as displayingno defense. Risk assessment helps the animalto determine which defensive behavior to ex-press. Thus, risk assessment is crucial to allof the cognitive and decision-making aspectsof defense.

THE ROLE OF "EXPEDITING"STIMULI IN CONTROLLING

SPECIFIC DEFENSES

As noted earlier, features of the social andphysical environment in which the threat is

Page 355: 0195162854

338 DEFENSE AND SOCIAL BEHAVIOR

encountered may affect the type of defensethat is offered. The presence of an exit routefacilitates flight (Blanchard RJ et al., 1989), ma-nipulable substrate facilitates defensive bury-ing (Pinel and Treit, 1979), shelters promotehiding and tunnel guarding (Dielenberg et al.,1999; Blanchard RJ et al., ZOOlb), and both ashelter and the presence of conspecifics en-hances antipredator alarm vocalizations (Blan-chard RJ et al., 1991). While there is nothingto prevent an animal from performing the ac-tions associated with a particular defense inthe absence of the relevant support stimulus,these defensive behaviors tend not to occurwithout their particular support stimuli.

Most laboratory studies of the responsivityof rats to threat stimuli include none of the pre-ceding features: not an exit from the test situa-tion nor a shelter, no conspecifics, not even aspecific, embodied threat stimulus. These omis-sions sharply reduce levels of flight, hiding, de-fensive threat/attack, burying, and alarm vo-calizations, which leaves freezing and riskassessment. The latter is typically not measured,which leaves freezing. This situation is partly re-sponsible for the widespread treatment of freez-ing as the primary, or even sole, measure of de-fensiveness in the rat.

THE ROLE OF THREAT INTENSITY INCONTROLLING SPECIFIC DEFENSES

An alternative explanation for how defensivebehaviors are determined was offered byFanselow and his coworkers. Fanselow andLester (1988) conceptualized a threat intensityor stimulus imminence-based differentiationamong three sets of defensive behaviors: (1)preencounter defenses (represented by com-promises in normal activities), (2) post-encounter defenses (freezing), and (3) circa-strike defenses (flight, vocalization, biting, andhigh activity responses such as those toshock), further suggesting (Fanselow, 1994)that this intensity/imminence parameter de-termines the form of defense offered in par-ticular situations.

Threat stimulus intensity does affect themagnitude of defensive behavior. Largepieces of a cloth rubbed on a cat elicit moredefensive behavior in rats than do small piecesof this cloth, even when both are hidden incontainers such that only cat odor remains asthe effective stimulus (L. Takahashi, personalcommunication). While Zangrossi and File(1992) found that exposure to cat odor pro-duces anxiety-like responses on the elevatedplus-maze for an hour, but not a day, after-wards, Adamec & Swallow (1993) reported amuch longer-lasting anxiety-like plus-mazeresponse after cat exposure, a difference thatlikely reflects greater threat intensity for a cat,compared to cat odor alone. However, it isalso clear that specific changs in environmen-tal stimuli can rapidly and dramatically alterthe expression of specific defenses, even whenthe threat stimulus itself is unchanged. For ex-ample, closing a door to block an escape routeinstantaneously changes flight to freezing inwild rats (Blanchard DC et al., 1981). Whilethis change might possibly be interpreted asincreasing threat intensity, in the stimulus in-tensity/imminence schema an increase in in-tensity changes freezing to flight, not the op-posite. Such considerations suggest that thedeterminants of defensive behaviors are morecomplex than a single intensity dimension.They agree with data indicating that that thepatterning of defense is responsive to "expe-diting" features of both the threat stimulusand the local environment.

OUTCOMES OFDEFENSIVE BEHAVIORS

The outcomes of defense provide the mecha-nism by which these behaviors evolve. Ratsare fossorial, nocturnal, and, to some degree,colonial. The fossorial (burrow-dwelling) featurefacilitates the adaptive value of running awayor hiding from danger, particularly if a tunnelof some sort is available. The nocturnal featurepromotes the use of nonvisual senses, and ol-faction in the rat is especially useful in the con-

Page 356: 0195162854

Chapter 31. Antipredator Defense 339

text of defense. Colonial rodents are particu-larly noted for their alarm cries, and these areprominent in rats.

A intriguing aspect of evolutionary analy-sis of behavior relates to extended responsiv-ity to olfactory threat stimuli in the rat. Blan-chard DC et al. (2003b) suggest that aversiveodorants may elicit different patterns of de-fensive behavior, depending on the ability ofthe odor to predict the presence of a preda-tor. Cat fur and skin odors (obtained by rub-bing a cat with a cloth or by using a collarworn by a cat) elicit a range of defenses, in-cluding avoidance, freezing, and risk assess-ment. When exposed later to the stimulus orsituation where the odor was initially en-countered but is now absent, rats show de-fensive behaviors to these conditioned stimuli(e.g., Blanchard RJ et al., 200Ib; Dielenberg etal., 2001; McGregor et al., 2002). When freshfecal material from the same cat that donatedthe fur or skin odor is presented, the imme-diate defensive behaviors are virtually identi-cal to those seen to the fur or skin odor. How-ever, no conditioned defensiveness occurswhen the rat is again confronted by the situ-ational and specific stimuli with which theodors of feces had originally been associated.This is consonant with findings that syntheticfeces/anal gland odors also fail to serve as ef-fective unconditioned stimuli for Pavloviandefensive conditioning (Blanchard DC et al.,2003b). These differences are interpreted asreflecting that fur and skin odors dissipatequickly and thus have much greater ability topredict the presence of a predator than do fe-ces, which, along with its odors, dissipatesvery slowly. Thus, the ability of odors to serveas unconditioned stimuli for defensive condi-tioning may reflect whether these odors ac-curately predict danger.

LABORATORY MODELS OF RATDEFENSIVE BEHAVIORS

There are many laboratory tasks designed tomeasure some aspect of behavior potentially

related to defense. Most "anxiety models" ortasks measuring "anxiety-like behaviors" (seeRodgers, 1997, for a review) tap some aspectof defensiveness, as do many of the tasks cre-ated to measure depression (Willner, 1991).Although these may or may not be adequatefor the purposes for which they were devised,in preclinical testing of drugs, such tasks donot provide a precise analysis of how theirmeasures relate to the range of natural be-haviors shown by rats. Attempts to elicit andmeasure defensive behaviors are relativelyrecent. Nevertheless, early reports (Yerkes,1913; Stone, 1932) detailing the response ofwild and laboratory rats to human touch andhandling documented that wild rats flee orbite if picked up, whereas laboratory rats areless reactive. Curti (1935) reported that expo-sure to a cat is capable of eliciting some fear-related responses in the rat, although her con-clusion that the odor of a cat was not anadequate specific stimulus for these responsessuggests that her test situation was not a sen-sitive one. Defensive behaviors of rats to con-specific attack (which involve some specific el-ements not found in antipredator defense)were well described by Grant and his col-leagues (Grant and Chance, 1958; Grant, 1963;Grant and MacKintosh, 1963). Although freez-ing to a cat or to shock stimuli had earlier beendescribed in some detail (e.g. Curti, 1935;Blanchard and Blanchard, 1969; Fanselow,1980), the first specific attempt to outline thearray of defensive responses to an oncomingpredator involved wild rats, in the Fear/Defense Test Battery (FDTB).

FEAR/DEFENSE TEST BATTERY

The FDTB was run in an oval runway with ahigh barrier down the center of its long axisthat did not extend to either end wall. Thiscreated an endless runway of 5 or 6 meters ona side that enabled the animal to run aroundthe end of the barrier and into the next straightside, indefinitely, while being chased by a hu-man experimenter. The closing of a door toblock the runway at one end trapped the rat,

Page 357: 0195162854

340

enabling its response to the oncoming exper-imenter to be evaluated at varying distancesof 5, 4, 3, 2, 1, 0.5, and 0 (contact) meters. Re-sponse to attempted pick-up was also meas-ured (see Blanchard DC, 1997, for a review ofprocedures and findings).

In this test, the responses of wild ratswere extremely consistent. In several studies,more than 97% of wild rats fled from the on-coming experimenter when the runway wasopen, permitting flight. The average distancebetween the experimenter and the subjectwhen the subject turned to move away wasabout 2.7 meters. When the door was closed,precluding flight, the rat quickly froze, typi-cally in an upright posture and always facingthe experimenter. Freezing was seen duringabout 80% of observations from 5 meters toabout 2 meters, but the rats showed a startleresponse to sudden sounds (hand clap or fir-ing of starter's pistol) that increased in ampli-tude as the experimenter approached, sug-gesting that this freezing response alsoinvolves preparation for violent action thatincreases as contact becomes imminent. Atabout 1 meter, defensive threat screams sud-denly began to occur, and at about 0.5 meter,some animals jumped at or past the experi-menter, with biting should contact occur.When the experimenter attempted to pick upthe subjects, 100% of wild rats bit. These re-lationships are schematized in Figure 31-1.

Laboratory rats in the FDTB showed lessof every response than did wild rats, except forfreezing, differences that were paralleled pre-cisely in rats bred over 35 generations for "tame-ness" or "wildness" from wild-trapped stock inNovosibirsk, Siberia (Blanchard DC et al., 1994).This wild-laboratory rat difference appears toreflect selective breeding of rats for failure toshow active defenses such as defensivethreat/attack and flight, during the process ofdomestication. The same general findings werenoted to human handling (Takahashi and Blan-chard, 1982) as those in response to an attack-ing conspecific. Moreover, the responses of wildrats reared in a laboratory setting were gener-

DEFENSE AND SOCIAL BEHAVIOR

Figure 31-1. Schematic of the intensity of defensive be-haviors (flight, freezing, defensive threat vocalizations, anddefensive attack) as a function of distance between an ani-mal and a threat stimulus, in situations where escape isavailable or not available. Based on responses of wild R.norvegicus to an oncoming experimenter in the FDTB

ally more similar to those of wild-trapped wildrats than to those of laboratory rats, suggestingthat the wild-laboratory rat difference is largelygenetic (Blanchard DC, 1997).

ANXIETY/DEFENSE TEST BATTERY

What the F/DTB does not do very well is toelicit risk assessment. It uses a very embodiedthreat stimulus, the human experimenter, andfocuses largely on the specific defenses offlight, freezing, and defensive threat/attack.In contrast, the Anxiety/Defense Test Battery(A/DTB) was devised specifically to elicit andpermit measurement of risk assessment,with focuses on stretch/attend and stretch/approach behaviors and on reductions of nor-mal activities in response to the presence ofthreat. The term anxiety in the name of thebattery reflects that risk assessment activitiesare extremely close to specific behaviors as-sociated with generalized anxiety disorder("vigilance and scanning" [American Psychi-atric Association, 1987]) and that anxiety hastraditionally been regarded as associated withambiguity of threat and danger cues ratherthan with the clear presence of danger (Freud,1930; Estes and Skinner, 1941). Various tests

Page 358: 0195162854

Chapter 31. Antipredator Defense 341

were devised to enable measurement of riskassessment and also were used to evaluate theeffects of various pharmacological agents onthis measure. The most often used test fromthe original A/DTB is a test of responsivity tocat odors, with a parallel test involving a livecat as the stimulus.

Cat Odor and Cat Exposure TestsBecause of the difficulties of keeping a live catin a laboratory setting, cat odor tests havecome to be more commonly used than cat ex-posure. The cat odor test involves a 1 meterlong alley with a cat odor stimulus (obtainedby rubbing a cloth on a cat) at one end, inwhich avoidance of the odor stimulus, andrisk assessment are the major measures. If ahide box is added, cat odor (from a cat collar)elicits hiding as well as avoidance and risk as-sessment. A single 10 minute exposure to odoris sufficient to produce conditioned defen-siveness in the situation, 24 hours later (Mc-Gregor et al., 2002). Newton Cameras (per-sonal communication) has further modifiedthis test to elicit strong and prolonged risk as-sessment in a situation in which a cat was pre-viously encountered, and he is using this pro-longed risk assessment to investigate the neuralsystems involved in this specific behavior.

RELEVANCE TOUNDERSTANDING EMOTION

Tests of defensive behavior are relevant toan understanding of emotional responses todanger on a behavioral and neural level. Be-haviorally, they outline a system of innate,unconditioned responses to various types ofthreat that may also be conditioned to ap-propriate stimuli in a single, brief exposure(Blanchard RJ et al., 200ib; Blanchard DC etal., 2003b; Dielenberg et al., 2001; McGregoret al., 2002). Although the status of these re-sponses as innate has not been researched inhumans, recent scenario studies (BlanchardDC et al., 200Ib) suggest that human re-

sponsivity to threat includes all of the de-fenses outlined in rats, and more, and thatthese conserved defensive behaviors occur inmuch the same situations as those analyzedfor rats.

In neural investigations, Canteras and hiscolleagues used combinations of c-fos, tract-tracing, and lesioning techniques to outline aset of potential neural circuits involved in re-sponsivity to a cat (Canteras, 2002). They sug-gested the importance of particular hypotha-lamic structures in responsivity to a cat.Lesions of one such area, the dorsal premam-millary nucleus, dramatically reduces reac-tions to a cat or cat odor (Canteras et al., 1997;Blanchard DC et al., 2003c) but not to footshock (Blanchard DC et al., 2003c). The dor-sal premammillary nucleus, in turn, projectsboth directly and indirectly to the periaque-ductal gray, an area in which a number of de-fensive behaviors are differentially elicitableby electrical or excitatory amino acid stimu-lation (Depaulis and Bandler, 1991). Thus, thecircuitry underlying many defensive behav-iors provides justification for their conceptu-alization as differentiable entities rather thanas behaviorally equivalent components of ansingle motivational system.

Beginning with rat defensive behaviors,and later with similar defensive behaviors inmice (the Mouse Defense Test Battery, whichincorporates features of both the F/FDTB andthe A/FDTB), a body of evidence has beengathered that suggests particular defensive be-haviors may show greater or lesser responsiv-ity to drugs effective against particular anxietydisorders. Risk assessment and defensive threat/attack respond selectively to drugs effectiveagainst generalized anxiety disorder, whereasantipanic drugs reduce flight, propanic drugs en-hance flight, and drugs with no effect on panicalso have no effect on flight (Blanchard DC etal., 2001A, 2003A). These findings, in combina-tion with systems analyses of defense, raise thepossibility of much more specific informationon the brain substrate of defensive behaviors,suggesting they may facilitate precision and se-

Page 359: 0195162854

342 DEFENSE AND SOCIAL BEHAVIOR

lectivity in the design of physiologically actingtreatments for psychopathologies.

Defensive behaviors also respond to, andchange with, experience. Their strong andrapid conditioning to some types of stimuli as-sociated with threat may be a factor in the eti-ology of a number of different anxiety syn-dromes. Because conditioning of defensivebehaviors, with the exception of freezing, isonly beginning to be investigated, it is notclear how these phenomena might be relatedto the development of threat-related learning(either "normal" or "abnormal") or how ex-perience-based therapies might best be usedto modulate these phenomena. Informationon conditioning of defensive behaviors usingboth full and partial predator stimuli and con-ditioned stimuli of different types, may be use-ful in understanding behavioral differences inhuman anxiety disorders and the possibilitiesof experiential as well as drug treatments forthese conditions.

REFERENCES

Adamec RE and Shallow T (1993) Lasting effects on ro-dent. Anxiety of a single exposure to a cat. Physiol-ogy and Behavior 54:101-109.

American Psychiatric Association (1987) DSM-HIR: Di-agnostic and statistical manual of mental disorders,3rd edition (revised). Washington, D.C.: The Asso-ciation.

Blanchard DC (1997) Stimulus and environmental con-trol of defensive behaviors. In: The functional be-haviorism of Robert C. Bolles: Learning, motivationand cognition. (Bouton M and Fanselow M, eds.),pp. 283-305. Washington, D.C.: American Psycho-logical Association.

Blanchard DC, Blanchard RJ, Rodgers RJ (1991) Risk as-sessment and animal models of anxiety. In: Animalmodels in psychopharmacology (Olivier B, Mos J,Slangen JL, eds.), pp. 117-134. Basel: BirkhauserVerlag AG.

Blanchard DC, Griebel G, Blanchard RJ (200la) Mousedefensive behaviors: Pharmacological and behav-ioral assays for anxiety and panic. Neuroscience andBiobehavioral Reviews 25:205-218.

Blanchard DC, Griebel G, Blanchard RJ (2003a) The

mouse defense test battery: Pharmacological andbehavioral assays for anxiety and panic. EuropeanJournal of Psychology 463:97-116.

Blanchard DC, Hynd AL, Minke KA, Blanchard RJ(200 Ib) Human defensive behaviors to threat sce-narios show parallels to fear- and anxiety-relateddefense patterns of nonhuman mammals. Neuro-science and Biobehavioral Reviews 25:761-770.

Blanchard DC, Lee EMC, Williams G, Blanchard RJ(1981) Taming of Rattus norvegicus by lesions ofthe mesencephalic central gray. Physiological Psy-chology 9:157-163.

Blanchard DC, Li C-I, Hubbard D, Markham C, YangM, Takahashi LK, Blanchard RJ (2003c) Dorsal pre-mammillary nucleus differentially modulates de-fensive behaviors induced by different threat stim-uli. Neuroscience Letters 345:145-148.

Blanchard DC, Markham C, Yang M, Hubbard D,Madarang E, Blanchard RJ (2003b) Failure to pro-duce conditioning with low-dose TMT, or, cat fe-ces, as unconditioned stimuli. Behavioral Neuro-science 117:360-368.

Blanchard DC, Popova NK, Plyusnina IZ, Velichko IV,Campbell D, Blanchard RJ, Nikulina J, Nikulina EM(1994) Defensive behaviors of "wild-type" and "do-mesticated" wild rats in a fear/defense test battery.Aggressive Behavior 20:387-398.

Blanchard DC, Rodgers RJ, Hori K, Hendrie CA, Blan-chard RJ (1989) Attenuation of defensive threat andattack in wild rats (Rattus rattus) by benzodi-azepines. Psychopharmacology 97:392-401.

Blanchard RJ and Blanchard DC (1969) Crouching as anindex of fear. Journal of Comparative PhysiologicalPsychology 67:370-375.

Blanchard RJ and Blanchard DC (1989) Anti-predator de-fensive behaviors in a visible burrow system. Jour-nal of Comparative Psychology 103:70-82.

Blanchard RJ, Blanchard DC, Agullana R, Weiss SM(1991) Twenty-two kHz alarm cries to presentationof a predator, by laboratory rats living in visi-ble burrow systems. Physiology and Behavior 50:967-972.

Blanchard RJ, Blanchard DC, Hori K (1989) Ethoexper-imental approaches to the study of defensive be-havior. In: Ethoexperimental approaches to thestudy of behavior (Blanchard RJ, Brain PF, Blan-chard DC, Parmigiani S, eds.), pp. 114-136. Dor-drecht: Kluwer Academic Publishers.

Blanchard RJ, Blanchard DC, Weiss SM, Mayer S (1990)Effects of ethanol and diazepam on reactivity topredatory odors. Pharmacology Biochemistry andBehavior 35:775-780.

Blanchard RJ, Dulloog L, Markham C, Nishimura O,Compton JN, Jun A, Han C, Blanchard DC (200la)Sexual and aggressive interactions in a visible bur-

Page 360: 0195162854

Chapter 31. Antipredator Defense 343

row system with provisioned burrows. Physiologyand Behavior 72:245-254.

Blanchard RJ, Kleinschmidt CF, Fukunaga-Stinson C,Blanchard DC (1980) Defensive attack behavior inmale and female rats. Animal Learning and Behav-ior 8:177-183.

Blanchard RJ, Yang M, Li C-I, Garvacio A, Blanchard DC(200 ib) Cue and context conditioning of defensivebehaviors to cat odor stimuli. Neuroscience andBiobehavioral Reviews 26:587-595.

Cameras NS (2002) The medial hypothalamic defensivesystem: Hodological organization and functionalimplications. Pharmacology Biochemistry and Be-havior 71:481-491.

Cameras NS, Chiavegatto S, Valle LE, Swanson LW(1997) Severe reduction of rat defensive behavior toa predator by discrete hypothalamic chemical le-sions. Brain Research Bulletin 44:297-305.

Curti MW (1935) Native responses of white rats in thepresence of cats. Psychological Monographs 46:76-98.

Depaulis A and Bandler R (1991) The midbrain peri-aqueductal grey matter: Functional, anatomical andimmunohistochemical organization. NATO ASI Se-ries A Vol 213. New York: Plenum.

Dielenberg RA, Arnold JC, McGregor IS (1999) Low-dose midazolam attenuates predatory odor avoid-ance in rats. Pharmacology Biochemistry and Be-havior 62:197-201.

Dielenberg RA, Carrive P, McGregor IS (2001) The car-diovascular and behavioral response to cat odor inrats: Unconditioned and conditioned effects. BrainResearch 897:228-237.

Estes WK, Skinner BF (1941) Some quantitative proper-ties of anxiety. Journal of Experimental Psychology29:390-400.

Freud S (1930) Inhibitions, symptoms, and anxiety. Lon-don: Hogarth Press.

Grant EC (1963) An analysis of the social behaviour ofthe male laboratory rat. Behaviour 21:260-281.

Grant EC and Chance MRA (1958) Rank order in cagedrats. Animal Behavior 6:183-194.

Grant EC and MacKintosh JH (1963) A comparison ofthe social postures of some common laboratory ro-dents. Behaviour 21:246-259.

Fanselow MS (1980) Conditioned and unconditionalcomponents of post-shock freezing. Pavlovian Jour-nal of Biological Science 15:177-182.

Fanselow MS (1994) Neural organization of the defen-

sive behavior system responsible for fear. Psycho-nomic Bulletin and Review 1:429-438.

Fanselow MS and Lester LS (1988) A functional behav-ioristic approach to aversively motivated behavior:Predatory imminence as a determinant of the to-pography of defensive behavior. In: Evolution andlearning (Bolles RC and Beecher MD, eds.), pp.185-211. Hillsdale, NJ: Erlbaum.

Griebel G, Blanchard DC, Jung A, Blanchard RJ (1995)A model of a antipredator defense in Swiss-Webstermice: Effects of benzodiazepine receptor ligandswith different intrinsic activities. Behavioural Phar-macology 6:732-745.

McGregor IS, Schrama L, Ambermoon P, DielenbergRA (2002) Not all 'predator odours' are equal: Catodour but not 2,4,5 trimethylthiazoline (TMT; foxodour) elicits specific defensive behaviours in rats.Behavioral Brain Research 129:1-16.

Pinel JPJ, Mana M, Ward J'AA (1989) Stretched-approach sequences directed at a localized shocksource by Rattus norvegicus. Journal of Compara-tive Psychology 103:140-148.

Pinel JPJ and Treit D (1978) Burying as a defensive re-sponse in rats. Journal of Comparative and Physio-logical Psychology 92:708-712.

Pinel JPJ and Treit D (1979) Conditioned defensive bury-ing in rats: Availability of burying materials. AnimalLearning and Behavior 7:392-396.

Rodgers RJ (1997) Animal models of 'anxiety': Wherenext? Behavioral Pharmacology 8:477-496.

Stone CP (1932) Wildness and savageness in rats of dif-ferent strains. In: Studies in the dynamics of behav-ior (Lashley KS, ed.), pp. 3-55. Chicago: Universityof Chicago Press.

Takahashi LK and Blanchard RJ (1982) Attack and de-fense in laboratory and wild Norway and black rats.Behavioral Processes 7:49-62.

Wilkie DM, MacLennan AJ, Pinel JP (1979) Rat defen-sive behavior: Burying noxious food. Journal of theExperimental Analysis of Behavior 31:299-306.

Willner P (1991) Behavioural models in psychopharma-cology: Theoretical, industrial and clinical perspec-tives. Cambridge: Cambridge University Press.

Yerkes RM (1913) The heredity of savageness and wild-ness in rats. Journal of Animal Behavior 3:286-296.

Zangrossi H and File SE (1992) Behavioral consequencesin animal tests of anxiety and exploration of exposureto cat odor. Brain Research Bulletin 29:381-388.

Page 361: 0195162854

Aggressive, Defensive, andSubmissive Behavior 32KLAUS A. MICZEK AND SIETSE F. DE BOER

The statements that "rats have very little so-cial life" and "are not particularly influencedby each other's action" in the previous Hand-book of Psychological Research on the Rat (Munn1950) has been corrected. Some 40 years ago,Barnett (1963) gave a succinct account of thesocial interactions of wild rats (Rattus norvegi-cus) in the classic work The Rat: A Study in Be-havior. When food supplies, nesting opportu-nities, and infectious and predatory pressuresare favorable, wild rats crowd together incolonies that may number many hundreds.

Nevertheless, life in the colony is not sim-ple. Colonies crash periodically due to intensesocial conflict despite conducive environmen-tal conditions. Usually, one adult male ratdominates a small group of females and youngrats by defending the region around their feed-ing and nest sites from intruders, and, in thissense, Barnett defined the region that is de-fended as territory. Outside the territories,neutral areas exist where fighting is minimaland avoidance occurs. In general, a colonycomprises a number of territories and neutralareas. Wild resident rats patrol their territoryand mark these landmarks by urine deposition(Eibl-Eibesfeldt, 1950; Telle, 1966). Althoughit is mainly the dominant male who drivesaway male intruders, lactating female rats de-fend their nest, against both males and fe-males. The forces of cohesion and dispersalare apparent in various developmental and re-productive phases of the rat: It is instructiveto see rats huddle and sleep together in agroup and to engage in chases, threats, attackleaps, and bites, prompting escapes and even-

tual emigration of the loser from the group,particularly when there are definite bound-aries to their territory such as under captiveconditions. Barnett (1963) contrasted the so-cial and aggressive interactions among feralrats (R. norvegicus) with those of laboratoryrats, characterizing the latter as being "un-likely to attack conspecifics with the vigor dis-played by wild rats" and to fail "to behave ina normal way."

Subsequent analyses of the salient socialsignals, aggressive acts, and postures duringsituations of conflict in small breedingcolonies of laboratory rats have found the dif-ferences in aggressive behavior between lab-oratory and feral rats to be mostly in degreerather than in kind (Grant and Mackintosh,1963; Luciano and Lore, 1975; Zook andAdams, 1975; Blanchard et al., 1977; Miczek,1979; Boice, 1981; de Boer and Koolhaas,2003) (Fig. 32-1). As among feral rats(Steiniger, 1950), within breeding colonies oflaboratory rats, a dominant, or alpha, rat typ-ically is defined by prevailing most often overrival, or beta, males and over subordinate, oromega, animals in aggressive confrontations.Even in small breeding colonies of laboratoryrats that are provisioned with clumpedsources for food and water, subordinate mem-bers need to be rescued periodically to ensuretheir survival (Blanchard et al., 1985). Re-peated conflict in unstable social groups of lab-oratory rats, as for feral rats (Calhoun, 1948),increases the risk of injuries, compromises theimmune system, diverts energy from repro-ductive activities and foraging, disrupts circa-

344

Page 362: 0195162854

Chapter 32. Aggressive, Defensive, and Submissive Behavior 345

Figure 32-1. Sonogram of 50 to 60 and 20 to 25 kHz vo-calizations during a confrontation between a resident andan intruder rat.

dian and physiological rhythms, places pro-longed demands on endocrine functions thatresult in gonadal atrophy and adrenal hyper-trophy, and ultimately shortens the life span(Fleshner et al., 1989; Stefanski, 2001). Notsurprisingly, social instability and/or chronicsubordination in rats and other animals is usedas a chronic social stress model in fundamen-tal research of the mechanisms underlyingstress pathology in humans (see Chapter 22.Koolhaas et al. in this book). Intermittent ex-posure to aggressive behavior that results inbrief episodes of social defeat stress sensitizesintruder rats, resulting in their increased drugself-administration (Miczek and Mutschler,1996; Covington and Miczek, 2001).

Aggressive behavior, although infre-quent, is part of life in a colony, particularlyduring its formation, and these interactionswithin the group are referred to as dominanceor within-group aggressive behavior. As in theirferal counterparts, the most potent trigger foraggressive behavior in resident male labora-tory rats is the intrusion by an unfamiliar adultmale, and these interactions are called resi-dent-intruder aggression. In some sense, the ag-gression directed toward an intruder can be re-ferred to as territorial, because it is more likelyin marked surroundings than in unfamiliar lo-cales. The probability of aggressive behaviorincreases when the resident male cohabitateswith a female, although the female does nothave to be present during the actual con-

frontation (Barnett et al., 1968; Flannelly andLore, 1977). Resident-intruder confrontationstypically occur between the male resident anda male intruder or, less frequently, between afemale resident and a female intruder. Ag-gressive behavior by a female resident is morelikely in the initial postpartum period, at whichtime the dam attacks both male and femaleintruders (maternal aggression). Prior positive(winning) or negative (losing) aggressive ex-periences promote the subsequent display ofoffensive or defensive/submissive aggressivebehavior, respectively.

PROVOCATIVE SIGNALS

When the resident has been instigated by amale intruder that is protected from attacks,the intensity and frequency of subsequent ag-gressive behavior may escalate to high levels.This instigation results from exposure toprovocative olfactory, visual, auditory, andtactile signals originating from the intrudingopponent and has been attributed to increasedaggressive arousal (Potegal, 1992).

The initial contact usually occurs via ol-factory signals (pheromones) that provide in-formation about the rat's sex, age, reproduc-tive status, and recent nutritional history andother relevant events. Both the resident andthe intruder visibly move their whiskers, in-dicating sniffing, and elongate their neck in astretched attend posture. The resident and in-truder explore each other with nasonasal con-tact (Schnauzenkontrolle). This special type ofsniffing and anogenital contact or inspectionserve to identify the olfactory signature of theopponent. Pheromones that convey informa-tion about the breeding status of the opponentguide subsequent interactions with the in-truder. Intruders that are mature breedingmales are attacked most readily, whereas at-tacks toward juvenile males or weanlings areless likely. The duration of actively exploringthe intruder can serve as a quantitative indexof social memory of previous encounters.

Page 363: 0195162854

346 DEFENSE AND SOCIAL BEHAVIOR

After olfactory identification, the residentmale may crawl under the intruder, who inturn freezes. Occasionally, the intruder can beseen to crawl under the resident. This type oftactile stimulation may inhibit further escala-tion to fighting, although there is evidence forfacilitative effects of tactile contact on initiat-ing attack behavior. Walking over represents asecond type of tactile signal, sometimes ac-companied by the deposition of urine.

In the minutes leading up to intense at-tack bites, the resident rat emits brief pulsesof ultrasonic vocalizations in the 50 kHz rangethat may reflect high excitement. During thispreattack phase, tooth-chattering frequentlyoccurs as well. Different types of ultrasonicvocalizations are emitted by the intruder, al-though these calls are more prominent in laterphases of the aggressive encounters. The in-truder's vocalizations are more prolonged andmonotonous and have most of their energy inthe 20 to 25 kHz range (Fig. 32-1).

INITIATION OF AGGRESSIVE BOUT

The attack bite is frequently preceded by al-logrooming, especially of the neck region of theintruder; this is also referred to as aggressiveneck grooming (Fig. 32-2). Slow motion videoanalysis reveals that the resident seizes foldsof neck skin, while the intruder remains im-mobile in a crouch posture. Any sudden move-ment by the intruder triggers a bite by theresident, often accompanied by kicking move-ments of the rear legs. The probability that ag-gressive neck grooming leads to subsequentattack exceeds chance significantly, but it isnot certain.

A potent trigger for the resident's attackis rapid locomotion by the intruder, andthis locomotion may occur in the form of anescape. Even when the experimenter hasarranged a large area for the resident rats toestablish as a colony and for the intruder toexplore, the captive nature of the environ-ment is bound to alter the behavior of the in-

Figure 32-2. Allogrooming (aggressive neck grooming) ofthe intruder rat by an aggressive resident.

truder. Genuine flight will be less likely andmore passive coping strategies will be evidentin the form of crouch postures (Fig. 32-3). Allfour feet are planted on the substrate and theintruder crouches while completely immo-bile, occasionally moving the head andwhiskers slightly.

The resident displays the sideways threat(Fig. 32-3), sometimes referred to as lateral at-tack, by arching its back, extending the rearlegs, with clear signs of piloerection (Im-poniergehabe'). The orientation of the residentis mostly in a right angle or in parallel to theintruder, accompanied by kicking of the rearleg that is closest to the intruder. The residentmay move toward and away from the in-truder while maintaining the sideways threatposture, and these movements may representbehavioral ambivalence. When the residentengages in the sideways threat more persist-ently, it may actually encircle the intruder,which assumes either the crouch posture orthe defensive upright posture. This defensiveupright posture by the intruder is also dis-played in reaction to the aggressive upright pos-ture of the resident, in which the rat stands onits hind legs in a half-erect posture. Fre-quently, both the resident and the intruderface each other in this upright position withvertical movements of the forepaws (mutual

Page 364: 0195162854

Chapter 32. Aggressive, Defensive, and Submissive Behavior 347

Figure 32-3. Resident rat (right) displaying the sidewaysdireat posture toward an intruder (left) in the defensive up-right posture.

upright posture) (Fig. 32-4). Piloerection andteeth-chattering are visible signs of intensesympathetic motor activity during the side-ways and upright threat postures and may beinterpreted to signify high arousal.

When the intruder moves away, it is fol-lowed or pursued by the resident who keepson sniffing the intruder and engages in neckgrooming, once the intruder arrests in the for-

ward motion. Rapid pursuits or chases arecharacteristically seen at the very beginning ofa fight or its termination.

AGGRESSIVE BOUT

The sine qua non criterion element of an ag-gressive encounter is the attack bite (Fig. 32-5).As evidenced by analysis of wounding pat-terns, the bites are most often directed towardthe neck and back region of the opponent andinclude quick closures of the jaws puncturingthe skin. It is possible that multiple bites aredelivered in rapid succession. In the case of in-tense biting and concurrent evasive action bythe opponent, the shearing action of the bitecan result in lacerations. This type of injury ispossible due to the kicking movements by theresident's rear legs while rapid jaw closuresoccur.

In its most intense form, the attack biteis preceded by an attack jump. The residentleaps at the intruder who typically attempts toescape. During the attack jump, the resident'slegs are completely off the ground, and theresident lands on the intruder's back, whichin turn evades and rapidly assumes a supineposture.

In the sequence of aggressive acts, thebite is most often preceded by a sidewaysthreat posture and followed by an aggressiveposture, sometimes referred to as "pinning" or

Figure 32-4. Mutual upright posture.Figure 32-5. Attack bite by a resident male rat and evasiveaction by the intruder rat.

Page 365: 0195162854

348 DEFENSE AND SOCIAL BEHAVIOR

"on top" (Fig. 32-6). When the resident as-sumes the aggressive posture, it bends overthe intruder, which concurrently displaysthe submissive supine posture. The anglebetween the resident's aggressive postureand the intruder's supine posture may be ei-ther orthogonal or in parallel. These pos-tures may be maintained for a few secondsor extend for minutes. By assuming thesupine posture, the intruder prevents accessto the neck and back, which are the primarytargets for the resident's bites. During thedisplay of the full aggressive posture, the res-ident rat continues to show piloerection,tooth-chattering, and stiff forelimbs, oftenattempting to reach the back of the neck asa target for bites. When successfully apply-ing bites, the jaws are seen to be closed re-peatedly, while the intruder is completelylimp.

Aggressive bouts are typically composedof 5 to 10 acts and postures, which impliesthat the resident loops through several cyclesof two to four behavioral elements, com-prising pursuit, sideways threat, attack biteand aggressive postures. In order to be con-sidered as part of an aggressive bout, theseacts and postures follow each other withinca. 6 seconds. The duration of a bout is quitevariable, but usually does not extend formore than 30 seconds.

Figure 32-6. A resident rat displaying the aggressive pos-ture above an intruder in the submissive-supine posture.

TERMINATION OFAGGRESSIVE BOUT

Termination of the aggressive bout comesabout by the intruder's escape, although thisis less successful under captive laboratory con-ditions. More often, the resident stops dis-playing the aggressive posture and movesaway, while the intruder maintains the sub-missive supine posture, sometimes for min-utes after the resident rat has departed. Thecontinued display of the supine posture by theintruder and the display of the crouch posture,with minimal movements, are the most fre-quent behavioral elements for terminating at-tacks and aggressive postures by the residentunder laboratory conditions. Michael Chance(1962) also described an upright posture,termed "sensory cut-off," in which the in-truder orients away from the resident rat, andthis posture renders attacks by the residentless likely.

Whether the increasingly passive modeof behavior by the intruder is the primary de-terminant for the termination of an aggressivebout or the inhibition of aggressive behaviororiginates within the aggressive resident re-mains to be distinguished. It is evident that in-active intruder rats that emit 22 kHz ultra-sonic vocalizations are less likely to beattacked by the resident than are active in-truders. At the same time, it is evident thatthe duration and composition of aggressivebursts are predictable parameters that may bebased on endogenous control mechanisms.

SEQUENTIAL STRUCTURE

The salient acts and postures displayed by arat when confronting an opponent are organ-ized in terms of time and sequence. The tem-poral and sequential structure of aggressivebehavior becomes apparent when the proba-bility of transition from one act to the nextone is analyzed with such tools as the lag se-

Page 366: 0195162854

Chapter 32. Aggressive, Defensive, and Submissive Behavior 349

quential analysis, cluster analysis, and log sur-vivor analysis (van der Poel et al., 1989;Miczek et al., 2002). The probability of a pur-suit leading to a sideways threat is more thantwice as likely as chance, and the probabilityof a sideways threat being followed by an at-tack bite is even higher (Miczek et al., 1989,1992) (Fig. 32-7).

This high-probability sequence is charac-teristic of aggressive behavior for a particularstrain of laboratory rats, and it can serve as atemplate for comparison to detect potentiallydeviant and excessive types of aggressive be-havior. Like the operational definitions of actsand postures, the temporal and sequential pat-terns of these behaviors are species-normativecharacteristics amenable to quantitative analysis.

LABORATORY RESIDENT-INTRUDER TEST

The display of offensive and defensive ag-gressive behaviors during a social conflict be-tween a resident and an intruder rat can beevoked within the laboratory using theresident-intruder aggression test (i.e., Olivier,1977; Olivier et al., 1994; Miczek, 1979; Kool-haas et al., 1980). In this test, male and femalerats are typically housed in pairs with matesof the same strain for 21 days, usually in larger

cages with unrestricted supply of food and wa-ter and objects for marking. Males or femalesare selected for consistent attack behaviorthrough the following procedure. After re-moval of the cagemate and pups, a stimulusanimal of the same strain and sex, usually oflesser weight and without a history of fight-ing, is introduced for a confrontation betweenthe resident and the intruder. The character-istics of the intruder animal are closely definedin terms of its age, size, and behavioral historyto provide a standard stimulus. During the ini-tial encounters, the resident-intruder con-frontation is terminated when (1) the residentrat has delivered 10 attack bites, (2) the in-truder displays the supine posture for 5 con-secutive seconds and emits ultrasonic vocal-izations, or (3) 5 minutes has elapsed.Typically, intruder rats display submissionwithin 90 seconds after being attacked. Resi-dent rats engage in distinctive species-typicalagonistic behaviors as described earlier, in-cluding pursuit, threats, and attacks, whereasintruders show escape and defensive behav-ior. In addition, intruder rats emit ultrasonicdistress vocalizations (Olivier, 1981; Thomaset al., 1983; van der Poel and Miczek, 1991).With increasing experience, the latency to thefirst attack by resident rats confronting an in-truder becomes very short and is less inform-ative as an index of readiness to display this

Figure 32-7. Lag sequential analysisof sideways threat, attack bite, andaggressive posture. The probabilityof each specific behavior following(lag +1, +2, +3, . . .) or preceding(lag —1, —2, — 3,...) as the first, sec-ond, third, fourth, or fifth element toanother specific behavioral elementis shown by the vertical bars. The ex-pected level of a random sequence isshown by the stippled horizontalbands. (Adapted from Miczek et al.,1992.)

Page 367: 0195162854

350 DEFENSE AND SOCIAL BEHAVIOR

behavior. During the encounter with the in-truder, the full range of behavioral elementsdisplayed by both animals is videotaped andsubsequently analyzed in detail using com-puterized behavioral recording and analysissystems. By recording the frequencies, dura-tions, latencies, and temporal and sequentialpatterns of all of the observed behavioral actsand postures, a detailed quantitative picture(ethogram and/or sequential structure) of ag-onistic behavior is obtained.

DEVELOPMENT OF PATHOLOGICALOR DEVIANT FORMS OF

RESIDENT-INTRUDER AGGRESSION

A considerable part of our current knowledgeof the ethology, pharmacology, and neurobi-ology of normal and functional forms ofhuman aggression is based on experimentalresident-intruder aggression in rats and otheranimals. Despite this wealth of data and pub-lications on aggression research in the rat, thesocial and neural determinants of pathologi-cal or deviant forms of human aggression (i.e.,impulsive violence) remain poorly under-stood. One important reason for this gap inour knowledge is the lack of good and rele-vant animal models of pathological aggres-sion. Ideally, such models should demonstrateexcessive, injurious, and impulsive aggressivebehavior that exceeds and/or deviates fromnormal species-typical levels or patterns(Miczek et al, 2002; de Boer and Koolhaas,2003). To date, in virtually all laboratory in-bred or outbred rat strains, the intensity andvariation of aggressive behavioral traits havebeen dramatically compromised as a result ofselection and inbreeding during the course ofthe domestication process (de Boer et al.,2003). Consequently, to obtain appreciablelevels of aggression in these placid and docilelaboratory strains, several procedural manip-ulations have been used such as prolonged so-cial isolation, brief social provocations, appli-cation of aversive stimuli, electrical brain

stimulation, administration of pharmacologi-cal agents, and, more recently, deletions ofspecific genes. Although these experimentallyheightened forms of aggressive behavior mayto some extent resemble more intense formscompared with their species-typical rates ofaggression, they may still fall into the norma-tive range compared with the patterns andlevels of their wild ancestors. Indeed, higherlevels and wider ranges of spontaneous in-traspecific aggression are encountered in feralor seminatural populations of rats comparedwith their laboratory-bred conspecifics (deBoer et al., 2003). Therefore, an increase inthe intensity of aggression is just one compo-nent of pathological behavior. Productive andrelevant animal models of pathological formsof aggression should demonstrate intense andinjurious aggression that exceeds normalspecies-typical levels and patterns—that is,forms of aggressive behavior that are nolonger subject to inhibitory control and havelost their function in social communication.

Loss of the social communicative natureof the aggressive interaction may be expressedin (1) the disappearance of the normal investi-gatory and threatening sequence of acts andpostures, (2) persistence in the attack-bitingmode even though the intruder displays thesubmissive supine posture, (3) severe wound-ing and eventually death of the intruder if notstopped by the experimenter, and (4) loss of theability to distinguish male from female intrud-ers, resulting in attacking the latter and/ or evenanesthetized opponents. The development ofrat aggression models with these behavioral ab-normalities may expand knowledge about theneurobiology of pathological and violent formsof human aggression.

IMPLICATIONS FORNEUROBIOLOGICAL RESEARCH

A prerequisite for productive neurobiologicalinquiries into the mechanisms mediating andmodulating aggressive behavior is an ade-

Page 368: 0195162854

Chapter 32. Aggressive, Defensive, and Submissive Behavior 351

quate methodology for inducing, measuring,and analyzing the pattern of aggressive actsand postures, in their species-typical as well aspathological forms. Much of neurobiologicalaggression research is conducted in residentrats confronting an intruder, primarily be-cause the neuroanatomy, neurochemistry,and neuropharmacology of rats continue tobe elucidated with all molecular and cellulartools of neuroscience. Among the most prom-ising lines of inquiry are those that focused ontargeting subtypes of serotonergic, GABAer-gic, glutamatergic, dopaminergic, and severalneuropeptidergic receptors and their genes aspotential targets for pharmacotherapeutic in-terventions (Miczek et al., 2002). Rats have af-forded the opportunity to measure aminergicactivity during the initiation, execution, andtermination of aggressive bouts, as well as inanticipation of such confrontations via in vivomicrodialysis in real time (van Erp andMiczek, 2000; Ferrari et al., 2003). Thismethodology promises to be relevant to thefundamental issue of integrating serotonin de-ficiency that characterizes certain violence-prone individuals with the phasic changes inserotonin that are triggered by aggressive be-havior itself. The enduring neuroadaptivechanges that result from brief aggressiveepisodes may encompass rapid release ofamines and peptides, lead to receptor upreg-ulation and downregulation for several days,induce functional cellular activation as seenvia immediate early gene expression, and en-gender neurogenesis (Miczek et al., 2004).

ACKNOWLEDGMENTS

The author would like to thank J. Thomas Sopko for his ex-ceptional technical assistance. Preparation of this review andthe original research from our own laboratory were supportedby U.S. Public Health Service research grants AA13983 andDA02632 and grants from the Alcoholic Beverage Medical Re-search Foundation (K.A.M., P.I.).

REFERENCES

Barnett SA (1963) The rat. A study in behavior. Chicago:Aldine.

Barnett SA, Evans CS, Stoddart RC (1968) Influence offemales on conflict among wild rats. Journal ofZoology 154:391-396.

Blanchard RJ, Blanchard DC, Flannelly KJ (1985) Socialstress, mortality and aggression in colonies and bur-rowing habitats. Behavioural Processes 11:209-213.

Blanchard RJ, Blanchard DC, Takahashi T, Kelley MJ(1977) Attack and defensive behaviour in the albinorat. Animal Behaviour 25:622-634.

Boice R (1981) Behavioral comparability of wild and do-mesticated rats. Behavioral Genetics 11:545-553.

Calhoun JB (1948) Mortality and movement of brownrats (Rattus norvegicus) in artificially supersaturatedpopulations. Journal of Wildlife Management 12:167-172.

Chance MRA (1962) An interpretation of some agonis-tic postures: The role of "cut-off' acts and postures.Symposium of the Zoological Society of London8:71-89.

Covington HE III and Miczek KA (2001) Repeatedsocial-defeat stress, cocaine or morphine: effects onbehavioral sensitization and intravenous cocaineself-administration "binges." Psychopharmacology158:388-398.

de Boer SF, van der Vegt BJ, Koolhaas JM (2003) Indi-vidual variation in aggression of feral rodent strains:A standard for the genetics of aggression and vio-lence. Behavior Genetics 33:481-497.

Eibl-Eibesfeldt I (1950) Beitrage zur Biologic der Haus-und der Ahrenmaus nebst einigen Beobachtungenan anderen Nagern. Zeitschrift fur Tierpsychologie7:558-587.

Ferrari PF, van Erp AMM, Tornatzky W, Miczek KA(2003) Accumbal dopamine and serotonin in antic-ipation of the next aggressive episode in rats. Euro-pean Journal of Neuroscience 17:371-378.

Flannelly K and Lore R (1977) Observations of the sub-terranean activity of domesticated and wild rats(Rattus norvegicus): A descriptive study. Psychologi-cal Record 2:315-329.

Fleshner M, Laudenslager ML, Simons L, Maier SF(1989) Reduced serum antibodies associated withsocial defeat in rats. Physiology and Behavior45:1183-1187.

Grant EC and Mackintosh JH (1963) A comparison ofthe social postures of some common laboratory ro-dents. Behaviour 21:246-295.

Koolhaas JM, Schuurman T, Wiepkema PR (1980) Theorganization of intraspecific agonistic behaviour inthe rat. Progress in Neurobiology 15:247-268.

Luciano D and Lore R (1975) Aggression and social ex-perience in domesticated rats. Journal of Compara-tive and Physiological Psychology 88:917-923.

Miczek KA (1979) A new test for aggression in rats with-out aversive stimulation: Differential effects of

Page 369: 0195162854

352 DEFENSE AND SOCIAL BEHAVIOR

d-amphetamine and cocaine. Psychopharmacology60:253-259.

Miczek KA, Covington HE III, Nikulina EM, HammerRP Jr (2004) Aggression and defeat: persistent effectson cocaine self-administration and gene expressionin peptidergic and aminergic mesocorticolimbic cir-cuits. Neuroscience and Biobehavioral Review27:787-802.

Miczek KA, Fish EW, De Bold JF, de Almeida RMM(2002) Social and neural determinants of aggressivebehavior: Pharmacotherapeutic targets at sero-tonin, dopamine and •y-aminobutyric acid systems.Psychopharmacology 163:434-^58.

Miczek KA, Haney M, Tidey J, Vatne T, Weerts E, De-Bold JF (1989) Temporal and sequential patterns ofagonistic behavior: Effects of alcohol, anxiolyticsand psychomotor stimulants. Psychopharmacology97:149-151.

Miczek KA and Mutschler NH (1996) Activational effectsof social stress on IV cocaine self-administration inrats. Psychopharmacology 128:256-264.

Miczek KA, Weerts EM, Tornatzky W, DeBold JF,Vatne TM (1992) Alcohol and "bursts" of aggressivebehavior: Ethological analysis of individual differ-ences in rats. Psychopharmacology 107:551-563.

Munn NL (1950) Handbook of psychological researchon the rat. An introduction to animal psychology.Boston: Houghton Mifflin.

Olivier B (1977) The ventromedial hypothalamus andaggressive behaviour in rats. Aggressive Behavior3:47-56.

Olivier B (1981) Selective anti-aggressive properties ofDU 27725: Ethological analyses of intermale and ter-ritorial aggression in the male rat. Pharmacology,Biochemistry and Behavior 14-Sl:61-77.

Olivier B, Molewijk E, van Oorschot R, van der Poel G,

Zethof T, van der Heyden J, Mos J (1994) New an-imal models of anxiety. European Neuropsy-chopharmacology 4:93-102.

Potegal M (1992) Time course of aggressive arousal infemale hamsters and male rats. Behavioral andNeural Biology 58:120-124.

Stefanski V (2001) Social stress in laboratory rats: Be-havior, immune function, and tumor metastasis.Physiology and Behavior 73:385-391.

Steiniger F (1950) Beitrag zur Soziologie und sonstigenBiologic der Wanderratte. Zeitschrift fur Tierpsy-chologie 7:356-379.

Telle HJ (1966) Beitrag zur Kenntnis der Verhal-tensweise von Ratten, vergleichend dargestellt beiRattus norvegicus und Rattus rattus. Zeitschrift furangewandte Zoologie 53:129-196.

Thomas DA, Takahashi LK, Barfield RJ (1983) Analysisof ultrasonic vocalizations emitted by intrudersduring aggressive encounters among rats (Rattusnorvegicus). Journal of Comparative Biology 97:201-206.

van der Poel AM and Miczek KA (1991) Long ultrasoniccalls in male rats following mating, defeat and aver-sive stimulation: Frequency modulation and boutstructure. Behaviour 119:127-142.

van der Poel AM, Noach EJK, Miczek KA (1989) Tem-poral patterning of ultrasonic distress calls in theadult rat: Effects of morphine and benzodiazepines.Psychopharmacology 97:147-148.

van Erp AMM and Miczek KA (2000) Aggressive be-havior, increased accumbal dopamine and de-creased cortical serotonin in rats. Journal of Neuro-science 20:9320-9325.

Zook JM and Adams DB (1975) Competitive fighting inthe rat. Journal of Comparative and PhysiologicalPsychology 88:418-423.

Page 370: 0195162854

Defensive Burying

DALLAS TREIT AND JOHN J. P. PINEL

In 1970, one of the authors was developing ananimal model of petit mal epilepsy (Pinel andChorover, 1972). The rats that served as sub-jects in this study were individually housed inPlexiglas cages, which were aligned againstthe wall on a long table in the laboratory. Asit turned out, this alignment and the archaicdesign of the cages played important roles inour first observation of the conditioned de-fensive burying paradigm. The cages were ba-sically closed cubes of Plexiglas with a single,small hinged door on the front wall providingthe only access.

Given the alignment of the cages againstthe wall, the consistency of the rats' initial be-havior was visually striking: during the habit-uation period, every rat constructed a nestfrom the bedding material on the floor of itscage, and every rat located its nest at the backof the cage, as far as possible from the activi-ties of the laboratory.

The induction of the petit mal state begana few days later with the first of a series of toxicintraperitoneal injections of chlorambucil. Eachrat was removed with difficulty from its cagethrough the small door, injected with chloram-bucil, and then returned to its cage. A few hourslater, the rats seemed in reasonable health, butthe topography of the bedding on the floor ofthe cages had changed in a way that was madeblatantly obvious by the alignment of the cages.Every rat had moved the accumulation of bed-ding at the back of its cage to the front, whereit had used it to bury the entrance. It seemedthat every rat had tried to block access to itscage of the "evil injecting hand."

Defensive burying had been described inonly one previous report of aversive condi-tioning in rats, but, unfortunately, no meas-ures of the behavior were provided (Hudson,1950). Accordingly, in the mid 1970s, moti-vated by our serendipitous observation ofburying behavior, we commenced a programof research that focused first on the behavioritself and then on its applications to neurosci-entific research.

DEVELOPMENT OF DEFENSIVEBURYING PARADIGMS

The paradigm that we developed for the studyof the defensive burying response was semi-natural in two key respects (Pinel and Treit,1978). First, the floor of the test environmentwas covered with a paniculate material (usu-ally bedding material), and second, the"threatening" stimulus was always an ob-ject—rather than an ethereal stimulus such asa light or a tone. In the wild, painful stimulitypically emanate from dangerous objects,and this spatial contiguity is likely a critical fac-tor in making it easy for animals to learn torecognize sources of danger.

In our typical study of defensive burying,rats are confronted on the test day with an un-familiar wire-wrapped dowel, often referredto as a shock prod, mounted on the wall of afamiliar test chamber. When a rat contacts theshock prod, it receives a single brief shock andreflexively withdraws. After a period of im-mobility, the rat moves forward toward the

353

33

Page 371: 0195162854

354

prod, pushing and spraying the ground mate-rial with its snout and rapid, alternating move-ments of its forepaws. Figure 33-1 providesan illustration of defensive burying behavior.

In our first series of experiments (Pineland Treit, 1978), we found that defensiveburying behavior conditioned to a shock prodwas well retained after only a single shock.Rats in the experimental group were shockedonce by the shock prod and were immediatelyremoved from the chamber for 10 seconds, 5

minutes, 5 hours, 3 days, or 20 days, at whichtime they were returned to the chamber withthe shock prod still in place but disconnectedfrom the shock source. At all intervals, the ex-

perimental rats buried the shock source sig-nificantly more often than did nonshockedcontrol rats. Indeed, few shocked rats movedmaterial in any direction other than towardthe shock prod. When defensive behavior isdirected at a previously neutral object that hasbeen the source of aversive stimulation, it hasbeen termed conditioned defensive burying(Pinel and Treit, 1978).

In another experiment in the same series,two identical shock prods were mounted on

opposite walls of the test chamber, and therats received a single shock from one of them.Virtually all of the defensive burying was di-rected at the shock source rather than the

DEFENSE AND SOCIAL BEHAVIOR

identical control object. When defensiveburying is selectively directed at a source ofaversive stimulation in the presence of a sim-ilar control object, it has been termed dis-criminated defensive burying (Pinel and Treit,1983).

UNCONDITIONEDDEFENSIVE BURYING

Although defensive burying is most com-monly studied as a conditioned response, italso occurs as an unconditioned response. Forexample, in one experiment, one of four dif-ferent sources of aversive stimulation wasmounted on the wall of the test chamber: ashock source, a length of polyethylene tubing,a flashbulb, or a mousetrap. When each ex-perimental subject touched one of thesesources with a forepaw, it was to receive ashock, an air blast, a flash, or a physical blow,respectively. The results in the shock and airblast conditions were as expected: control ratsengaged in little or no burying, whereas al-most every experimental rat buried the prodor the air tube. However, in the mousetrapand flashbulb conditions, rats began buryingthe test objects before the aversive stimuluscould be administered. Habituation to the

Figure 33-1. A rat burying a wall-mounted prod from which it has justreceived a single, brief electric shock.

Page 372: 0195162854

Chapter 33. Defensive Burying 355

mousetrap or the flashbulb eliminated this un-conditioned burying (Pinel and Treit, 1983).

In another study of unconditioned bury-ing, rats sprayed bedding material over thebodies of decaying dead conspecifics but notanethetized rats or fresh corpses (Pinel et al.,1981). The authors hypothesized that this un-conditioned burying behavior was elicited bythe odor of putrescine or cadaverine, twochemical stimuli associated with decaying tis-sue. Indeed, rats buried anesthetized con-specifics or wooden dowels that had beenpretreated with putrescine or cadaverine,whereas rats rendered anosmic by intranasalinjections of zinc sulfate did not.

The first few laboratory studies of defen-sive burying established two things: that ratsenter the experimental environment with anestablished tendency to bury some objects butnot others and that they readily learn to buryany object that has been the source of aver-sive stimulation.

stricted to the stereotypical response seenwith homogeneous particulate materials suchas wood shavings or ground corncob.

Conditioned defensive burying also oc-curs in a variety of test environments. The du-ration of burying in rats decreases with in-creases in the size of the test chamber;however, it still occurs in very large chambers,where rats are not required to stay in the vicin-ity of the shock prod.

Conditioned defensive burying occurs re-liably even in two-compartment boxes, whichprovide rats with the ability to escape thecompartment that contains the shock prod(Pinel et al., 1980). The two-compartment ap-paratus is particularly useful for observing ormanipulating the active (e.g., burying) andpassive (e.g., spatial avoidance) componentsof the rats' defensive responses toward theshock prod, by opening or closing a partitionseparating the two sides of the chamber (Treitet al., 1986).

CHARACTERISTICS AND GENERALITYOF DEFENSIVE BURYING

Subsequent studies of defensive burying es-tablished its generality (Pinel and Treit, 1983).For example, burying behavior has been ob-served using a variety of bedding materials,including sand, sawdust, wood shavings,ground corncob, and even wooden blocks(Pinel and Treit, 1979). The wooden blockconditions were particularly informative be-cause the blocks were large enough to pre-clude the burying of a wall-mounted shockprod with the typical burying response (i.e.,forelimb spraying). Instead, the rats picked upblocks in their teeth and placed or threw themover the prod. In one wooden block condi-tion, the wooden blocks were all placed in apile at the opposite end of the chamber fromthe shock prod. The rats in this condition firstcarried, threw, or pushed the blocks to thevicinity of the shock prod before starting tobury it. Clearly, burying behavior is not re-

BURYING BEHAVIOR:ORGANISMIC VARIABLES

SPECIES

Burying behavior has been observed in a va-riety of different rodent species; however,there has been little systematic research inspecies other than rats. In general, rats tendto bury a shock prod for longer periods of timethan do either mice or ground squirrels, andburying is only rarely observed in gerbils andhamsters. In one study, conditioned defensiveburying of a shock prod was directly com-pared in Richardson's ground squirrels, thir-teen-lined ground squirrels, and Long-Evansrats. Defensive burying was observed in allthree rodent species, but the topography ofthe response was different and the durationwas less in the two species of ground squirrelsthan in the rats (Heynen et al., 1989). Com-parisons of conditioned defensive burying invarious rodent species are complicated by the

Page 373: 0195162854

356 DEFENSE AND SOCIAL BEHAVIOR

fact that they tend to be based on paradigmsinitially developed to produce robust buryingin rats.

STRAINS

Several studies have compared conditioneddefensive burying in different strains—in bothrats and mice. In one study of burying in dif-ferent strains of rats (Treit et al., 1980), Fisherrats were found to bury a shock prod morethan Wistar rats, which buried more oftenthan Long-Evans rats. In another study (Pareet al., 1992), Fisher and Wistar rats were foundto bury more than were Wistar-Kyoto rats.

In one comparison of burying in threestrains of mice, CF-1 mice were found to burymore often than CD-I and BALB/c mice(Treit et al., 1980).

SEX AND AGE

In rats, burying has been observed in bothsexes and a wide range of ages. There do notappear to be differences between male ratsand nulliparous female rats in the degree towhich they bury shock sources. In contrast,age has a substantial effect on burying. Treitet al. (1980) compared male rats that were 30,60, or 90 days old and found the 60-day oldrats to engage in significantly more defensiveburying.

BURYING AS A DEFENSIVERESPONSE IN THE WILD

Defensive burying has not been frequentlystudied in the wild, and burying has beenstudied in only a handful of the hundreds ofknown wild species of rodents. Indeed, theonly systematic ethological studies of defen-sive burying were conducted in groundsquirrels by Owings and Coss (1977). Theyfound that ground squirrels drive off preda-tory snakes by using the defensive buryingresponse to spray sand at them. Owings and

Coss also reported that ground squirrels usethe burying response to construct walls intheir burrows to block the movement ofsnakes. Similarly, Calhoun (1962) noted thatlower-status wild Norway rats exposed toconspecific threat plugged the entrance holesto their underground nests, and Johnston(1975) observed a male golden hamsterblocking the entrance to its chamber withwood shavings after it had been defeated bya higher-ranking male in a seminatural envi-ronment.

CONDUCTING CONDITIONEDDEFENSIVE BURYING EXPERIMENTS

SUBJECTS

Although conditioned defensive burying canbe readily observed in most rats and manyother rodents, young adults tend to displaythe most burying (Treit et al., 1980). Femaleswith litters also display particularly high lev-els (Pinel et al., 1990).

HOUSING

It is important that the subjects be reared andhoused on bedding material. Pinel et al. (1989)reared rats with no opportunity to interactwith particulate matter. When exposed to theconditioned defensive burying paradigm asadults, they attempted to bury the shocksource but their burying behavior was spo-radic, uncoordinated, and poorly directed.

HANDLING AND HABITUATION

In most aversive conditioning paradigms, it isimportant that the test environment does notinduce confounding defensive responses.Thus, in studies of conditioned defensiveburying, the subjects are typically handled for3 days and then habituated to the test box(without the shock source) for each of the next4 days (Treit and Fundytus, 1988).

Page 374: 0195162854

Chapter 33. Defensive Burying

TEST CHAMBER

Any test chamber will suffice, but burying ismost robust in small chambers, where sub-jects are forced to stay in the vicinity of theshock source. The typical test chamber is a40 X 30 X 40 cm Plexiglas chamber with a 5cm layer of bedding material, but smallerchambers result in even more burying (Pinelet al, 1980).

SHOCK SOURCE

Any aversive stimulus and source can be usedin conditioned defensive burying experi-ments, but electric shock delivered from ashock prod (two wires wrapped around awooden dowel) has been used most com-monly. The dowel is typically mounted on thewall, 2 cm above the level of the bedding (seeFig. 33-1). More complex stimuli may elicitunconditioned burying.

SHOCK PARAMETERS

All shocks in conventional conditioned defen-sive burying experiments are very brief (about0.1 second), as determined by the latency ofthe withdrawal reflex. Because there is con-siderable variability in the contacts made byvarious animals, constant current shockersshould be used, and both current flow and theanimals' reactions should be monitored. Cri-teria should be established for excludinganimals that do not experience a reasonableshock. Current intensities should be selectedwith caution. The duration of burying by ratswas shown to increase monotonically from0.5 to 10 mA (Treit and Pinel, 1983). How-ever, in some studies, the objective is notmerely to produce a lot of burying behaviorbut rather to assess the effects of various treat-ments on burying behavior. In such cases,extremely robust burying can prove to be in-sensitive (see the later discussion of anxiolyt-ics). In most conditioned defensive buryingexperiments, only one shock is delivered, but

357

in others, the shock source remains activatedduring the entire test period.

BEHAVIORAL MEASURES

All experimental sessions should be video-taped, to facilitate the detailed assessment ofthe rat's behavior. Specific measures have in-cluded the following:

• Duration of burying (i.e., total time each ratsprays bedding material toward test object)

• Frequency of burying bouts• Latency to burying• Number of cautious approach sequences• Number of contacts with the test object• Behavioral reaction to the aversive stimulus

(e.g., reaction to shock has been measuredon 4-point scale [Degroot and Treit, 2003])

• Duration of freezing behavior• Height of bedding material over the prod at

the end of the test session

DURATION OF THE TEST

Longer tests are associated with more bury-ing; however, in most studies, the tests last 10or 15 minutes.

MAINTAINING THE APPARATUS

After each animal has been removed from thechamber, the bedding material should becleaned of feces and smoothed to a uniformheight, and the shock prod should be cleanedof any moisture, dust, or debris. The entireshock circuit should also be periodically testedwith a multimeter to ensure no change in con-ductivity has occurred.

USE OF DEFENSIVE BURYINGPARADIGMS IN

NEUROSCIENCE RESEARCH

Defensive burying paradigms have been usedin neuroscience for a variety of purposes;however, they have most commonly been

Page 375: 0195162854

358 DEFENSE AND SOCIAL BEHAVIOR

used to screen anxiolytic drugs and to studythe roles of the septum, amygdala, and hip-pocampus in fear and anxiety. These four linesof research are discussed in the followingsections.

SCREENING ANXIOLYTIC DRUGS

Several studies have shown that anxiolyticdrugs (e.g., diazepam) produce a dose-de-pendent suppression of prod burying, with arelative potency that is consistent with theirclinical effectiveness in the treatment of hu-man anxiety. For reviews, see Treit andMenard (1998) and Treit et al. (2003).

The drug class specificity of the test is sen-sitive to procedural variations (Treit andMenard, 1998). For example, anxiolytics re-duce shock prod burying at intermediate, butnot at high, shock intensities. Suppression ofdefensive burying by benzodiazepine-typeanxiolytics is not secondary to a general mo-tor impairment, associative learning deficits,or analgesia; it can be blocked by benzodi-azepine receptor antagonists such as flumaze-nil (Treit et al., 2003). Conversely, drugs thatincrease anxiety in humans (i.e., anxiogenicagents such as yohimbine) increase theamount of time that rats bury a shock source.These effects of anxiolytic and anxiogenicagents on defensive burying support the viewthat burying shock sources is a "fear" reactionin rats (Treit and Menard, 1998).

One of the strengths of shock prod bury-ing as a screening test for anxiolytics is thatseveral different "fear" responses can be mea-sured within the same setting. This has provedto be particularly valuable for the detection ofserotonergic-type anxiolytics, such as bus-pirone, which have often been difficult to de-tect in other screening tests (Treit et al., 2003).In one study, for example, Treit and Fundy-tus (1988) compared the effects of chlor-diazepoxide and buspirone in a modified bury-ing test in which the shock source remainedcontinuously electrified. Both buspirone andchlordiazepoxide decreased the amount of

time rats buried the shock prod and con-comitantly increased the number of contact-induced shocks that rats received from theprod. These bidirectional effects on prodburying and on the number of prod shocksprovide convergent evidence of anxiolytictreatment effects.

THE NEURAL MECHANISMSOF FEAR AND ANXIETY

The SeptumNumerous studies have shown that ablation orpharmacological inhibition of the septum pro-duces anxiolytic effects in the shock-prod bury-ing test (for reviews, see Menard and Treit,1999; Treit and Menard, 2000). Briefly, elec-trolytic or excitotoxic lesions of the septum de-creased burying of a constantly electrified prod,without concomitant effects on general activ-ity, reactivity to handling, reactivity to shock,or shock source avoidance. A similar pattern ofeffects was produced when septal activity wasinhibited using intraseptal microinfiisions ofthe benzodiazepine-type anxiolytic midazolam(Menard and Treit, 1999), the direct actingGABAA agonist muscimol (Degroot and Treit,2003), the 5-hydroxytryptaminela (serotoninla

agonist (R)(+)-8-hydroxy-2-(di-n-propylamino)tetralin, or by both N-methyl-D-aspartate(NMDA) D(—)-2-amino-5-phosphonopentanoicacid (AP-5) and non-NMDA receptor antago-nists (6-cyano-7-nitroquinoxaline; Menard andTreit, 1999, 2000).

In most of these studies, the anxiolytic ef-fects of septal suppression were replicated inthe elevated plus-maze. In the elevated plus-maze, untrained rats avoid the open arms ofthe elevated maze and remain in the enclosedarms (Fellow et al., 1985). The combined useof the plus-maze and the defensive buryingtest for studying the neural mechanisms offear and anxiety is important for three reasons.First, the fear-inducing stimuli in the two testsare distinctly different (i.e., painful electricshock versus open elevated spaces). Second,fear reduction is primarily indicated by an in-

Page 376: 0195162854

Chapter 33. Defensive Burying

crease in a specific activity in the plus-maze(i.e., open-arm exploration) and by a decreasein a specific activity in the defensive buryingtest (i.e., shock prod burying). Thus, reduc-tions in "anxiety" seen in both tests are diffi-cult to explain in terms of nonspecific effectson general activity, arousal, pain sensitivity,or behavioral inhibition. Third, neither testinvolves an explicit memory requirement, afactor that can complicate the interpretationof drug or lesion effects in other paradigms(Treit, 1985).

The AmygdalaThe amygdala has long been implicated in fearand anxiety (e.g., Davis, 1992; LeDoux, 1996).What is the relative contribution of the amyg-dala in anxiety compared with other limbicstructures, such as the septum? To address thisquestion, Treit and colleagues compared the ef-fects of amygdala lesions with those of septal

359

lesions in the defensive burying and plus-mazetests (for a review, see Treit and Menard, 2000).As in previous studies, septal lesions decreasedshock prod burying and increased open-arm ex-ploration, without producing effects on generalactivity, reactivity to handling, reactivity toshock, or shock prod avoidance. Interestingly,however, amygdalar lesions had no effects onburying behavior or plus-maze behavior butdramatically increased shock prod contacts.This selective effect of amygdalar lesions wasfound in all experiments, across a variety of dif-ferent lesion parameters, and in the absence ofany effects on general activity or shock reac-tivity. In addition, this selective effect of amyg-dalar lesions did not appear to reflect a generaldeficit in response inhibition or passive avoid-ance because the lesioned rats avoided the openarms of the plus-maze to the same extent assham-lesion controls (Treit and Menard, 2000)(Table 33-1).

Table 33-1. Summary of Drug and Lesion Effects (see Text)

Manipulation

Injection ofanxiolytic drugs(e.g., midazolam)

Injection ofanxiogenic drugs(e.g., yohimbine)

LesionMicroinfusion of

midazolamLesionMicroinfusion of

midazolamMicroinfusion of

physostigmine(20 /*g)

Microinfusion ofmuscimol (10 ng)

Microinfusion ofcombined, subeffectivedoses of muscimol(2.5 ng) in septumand physostigmine(5 /Ag) in hippocampus

Site

Systemic (e.g.,intraperitoneally)

Systemic (e.g.,intraperitoneally)

Septum

Amygdala

Hippocampus

Septum

Septum andhippocampus

Plus-Maze

Open Arm.

Exploration

Increased

Decreased

IncreasedIncreased

No effectNo effect

Not tested

Not tested

Not tested

Shock Prod

Contacts

Increased athigh doses

No effect

No effectNo effect

IncreasedIncreased

No effect

No effect

No effect

Shock Prod

Burying

Decreased

Increased

DecreasedDecreased

No effectNo effect

Decreased

Decreased

Decreased

Page 377: 0195162854

360 DEFENSE AND SOCIAL BEHAVIOR

A possible interpretation of the effects ofamygdala lesions on defensive burying is thatthe rats could not learn or remember the as-sociation between the electric shock and theshock source. However, this interpretation isinconsistent with the prod-burying behaviorof amygdala-lesion rats, which was well di-rected and indistinguishable from sham-lesioncontrols, a finding that has been replicated inother laboratories (Treit and Menard, 2000). Inaddition, neurotoxic or reversible tetrodotoxinlesions of the amygdala did not impair the abil-ity of the rats to subsequently avoid the shockprod after a delay of several days was imposedbetween the initial shock and a retention test(Lehmann et al., 2000, 2003).

Taken together, the effects of septal andamygdalar lesions suggest that the amygdalaand septum independently control the ex-pression of different fear reactions. Subse-quent studies have reinforced this generalconclusion. Microinfusions of midazolam intothe septum increased open arm exploration inthe plus maze and decreased defensive bury-ing in the shock prod test, whereas amygdalarinfusions produced neither of these effects.Amygdalar infusions did, however, dramati-cally impair the shock prod avoidance of rats,an anxiolytic effect not found after septal in-fusions. Coinfusions of the benzodiazepine re-ceptor antagonist flumazenil blocked each ofthese specific anxiolytic effects without pro-ducing any intrinsic activity by itself. These re-sults suggest that benzodiazepine receptorsystems within the amygdala and septum dif-ferentially mediate specific fear reactions(Treit and Menard, 2000) (Table 33-1).

The HippocampusAnatomically, the septum is extensively con-nected to the hippocampus (e.g., Risold andSwanson, 1997); together they form a sub-stantial part of the limbic system. Function-ally, according to Gray's (1982) theory, theseptum and hippocampus act in concert tocontrol fear and anxiety, as evidenced in partby the correspondence between the effects of

septal or hippocampal lesions in traditionalaversive learning paradigms and the effects ofanxiolytic drugs in these same paradigms(Gray, 1982).

There is evidence that hippocampalcholinergic systems may be particularly im-portant in the modulation of anxiety. For ex-ample, increases in the fear reactions of ratshave been observed in a variety of tests afterintrahippocampal infusions of cholinergic an-tagonists (e.g., File et al., 1998). One expecta-tion, based on these antagonist studies, is thatproducing upregulation of cholinergic systems,for example, with the acetylcholinesterase in-hibitor physostigmine, might reduce anxiety.Furthermore, given the connections betweenthe septum and hippocampus, it seemed thatseptal GABAergic systems might interact withhippocampal cholinergic systems in the con-trol of anxiety.

To test these hypotheses, Degroot andTreit (2003) examined the independent andcombined effects of stimulating septalGABAergic systems and hippocampal cholin-geric systems using the defensive burying test.They found the following: (1) that a 10 ng in-fusion of muscimol into the septum produceda significant suppression of shock prod bury-ing, whereas lower doses (2.5 and 5.0 ng) didnot, (2) that burying was significantly reducedafter a 20 /xg infusion of physostigmine intothe hippocampus but not after a lower dose(5 and 10 /Jig), and (3) that the combination ofsubthreshold doses of physostigmine (5 /Ag)and muscimol (2.5 ng) significantly reducedburying. These results generally supportGray's theory and suggest that hippocampalcholinergic and septal GABAergic systems canact synergistically in the modulation of fearreactions.

CONCLUSION

Various forms of the defensive burying para-digm have proved useful in neuroscientificresearch. Because the defensive burying re-

Page 378: 0195162854

Chapter 33. Defensive Burying 361

sponse is reliable, highly directed, and specificto a variety of aversive situations, large dif-ferences in burying between experimental andcontrol subjects are typical. In addition, de-fensive burying requires no pretraining, it canbe studied as an unconditioned response, itcan be conditioned in a single trial, and theconditioned form is well retained. Finally, de-fensive burying is not restricted to a particu-lar type of aversive stimulus, test environ-ment, burying material, or species, strain, sex,or age of rodent. Most important, it changesin predictable ways to anxiolytic and anxio-genic drugs and brain lesions.

REFERENCES

CalhounJ (1962) The ecology and sociology of the Nor-way rat. Bethesda: U.S. Department of Health, Ed-ucation and Welfare.

Davis M (1992) The role of the amygdala in fear and anx-iety. Annual Review of Neuroscience 15:353-375.

Degroot A and Treit D (2003) Septal GABAergic and hip-pocampal cholinergic systems interact in the mod-ulation of anxiety. Neuroscience 117:493-501.

File SE, Gonzalez LE, Andrews N (1998) Endogenousacetylcholine in the dorsal hippocampus reducesanxiety through actions on nicotinic and mus-carinic receptors. Behavioral Neuroscience 112:352-359.

Gray JA (1982) The neuropsychology of anxiety: An en-quiry into the function of the septo-hippocampalsystem. Oxford: Oxford University Press.

Heynen AJ, Sainsbury RS, Montoya CP (1989) Cross-species responses in the defensive burying para-digm: A comparison between Long-Evans rats(Rattus norvegicus), Richardson's ground squirrels(Spermophilus richardsonii), and Thirteen-Linedground squirrels (Catellus tridecemlineatus). Jour-nal of Comparative Psychology 103:184-190.

Hudson BB (1950) One-trial learning in the domestic rat.Genetic Psychology Monographs 41:99-145.

Johnston RE (1975) Scent marking by male goldenhamsters (Mesocricetus auratus), III: Behaviorin a seminatural environment. Z Tierpsychol 37:213-221.

LeDoux J (1996) Emotional networks and motor con-trol: A fearful view. In: Progress in brain research(Holstege G, Bandler R, Saper CB, eds.), pp.437—446. Amsterdam: Elsevier Press.

Lehmann H, Treit D, Parent MB (2000) Amygdala le-

sions do not impair shock-probe avoidance reten-tion performance. Behavioral Neuroscience 114:107-116.

Lehmann H, Treit D, Parent MB (2003) Spared antero-grade memory for shock-probe fear conditioning af-ter inactivation of the amygdala. Learning andMemory 10:261-269.

Menard J and Treit D (1999) Effects of centrally admin-istered anxiolytic compounds in animal models ofanxiety. Neuroscience and Biobehavioral Reviews23:591-613.

Menard J and Treit D (2000) Intra-septal infusions of ex-citatory amino acid receptor antagonists have dif-ferent effects in two animal models of anxiety. Be-havioural Pharmacology 11:99-108.

Owings DH and Coss RG (1977) Snake mobbing byCalifornia ground squirrels: Adaptive variation andontogeny. Behavior 62:50-69.

Pare WP (1992) The performance of WKY rats on threetests of emotional behavior. Physiology and Behav-ior 51:1051-1056.

Pellow S, Chopin P, File SE, Briley M (1985) Validationof open: Closed arm entries in an elevated plus-maze as a measure of anxiety in the rat. Journal ofNeuroscience Methods 14:149-167.

Pinel JPJ and Chorover SL (1972) Inhibition of arousalof epilepsy induced by chlorambucil in rats. Nature236:232-234.

Pinel JPJ and Treit D (1978) Burying as a defensive re-sponse in rats. Journal of Comparative and Physio-logical Psychology 92:708-712.

Pinel JPJ and Treit D (1979) Conditioned defensive bury-ing in rats: Availability of burying materials. AnimalLearning and Behavior 7:392-396.

Pinel JPJ and Treit D (1983) The conditioned defensiveburying paradigm and behavioral neuroscience. In:Behavioral approaches to brain research (RobinsonT, ed.), pp. 212-234. New York: Oxford UniversityPress.

Pinel JPJ, Gorzalka BB, Ladak F (1981) Cadaverine andputrescine initiate the burial of dead conspecifics byrats. Physiology and Behavior 27:819-824.

Pinel JPJ, Petrovic DM, Jones CH (1990) Defensive bury-ing, nest relocation, and pup transport in lactatingfemale rats. The Quarterly Journal of ExperimentalPsychology 426:401-411.

Pinel JPJ, Symons LA, Christensen BK, Tees RC (1989)Development of defensive burying in Rattusnorvegicus: Experience and defensive responses.Journal of Comparative Psychology 103:359-365.

Pinel JPJ, Treit D, Ladak F, Maclennan AJ (1980) Con-ditioned defensive burying in rats free to escape. An-imal Learning and Behavior 8:477-451.

Resold PY and Swanson LW (1997) Connections of the

Page 379: 0195162854

362 DEFENSE AND SOCIAL BEHAVIOR

rat lateral septal complex. Brain Research Reviews24:115-195.

Treit D (1985) Animal models for the study of anti-anx-iety agents: A review. Neuroscience and Biobehav-ioral Reviews 9:203-222.

Treit D and Fundytus M (1988) A comparison of bus-pirone and chlordiazepoxide in the shock-probe/burying test for anxiolytics. Pharmacology Bio-chemistry and Behavior 30:1071-1075.

Treit D and Menard J (1998). Animal models of anxietyand depression. In: Neuromethods. Vol 32, In vivoneuromethods (Boulton A, Baker G, Bateson A,eds.), pp. 89-148. Totowa, NJ: Humana Press.

Treit D and Menard J (2000) The septum and anxiety.In: The behavioral neuroscience of the septal region

(Numan R, ed.), pp. 210-223. New York: Springer-Verlag Inc.

Treit D, Degroot A, Shah A (2003) Animal models ofanxiety and anxiolytic drug action. In: Handbook ofdepression and anxiety, 2nd edition (Kasper S, denBoer JA, Sitsen JMA, eds.), pp. 681-702. New York:Marcel Dekker.

Treit D, Lolordo VM, Armstrong DE (1986) The effectsof diazepam on "fear" reactions in rats are modu-lated by environmental constraints on the rat's de-fensive repertoire. Pharmacology Biochemistry andBehavior 25:561-565.

Treit D, Terlecki LJ, Pinel JPJ (1980) Conditioned de-fensive burying in rodents: Organismic variables.Bulletin of the Psychonomic Society 16:451—454.

Page 380: 0195162854

Social Learning

BENNETT G. GALEF, ]R.34

Systematic observation of free-living mam-mals and birds often reveals differences in thebehavior of species members that live in dif-ferent areas. Such geographic variation in thebehavior of chimpanzees and orangutans isparticularly well documented (Whiten et al.,1999; van Schaik et al., 2003) and is widelyknown because of the attention it has receivedin the popular press. However, before the re-cent dramatic increase in field studies of thegreat apes, it was not unreasonable to pro-pose, as did Steiniger (1950, p. 369), that "the[Norway] rat appears especially able to de-velop local traditions, more so perhaps thanother more-dosely examined mammals, pos-sibly including the anthropoids."

NORWAY RATS

Norway rats are arguably the most successful,and surely the most widely distributed, non-human mammals on Earth. Breeding popula-tions have been reported from Nome, Alaska,at 60 degrees North latitude, where rats feedon human garbage, to South Georgia Island,at 55 degrees South latitude, where tussockgrass, beetles, and ground-nesting birds pro-vide sustenance for colonies of Norway rats.

As the preceding two examples suggest,much of the success that rats enjoy resultsfrom the extraordinary range of foods thatthey are able to exploit, and as in the greatapes, much of the known variation in behav-ior in free-living Norway rats involves forag-ing behavior. Rats in West Virginia catch and

eat fingerling fish in trout hatcheries, whereasthose living on Norderoog island in the NorthSea stalk and kill ducks and sparrows. Yetother R. norvegicus living along the banks ofthe Po River in Italy dive for and feed on mol-lusks living on the bottom of the river, whiletheir fellow rats in Japan scavenge dead fishthat wash up on the seashore. Such naturallyoccurring variability in feeding behavior hasbeen the focus of most experimental studiesof social learning in the species.

PREVIEW

I begin the present brief review of the litera-ture on social influences on food choices ofNorway rats with a description of fieldworkstrongly suggesting that interactions betweenadult free-living rats and their young can de-termine which foods the young come to eat.I then describe very briefly several behavioralprocesses that have been shown in the labo-ratory to be sufficient to influence food choicein young rats. Last, I describe in somewhatgreater detail a type of social influence on rats'food preferences that has already proved to beuseful in studies of the physical substrates oflearning and memory.

FIELD OBSERVATIONSOF NORWAY RATS

Fritz Steiniger, an applied ecologist whoseprofessional interest lay in enhancing the effi-

363

Page 381: 0195162854

364 DEFENSE AND SOCIAL BEHAVIOR

ciency with which rodent pests could be ex-terminated, was the first to report difficultiesin controlling pest populations of Norway ratsusing the economically desirable method ofplacing permanent stations containing poi-soned bait in rat-infested areas (Steiniger,1950). Steiniger found that although rats ateample amounts of poison bait and died inlarge numbers when a permanent bait stationwas first introduced into their colony's terri-tory, later acceptance of the bait by colonymembers was very poor, and colonies tar-geted for extermination soon returned to theirinitial sizes.

Steiniger reported that permanent baitstations failed because young rats, born tocolony members that had survived their ini-tial contact with the poisoned bait and hadlearned to avoid eating it, refused to even tastethe bait that the adults of their colony wereavoiding.

A LABORATORY ANALOGUE

Avoidance by young wild rats of a food thatadults of their colony have learned to avoideating is a robust phenomenon that is easilyobserved in rats transferred from their natu-ral habitats to laboratory enclosures. We cap-tured adult wild rats (R. norvegicus) on garbagedumps in southern Ontario, transferred themto our laboratory, and placed them in groupsof five or six in 2 m2 enclosures that each con-tained nesting boxes and nesting materials andprovided ad libitum access to water. For 3hours each day, we offered each colony twofoods that differed in taste, smell, texture, andcolor (Galef and Clark, 1971b).

To begin a typical experiment, we intro-duced a sublethal concentration of toxin intoone of the two foods that we gave our captivesto eat daily. The rats soon learned to avoid eat-ing the poisoned food, and for weeks there-after, they avoided eating the food that hadbeen noxious, even when we gave them un-contaminated samples of it (Garcia et al., 1966).

After we had trained our colonies toavoid one of the two foods that we placed intheir enclosure each day, we waited for femalecolony members to give birth and for theiryoung to grow to weaning age. As the youngapproached independence, we started to ob-serve their colony on closed-circuit televisionthroughout daily feeding periods. When theyoung started to eat solid food, we recordedthe frequency with which they ate each of thetwo foods in their cage: one that adult colonymembers were eating and the other that theadults had learned to avoid.

We found, without exception, that wean-ing rats ate only the food that the adults oftheir colony were eating and totally avoidedthe alternative food that the adults hadlearned to avoid. Even after we removed pupsfrom their natal enclosures, housed them in-dividually, and offered them the same twofoods that had been available when they werein their colony cages, pups continued to eatonly the food that the adults of their colonyhad eaten (Galef and Clark, 1971b) (Fig. 34-1).

ANALYSIS OF THE PHENOMENON

My students and I have spent much of the past30 years determining how the food choices ofadult rats might influence those of the youngthey rear (see reviews in Galef, 1977, 1988,1996a, 1996b). Over those years, those work-ing in my laboratory and in other laboratoriesas well have discovered many different waysin which the food choices of young rats areaffected by social interactions with conspecificadults.

Prenatal EffectsFetal rats exposed to a flavor while still in theirmother's womb (through injection of a fla-vored solution into the dam's amniotic fluid)will, when grown, drink more of a solutioncontaining that flavor than will control ratsthat lack prenatal exposure to it (Smother-man, 1982). Even feeding a food with a strongodor to a female rat while she gestates a litter

Page 382: 0195162854

Chapter 34. Social Learning 365

Figure 34-1. Rat pups born into colonies trained to avoideating either diet A or diet B are offered a choice for 3 hoursper day between diet A and diet B. Abscissa shows the dayssince pups started to eat solid food; ordinate, relative fre-quency with which pups from the two types of colony atediet A, Pup diet choice while still in their natal colonies (left)and the amount of diet A eaten, as a percentage of totalamount eaten, by pups after transfer to individual cages andoffer of diet A and diet B for 9 hours per day (right). (Datafrom Galef and Clark [1971].)

suffices to enhance her postnatal preferencesof her young for the odor of that food (Hep-per, 1988).

Effects during SucklingFlavors of foods that a rat dam eats while lac-tating affect the flavor of her milk, and expo-sure to milk flavored by the foods that a lac-tating dam eats while rearing her youngaffects the food preferences of her pups atweaning (e.g., Galef and Sherry, 1973).

Effects during WeaningGalef and Clark (197la) used time-lapse videorecordings to observe each of nine wild ratpups that had ad libitum access to solid foodtake its first meal. All nine pups ate solid foodfor the first time under the same circum-stances. Each ate at the same time that anadult member of its colony was feeding,which was highly unlikely given the temporaldistribution of adult meals, and each ate at thesame place an adult was feeding, not at an al-ternative feeding site a short distance away.

Even an anesthetized adult rat placed near oneof two otherwise identical feeding sites madethat site far more attractive to pups than onewithout an adult present (Galef, 1981).

By comparing the circumstances inwhich intact and visually deprived ratsweaned, we found that intact pups use visualcues to approach adults from a distance whenselecting a place to eat solid food for the firsttime.

Effects of Snatching Food from AdultsYoung rats, like the young of many othermammalian species, seem to be especially in-terested in the particular piece of food thatsomeone else is eating. Juvenile rats will walkacross a cage floor carpeted with food pelletsand steal an identical pellet from the mouthor paws of an adult or a peer that is eating it.Young rats that have stolen a pellet of unfa-miliar food from the mouth of a conspecificsubsequently show a greater preference forthat food than do young rats that have eatenan identical food pellet taken from the floorof their cage (Galef et al., 2001).

Effects of Scent Marks and Scent TrailsWhile feeding, adult rats deposit olfactorycues both on and around a food they are eat-ing (Galef and Beck, 1985). Such residualodors attract pups and, like the physical pres-ence of an adult rat at a feeding site, causeyoung rats to prefer marked sites. Further,when an adult has finished eating and travelsback to its burrow, it deposits a scent trail thatdirects young rats seeking food to the locationat which the adult ate (Galef and Buckley,1996).

IMPLICATIONS OF REDUNDANCY

Redundancy in the behavioral processes thatsupport social influences on food choice inrats is in itself important. Such redundancysuggests that for rats, as for the honeybeesstudied by Karl von Frisch (1967), socially ac-quired information substantially increases for-

Page 383: 0195162854

366 DEFENSE AND SOCIAL BEHAVIOR

aging efficiency. Indeed, it is easy to demon-strate that, for naive rats residing in an envi-ronment where foods containing needed nu-trients are difficult to identify, the presence ofconspecifics that have already learned to se-lect an appropriate diet can make the differ-ence between life and death. Young rats thatwould have died because of an inability tolearn independently to focus their intake onthe sole protein-rich food available among acafeteria of foods available to them learnedrapidly to eat that food when caged with adultconspecifics trained to do so (Beck and Galef,1989).

IS THERE ANYTHING SPECIALABOUT SOCIAL LEARNING?

Our analyses have indicated that in most in-stances of social influence on the food choicesof young rats, interaction between adult andyoung rats has resulted in introduction of theyoung to one food rather than another. Adultsbias young either to initiate feeding on foodsthat the adults are eating, rather than on al-ternative foods, or to start to feed at feedingsites that the adults are visiting, rather than atalternative sites. Differences in the responsesof young rats to familiar and unfamiliar foodsand locations are then responsible for muchof the influence of adults on the choices ofjuveniles with which they interact (Galef,1971b).

Such effects of socially induced familiar-ity on food choice are particularly pronouncedin genetically wild Norway rats that are ex-tremely hesitant to eat unfamiliar foods (Bar-nett, 1958). The extreme neophobia of wildrats makes introduction of juveniles to onefood rather than another a critical event in thedevelopment of their feeding repertoires(Galef and Clark, 1971b).

However, not all of the social influenceson the food choices of rats reflect a simple so-cial biasing of naive young rats to eat one foodrather than another together with neophobia.In the case discussed in the next section, so-

cially induced food preference seems to resultfrom a behavioral process that directly altersthe affective response of young rats to foodsexperienced in a social context (Galef et al.,1997).

FLAVOR CUES ONTHE BREATH OF RATS

In the early 1980s, scientists in several labora-tories demonstrated that after a naive "ob-server" rat interacts with a recently fed con-specific "demonstrator," the observer exhibitsa substantial enhancement of its preferencefor whatever food its demonstrator ate (Galefand Wigmore, 1983; Strupp and Levitsky,1984). For example, after naive observer ratsinteracted briefly with conspecific demonstra-tors fed either a cinnamon- or a cocoa-flavored diet, the former group of observerspreferred cinnamon-flavored food, whereasthe latter preferred cocoa-flavored food, if of-fered a choice between the two (Fig. 34-2).

Gin Coc

Demonstrator's diet

Figure 34-2. Amount of cinnamon-flavored diet (Diet Cin)eaten, as a percentage of total intake over 22 hours, by ob-server rats that interacted with demonstrators fed eitherDiet Cin or cocoa-flavored diet (Diet Coc). Error bars show1 SEM. (Data from experiments like those described in Galefand Wigmore [1983].)

Page 384: 0195162854

Chapter 34. Social Learning 367

The effects of a single brief exposure torecently fed demonstrator rats on the foodchoices of their observers are both surpris-ingly powerful and surprisingly long lasting.Many observer rats taught to totally avoid in-gesting a food by following its ingestion withan injection of toxin, and then placed withdemonstrator rats that have eaten the foodthat their observers had learned to avoid, to-tally abandoned their aversions. Similarly,most observer rats that interacted with ademonstrator fed a diet adulterated withcayenne pepper (an inherently unpalatabletaste to rats) subsequently preferred pepperedto unadulterated diet (Galef, 1986b). Such ef-fects of demonstrator rats on the food choicesof their observers can be seen a month ormore after a demonstrator and observer in-teract (Galef and Whiskin, 2003).

ANALYSIS

The behavioral process that produces such so-cial influence on the food choices of observerrats is now quite well understood. Olfactorycues passing to observer rats from demonstra-tors cause observers to increase their prefer-ences for the foods that their respective dem-onstrators ate (Galef and Wigmore, 1983).Observers sniff at the mouths of demonstra-tors, and this sampling of a demonstrator'sbreath is both necessary and sufficient fordemonstrators to influence the later foodchoices of observers (Galef and Stein, 1985).

Both food-related odors escaping fromthe digestive tract of a demonstrator and thescent of bits of food clinging to a demonstra-tor's fur and vibrissae allow observers to iden-tify the food that a demonstrator has recentlyeaten. And after an observer rat experiencessimultaneously the scent of a food and ratbreath, the observer shows an enhanced pref-erence for the food the scent of which it ex-perienced together with rat breath (Galef andStein, 1985).

Gas chromatography performed on sam-ples of rat breath has shown that it contains

two sulfur compounds: carbon disulfide andcarbonyl sulfide. Rats exposed to a fooddusted onto either the head of an anesthetizedconspecific or a piece of cloth moistened witha dilute solution of carbon disulfide subse-quently show an enhanced preference for thatfood. To the contrary, rats exposed to a food,that had been dusted onto the head of a deadconspecific, onto the rear of a live conspecificor onto a piece of cloth moistened with dis-tilled water do not develop a similar prefer-ence (Galef et al., 1988) (Fig. 34-3). Thus,experience of carbon disulfide, a natural con-stituent of rat breath, in conjunction with afood odor, like experience of rat breath in con-junction with a food odor, is sufficient to en-hance preference for the food.

SYNTHESIS

The breath of humans, like the breath of rats,contains trace quantities of carbon disulfide.As would be expected on the hypothesis thatexperience of food odors together with car-

Figure 34-3. Observer rats interacted with either an anes-thetized demonstrator rat or a cloth "surrogate" demon-strator. The demonstrator with which each observer inter-acted had been powdered with either cinnamon- orcocoa-flavored diet. Surrogates were moistened with eithera dilute aqueous solution of CS2 or an equal amount of dis-tilled water. The figure indicates the mean percent of eachobserver's total intake that was the diet with which itsdemonstrator or surrogate had been powdered. Error barsshow 1 SEM. (Data from Galef et al. [1988].)

Page 385: 0195162854

368 DEFENSE AND SOCIAL BEHAVIOR

bon disulfide induces food preferences in rats,when a human "demonstrator" eats a flavoredfood and breathes on a rat, the rat's prefer-ence for the food that its human demonstra-tor ate is markedly enhanced (Galef, 2001).

LIMITATIONS

Surprisingly, rats do not learn to avoid a foodby interacting with a sick or an unconsciousdemonstrator that has eaten it. To the con-trary, rats show an increased preference for afood that was eaten by an ill conspecific withwhich they interacted (Galef et al., 1990).

Further, exposure to an odor in conjunc-tion with a conspecific does not enhance thegeneral affinity of a rat for that odor; exposureto an odor in a social context that profoundlyaffects food preference has no effect on theodor preferences of rats in other contexts. Forexample, rats that have interacted with a con-specific that has eaten a cinnamon-flavoreddiet prefer cinnamon-flavored food but showno enhancement of their preference for cin-namon-scented nest materials or cinnamon-scented nest sites (Galef and Iliffe, 1994). Suchfindings suggest that social induction of foodpreference is a learning process evolved specif-ically to facilitate foraging rather than otheractivities of rats.

EXTENTIONS

Rats can use information concerning foodsthat other rats have eaten in some interestingways. For example, after "observer" rats hadan opportunity to learn where in a three-armmaze each of three distinctively flavored foodswere to be found, we let each observer rat in-teract briefly with a demonstrator rat that hadeaten one of those three foods. Without anyspecific training, the observers went directlyto the arm of the maze where they hadlearned that the food that their demonstratorhad eaten was usually located (Galef and Wig-more, 1983). Obviously, rats can integratetheir cognitive map of food distribution with

socially acquired information about the cur-rent availability of foods to increase the effi-ciency with which they forage.

APPLICATION TO STUDIES OFNERVOUS SYSTEM FUNCTION

Socially induced enhanced diet preferenceprovides an efficient and reliable way to in-duce a learned appetitive behavior in rats (ormice, gerbils, hamsters, voles or bats) that,like other types of learned behavior, can serveas a dependent variable in studies of brainfunction. Neuroscientists have used the so-cially induced change in food preference de-scribed here to study the effects of manipula-tions of the neural substrate on learning andmemory (Burton et al., 2000; Winocur et al.,2001; see Galef, 2002, for further references).As one might expect, both direct and geneticmanipulations of the nervous system affectsocial learning of food preferences.

There are several advantages in using so-cially learned food preference as a dependentmeasure in studies of brain function: (i) learn-ing occurs in a single trial, (2) little or no skillis needed to train subjects, (3) no specialequipment is needed to train subjects, and (4)subjects need never be deprived or stressed.The procedure for inducing social enhance-ment of food preference consists of threestraightforward steps. First, a demonstratorrat is placed on a feeding schedule and givenone of two distinctively flavored foods to eat.Second, each demonstrator is placed togetherwith an observer, and demonstrator and ob-server rats are allowed to interact for 15 min-utes or longer. During this period of interac-tion, observers have the opportunity to smellthe scented food on the breath of their re-spective demonstrators. Last, each observer isgiven a choice between the two distinctivelyflavored foods that were offered to demon-strators in the first step (Galef, 2002) (Fig.34-4). In the third step, observers invariablyshow an enhanced preference for whichever

Page 386: 0195162854

Chapter 34. Social Learning 369

REFERENCES

Figure 34-4. Schematic of the three stages of an experimentdemonstrating social influence on the diet preferences ofobserver rats'. In stage 1, each demonstrator rat ate one oftwo distinctively flavored foods. (From Galef [2002].Reprinted with permission of John Wiley & Sons, Inc.)

flavored food was eaten by their respectivedemonstrators.

The effect is robust. Demonstrator andobserver can be male or female, young or old,previously familiar or unfamiliar with one an-other, and genetically related or unrelated toone another (Galef et al, 1984). Demonstra-tors can ingest almost any scented liquid orsolid before interacting with their observers.There can be a delay of several hours betweenwhen a demonstrator is fed and when it in-teracts with its observer. Demonstrators canbe separated from their observers by ahardware-cloth screen while they interact,and interaction can take place in the homecage of demonstrator or observer or in a neu-tral arena. There can be a delay of weeks be-tween when demonstrator and observerinteract and when the observer is tested. In-variably, if demonstrators that have recentlyingested a distinctively flavored substance areplaced for a few minutes together with ob-servers that are otherwise unfamiliar with theflavor of the food that was eaten by their re-spective demonstrators, the observers subse-quently show significant enhancement oftheir relative intake of that food.

Barnett SA (1958) Experiments on "neophobia" in wildand laboratory rats. British Journal of Psychology49:195-201.

Beck M and Galef BG Jr (1989) Social influences on theselection of protein-sufficient diet by Norway rats.Journal of Comparative Psychology 103:132-139.

Burton s, Murphy D, Qureshi U, Suton P, O'Keefe J(2000) Combined lesions of hippocampus andsubiculum do not produce deficits in nonspatial so-ciallearning. Journal of Neuroscience 20:5468-5475.

Galef BG Jr (1977) Mechanisms for the social transmis-sion of food preferences from adult to weanling rats.In: Learning mechanisms in food selection (BarkerLM, Best M, Domjan M, eds.), pp. 123-150. Waco,TX: Baylor University Press.

Galef BG Jr (1981) The development of olfactory controlof feeding site selection in rat pups. Journal of Com-parative and Physiological Psychology 95:615-622.

Galef BG Jr (1986) Social interaction modifies learnedaversions, sodium appetite, and both palatabilityand handling-time induced dietary preference in rats(Rattus norvegicus'). Journal of Comparative Psy-chology 100:432-439.

Galef BG Jr (1988) Communication of information con-cerning distant diets in a social, central-place forag-ing species (Rattus norvegicus}. In: Social learning:psychological and biological perspectives (ZentallTR and Galef BG Jr, eds.) pp. 119-140. Hillsdale, NJ:Erlbaum.

Galef BG Jr (1992) The question of animal culture. Hu-man Nature 3:157-178.

Galef BG Jr (1996a) Social enhancement of food prefer-ences in Norway rats. In: Social learning and imita-tion: the roots of culture (Heyes CM and Galef BGJr, eds.) pp. 49-64. New York: Academic Press.

Galef BG Jr (1996b) Social influences on food preferencesand feeding behaviors of vertebrates. In: Why weeat what we eat (Capaldi E, ed.) pp. 207-232. Wash-ington, D.C.: American Psychological Association.

Galef BG Jr (2001) Analyses of social learning processesaffecting animals' choices of foods and mates. Mex-ican Journal of Behavior Analysis 27:145-164.

Galef BG Jr and Whiskin EE (2003) Socially transmittedfood preferences can be used to study long-termmemory in rats. Learning and Behavior 31:160-164.

Galef BG Jr and Allen C (1995) A new model systemfor studying animal traditions. Animal Behaviour50:705-717.

Galef BG Jr and Beck M (1985) Aversive and attractivemarking of toxic and safe foods by Norway rats. Be-havioral and Neural Biology 43:298-310.

Galef BG Jr and Buckley LL (1996) Use of foraging trailsby Norway rats. Animal Behaviour 51:765-771.

Page 387: 0195162854

370 DEFENSE AND SOCIAL BEHAVIOR

Galef BG Jr and Clark MM (1971a) Parent-offspring in-teractions determine time and place of first inges-tion of solid food by wild rat pups. PsychonomicScience 25:15-16.

Galef BG Jr and Clark MM (1971b) Social factors in thepoison avoidance and feeding behavior of wild anddomesticated rat pups. Journal of Comparative andPhysiological Psychology 25:341-357.

Galef BG Jr and Iliffe CP (1994) Social enhancement ofodor preference in rats: is there something specialabout odors associated with foods? Journal of Com-parative Psychology 108:266-273.

Galef BG Jr (2002) Social learning of food preferences inrodents: rapid appetitive learning. Current Proto-cols in Neuroscience. 8.5D1-8.5D8.

Galef BG Jr, Kennett DJ, Wigmore SW (1984) Transferof information concerning distant foods in rats: a ro-bust phenomenon. Animal Learning and Behavior12:292-296.

Galef BG Jr, Marczinski CA, Murray KA, Whiskin EE(2001) Studies of food stealing by young Norwayrats. Journal of Comparative Psychology 115:16-21.

Galef BG Jr, Mason JR, Pretti G, Bean, NJ (1988) Car-bon disulfide: a semiochemical mediating socially-induced diet choice in rats. Physiology and Behav-iour 42:119-124.

Galef BG Jr, McQuoid LM, Whiskin EE (1990) Furtherevidence that Norway rats do not socially transmitlearned aversions to toxic baits. Animal Learningand Behavior 18:199-205.

Galef BG Jr and Sherry DF (1973) Mother s milk: amedium for transmission of information aboutmother's diet. Journal of Comparative and Physio-logical Psychology 83:374-378.

Galef BG Jr and Stein M (1985) Demonstrator influenceon observer diet preference: analyses of critical so-

cial interactions and olfactory signals. Animal Learn-ing and Behavior 13:131-138.

Galef BG Jr, Whiskin EE, Bielavska E (1997) Interactionwith demonstrator rats changes their observers' af-fective responses to flavors. Journal of ComparativePsychology 111:393-398.

Galef BG Jr and Wigmore SW (1983) Transfer of infor-mation concerning distant foods: a laboratory in-vestigation of the information-centre" hypothesis.Animal Behaviour 31:748-758.

Garcia J, Ervin FR, Koelling RA (1966) Learning withprolonged delay of reinforcement. PsychonomicScience 5:121-122.

Hepper PG (1988) Adaptive fetal learning: prenatal ex-posure to garlic affects postnatal preference. AnimalBehaviour 36:935-936.

Smotherman WP (1982) Odor aversion learning by therat fetus. Physiology and Behavior 29:769-771.

Steiniger von F (1950) Beitrage zur Soziologie und son-stigen Biologic der Wanderratte. Zeitschrift furTierpsychologie 7:356-379.

Strupp BJ and Levitsky DA (1984) Social transmission offood preferences in adult hooded rats (Rattusnorvegicus). Journal of Comparative Psychology98:257-266.

Van Schaik CP, Ancrenaz M, Borgen G, Galdikas B, Sin-gleton I, Suzuki A, Utami SS, Merill M (2003) Orang-utan cultures and the evolution of material cultureimplications. Science 299:102-105.

Von Frisch K (1967) The dance language and orienta-tion of bees. Cambridge, Mass.: Belknap Press.

Whiten A, Goodall J, McGrew WC, Nishida T, ReynoldsV, Sugiyama Y, Tutin CEG, Wrangham RW, BoeschC (1999) Culture in chimpanzees. Nature 399:682-685.

Winocur G, McDonald RM, Moscovitch M (2001). An-terograde and retrograde amnesia in rats with largehippocampal lesions. Hippocampus 11:18-26.

Page 388: 0195162854

Vocalization

GRETA SOKOLOFFAND MARK S. BLUMBERG

35

Throughout their life span, rats emit vocaliza-tions in a variety of environmental and socialcontexts, including nest separation in infantsand sexual behavior in adults. Rats, being smalland therefore having a small vocal apparatus,produce vocalizations predominantly at ultra-sonic frequencies, that is, at frequencies abovethe range detectable by the human ear (>20kHz). This chapter provides an overview of thecontexts, mechanisms, and suggested functionsof rat vocalizations. The proposed use of thesevocalizations as a model for investigations ofanxiety and depression is also discussed.

FREQUENCY ANDTEMPORAL CHARACTERISTICS

Rats emit ultrasonic vocalizations across a fre-quency range of 20 to 70 kHz (Table 35-1). Dur-ing early postnatal development, infants pro-duce vocalizations with a dominant frequencyof approximately 40 kHz when isolated fromthe nest. As the pup, and its vocal apparatus, in-creases in size, the dominant frequency of thisvocalization decreases progressively to 25 kHzby 20 days of age (Blumberg et al., 2000a). In-fant vocalizations are relatively pure tones pro-duced by expiring air under high pressureagainst constricted laryngeal folds, as is also thecase for the adult's 22-kHz vocalization(Roberts, 1975b, 1975c; Sanders et al., 2001).The 22-kHz vocalization, associated with a va-riety of contexts, including sexual behavior andaggression (Sales, 1972a, 1972b), has been re-ferred to as a long call because it is produced dur-

ing prolonged expirations (>1 second). In con-trast, a third category of rat ultrasonic vocaliza-tion has been referred to as a short call becauseof its relatively brief duration (<65 millisec-onds); these chirp-like vocalizations, emitted byweanlings, juveniles, and adults at high fre-quencies (35 to 70 kHz), are associated with con-texts that include vigorous activity, high levelsof arousal, and social contact. Although it hasbeen suggested that some instances of these vo-calizations may be produced as a biomechani-cal byproduct of activity (Blumberg, 1992),many instances of this vocalization appear to beindependent of activity (Knutson et al., 2002).Very little is known regarding the mechanismsthat produce these vocalizations.

Just as small animals produce higher-frequency sounds, so are they able to hearhigher-frequency sounds. The auditory systemof adult rats exhibits a peak sensitivity range of10 to 50 kHz (Crowley et al., 1965; Gourevitchand Hack, 1966; Brown, 1973). Similarly, pre-weanling rats detect and respond to auditorystimuli in the range of 1 to 70 kHz, with a peaksensitivity of approximately 40 kHz (Crowley etal., 1965; Crowley and Hepp-Reymond, 1966).

ENVIRONMENTAL CONTEXTSASSOCIATED WITH

ULTRASONIC VOCALIZATIONS

INFANTS

Infant rats emit broadband vocalizations com-prising audible and ultrasonic frequencies dur-

371

Page 389: 0195162854

372 DEFENSE AND SOCIAL BEHAVIOR

Table 35-1. Four General Categories of Rat Vocalizations

Frequency Age Examples of Contexts and Stimuli

Broadband25 to 45 kHz (depending

on age and size)22kHz

35 to 70 kHz

Infancy to adulthoodInfancy

Adulthood

Juvenile period to adulthood

Painful stimuli, such as tail pinch and foot shockIsolation from the nest; cooling

After ejaculation, after defeat in aggressiveencounter, and after termination of foot shock

Arousal, such as during play, copulation, andaggression

ing maternal handling (e.g., grooming) and in-tense tactile stimulation (e.g., tail pinching).In contrast, ultrasonic vocalizations areevoked by isolation from the nest (Noirot,1972). Cold is the primary feature of isolationthat elicits the vocalization, as isolated pupsthat are kept warm do not vocalize (Allin andBanks, 1971; Okon, 1971; Blumberg et al.,1992b). Furthermore, as pups mature and arebetter able to thermoregulate, isolation re-sults in fewer ultrasonic emissions (Okon,1971; Blumberg et al., 1992a).

Infant ultrasound production increasesduring the first 2 postnatal weeks and declinesthereafter (Noirot, 1968; Sewell, 1970; Okon,1971; Noirot, 1972). During development, itappears that factors capable of evoking andmodulating ultrasound production diversifyand become more complex. For example, ol-factory stimuli begin to play a role in the at-tenuation of ultrasound production duringcold exposure. Specifically, odors associatedwith the nest (e.g., dam, siblings, home cagebedding) effectively reduce isolation-inducedvocalizations (Oswalt and Meier, 1975; Hoferand Shair, 1987), as does exposure to an un-familiar adult male rat (Takahashi, 1992).

One phenomenon that illustrates theemergence of complex controls of ultrasoundproduction during the second postnatal weekis maternal potentiation (Hofer et al., 1998;Kraebel et al., 2002; Shair et al., 2003). Theparadigm for studying maternal potentiationbegins with an initial period of isolation. Thepup is then briefly reunited with the dam. Fi-nally, the pup is again isolated, resulting in

elevated levels of ultrasound production. Aswith other contexts in which infant vocaliza-tions are studied, air temperature plays amodulatory role in the expression of mater-nal potentiation (Kraebel et al., 2002; Shair etal., 2003).

ADULTS

Ultrasonic vocalizations can be detected incolonies of wild rats as well as in group-housed or isolated laboratory rats (Calhoun,1962; Francis, 1977). Although these adult vo-calizations are primarily associated with socialinteractions, isolated rats vocalize sponta-neously in a circadian fashion, with peak vo-calization rates occurring during the middleof the dark period (Francis, 1977).

High-frequency ultrasonic "chirps" pre-dominate during social interactions associatedwith high levels of arousal. In adult rats, chirpsare emitted when rats are placed together aswell as when an individual rat is placed in anarea previously visited by other rats (Brudzyn-ski and Pniak, 2002). Furthermore, juvenilesvocalize in social situations that have previ-ously been associated with play (Knutson etal., 1998), and adult male rats vocalize whenplaced in empty cages where social contacthas previously occurred (Bialy et al., 2000;Brudzynski and Pniak, 2002).

Ultrasonic vocalizations also occur dur-ing reproductive behavior. Males and femalesemit 50- to 70-kHz vocalizations primarilyduring genital investigation, chasing, andmounting (Sales, 1972a, 1972b; Barfield et al.,

Page 390: 0195162854

Chapter 35. Vocalization 373

1979; Thomas and Barfield, 1985). The 22 kHzvocalization of males reliably occurs duringthe postejaculatory refractory period (Barfieldet al., 1979).

When exposed to a natural predator, likea cat, rats housed in groups emit 22-kHz vo-calizations. Experiments using artificial bur-rows consisting of multiple rats have indicatedthat 22-kHz vocalizations are produced whenthe cat is introduced and continue to be emit-ted after the cat is removed (Blanchard et al.,1991). This is in contrast to individual ratsthat, when tested in the presence of a cat, donot vocalize regardless of whether they areable to escape (Blanchard et al., 1991).

Aggressive encounters are another con-text in which ultrasonic vocalizations are reli-ably evoked. Initially, when two male rats areintroduced and during the initial phases of ag-gressive behavior, 50- to 70-kHz vocalizationsare emitted by both animals. After the ag-gressive encounter, however, 22-kHz vocal-izations are emitted exclusively by the sub-missive rat while exhibiting the belly-upsubmissive posture (Sewell, 1967; Sales,1972b).

ANATOMICAL CONSIDERATIONS

Audible rat vocalizations are produced by vi-bration of the laryngeal folds, as with humanvoiced speech (Roberts, 1975a). They areelicited by stimulation of Ad- and C-fibres (Ar-did et al., 1993) as well as by stimulation ofthe trigeminal spinal tract nucleus (Yajima etal., 1981), suggesting that audible vocaliza-tions are produced in response to noxious orpainful stimulation. Although both audibleand ultrasonic vocalizations are produced bythe larynx and occur during the expiratoryphase of respiration, the 40-kHz infant vocal-ization and the 22-kHz adult vocalization ap-pear unique in that they are produced using awhistle-like mechanism that entails forced ex-piration through a constricted and nonvibrat-ing larynx (Roberts, 1975b, 1975c). The 50- to

70-kHz vocalization is acoustically more com-plex and is likely produced by vibration of thelaryngeal folds.

Audible vocalizations become weakerand ultrasonic vocalizations are virtually abol-ished by nerve cuts that denervate the laryn-geal musculature (Roberts, 1975b). Specifi-cally, unilateral or bilateral transection of theinferior laryngeal nerve, a branch of the vagusnerve, abolishes ultrasound in infant rats. Inaddition, transection of the superior laryngealnerve, also a branch of the vagus nerve,changes the sound pressure and frequency ofthese vocalizations as well as vocalization rate(Wetzel et al., 1980).

The neural circuit that controls the lar-ynx in rats and other species includes the mid-brain periaqueductal gray (PAG). In fact, stim-ulation of the PAG evokes species-specificvocalizations in many animals (Zhang et al.,1994), including rats (Yajima et al., 1980). Inrats, stimulation of a pathway emanating fromthe dorsal region of the thalamus and termi-nating in the dorsomedial region of the PAGelicits 22-kHz vocalizations (Yajima et al.,1980). From the PAG, projections to the dor-somedial aspect of the medullary reticularformation are also involved in ultrasound pro-duction, including a number of cranial nervenuclei (e.g., facial, hypoglossal, and vagal)(Yajima et al., 1981). The inferior and superiorlaryngeal nerves, necessary for the productionof ultrasound, arise from the dorsal and ven-tral regions of the medullary nucleus am-biguus, respectively (Wetzel et al., 1980). Au-dible vocalizations are evoked by stimulationof numerous regions of the hypothalamus aswell as the ventromedial PAG (Yajima et al.,1980).

FUNCTIONAL SIGNIFICANCE OFULTRASONIC VOCALIZATIONS

Many theories exist concerning the functionalsignificance of rat ultrasonic vocalizations.The earliest theories focused on communica-

Page 391: 0195162854

374 DEFENSE AND SOCIAL BEHAVIOR

tory functions. Extensive work on the effectsof thermal stimuli on ultrasound productionin infant rats led to physiological theories ofultrasound production. More recently, moti-vational theories of ultrasound productionhave emerged that focus on the emotionalcontexts in which the vocalizations are pro-duced (Table 35-2).

COMMUNICATORY HYPOTHESES

The fact that infant rats vocalize when isolatedfrom the nest and the fact that dams retrieveisolated pups suggests an important commu-nicatory role for these vocalizations. Today,infant rat vocalizations are commonly re-ferred to as distress or separation calls (Oswaltand Meier, 1975; Hofer et al., 1994). In fact,even the absence of vocalizations in isolatedinfant rats has been ascribed a communicatoryfunction, namely, to prevent the alerting ofpredators to an infant's location (Takahashi,1992; Hofer et al., 1994).

During reproductive behavior, 50- to 70-kHz vocalizations emitted by males have beenproposed to serve the communicatory func-tion of increasing female solicitation behav-

iors such as hopping and darting (Sales, 1972a;Barfield et al., 1979). In contrast, the poste-jaculatory 22-kHz vocalization has been hy-pothesized to communicate to the female areduction in the male's sexual motivation(Barfield et al., 1979). Modulation of this vo-calization by the female's presence during thepostejaculatory interval has been interpretedas evidence in favor of a communicatory func-tion (Sachs and Bialy, 2000).

Still other communication hypotheseshave been posited for 22-kHz vocalizations.With respect to a male defeated in an aggres-sive encounter, it has been proposed that thevocalization serves to prevent further aggres-sion by the dominant male (Sales, 1972b).With respect to group-housed rats that vocal-ize when exposed to a cat, it has been hy-pothesized that these vocalizations serve toalert conspecifics to the predator's presence(Blanchard et al., 1991). Thus, the 22-kHz vo-calizations of adults, like those of infants, havebeen deemed distress or alarm calls.

As stated, infant ultrasonic vocalizationselicit maternal retrieval and the ultrasonic vo-calizations of male rats during reproductivebehavior alter female behavior (Allin and

Table 35-2. Theories of Ultrasound Production in Rats

Theory Category Example Citations

Communicatory

Physiological

Motivational / emotional

Infant

50 to 70 kHz

22kHz

Infant

22kHz

Infant

~50 to 70 kHz

22kHz

Eliciting maternal retrievalafter separation from thenest

Facilitation of femalesexual behavior

Appeasement afteraggressive encounter

Relationship tocardiopulmonary function

Relationship to braintemperature

Anxiety/distress

Expectation of play or socialcontact

Withdrawal from substanceabuse

Allin and Banks, 1972; Farrelland Alberts, 2002a

Barfield et al., 1979

Sales, 1972a

Blumberg and Alberts, 1990;Blumberg and Sokoloff,2001

Blumberg and Moltz, 1987

Shair et al., 2003; Winslowand Insel, 1991

Knutson et al., 1998;Brudzynski and Pniak, 2002

Vivian and Mizcek, 1993

Page 392: 0195162854

Chapter 35. Vocalization 375

Banks, 1972; Noirot, 1972; Sachs and Bialy,2000; Farrell and Alberts, 2002a). Such demon-strations of communicatory effect, however,are not equivalent to demonstrations of com-municatory function (Blumberg and Alberts,1992, 1997). Furthermore, although evidenceexists that seems to fit easily within a com-municatory framework, other evidence doesnot fit so easily. For example, the apparentsuppressive effect of littermates on the vocal-ization can be overcome by decreasing tem-perature until the huddle's thermoregulatorycapabilities are exceeded (Blumberg et al.,1992a; Sokoloff and Blumberg, 2001). Fur-thermore, studies in which pup odor andacoustic cues are manipulated demonstratethat the vocalization alone is not sufficient toevoke maternal retrieval (Smotherman et al.,1974; Farrell and Alberts, 2002b). As anotherexample, during agonistic encounters, the ag-gressive behavior of dominant males is notnecessarily reduced by the vocalizations ofsubordinates, and the freezing behavior andvocalizations of the subordinate males are notreduced by the vocalizations of the dominantmales (Takahashi et al., 1983). Taken to-gether, these examples highlight the need forcaution when assessing the communicatorysignificance of rat vocalizations.

PHYSIOLOGICAL HYPOTHESES

Many investigators have appreciated the sig-nificance of cold exposure for eliciting infantvocalizations during isolation from the nest(Allin and Banks, 1971; Okon, 1971; Oswaltand Meier, 1975). In fact, some level of coldexposure remains as an integral component ofmost methodological approaches in the field,even when the primary focus of a study is theability of pharmacological agents or nest-related cues to attenuate the vocalization(Kraebel et al., 2002). The question is whethertemperature is merely a cue to the pup that itis isolated from the nest or a physical stimu-lus that evokes a significant change in thepup's physiological functioning.

Infant rats are capable of endogenousheat production, but their small size results inrapid heat loss when exposed to standardroom temperatures (i.e., 22° C). When ob-served at air temperatures in which brownadipose tissue thermogenesis is sufficient tomaintain elevated body temperatures, how-ever, infant rats do not vocalize (Blumbergand Stolba, 1996). Of importance is that phar-macological manipulations that augment orattenuate brown adipose tissue thermogene-sis result in decreases or increases in infant vo-calization rates, respectively (Blumberg et al.,1999; Farrell and Alberts, 2000).

The effects of cold exposure on the car-diopulmonary system of infant rats have led tospecific hypotheses concerning the physiolog-ical mechanisms underlying ultrasound pro-duction (Blumberg and Alberts, 1990; Blum-berg and Sokoloff, 2001). Infants do notvocalize during cold exposure unless the cool-ing is severe enough to cause a decrease in car-diac rate (Blumberg et al., 1999). In addition,at these same temperatures, blood viscosityincreases significantly, further compromisingcardiopulmonary function. Therefore, it hasbeen hypothesized that infant vocalizations areacoustic byproducts of a physiological maneu-ver that serves to maintain cardiopulmonaryfunction (Blumberg and Sokoloff, 2001).

In addition to cold exposure, administra-tion of the a2-adrenoceptor agonist clonidineresults in prolonged and robust vocalizationresponses. The effect of clonidine is so pro-found that returning the pup to the nest is notsufficient to attenuate the vocalization (Kehoeand Harris, 1989). Interestingly, clonidine alsoproduces a profound bradycardia and whenthe jSi-adrenoceptor agonist prenalterol wasused to inoculate pups against clonidine-induced bradycardia, ultrasound productionwas significantly attenuated (Blumberg et al.,2000b).

Physiological correlates of adult vocal-izations have also been reported. For exam-ple, the 22-kHz vocalization occurs during thechill phase of fever evoked by central admin-

Page 393: 0195162854

376 DEFENSE AND SOCIAL BEHAVIOR

istration of prostaglandin E2; in addition, thepostejaculatory vocalization is virtually abol-ished by prior administration of sodium salicy-late, a drug that decreases body temperature(Blumberg and Moltz, 1987). The significanceof this relationship between temperature andemission of vocalization is not yet dear.

MOTIVATIONAL ANDEMOTIONAL THEORIES

Rat vocalizations are often suggested to be ex-pressions of emotion. According to one the-ory, 50- to 70-kHz vocalizations are an indexof positive affect, whereas 22-kHz vocaliza-tions are an index of negative affect (Knutsonet al., 2002). The view that infant vocalizationsmodel human distress and anxiety has devel-oped in lock-step with the use of infant vo-calizations to examine the efficacy of phar-macological agents developed for thetreatment of human psychological disorders(Miczek et al., 1995).

The contexts and manipulations thatelicit ultrasound production provide the foun-dation for the affective classification of the dif-ferent categories of vocalizations. As stated,50- to 70-kHz vocalizations occur during so-cial contact, such as reproductive behaviorand play (Sales, 1972a; Barfield et al., 1979;Knutson et al., 1998). Electrical brain stimula-tion, which is known to be reinforcing, alsoelicits these vocalizations (Burgdorf et al.,2000), a finding that is consistent with the no-tion that these vocalizations are an index ofpositive affect (Burgdorf et al., 2000; Knutsonet al., 2002). In juvenile rats, "tickling" by hu-man handlers results in 50-kHz chirps thathave been described as laughter (Pankseppand Burgdorf, 2000).

In contrast to 50- to 70-kHz chirps, 22-kHz vocalizations are typically associatedwith physiological and psychological stres-sors, such as fever (Blumberg and Moltz, 1987)and aggressive encounters (Sales, 1972b).Aversive conditioning paradigms (i.e., fearconditioning; Lee et al., 2001) and withdrawal

from morphine dependence (Vivian andMiczek, 1991) also elicit these vocalizations.These findings are consistent with the notionthat the 22-kHz vocalization is an index ofnegative affect (Miczek et al., 1995; Knutsonet al., 2002).

There are, however, pieces that do not fitso easily into the emotional framework justpresented. For example, as already noted, 50-to 70-kHz vocalizations accompany agonisticencounters just as readily as they do sexualand playful ones. Are we then to suppose thatrats experience positive affect while fighting?Similarly, are we to assume that the postejac-ulatory male, emitting the 22-kHz vocaliza-tion, experiences a negative affective stateakin to that experienced by a male that hasjust been defeated in an aggressive encounterwith a conspecific? These and other incon-gruities pose difficulties for any unidimen-sional affective theory of rat vocalizations.

THE VOCALIZING RAT ASA MODEL OF HUMAN

PSYCHOLOGICAL DISORDERS

Although there remains much uncertaintyconcerning the mechanisms and functionalsignificance of rat vocalizations, it is widelybelieved that they can be effectively used toinvestigate anxiety, distress, fear, and drugabuse. For this reason, psychopharmacologi-cal approaches to studying rat vocalizationshave become very popular. The hope is thatincreased knowledge concerning the effects ofpsychoactive drugs on infant and adult vocal-izations will increase understanding of humanpsychiatric disorders and their treatment(Miczek et al., 1995).

Because 22-kHz vocalizations are some-times associated with aversive events and be-haviors associated with fear (e.g., freezing), itis suggested that rats producing this vocaliza-tion can be used as a model for depressionand anxiety in humans (Miczek et al., 1995;Schreiber et al., 1998). Similarly, using the iso-

Page 394: 0195162854

Chapter 35. Vocalization 377

lation paradigm, ultrasonic vocalizations in in-fant rats have been proposed as a model ofseparation anxiety (Winslow and Insel, 1991).Support for this view comes from pharmaco-logical studies in which anxiolytic agents areshown to attenuate ultrasound productionunder some conditions. For example, selec-tive serotonin reuptake inhibitors (SSRIs) andbenzodiazepines have been shown to reducevocalizations in adult and infant rats (Insel etal., 1986; Olivier et al, 1998; Schreiber et al,1998). Opioids also reduce the occurrence of22-kHz vocalizations in adults in response totail shock (van der Poel et al., 1989). Finally,anxiolytic agents attenuate ultrasonic vocal-izations during morphine withdrawal (Vivianand Miczek, 1991).

Although the vocalizing rat, infant andadult, has been proposed as a useful model fortesting anxiolytic drugs (Olivier et al., 1998),some research indicates that it is not a robustmodel. First, at least one anxiolytic agent,clonidine, increases ultrasound in infants (Ke-hoe and Harris, 1989; Blumberg et al., 1999).Second, benzodiazepines do not reduce pre-stimulus ultrasonic vocalizations after fearconditioning (van der Poel et al., 1989). Third,only antidepressants working via the seroton-ergic system attenuate ultrasound production,whereas other antidepressant compounds act-ing on the noradrenergic system result in anx-iogenesis, as measured by increased ultra-sound production, and still other compounds(e.g., amitriptyline) have no effect (Borsini etal., 2002). Regardless, we should not be sur-prised if the diversity of rat vocalizations donot fit neatly into conceptual categories de-veloped for the diagnosis of human clinicaldisorders (Blumberg and Sokoloff, 2001).

MEASURING ULTRASONICVOCALIZATIONS INTHE LABORATORY

The ease of making ultrasonic vocalizationsaudible using a bat detector, originally in-

vented to study echolocation in bats, has stim-ulated a growth industry in the area of rat vo-calizations. Today, in numerous laboratoriesacross the world, infant and adult vocaliza-tions are used for studies of basic physiology,separation responses, and psychopharmacol-ogy. Reflecting this surge in interest, there area number of companies that produce ultra-sound-sensitive detectors with an array of fea-tures and capabilities.

The analysis of ultrasonic vocalizationscan be as simple as manually counting the to-tal number of vocalizations, either during datacollection or afterward from an audio record-ing. Skilled listeners typically exhibit high in-terrater reliabities even when counting vocal-izations in infant rats that can occur manytimes each second. Alternatively, automaticscoring of vocalizations can be accomplishedby digitally recording the vocalizations usinga data acquisition system and counting burststhat exceed a threshold value; some compa-nies now offer systems that are designedspecifically for this task. For more complexacoustic analyses (e.g., frequency modulation,peak frequency, amplitude, duration of indi-vidual calls), investigators can use any of anumber of analysis programs available.

CONCLUSIONS

Ultrasonic vocalizations accompany a widearray of social behavior in the rat, from theisolation-induced vocalization of the infants tothe postejaculatory vocalization of the adultmales. Interestingly, there has of yet been noserious attempt to synthesize the various the-ories and perspectives in the field. For exam-ple, communicatory and motivational theo-ries of ultrasound production have focusedprimarily on the function of this behavior (i.e.,a signal for maternal retrieval or as an indexof emotional state). In contrast, physiologicaltheories have focused primarily on mecha-nisms underlying the behavior (i.e., reflexivecardiopulmonary compensations). Therefore,

Page 395: 0195162854

378 DEFENSE AND SOCIAL BEHAVIOR

there exists a large gap in our understandingof these vocal behaviors throughout develop-ment and across the different environmentalcontexts that they are expressed.

The 40-kHz vocalization of the isolatedinfant rat and the 22-kHz vocalization of theadult rat are produced by a similar laryngealmechanism, thus suggesting that the two vo-calizations are homologous (Blumberg andAlberts, 1991; Blumberg et al., 2000a). As dis-cussed, these vocalizations are not elicited insimilar contexts or by similar stimuli, nor arethey similarly modulated by pharmacologicalagents. For example, although clonidineevokes ultrasound production in infant rats(Kehoe and Harris, 1989; Blumberg et al.,2000a), it attenuates conditioned 22-kHz vo-calizations in adults (Molewijk et al.). In con-trast, cholinergic agonists increase 22-kHz vo-calizations in adults rats (Brudzynski, 2001)but not 40-kHz vocalizations in infants (Ke-hoe et al., 2001). These two developmentaldifferences alone point to the gaps in our un-derstanding of the origins and mechanisms ofrat ultrasound. Focusing on developmentalchanges in these vocalizations may be an in-formative approach for elucidating their un-derlying mechanisms.

Perhaps a synthesis of the various theo-retical viewpoints will be developed that de-scribes and explains the causes and functionsof these vocalizations in all their diversity. Inthe meantime, attempts to promote thesevocalizations as models for human psycho-logical conditions and psychiatric disordersshould be met with healthy skepticism.

REFERENCES

Allin JT and Banks EM (1971) Effects of temperature onultrasound production by infant albino rats. Devel-opmental Psychobiology 4:149-156.

Allin JT and Banks EM (1972) Functional aspects of ul-trasound production by infant albino rats (Rattusnorvegicus). Animal Behaviour 20:175-185.

Ardid D, Jourdan D, Eschalier A, Arabia C, Bars DL(1993) Vocalization elicited by activation of AS- andC-fibres in the rat. Neuroreport 5:105-108.

Barfield RJ, Auerbach P, Geyer LA, Mclntosh TK (1979)Ultrasonic vocalizations in rat sexual behavior.American Zoologist 19:469-480.

Bialy M, Rydz M, Kaczmarek L (2000) Precontact 50-kHz vocalizations in male rats during acquisition ofsexual experience. Behavioral Neuroscience 114:983-990.

Blanchard RJ, Blanchard DC, Agullana R, Weiss SM(1991) Twenty-two kHz alarm cries to presentationof a predator, by laboratory rats living in a visibleburrow system. Physiology and Behavior 50:967-972.

Blumberg MS (1992) Rodent ultrasonic short calls: lo-comotion, biomechanics, and communication.Journal of Comparative Psychology 106:360-365.

Blumberg MS, Alberts JR (1990) Ultrasonic vocalizationsby rat pups in the cold: an acoustic by-product oflaryngeal braking? Behavioral Neuroscience 104:808-817.

Blumberg MS and Alberts JR (1991) On the significanceof similarities between ultrasonic vocalizations ofinfant and adult rats. Neuroscience and Biobehav-ioral Reviews 50:95-99.

Blumberg MS and Alberts JR (1992) Functions and ef-fects in animal communication: reactions to Guil-ford & Dawkins. Animal Behaviour 44:382-383.

Blumberg MS and Alberts JR (1997) Incidental emis-sions, fortuitous effects, and the origins of commu-nication. In: Perspectives in ethology (ThompsonNS, ed.), pp. 225-249. New York: Plenum Press.

Blumberg MS, Efimova IV, Alberts JR (1992a) Ultrasonicvocalizations by rats pups: the primary importanceof ambient temperature and the thermal signifi-cance of contact comfort. Developmental Psy-chobiology 25:229-250.

Blumberg MS, Efimova IV, Alberts JR (1992b) Thermo-genesis during ultrasonic vocalization by rat pupsisolated in a warm environment: a thermographicanalysis. Developmental Psychobiology 25:497-510.

Blumberg MS, Kreber LA, Sokoloff G, Kent KJ (2000b)Cardiovascular mediation of clonidine-induced ul-trasound production in infant rats. Behavioral Neu-roscience 114:602-608.

Blumberg MS and Moltz H (1987) Hypothalamic tem-perature and the 22 kHz vocalization of the malerat. Physiology and Behavior 40:637-640.

Blumberg MS, Sokoloff G, Kent KJ (1999) Cardiovascu-lar concomitants of ultrasound production duringcold exposure in infant rats. Behavioral Neuro-science 113:1274-1282.

Blumberg MS, Sokoloff G, Kent KJ (2000a) A develop-mental analysis of clonidine's effects on cardiac rateand ultrasound production in infant rats. Develop-mental Psychobiology 36:186-193.

Blumberg MS and Sokoloff G (2001) Do infant rats cry?Psychological Review 108:83-95.

Page 396: 0195162854

Chapter 35. Vocalization 379

Blumberg MS and Stolba MA (1996) Thermogenesis,myoclonic twitching, and ultrasonic vocalization inneonatal rats during moderate and extreme cold ex-posure. Behavioral Neuroscience 110:305-314.

Borsini F, Podhorna J, Marazziti D (2002) Do animalmodels of anxiety predict anxiolytic-like effects ofantidepressants? Psychopharmacology 163:121-141.

Brown AM (1973) High frequency peaks in the cochlearmicrophonic response of rodents. Journal of Com-parative Physiology 83:377-392.

Brudzynski SM (2001) Pharmacological and behavioralcharacteristics fo 22 kHz alarm calls in rats. Neuro-science and Biobehavioral Reviews 25:611-617.

Brudzynski SM and Pniak A (2002) Social contacts andproduction of 50-kHz short ultrasonic calls in adultrats. Journal of Comparative Psychology 116:73-82.

Burgdorf J, Knutson B, Panksepp J (2000) Anticipationof rewading electrical brain stimulation evokes ul-trasonic vocalizations in rats. Behavioral Neuro-science 114:320-327.

Calhoun JB (1962) The ecology and sociology of theNorway rat. Bethesda: U.S. Department of Health,Education, and Welfare.

Crowley DE, Hepp-Reymond M-C (1966) Developmentof cochlear function in the ear of the infant rat. Jour-nal of Comparative and Physiological Psychology63:427-432.

Crowley DE, Hepp-Raymond M-C, Tabonite D, Palin J(1965) Cochlear potentials in the albino rat. Journalof Auditory Research 5:307-316.

Farrell WJ and Alberts JR (2000) Ultrasonic vocalizationsby rat pups after adrenergic manipulations of brownfat metabolism. Behavioral Neuroscience 114:805-813.

Farrell WJ and Alberts JR (2002a) Maternal responsive-ness to infant Norway rat (Rattus norvegicus) ultra-sonic vocalizations during the maternal behavior cy-cle and after steroid and experiential inductionregimens. Journal of Comparative Psychology 116:286-296.

Farrell WJ and Alberts JR (2002b) Stimulus control ofmaternal responsiveness to Norway rat (Rattusnorvegicus) pup ultrasonic vocalizations. Journal ofComparative Psychology 116:297-307.

Francis RL (1977) 22-kHz calls by isolated rats. Nature265:236-238.

Gourevitch G and Hack M (1966) Audibility in the rat.Journal of Comparative and Physiological Psychol-ogy 62:289-291.

Hofer MA and Shair HN (1987) Isolation distress in2-week-old rats: influence of home cage, social com-panions, and prior experience with littermates. De-velopmental Psychobiology 20:465-476.

Hofer MA, Brunelli SA, Shair HN (1994) Potentiation ofisolation-induced vocalization by brief exposure ofrat pups to maternal cues. Developmental Psy-chobiology 26:81-95.

Hofer MA, Masmela JR, Brunelli SA, Shair HN (1998)The ontogeny of maternal potentiation of the infantrats' isolation call. Developmental Psychobiology33:189-201.

Insel TR, Hill JL, Mayor RB (1986) Rat pup ultrasonicisolation calls: possible mediation by the benzo-diazepine receptor complex. Pharmacology Bio-chemistry and Behavior 24:1263-1267.

Kehoe P and Harris JC (1989) Ontogeny of noradrener-gic effects of ultrasonic vocalizations in rat pups. Be-havioral Neuroscience 103:1099-1107.

Kehoe P, Callahan M, Daigle A, Malinson K, Brudzyn-ski S (2001) The effect of cholinergic stimulation onrat pup ultrasonic vocalizations. DevelopmentalPsychobiology 38:92-100.

Knutson B, Burgdorf J, Panksepp J (1998) Anticipationof play elicits high-frequency ultrasonic vocaliza-tions in young rats. Journal of Comparative Psy-chology 1:65-73.

Knutson B, Burgdorf J, Panksepp J (2002) Ultrasonic vo-calizations as indices of affective states in rats. Psy-chological Bulletin 128:961-977.

Kraebel KS, Brasser SM, Campbell JO, Spear LP, SpearNE (2002) Developmental differences in temporalpatterns and potentiation of isolation-induced ul-trasonic vocalizations: influence of temperaturevariables. Developmental Psychobiology 40:147-159.

Lee HJ, Choi J-S, Brown TH, Kim JJ (2001) AmygdalarNMDA receptors are critical for the expression ofmultiple conditioned fear responses. Journal ofNeuroscience 21:4116-4124.

Miczek KA, Weerts EM, Vivian JA, Barros HM (1995)Aggression, anxiety and vocalizations in animals:GABAA and 5-HT anxiolytics. Psychopharmacology121:38-56.

Molewijk HE, van der Poel AM, Mos J, van der HeydenJAM, Olivier B (1995) Conditioned ultrasonic dis-tress vocalizations in adult male rats as a behav-ioural paradigm for screening anti-panic drugs. Psy-chopharmacology 117:32-40.

Noirot E (1968) Ultrasounds in young rodents. II.Changes with age in albino rats. Animal Behaviour16:129-134.

Noirot E (1972) Ultrasounds and maternal behavior insmall rodents. Developmental Psychobiology 5:371-387.

Okon EE (1971) The temperature relations of vocaliza-tion in infant Golden hamsters and Wistar rats.Journal of Zoology London 164:227-237.

Olivier B, Molewijk HE, van der Heyden JAM, vanOorschot R, Ronken E, MosJ, Miczek KA (1998) Ul-trasonic vocalizations in rat pups: effects of sero-tonergic ligands. Neuroscience and BiobehavioralReview 23:215-227.

Oswalt GL and Meier GW (1975) Olfactory, thermal,

Page 397: 0195162854

380 DEFENSE AND SOCIAL BEHAVIOR

and tactual influences on infantile ultrasonic vocal-ization in rats. Developmental Psychobiology 8:129-135.

Panksepp J and Burgdorf J (2000) 50-kHz chirping (laugh-ter?) in response to conditioned and unconditionedtickle-induced reward in rats: effects of social hous-ing and genetic variables. Behavioural Brain Re-search 115:25-38.

Roberts LH (1975a) The functional anatomy of the ro-dent larynx in relation to audible and ultrasonic cryproduction. Zoological Journal of the Linnaean So-ciety 56:255-264.

Roberts LH (1975b) Evidence for the laryngeal sourceof ultrasonic and audible cries of rodents. Journal ofZoology, London 175:243-257.

Roberts LH (1975c) The rodent ultrasound productionmechanism. Ultrasonics 13:83-85.

Sachs BD and Bialy M (2000) Female presence duringpostejaculatory interval facilitates penile erectionand 22-kHz vocalization in male rats. BehavioralNeuroscience 114:1203-1208.

Sales GD (1972a) Ultrasound and mating behaviour inrodents with some observations on other behav-ioural situations. Journal of Zoology, London 168:149-164.

Sales GD (1972b) Ultrasound and aggressive behaviourin rats and other small mammals. Animal behaviour20:88-100.

Sanders I, Weisz DJ, Yang BY, Fung K, Amirali A (2001)The mechanism of ultrasonic vocalization in the rat.In: Society for Neuroscience. San Diego, CA. No-vember 10-15, 2001.

Schreiber R, Melon C, Vry JD (1998) The role of 5-HTreceptor subtypes in the anxiolytic effects of electiveserotonin ruptake inhibitors in the rat ultrasonic vo-calization test. Psychopharmacology 135:383-391.

Sewell GD (1967) Ultrasound in adult rodents. Nature215:512.

Sewell GD (1970) Ultrasonic signals from rodents. Ul-trasonics 8:26-30.

Shair HN, Brunelli SA, Masmela JR, Boone E, Hofer MA(2003) Social, thermal, and temporal influences onisolation-induced and maternally potentiated ultra-sonic vocalizations of rat pups. Developmental Psy-chobiology 42:206-222.

Smotherman WP, Bell RW, Starzec J, Elias J, ZachmanTA (1974) Maternal responses to infant vocaliza-tions and olfactory cues in rats and mice. Behav-ioural Biology 12:55-66.

Sokoloff G and Blumberg MS (2001) Competition andcooperation among huddling infant rats. Develop-mental Psychobiology 39:1-9.

Takahashi LK (1992) Developmental expression of de-fensive responses during exposure to conspecificadults in preweanling rats (Rattus norvegicus). Jour-nal of Comparative Psychology 106:66-77.

Takahashi LK, Thomas DA, Barfield RJ (1983) Analysisof ultrasonic vocalizations emitted by residents dur-ing aggressive encounters among rats (Rattusnorvegicus). Journal of Comparative and Physiolog-ical Psychology 97:207-212.

Thomas DA and Barfield RJ (1985) Ultrasonic vocaliza-tion of the female rat (Rattus norvegicus) duringmating. Animal Behavior 33:720-725.

van der Poel AM, Noach EJK, Miczek KA (1989) Tem-poral patterning pf ultrasonic distress calls in theadult rat: effects of morphine and benzodiazepines.Psychopharmacology 97:147-148.

Vivian JA and Miczek KA (1991) Ultrasounds during mor-phine withdrawal. Pyshcopharmacology 104:187-193.

Wetzel DM, Kelley DB, Campbell BA (1980) Centralcontrol of ultrasonic vocalizations in neonatal rats:I. Brain stem motor nuclei. Journal of Comparativeand Physiological Psychology 94:596-605.

Winslow JT and Insel TR (1991) Endogenous opioids:Do they modulate the rat pup's response to socialisolation. Behavioral Neuroscience 105:253-263.

Yajima Y, Hayashi Y, Yoshii N (1980) The midbrain graysubstance as a highly sensitive neural structure forthe production of ultrasonic vocalizations in the rat.Brain Research 198:446-452.

Yajima Y, Hayashi Y, Yoshii N (1981) Identification ofultrasonic vocalization substrates determined byelectrical stimulation applied to the medulla oblon-gata in the rat. Brain Research 229:353-362.

Zhang SP, Davis PJ, Bandler R, Carrive P (1994) Brainstem integration of vocalization: role of the mid-brain periaqueductal gray. Journal of Neurophysi-ology 72:1337-1356.

Page 398: 0195162854

Cognition X/ll

Page 399: 0195162854

This page intentionally left blank

Page 400: 0195162854

Object Recognition

DAVE G. MUMBY36

OBJECT-RECOGNITION PARADIGMS

Object-recognition memory is the ability to dis-criminate the familiarity of previously en-countered objects.1 People with normal mem-ory may engage this ability hundreds of timeseach day, but impaired recognition occurs inmany memory disorders, including those re-sulting from Alzheimer's disease, stroke,chronic alcoholism, encephalitis, and trau-matic brain injury. The ability to distinguishbetween an object one has encountered pre-viously and one that is new is so fundamentalto normal memory function that understand-ing its neural bases seems necessary to developa comprehensive picture of how the brain re-members things. Such knowledge may alsocontribute to better methods of diagnosingand treating certain memory disorders.

Rats also distinguish between objectsthey have previously encountered and onesthey have not. The extent to which object-recognition memory involves similar pro-cesses in rats and humans is not entirely dear.Standardized tasks for assessing object recog-nition in rats, the effects of various brain le-sions, and effects of drugs on this ability allsuggest similar processes.

Two paradigms are most often used toassess object recognition in rats: delayednonmatching-to-sample (DNMS) and novel-object-preference (NOP). On DNMS tasks, asample object is briefly presented, and after aretention delay, it is presented again, alongwith a novel object (i.e., one the rat has notpreviously encountered on the current ses-

sion). The rat is rewarded for selecting thenovel object. Reliably accurate performancerequires, among other things, that the rat canrecognize the sample object. Memory de-mands are controlled by varying duration ofthe delay or number of objects to rememberon each trial. There are several trials per ses-sion, each using a different sample and novelobject, so rats are consistently rewarded forselecting an unfamiliar object. Most DNMSprocedures use pseudo-trial-unique objects-particular objects never recur within a sessionbut may recur across multiple sessions widelyseparated in time. Rats also learn delayedmatching-to-sample with objects, but theymaster DNMS more quickly because thenonmatching response is consistent with theirnatural bias for selecting novel objects.

The NOP task takes advantage of the ten-dency of rats to investigate novel objects morethan familiar objects (Berlyne, 1950). Con-ventional procedures are similar to those de-scribed by Ennaceur and Delacour (1988): Arat is placed in an open-field arena and allowedto explore two identical sample objects for afew minutes. The rat is then removed for adelay, after which it returns to the arena withtwo new objects—one is identical to the sam-ple and the other is novel. Normal rats spendmore time exploring the novel object duringthe test, indicating that they recognize thesample object. With conventional procedures,rats can show a novel-object preference afterdelays of up to 24 hours and, with modifiedprocedures, up to several weeks (see "Retro-grade Object Recognition").

383

Page 401: 0195162854

384

SENSORY SYSTEMS

Investigators often make the tacit assumptionthat rats perform DNMS and NOP tasks bydiscriminating the familiar visual features ofsample objects. Rats have the opportunity tosee, feel, and smell the objects on conven-tional versions of both tasks, however. Care-ful observation of a rat's behavior may givefew clues to the type of sensory informationused in a particular trial. A rat may sniff andpalpate an object while investigating it, butthis does not mean the rat is using olfactoryor tactile information to make its choice. If,however, a rat performing a DNMS task con-sistently veers toward the correct object whileapproaching it and is still several centimetersaway, it is likely using vision.

DELAYED NONMATCHING-TO-SAMPLE

TASK VARIANTS

Figure 36-1 illustrates a Y-maze procedure inwhich distinctive goal boxes containing ob-

Figure 36-1. Y-maze DNMS task. Pluses indicate whicharms the rat will be rewarded for entering. A, The first trialbegins with the rat in one arm, from which it can enter ei-ther of the other two arms, both containing identical goalboxes. B to D, Each successive choice is between a box iden-tical to the one containing the rat and a box the rat has notencountered during the current session. (From Aggleton,1985.)

COGNITION

jects are inserted into the arms (Aggleton,1985). The rat is confined for 20 seconds in asample box and then placed in a "featureless"box for the retention delay, after which it re-ceives access to the other two maze arms.Each arm contains a distinctive box with ob-jects—one matches the sample box, and theother is novel. The rat is rewarded if it entersthe novel box, which then serves as the sam-ple for the next trial.

Other DNMS versions use discrete trialsconsisting of a sample phase and a choicephase. One method (Fig. 36-2A-D) uses a run-way with a start area separated from a goalarea by an experimenter-controlled door(Rothblat and Hayes, 1987). The goal areacontains food wells over which objects are po-sitioned. For the sample phase, the door to thestart area is raised, and the rat approaches anddisplaces a sample object, for which it receivesfood reward. The rat is picked up and placedback in the start area for the delay, after whichit again has access to the goal area, which nowcontains a duplicate of the sample and a novelobject. A version in which the sample phaseis located at one end of the apparatus and thechoice phase is located at the other end, andrats are not handled between or within trials(Kesner et al., 1993) (Fig. 36-2E-H). The ver-sion shown in Figure 36-3 is similar but has acentral start compartment, a pair of food wellsat either end, and the end of the apparatuswhere the sample and choice phases occurchanges randomly across trials (Mumby et al.,1990).

Before DNMS training, rats are familiar-ized with the apparatus and shaped to displaceobjects from food wells. This can be accom-plished by training on a simple object-discrimination task, on which the same twoobjects are repeatedly presented together andselection of one of them is rewarded (Mumbyet al., 1990). Rats learn object discriminationsquickly, probably because manipulation ofsmall objects is a natural behavior of ratsseeking food (Barnett, 1956). The acquisitionphase of DNMS training ensues, using a brief

Page 402: 0195162854

Chapter 36. Object Recognition 385

Figure 36-2. A to D, DNMS proce-dures similar to those described byRothblat and Hayes (1987). The ratis picked up and returned to the startarea between the sample and choicephases of each trial. E to H, Proce-dures used by Kesner et al. (1993). Acentral door is raised and lowered bythe experimenter to control the re-tention delay and the rat's opportu-nity to shuttle back and forth be-tween the sample and choice ends ofthe runway. The rat is not handledbetween or within trials.

retention delay of only a few seconds, untilthe rat reaches some criterion level of accu-racy, establishing that is has learned the non-matching rule. Thereafter, recognition is as-sessed with longer delays, up to severalminutes. Acquisition is measured by the num-ber of trials or errors required to reach the per-formance criterion at the brief delay, and per-formance at longer delays is expressed aspercentage of trials that are correct. Proce-dures that vary the number of objects to re-member on each trial can be found for Y-mazeDNMS (Steele and Rawlins, 1993) and discretetrial versions (Mumby et al., 1995).

MINIMIZING EXPERIMENTER EFFECTS

Rats likely perceive humans as large, noisy,smelly potential predators. The experimenterplays an interactive role in DNMS testing, soit is essential that rats are first well-tamed. Arat that perceives the experimenter as themost interesting thing in the room will paymore attention to the experimenter than tothe task. This is often the most difficult andfrustrating aspect of DNMS testing for persons

inexperienced with rats and with how they re-act to movements and sounds.

Experimenters should also monitor theirown behavior for inadvertent cues that ratscould use to solve the task, such as uncon-scious movements the experimenter reliablymakes as the rat approaches the correct ob-ject. If either the sample or the novel objectis consistently handled more than the other,rats may learn to solve the task by discrimi-nating the relative strength of the experi-menter's scent on the objects (Mumby,Kornecook et al., 1995). The easiest solutionis to use two identical copies of an object asthe sample on each trial—one for the samplephase and the other for the test phase—so thatall of the objects can be positioned before thetrial begins.

FOOD DEPRIVATION

A common practice when using tasks moti-vated by food reward is to restrict the dailyfood intake of rats. Typically, rats are main-tained at body weights around 80% to 85% ofthe weight of age-matched, free-feeding rats.

Page 403: 0195162854

386 COGNITION

Figure 36-3. DNMS procedures described by Mumby et al.(1990). A, Before each trial, the rat is enclosed in the cen-tral area, and two objects are positioned over food wells,one at each end. B, A door is opened, and the rat ap-proaches, and displaces the sample, and receives a food pel-let. C, The sample is placed over the vacant food well atthe other end. The second door remains closed for the re-tention delay. D, The rat displaces an object, and receivesa food pellet if it is the novel object. The rat returns to thecentral area until the next trial; it is not handled betweenor within trials.

This level of deprivation is unnecessary andmay be counterproductive when training ratson DNMS. Hyperactivity and problems withresponse inhibition interfere with accuracy,and both increase with increasing food depri-vation. Deprivation to 85% of free-feedingweight may help initially, when rats are learn-ing basic task procedures, but once they arereadily approaching and displacing objects,they can be maintained at 90% to 100% offree-feeding weights.

Success in training rats on DNMS re-quires that they move about the apparatus ata pace that is neither too slow nor too fast andthat they do not show persistent side biases(e.g., always picking the object on the leftside). Rats that run too fast make frequent er-rors and may not take time to examine thesample at the beginning of a trial. Inexperi-enced experimenters may be pleased that theirrats run quickly and waste no time while fran-tically leaping into the goal area and knockingover the first object they see, but such be-havior significantly prolongs the acquisitionphase of DNMS training. Fast-running andside biases often appear together. Slowingdown a hasty rat is often all that is needed toeliminate a side bias, but additional measuresmay be required, such as remedial training ona simple object-discrimination task.

Slow rats are easily distracted betweentrials and during retention delays. This can befrustrating for impatient experimenters andmay lead to unwise improvisations in proto-col in attempts to get a rat back on task.Optimal performance requires adjustment ofeach rat's daily food ration in accordance withits current pace—rats that are too slow receiveslightly less food and rats that are too frenziedreceive slightly more. The objective is to op-timize and equalize the pace of the rats. Thisis more important for successful DNMS train-ing than making sure all rats eat the sameamount of food.

INTERPRETING TREATMENTEFFECTS ON DNMS TASKS

Before concluding that a DNMS deficit is dueto failure of object-recognition memory, sev-eral other possibilities must be considered.Some treatments produce nonspecific effectsthat interfere with successful DNMS per-formance, such as hyperactivity, an alteredstress response, or problems with response in-hibition. Nonspecific effects can last severaldays after highly invasive treatments, such asbrain surgery, but often they subside eventu-

Page 404: 0195162854

Chapter 36. Object Recognition

ally. If rats are not given a chance to overcometransient nonspecific effects, DNMS perform-ance may be impaired even if recognitionmemory is unaffected. One solution is to trainthem first on an object-discrimination task be-fore commencing with DNMS testing. Ratslearn object-discrimination problems rela-tively easily, even after some treatments thatimpair DNMS performance, and they canovercome various nonspecific treatment effectswhile performing an object-discriminationtask.

It is often assumed that after learning thenonmatching rule, rats will apply it on everytrial, so accuracy will depend mostly onwhether they recognize the sample object. Aclassic pattern of treatment effect on memorytests is the delay-dependent deficit, in which mag-nitude of the deficit increases as retention de-lay increases. The usual interpretation is thatperformance is normal at brief delays becausethe conditions do not sufficiently tax memoryprocesses affected by the treatment. But delay-dependent effects can also occur for nonm-nemonic reasons. Even after reaching asymp-totic performance at short delays, certain skillsmust be mastered to achieve good scores atlonger delays. With practice at long delays, ratsmay become better at attending to object fea-tures, avoiding distraction and frustration dur-ing the retention delay, withholding hastychoices, or discriminating the sample phasefrom the choice phase of a trial. Deficienciesin these skills diminish accuracy, even in ani-mals with normal object-recognition abilities.

Pretreatment training can reduce thelikelihood of seeing a treatment effect onDNMS performance (Mumby, 2001) but maymake it easier to interpret an effect when oneoccurs. If a treatment produces DNMS deficitsonly when rats are not pretrained, it is possi-ble that the treatment interferes with learninghow to perform well on the task but may ac-tually have no effect on object-recognitionmemory. To understand why this is so, keepin mind what rats learn during DNMS train-ing: They do not learn how to recognize

387

objects—they already possess that capacity be-fore the experiment begins. During training,rats learn the reward contingency (the non-matching rule) and various skills necessary forgood performance. Pretreatment trainingdoes not guarantee that acquired skills will beunaffected by a treatment, but if rats readilyapproach and displace objects and performwith better-than-chance accuracy on the firstfew posttreatment trials, they likely remem-ber the nonmatching rule and general taskprocedures.

NOVEL-OBJECT PREFERENCE

BASIC TEST PROCEDURES

The NOP task has become a popular methodfor assessing object recognition in rats, mainlyfor practical reasons: it does not require fooddeprivation or prolonged training, so experi-ments can be completed in considerably lesstime than with DNMS tasks. The NOP task isjust one of many procedures to assess howrats respond to various types of novelty. Dur-ing the 1950s, while most Russian and West-ern psychologists were preoccupied with un-derstanding how associations are learnedthrough Pavlovian and operant conditioningprocedures, Berlyne and some of his contem-poraries were interested in how animals re-spond to certain features of the environmentsimply because they are new.

In one influential study, Berlyne (1950) al-lowed rats to explore three copies of an ob-ject in an open field arena. When one objectwas replaced with a novel object, the ratsshowed an exploratory preference for thenovel object. It is not clear to what extent thisoccurred because that object was unfamiliaror because it was the odd object of the triplet,but the phenomenon inspired many otherstudies that examined conditions affectinghow rats respond to novelty. Most of the rel-evant work of this period has been reviewedby Berlyne (1960) and Fowler (1965). There

Page 405: 0195162854

388 COGNITION

was much theorizing about emotions anddrive states, such as surprise and curiosity,none of which is particularly relevant tounderstanding the processes of object-recognition memory. In many studies, how-ever, behaviors of interest included object ex-ploration, and much was learned about thevariables that affect this behavior. Knowledgeof some of these factors is essential to suc-cessful planning and interpretation of NOP ex-periments. Varieties of exploratory behaviorand procedures to evoke and measure it aredescribed by Berlyne (1960) and Fowler (1965)and in a recent analysis by Hughes (1997).

There is no standard operational definitionof object exploration, but most investigators findit sufficient to include criteria that are likely tocapture most of a rat's actual object investiga-tion, such as having the head oriented towardthe object and within a certain distance, usuallya few centimeters. Renner and his colleaguesprovided useful analyses of the structure and or-ganization of exploration and object investiga-tion in rats (Renner, 1987; Renner and Seltzer,1991, 1994). When using the NOP test to assessobject recognition, it is not necessary to analyzethe fine details of how rats interact with the ob-jects. It is essential, however, to have an opera-tional definition that can be applied easily andconsistently.

Time spent exploring each object duringthe test phase is used to calculate a measureof relative exploratory preference for the sam-ple and novel objects. For example, an explo-ration ratio reflects the proportion of total ob-ject exploration that was spent exploring thenovel object [Tnovei/(Tnovei + Tsample)]. An-cillary measures include time spent investi-gating the sample during the familiarizationphase and the difference in time spent ex-ploring the novel object and the sample on thetest phase. Some investigators use differencescores during the test as the primary depend-ent measure, but a proportional or ratio meas-ure has the advantage of controlling for indi-vidual differences in total object exploration.

ADDITIONAL FACTORS THATAFFECT OBJECT EXPLORATION

Whether rats approach or avoid a novelobject depends on the familiarity of the en-vironment in which it is encountered.When the environment is familiar, novelobjects evoke approach and investigation;when the environment is unfamiliar, ratsavoid novel objects (Besheer and Bevins,2000; Montgomery, 1955; Sheldon, 1969).Accordingly, rats should be familiarized tothe apparatus before NOP testing, by al-lowing them to explore it for a short period(10 to 15 minutes) on two or three occa-sions. Subsequently, during NOP tests, re-sponses evoked by background stimuli areless likely to overshadow how the rats re-spond to the objects.

The complexity of objects affects howmuch investigation they evoke (Berlyne,1955). Before NOP tests are performed, ob-jects should be screened by measuring howmuch investigation each one evokes in a non-choice situation. Objects that evoke either toomuch or too little investigation should be ex-cluded. This reduces variance in undiscrimi-nated object exploration and produces cleanerresults on NOP tests.

The duration of the preference test caninfluence the likelihood of detecting a novel-object preference. The preference tends tobe robust during only the first 1 or 2 min-utes of a test session and diminishes there-after, presumably because both objects be-come equally familiar as they are explored(Dix and Aggleton, 1999). A minute-by-minute assessment of differential explo-ration shows how preferences change overthe test phase, allowing one to focus on themost sensitive portion (Mumby et al., 2002).The inclusion of portions of the test thatcome after rats stop discriminating merelyadds noise to the data, which may obscurea significant preference that occurs in theearly portions of the test.

Page 406: 0195162854

Chapter 36. Object Recognition

INTERPRETING TREATMENT EFFECTS

ON NOP TASKS

Despite the simplicity of NOP procedures, itis a challenge to discern when a treatment ef-fect occurs because of impaired object recog-nition or because of some other reason. Whena treatment has no effect and rats spend moretime exploring novel objects than sample ob-jects, it is likely that object-recognition abili-ties are not grossly impaired. When the pref-erence is disrupted, however, and even ifperceptual impairments can be ruled out,there are several potential reasons for a fail-ure to spend more time exploring the novelobject. One possibility is that a rat does notrecognize the sample; another possibility isthat the instinctive bias for exploring novel ob-jects has been overshadowed or abolished.

A fundamental question that should beanswered before comparing the scores of twogroups is whether performance within eachgroup indicates a significant preference. Aone-sample t test can determine whether themean exploration ratio of a group is signifi-cantly different from the chance level of 0.5.(Most studies report ratios in control rats be-tween 0.6 and 0.7.) The first comparison be-tween two groups, therefore, is on the appro-priate nominal scale (i.e., preference versus nopreference). Comparisons of group meansshould follow, but it is important to note thatthose comparisons merely concern the degreeof preference. It is not clear to what extentstronger preference for the novel object in onerat relative to another rat should be taken asindicating superior recognition abilities.

RETROGRADE OBJECT RECOGNITION

Retrograde object recognition is the ability todiscriminate the familiarity of objects thatwere encountered before a treatment. Whenthe treatment is a brain lesion, tasks areneeded on which control animals show sig-

389

nificant retention after the period of post-surgery recovery. Neither the DNMS nor con-ventional NOP procedures are suitable. How-ever, rats can show a novel-object preferenceafter delays of several weeks if they are givenrepeated sample exposures (Mumby, Glenn,Nesbitt, & Kyriazis, 2002). To increase theamount of object exploration during the test,after such long retention delays, rats shouldfirst be rehabituated to the chamber beforethe test.

A NEW NOVEL-OBJECTPREFERENCE TASK

Evolutionary perspectives emphasize the im-portance of considering the behavior of ratsin their natural environment when designinglaboratory paradigms (Timber-lake, 1984). TheNOP task takes advantage of an innate ex-ploratory bias, but the conventional proce-dures may constrain certain niche-relatedexploratory responses. For example, when ex-ploring their natural environment, rats travelfrom place to place, investigating objectsalong the way, with a tendency to move tonew places and objects rather than to returnto ones they have just investigated. These as-pects of natural exploratory behavior are con-tinuously thwarted in the confined space of astandard open field with two objects.

A modified NOP procedure circumventsthese limitations. (Mumby et al., submitted).The apparatus is a circular track that can bedivided into different sectors (Fig. 36-4). Eachdivider has a door that can be set to open inonly one direction, so when rats pass into anew sector, they cannot return to the previ-ous sector. With all doors set to open in thesame direction, rats travel around the track inthat direction only, spending as much time asthey choose in each sector and becoming fa-miliar with several objects concurrently, asthey encounter different pairs of identical ob-jects in each sector. For the test, each sector

Page 407: 0195162854

390 COGNITION

Retention Delay

Figure 36-4. A, Circular-track for assessing novel-objectpreference. (See text for details.) B, Data from rats testedwith four object pairs, and retention delays of 15 and 60minutes (n = 12) or 24 hours (n = 12). Exploration ratio isthe proportion of total object exploration during the test di-rected toward novel objects.

contains a copy of the sample and a novel ob-ject, and exploration of each object is recordedas the rat travels around the track. Becauserats cannot return to a sector after they leave,several measurements of novel-object prefer-ence can be taken on each trial. Different ma-nipulations can be made in different sectors,enabling simultaneous assessment of a rat's re-sponse to different types of novelty.

CONCLUSION

Despite superficial similarities, DNMS andNOP tests engage different behavioral andmotivational systems and therefore entail dif-ferent procedural pitfalls and interpretationalchallenges. Still, findings from lesion experi-ments using either task have so far been fairlyconsistent, suggesting that the hippocampusis not critical for object-recognition memory(Mumby, 2001), whereas the perirhinal cortexplays a more significant role. These conclu-sions are consistent with findings of changesin single-unit responses and c-fos expression

within the perirhinal cortex produced by re-peatedly presented visual stimuli (Aggletonand Brown, 1999). Studies of the neural basesof object-recognition memory are beginningto focus on other structures and circuits, butit is too early to ascribe a particular role inobject-recognition memory to any of them.

NOTE

1. This chapter discusses behavior pertaining totests that require discriminating the familiarity of three-dimensional objects. Not considered are tasks that re-quire learning relationships among individual objectsand other events (e.g., reward).

REFERENCES

Aggleton JP (1985) One-trial object recognition by rats.Quarterly Journal of Experimental Psychology37B:279-294.

Aggleton JP and Brown MW (1999) Episodic memory,amnesia, and the hippocampal-anterior thalamicaxis. Behavioral Brain Science 22:425-444.

Barnett SA (1956) Behavior components in the feedingof wild and laboratory rats. Behaviour 9:24-43.

Besheer J and Bevins RA (2000) The role of environ-mental familiarization in novel-object preference.Behavioural Processes 50:19-29.

Berlyne DE (1950) Novelty and curiosity as determi-nants of exploratory behaviour. British Journal ofPsychology 41:68-80.

Berlyne DE (1960) Conflict, arousal, and curiosity (Har-low HF, ed.). New York: McGraw-Hill.

Berlyne DE (1955) The arousal and satiation of percep-tual curiosity in the rat. Journal of Comparative andPhysiological Psychology 48:238-246.

Dix SL and Aggleton JP (1999) Extending the sponta-neous preference test of recognition: evidence ofobject-location and object-context recognition. Be-havioural Brain Research 99:191-200.

Ennaceur A and Delacour J (1988) A new one-trial test forneurobiological studies of memory in rats: I. behav-ioural data. Behavioural Brain Research 31:47-59.

Fowler H (1965) Curiosity and exploratory behavior.New York: Macmillan.

Hughes RN (1997) Intrinsic exploration in animals: mo-tives and measurement. Behavioural Processes41:213-226.

Kesner RP, Bolland BL, Dakis M (1993) Memory for spa-tial locations, motor responses, and objects: triple

Page 408: 0195162854

Chapter 36. Object Recognition 391

dissociation among the hippocampus, caudate nu-cleus, and extrastriate visual cortex. ExperimentalBrain Research 93:462-470.

Montgomery KG (1955) The relation between fear in-duced by novel stimulation and exploratory behav-ior. Journal of Comparative and Physiological Psy-chology 48:254-260.

Mumby DG (0000) Reducing constraints on exploratorybehavior enhances novel-object preference in rats,(submitted to Learning and Memory).

Mumby DG (2001) Perspectives on object-recognitionmemory following hippocampal damage: lessonsfrom studies in rats. Behavioural Brain Research127:159-181.

Mumby DG, Kornecook TJ, Wood ER, Pinel JPJ (1995)The role of experimenter-odor cues in the per-formance of object-memory tasks by rats. AnimalLearning and Behavior 23:447-453.

Mumby DG, Gaskin S, Glenn MJ, Schramek TE,Lehmann H (2002) Hippocampal damage and ex-ploratory preferences in rats: memory for objects,places, and contexts. Learning and Memory 9:49-57.

Mumby DG, Pinel JPJ, Kornecook TJ, Shen MJ, RedilaVA (1995) Memory deficits following lesions of hip-

object-memory test battery. Psychobiology 23:26-36.

Mumby DG, Pinel JPJ, Wood ER (1990) Nonrecurring-items delayed nonmatching-to-sample in rats: a newparadigm for testing nonspatial working memory.Psychobiology 18:321-326.

Mumby DG, Glenn MJ, Nesbitt C, Kyriazis DA (2002)Dissociation in retrograde memory for object dis-criminations and object recognition in rats withperirhinal cortex damage. Behavioural Brain Re-search 132:215-226.

Renner MJ (1987) Experience-dependent changes in ex-ploratory behavior in the adult rat (Rattus norvegicus):overall activity level and interactions with objects.Journal of Comparative Psychology 101:94-100.

Renner MJ and Seltzer CP (1991) Molar characteristicsof exploratory and investigatory behavior in the rat(Rattus norvegicus). Journal of Comparative Psy-chology 105:326-339.

Renner MJ and Seltzer CP (1994) Sequential structure inbehavioral components of object investigation byLong-Evans rats. Journal of Comparative Psychol-ogy 108:335-343.

Rothblat LA and Hayes LL (1987) Short-term objectrecognition memory in the rat: nonmatching withtrial-unique stimuli. Behavioral Neuroscience 101:587-590.

Sheldon AB (1969) Preference for familiar versus novelstimuli as a function of the familiarity of the envi-ronment. Journal of Comparative and Physiologicalri«,,,,l»«1 A^.T ^-7.ri^ col

Steele K. and Kawlins JJNP (iyy3; me errects or mp-pocampectomy on performance by rats of a runningrecognition task using long lists of non-spatial items.Brain Research 54:1-10.

Timberlake W (1984) An ecological approach to learn-ing. Learning and Motivation 15:321-333.

Page 409: 0195162854

Piloting

ETIENNE SAVE AND BRUNO POUCET37

THE PERCEPTION OF SPACE

In the wild, rats restrict their activity to a homerange, which represents only a minor portionof the potentially useful environment (Bovet,1998). An important place within the homerange is the nest, which provides protectionagainst predators and climatic conditions.Nevertheless, rats must leave their nest tolook for food, water, and conspecifics, so theyspend a substantial amount of time in movingfrom one place to another. Their survival de-pends on their ability to memorize locations(places) and to use behavioral strategies tonavigate efficiently between their home baseand other places of interest.

The selection of appropriate navigationalstrategies is primarily determined by the per-ception of space, that is, by the nature of thecues that can be used for navigation. Thereare two main categories of spatial cues: allo-thetic and idiothetic. Allothetic cues are pro-vided by the environment and include visual,olfactory, and auditory information. Allotheticnavigation refers to the process of "determin-ing and maintaining a course or trajectoryfrom one place to another" (Gallistel, 1990) byusing environmental cues. Idiothetic cues arederived from the animal's own movements,including information provided by the vestib-ular, proprioceptive, and somatosensory sys-tems; efference copies of motor commands;and external motion-related information suchas optic flow. Idiothetic cues support a formof navigation called dead reckoning, in whichan animal locates its starting point relative to

its current position (see Chapter 38). In recenthypotheses, allothetic navigation and deadreckoning are assumed to be complementarysystems. The purpose of the present chapteris to describe allothetic navigation and themethods for studying allothetic navigation inthe laboratory. Because a remarkable array oftasks have been developed, this chapter re-views only those tasks that have receivedwidespread use and are important for under-standing the theoretical basis of allotheticnavigation.

THEORIES OF SPATIAL LEARNING

Several theories have been developed to ac-count for spatial learning. For stimulus-response (S-R) theorists, spatial learning con-sists of chaining together a number of motorresponses that link relevant external stimuli.This view was challenged by Tolman (1948),who argued that rats can learn the location ofa place where they have been rewarded inde-pendent of the specific movements necessaryto reach it. Tolman explained this ability byintroducing the notion of a cognitive map, amental representation held by the rat in whichit encodes the elements of a task and the spa-tial relationships between these elements.

The Tolman notion of a cognitive mapwas further developed by O'Keefe and Nadel(1978), who proposed a theoretical frame-work to account for spatial behavior and itsneural bases. They made a distinction be-tween two main processes used by rats to per-

392

Page 410: 0195162854

Chapter 37. Piloting 393

form spatial tasks. They proposed that rats canuse either a taxon system or a locale system.The taxon system is based on S-R associations,whereas the locale system is based on spatialrelationships in the form of a cognitive map.O'Keefe and Nadel further proposed that thebrain regions subserving these forms of spa-tial navigation differ and are at least in part in-dependent.

In contrast to O'Keefe and Nadel's (1978)notion of a cognitive map, associationists suchas Sutherland and Rudy (1989) propose thatall spatial behaviors can be described in termsof learned associations but that it is the formof the association, simple or configural, thatdetermines which brain structures will beused. Simple associations correspond to S-Rassociations in theories of conditioning. Con-figural associations combine the representa-tions of elementary stimulus events to buildglobal representations. The configural associ-ation system also stores associations betweenthe elementary events and the global repre-sentation. According to this theory, placelearning is assumed to involve the configuralassociation system, which encodes and storesthe configuration of spatial relationships be-tween cues that specifically define a location.

SPATIAL STRATEGIES

The taxon and locale systems enable the ratsto use different strategies, including cue nav-igation, guidance, and route strategies (taxonsystem), and place navigation strategy (localesystem).

CUE NAVIGATION

The most elementary navigational strategy isto move toward or away from a directly per-ceived cue. The use of such a system requiresa radial stimulation gradient field centered onthe source (Benhamou and Bovet, 1992). Thenotion of gradient field refers to the variation

of intensity of the source as a function of thedistance. In the simplest case, intensity variesmonotonically with the distance. The animalcan therefore reach the goal (i.e., the origin ofthe gradient) by navigating in the gradient di-rection or, conversely, avoid a place by navi-gating in the opposite direction. A gradientfield may originate from auditory and olfac-tory cues. Navigation toward a conspicuousvisual cue has similar characteristics. The vi-sual cue is considered to be at the center of aradial gradient where the gradient directioncorresponds to the direction in the light flux.Thus, reducing the distance from the visualcue (which leads to an increase in the appar-ent cue size) is equivalent to navigating alongthe axis of greatest intensity. The animal canalso use a similar mechanism to reach a hid-den goal that is closely associated to a salientlandmark (called a beacon). A landmark istherefore considered to be an intermediategoal whose spatial contiguity with the realgoal has been learned through an associativeprocess.

GUIDANCE

If the goal cannot be directly perceived and ifthere is no landmark closely associated to thegoal, the animal must rely on environmentallandmarks to navigate. One strategy is to uselandmarks as directional cues. Guidance,therefore, requires that the animal directs itsattention to particular landmarks and main-tain some egocentric (i.e., relative to the ani-mal) spatial relationships with respect to theselandmarks to reach the goal (O'Keefe andNadel, 1978). A place-recognition process isthen needed to identify the goal location. Ratsmay be guided by either individual or config-urations of landmarks. For example, they maybe able to memorize a view of the spatialarrangement of landmarks seen from the goal(snapshot) and so adjust movements until acurrent view matches a memorized view(Collett et al., 1986).

Page 411: 0195162854

394

ROUTES

To reach a remote goal, cue navigation andguidance may be elaborated into route learn-ing. In route learning, rats may memorize a se-quence of associations coupling specific land-marks and motor responses, such as, "afterreaching the rock, turn left." Each landmarkis considered as an intermediary goal and thesequence of these goals constitutes a route.Routes allow rapid navigation at the expense ofbehavioral flexibility. Environmental changesthat alter the sequence of landmarks may befatal to successful navigation. In addition, re-tracing the route cannot be achieved by sim-ply inverting the order of the landmarks butrequires active learning of a new sequence.

PLACE NAVIGATION

Place navigation requires formation and useof a map-like representation of the environ-ment (a cognitive map) that encodes the geo-metrical relationships between landmarks andplaces, independent of the animal's position.The use of a mapping strategy allows an ani-mal to reach a goal from a variety of points,including over trajectories that it has not pre-viously used. A rat thus is able to infer its lo-cation relative to particular places whose po-sition is encoded in the representation and tocompute optimal trajectories to reach theseplaces. Place navigation is flexible because if apossible path is obstructed or a landmark ismissing, the animal remains able to generatesuccessful trajectories by performing detoursor shortcuts and by adapting its navigationalbehavior to environmental changes.

PROCEDURAL, WORKING,AND REFERENCE MEMORY

Spatial learning requires different kinds ofmemory. Procedural memory is the memory ofhow to perform a task (e.g., visit an arm maze,get a reward, return to a place, etc.). Workingmemory is a short-term store that allows a rat

COGNITION

to remember that, for example, the goal is ata specific location during the current trial.Erasing the content of working memory aftera trial enables the animal to learn a new goallocation in the next trial. If the goal locationremains constant across trials, it is encoded inthe long-term reference memory.

BEHAVIORAL TESTS

In this section, we describe a number of be-havioral tests that have been developed toevaluate and study allothetic navigationstrategies. Nevertheless, it should be borne inmind that each of the tasks can be solved byone or more of the processes described ear-lier, and so special "probe" trials are neededto identify a particular strategy used.

THE T-MAZE

Rodents have a tendency to optimize their for-aging behavior so that they avoid enteringplaces that they have already visited, wherefood may have been depleted or is absent. Ina T- or Y-maze, this results in alternation be-havior. Two kinds of tasks can be conducted.First, the spontaneous alternation paradigmexploits the innate tendency of rodents to al-ternate arm choices on successive trials. Sec-ond, in the delayed forced-choice alternationtask, alternation is rewarded whenever the an-imal enters an arm that is different from thatspecified by the experimenter during a sam-ple trial (Fig. 37-1 A). The nature of the strat-egy used by the animals to solve these tasksremains unclear. Because the goal arm differsfrom one trial to another, rats cannot learn afixed association between a single responseand a location. There is evidence that theyuse a variety of cues, including extramaze spa-tial cues and movement-related cues (Dud-chenko, 2001). Performance in these tasks alsodepends on working memory, because a ratmust keep track of which arm has been vis-ited on the previous trial. In addition, in the

Page 412: 0195162854

Chapter 37. Piloting 395

Figure 37-1. A, Delayed forced alternation task in theT-maze. During the sample trial, the rewarded arm (+) isspecified by the experimenter (the other arm is blocked).After a delay, the animal is required to choose between thetwo arms (choice trial) and is rewarded if it chooses theother arm. B, Response-versus-place strategy in the crossmaze. During training, the rat is trained to enter the rightarm. During the test trial, the animal is released from theopposite arm and is required to choose between the rightarm and the left arm. A right turn indicates that the rat usesa response strategy based on stimulus-response association(association between a body turn and a place). A left turnindicates that the rat uses a place strategy based on a cog-nitive map.

delayed forced-choice alternation task versionof the task, the introduction of a delay be-tween the sample trial and the choice trial al-lows increased burden to be placed on theworking memory demand.

THE CROSS-MAZE

The cross-maze has been developed to exam-ine whether rats use a response strategy, thatis, a learned association between a specificbody turn and a place (taxon system) or aplace strategy (locale system), to reach a tar-get. The cross-maze consists of a sort of dou-ble T-maze constructed so that an animal can

reach the choice point connecting the twoarms from two opposite starting arms. Theapparatus is surrounded by numerous cues.During training, the animal is released fromone starting arm and has to enter a specifiedarm, for instance, the right arm, to obtain afood reward. For the test trial, the animal isplaced in the arm opposite the previous start-ing arm. If the rat performs a right turn toenter the goal arm, this suggests that it uses aresponse strategy based on learned S-R asso-ciation. This outcome supports the S-R the-ory. In contrast, if the rat performs a left turn,this suggests that it can use a place strategy,and this supports the cognitive map theory(Fig. 37-1B). It is possible that rats are able toaccess either strategy (Packard and McGaugh,1996).

THE RADIAL ARM MAZE

The radial arm maze is widely used to studyspatial behavior, and many versions of thetask have been developed (Foreman and Er-makova, 1998). The maze consists of a num-ber of elevated runways (most often, eight)that radiate from a central platform, like thespokes of a wheel. In the continuous versionof the task developed by Olton and Samuel-son (1976), the animal is released on the cen-tral platform and allowed to freely explore themaze to collect small pieces of food from theends of every arm. The optimal strategy con-sists of the rat visiting all of the arms withoutentering an already visited arm. A trial is com-plete when all food has been collected. Ratsare extraordinarily good at performing thetask; how they identify and remember the armlocations using the numerous cues surround-ing the maze is not completely understood.Whether each arm location is encoded as aplace in a cognitive map that guides the ani-mal's choice or is considered to be an inde-pendent item in a list remains a matter ofdebate. The task also requires proceduralmemory (food can be found at the end of anarm) and working memory (a given arm has

Page 413: 0195162854

396 COGNITION

been exploited). Due to the regular geometryof the maze., the task may be successfullysolved by making an egocentric response suchas "keep turning left," although rats seldomuse this strategy spontaneously. Nevertheless,to negate this strategy, only some arms maybe baited. The radial arm task can also be usedto differentiate the use of taxon strategies(some arms are marked by a local cue) and lo-cale strategies (other arms can be identifiedonly by distal cues).

THE PLACE PREFERENCE TASK

The place preference task combines randomsearch and goal-directed navigation (Rossieret al., 2000). On a circular elevated platform,the rat must learn to locate an unmarkedzone (the target zone) on the basis of roomcues. Entering the target zone produces therelease of a small (20 mg) food pellet froma feeder located above the arena. Becausethe pellet can land anywhere in the arena,the animal has to perform a pellet-chasingtask (i.e., a random search) to get the rewardwhile simultaneously learning that there isa target zone. This task involves place nav-igation abilities (locale system) but may beused to study dead reckoning if room cuesare concealed.

THE MORRIS WATER TASK

The Morris water maze provides a versatiletest to study navigation (Morris, 1981); unlikethe tasks described earlier, food and waterdeprivation is not used to motivate the ani-mals. In addition, the animals swim, a behav-ior at which rats are naturally adept. Briefly,the animal swims in a circular tank filled withopaque water to reach a submerged or visibleplatform, which is the only place to which itcan escape from the water. In the place ver-sion of the task, there is no cue closely asso-ciated to the platform. Localization of the hid-den platform can be performed and learnedonly by using distal room cues. To prevent

Figure 37-2. Two versions of the Morris water task. In theplace version, the animal has to learn the location and nav-igate toward a submerged platform on the basis of distalroom cues. This strategy is based on the locale system. Inthe cue version, the platform location is indicated by a bea-con which guides the rat's navigation. This strategy is basedon the taxon system.

the animal from using a local view for guid-ance, different starting places are used on suc-cessive trials. It is therefore assumed thatlearning results in elaboration of a cognitivemap using the locale system (Fig. 37-2).

Acquisition of the Morris navigation taskrequires procedural memory and workingmemory. Procedural memory allows the ratto learn that it must swim, find an escape plat-form, climb on the platform, and so on. Work-ing memory results in learning that the plat-form is at a specific location; rats can learn thisin a single trial after mastering the proceduraldetails of the task. The distinction betweenthese two memories is important when con-sidering the effects of brain damage on acqui-

Page 414: 0195162854

Chapter 37. Piloting

sition. For instance, poor performance may bewrongly attributed to spatial information pro-cessing impairment when it is due to proce-dural learning deficit (Whishaw et al., 1995).A matching-to-sample procedure can be usedto study working spatial memory. During asample trial, the animal has to locate the plat-form at a new location. During a matchingtrial, it can demonstrate that it knows that lo-cation. After one pair (or series) of trials, a newsample is presented.

Many variations of the Morris naviga-tion task have been developed, allowing thestudy of a large range of navigation prob-lems. For instance, if the platform surface isprotruding above the water or if the sub-merged platform is closely associated with abeacon, rats are prone to use a cue naviga-tion strategy rather than a place strategy.Guidance can also be used to solve the task.Animals can swim at a fixed distance of thewall corresponding to the distance of theplatform until they come across the plat-form. In principle, normal rats do not use thisstrategy, but rats with brain damage mayadopt it. Guidance may also be used if an an-imal is released from a constant start posi-tion. Navigation thus may be guided alongan axis linking a prominent room cue to thestarting point and including the platform.

The fact that an animal may use variousstrategies to solve water navigation problemsrequires the researcher to be aware of the an-imal's flexibility and allows the study of manydifferent problems.

THE EXPLORATION TASK

Exploration and navigation are intimately re-lated because navigation cannot occur with-out preliminary exploration of the environ-ment (Renner, 1990). One of the majorfunctions of exploration is information gath-ering, so it is necessary for the formation ofspatial representations (O'Keefe and Nadel,1978). Early studies have suggested that ex-ploration, and, in particular, object explo-

397

ration, is modulated by various factors suchas the level of familiarity of the environment.More recently, Thinus-Blanc and colleagues(Poucet et al., 1986; Thinus-Blanc et al., 1987)developed a paradigm based on the explo-ration of a group of objects to study the na-ture of the spatial information processed dur-ing exploration. This paradigm consists oftwo successive phases: habituation and spatialchange. First, repeated exposure to the sameconfiguration of objects results in habituationof exploratory activity, that is, a decrement inthe response with increasing familiarity. It ishypothesized that during this phase, rats ac-quire some knowledge of the spatial charac-teristics of the environment. Habituation thusreflects the elaboration of a spatial representa-tion. Once habituation is achieved, the animalsare exposed to a new configuration of the samefamiliar objects. A renewal of exploration fol-lowing this change is assumed to reflect thatthe animal (1) encodes certain spatial relation-ships and (2) is able to compare the previousand current situations. Such renewal thus is anadaptive behavior aimed at updating the spa-tial representation of the environment.

The exploratory response to the spatialchange depends on the nature of the change(Fig. 37-3). For instance, strong modificationof the configuration by displacing one of fourobjects induces intense reexploration of all ofthe objects (Fig. 37-3B). In contrast, a mod-erate change yields reexploration specificallydirected toward the displaced object (Fig.37-3C). Topological modifications (two ob-jects are switched [Fig. 37-3D]), also result inspecific reexploration of the displaced objects(Poucet et al., 1986). Interestingly, when themetric properties of the configuration aremodified without altering the affine proper-ties, for example, by merely expanding asquare arrangement (Fig. 37-3E), the rats donot exhibit any renewal of exploration(Thinus-Blanc et al., 1987). These results sug-gest that during exploration the animals en-code some but not all geometric properties ofthe environment.

Page 415: 0195162854

398 COGNITION

Figure 3 7-3. Effects of various mod-ifications (from a to e) of a configu-ration of objects on exploratory ac-tivity. A renewal of exploration (b, c,and d) indicates that the animals areable to detect the spatial change,thus suggesting that they have en-coded some spatial relationships. Incontrast, an absence of reexplorationis observed when the configurationis only expanded. (Redrawn fromPoucet et al., 1986, and Thinus-Blancet al., 1987.)

WHICH TASK SHOULD BE USED?

The tasks that have been presented differ inmany aspects. First, they differ in terms of mo-tivation. The Morris water tasks are escapetasks that are based on aversive motivation.In contrast, the t-maze and radial maze tasksare appetitive tasks. This is a critical aspectthat likely affects the rate of learning. Explo-ration tasks (including spontaneous alterna-tion) are a special case because they are basedon spontaneous motivation, providing an in-teresting alternative to other situations. Sec-ond, spatial tasks differ in terms of cognitivecapacities. Open environments like the watermaze or the place preference arena are moreappropriate than structured mazes to studynavigation per se because animals are free toselect a strategy, allowing the animal to dis-play its preferred navigational strategies.Third, the nature of the cues used by the an-imals differs from one task to another. In mosttasks (water maze task, place preference task,radial arm-maze task, cross-maze task, etc.),the apparatus is located in a room that pro-vides numerous cues and the rats have to usethese cues to locate the platform, or the baitedarms in the appetitive tasks. In these situa-tions, it is somewhat difficult to study whatkind of cues are used and how they are used

for navigation. One possibility therefore is totrain rats in a controlled environment. The ap-paratus is isolated from the room by curtains,for example, and the experimenter provides afew cues that can be manipulated. Another as-pect concerning the cues is that the process-ing of proximal cues (i.e., objects that areplaced in the rat's locomotor space [explo-ration]), may be different from the processingof room cues (water maze).

BRAIN SUBSTRATES OF NAVIGATION

Damage to almost any portion of the braincan result in impairments in spatial naviga-tion. Nevertheless, substantial evidence sug-gests that taxon navigation is subserved bybasal forebrain structures, whereas locale nav-igation is subserved by temporal lobe struc-tures, including the hippocampus (O'Keefeand Nadel, 1978). Hippocampal lesions pro-duce place learning impairment in the watermaze but do not affect navigation to a visibleplatform (Morris et al., 1982) or with a bea-con closely associated with the platform (Saveand Poucet, 2000). In the object explorationparadigm, hippocampal rats fail to detect thespatial change but are able to discriminate anovel object that replaces a familiar one (Save

Page 416: 0195162854

Chapter 37. Piloting 399

et al., 1992). An intact hippocampus is also re-quired to solve the radial arm-maze task(Olton et al., 1979) and the T-maze (Dud-chenko, 2001). However, animals with hip-pocampal damage that receive special trainingand testing can solve many spatial tasks, evenplace tasks. A number of other cortical struc-tures (entorhinal, perihinal, postrhinal, retro-splenial, and parietal cortices; Aggleton et al.,2000) and subcortical structures (anterior thal-amus, mammillary bodies, cerebellum) havebeen hypothesized to contribute to place nav-igation. The specific role of these structures inrelation to the hippocampus remains to beelucidated. In contrast, the taxon system (cue,guidance, and routes strategies) does not seemto be dependent on the hippocampal systembut involves the dorsal striatum and its alliedstructures (McDonald and White, 1994).

CONCLUSION

One major purpose of spatial laboratorywork is to understand how rats "naturally"process spatial information and what brainstructures they use to do so. Thus, a numberof behavioral situations have been developedto test their spatial abilities and to study theinvolvement of the neural structures. It mustbe acknowledged that the apparatuses usedand the environmental conditions (nature ofcues, size of the environment, type of moti-vation, etc.) may be very different from thereal conditions in which a rat navigation hasevolved. In addition, laboratory animalsthemselves are different from wild animalsand may exhibit behaviors or strategies thatdo not match their wild counterparts. Thus,researchers must be sensitive to the ecolog-ical validity of their results. On a positivenote, however, behavioral models do exploitvarious aspects of natural behaviors such asexploratory activity, place learning, naviga-tion, spatial memory, spontaneous alterna-tion, and so on. Thus, it is reasonable to thinkthat these main processes are common to

both laboratory and wild rats (e.g., Gaulinand Fitzgerald, 1989). In addition, it is rea-sonable to believe that the spatial strategiesand the neural processes used by rats arethose used by many other animal species, in-cluding humans.

REFERENCES

Aggleton JP, Vann SD, Oswald CJP, Good M (2000)Identifying cortical inputs to the rat hippocampusthat subserve allocentric spatial processes: a simpleproblem with a complex answer. Hippocampus10:466-474.

Benhamou S and Bovet P (1992) Distinguishing betweenelementary orientation mechanisms by means ofpath analysis. Animal Behaviour 43:371-377.

Bovet J (1998) Long-distance travels and homing: dispersal,migrations, excursions. In: Handbook of spatial re-search paradigms and methodologies, Volume 2: Clin-ical and Comparative studies (Foreman N and GillettR, eds.), pp. 239-269. Hove, U.K.: Psychology Press.

Collett TS, Cartwright BA, Smith BA (1986) Landmarklearning and visuo-spatial memories in gerbils.Journal of Comparative Physiology A 158:835-851.

Dudchenko PA (2001) How do animals actually solvethe T maze? Behavioral Neuroscience 115:850-860.

Foreman N and Ermakova I (1998) The radial arm maze:twenty years on. In: Handbook of spatial researchparadigms and methodologies, Volume 2: Clinicaland Comparative studies (Foreman N and Gillett R,eds.), pp. 87-143. Hove, U.K.: Psychology Press.

Gallistel CR (1990) The organization of learning. Cam-bridge, MA: The MIT Press.

Gaulin SJC and Fitzgerald RW (1989) Sexual selection forspatial-learning ability. Animal Behaviour 37:322-331.

McDonald RJ and White NM (1994) Parallel informa-tion processing in the water maze: evidence for in-dependent memory systems involving dorsal stria-tum and hippocampus. Behavioral and NeuralBiology 61:260-270.

Morris RGM (1981) Spatial localization does not requirethe presence of local cues. Learning and Motivation12:239-260.

Morris RGM, Garrud P, Rawlins JNP, O'Keefe J (1982)Place navigation impaired in rats with hippocampallesions. Nature 297:681-683.

O'Keefe J and Nadel L (1978) The hippocampus as a cog-nitive map. Oxford: Clarendon Press.

Olton DS and Samuelson RJ (1976) Remembrance ofplaces passed: spatial memory in rats. Journal of Ex-perimental Psychology: Animal Behavior Processes2:97-116.

Page 417: 0195162854

400 COGNITION

Olton DS, Becker JT, Handelmann GE (1979) Hip-pocampus, space and memory. Behavioral andBrain Sciences 2:313-365.

Packard MG and McGaugh JL (1996). Inactivation ofhippocampus or caudate nucleus with Lidocaine dif-ferentially affects expression of place and responselearning. Neurobiology of Learning and Memory65:65-72.

Poucet B, Chapuis N, Durup M, Thinus-Blanc C (1986)A study of exploratory behavior as an index of spa-tial knowledge in hamsters. Animal Learning andBehavior 14:93-100.

Renner MJ (1990) Neglected aspects of exploratory andinvestigatory behavior. Psychobiology 18:16-22.

Rossier J, Kaminsky Y, Schenk F, and Bures J (2000) Theplace preference task: a new tool for studying therelation between and place cell activity in rats. Be-havioral Neuroscience 114:273-284.

Save E and Poucet B (2000) Involvement of the hip-

pocampus and associative parietal cortex in the useof proximal and distal landmarks for navigation. Be-havioural Brain Research 109:195-206.

Sutherland RJ and RudyJW (1989) Configural associa-tion theory: the role of the hippocampa formationin learning, memory, and amnesia. Psychobiology17:129-144.

Thinus-Blanc C, Bouzouba L, Chaix K, Chapuis N, Du-rup M, Poucet B (1987) A study of spatial parame-ters encoded during exploration in hamsters. Jour-nal of Experimental Psychology: Animal BehaviorProcesses 13:418-427.

Tolman EC (1948) Cognitive maps in rats and men. Psy-chological bulletin 55:189-208.

Whishaw IQ, CasselJC, Jarrard LE (1995) Rats with fim-bria-fornix lesions display a place response in aswimming pool: a dissociation between gettingthere and knowing where. Journal of Neuroscience15:5779-5788.

Page 418: 0195162854

Dead Reckoning

DOUGLAS G. WALLACE ANDIAN Q. WHISHAW

38

Rats may use at least two navigational strate-gies that aid in the recovery of resources andprotection from predation: allothetic and id-iothetic (Gallistel, 1990). Allothetic navigationinvolves the use of external cues (visual, au-ditory, or olfactory cues). Idiothetic naviga-tion involves the use of uses cues generatedby self-movement (proprioceptive and ves-tibular cues, sensory flow, or efferent copiesof movement commands). A prominent formof idiothetic navigation is dead reckoning, inwhich a rat locates its present position in re-lation to a starting position, to which it canreturn. A central point to be made in this chap-ter is that dead reckoning is an innate, onlineform of behavior that is central to the rat's sur-vival. This chapter presents behavioral tech-niques used to investigate dead reckoning anddescribes a theoretical framework for deadreckoning.

FOOD HOARDING ANDDEAD RECKONING

The rat's proclivity to hoard large food itemsin a refuge has afforded researchers a tech-nique to examine the cues that organize nat-urally occurring spatial behavior (Whishaw etal, 1995). The food-hoarding paradigm in-volves placing a hungry rat in a refuge thatpermits access to a large arena such that it cansearch the arena for a randomly located foodpellet. We use a circular table with eight holeslocated around the perimeter of the table (Fig.

38-1A). Each hole has runners located belowthe table to which a small box, "a basementapartment," can be affixed. Rats placed in thebasement apartment are free to jump up ontothe table and to explore it. If a rat is hungryand if it finds a food item on the table, pro-vided the food item takes more than a few sec-onds to eat, the rat will carry the food itemback to the starting refuge. There, it eats thefood before making another trip out onto thetable. This procedure can be conducted undera variety of environmental conditions (Wal-lace et al., 2002b). This chapter discusses therationale for manipulating testing conditionsand the behaviors displayed by rats.

Methodologically, the simplest testingprocedure involves using the foraging table anda home base made visible by adding a second"upstairs apartment" above the basementapartment (Fig. 38-lB). The upstairs apartmentis a small box that has a hole in one side and inthe floor so that a rat can climb down into theunderlying refuge. Under this condition, ratscan potentially use three different strategies toreach the home base: (1) a cue response, the prox-imal cue associated with the visible upstairsapartment; (2) a. place response, two or more dis-tal cues whose relationships can indicate the lo-cation of the apartments; or (3) dead reckoning,cues generated by self-movement on an out-ward trip that can be used to calculate a directreturn to the starting point.

During the cued training, the outwardsearching behavior of rats is slow and cir-cuitous (Fig. 38-2, top). Rats frequently stop

401

Page 419: 0195162854

402 COGNITION

Figure 38-1. (A) Photograph of the room and circular tableused to test rats. (B) Training, cued: a schematic of the tableand a possible sequence of food pellet placements for oneday of training with the cued home base. Place probes (light[C] and dark [D]): the arrangement of the table under placeand dark probes. (E) Place probe, new location: the arrange-ment of the table and the home base when the rat was re-leased from a novel location. The circle that is drawn tan-gent to the inner portion of the holes was used to codewhen a return trip was terminated.

training with the upstairs apartment removed(Fig. 38-1C). The home base is now indicatedby the hole, which is indistinguishable fromthe other seven holes on the table, except forits relation to distal cues. Rats leave the "hid-den" home base in search of a randomly lo-cated food pellet. The searching behavior iscomparable in topography and kinematicsto that observed during cued training (Fig.38-2B). On finding the food pellet, the ratquickly orients toward the home base. Typi-

and make scanning movements or rear andthen continue to search the table. On rindingthe food pellet, the animal quickly orients to-ward the home base. The path back to thehome base is direct and is associated with arapid increase in speed relative to the outwardsearch behavior (Wallace et al., 2002b). Thedirect return may be guided the visible cue ofthe home base, but it could also be mediatedby a place response or dead reckoning, as isdescribed later.

PLACE PROBE

For the place probe, rats are released fromthe same location as experienced during cued

Figure 38-2. Each set of panels plot the kinematic and topo-graphic profile of a representative trip. Black dots and solidlines represent the outward trip topography and kinemat-ics, respectively, and white dots and dotted lines representthe homeward trip topography and kinematics, respec-tively. (A) Training, cued: One trip with a cued home base.(B) Place probe, light: one trip with a hidden home base.(C) Place probe, dark: one trip under dark conditions.

Page 420: 0195162854

Chapter 38. Dead Reckoning 403

cally, the rat accelerates as it approaches thehome base and then decelerates as it entersthe home base (Wallace et al., 2002b). Theplace probe demonstrates the rat's ability touse distal environmental cues to locate thehome base even though it had experienced the"cued" condition only during training.

DEAD RECKONING PROBE

The dead reckoning probe involves releasingthe rat from the same location as experiencedduring cued training and the place probe,except that all light sources are eliminated(Fig. 38-1D). Under these conditions, infraredrecording cameras and spotters are used toobserve the rat's movements. Because thereare no available distal or proximal cues, a ratcan return directly to its starting point onlythrough the use of self-movement cues gen-erated on the outward trip. The topographyof the searching and homing components issimilar to that observed under cued trainingor the place probe (Fig. 38-2C). A rat's search-ing components are circuitous but, again, thehomeward component is a direct path tothe home base. Although movements onboth outward and homeward components areslower under dark conditions, rats move fasteron the homeward component. It is worth not-ing that it is unlikely that they are using otherallothetic cues such as olfaction (the outwardand homeward portion of the trip are differ-ent) or other olfactory or auditory cues be-cause there are no other obvious allotheticcues in the test room. Nevertheless, a furthercontrol for the use of other surface or distalcues is embedded in the "new location" probe,which is described next.

NOVEL LOCATION PROBE

The way in which rats perform on the "place"and "dark" probe trials is relevant to the wayin which rats solve a novel-location probe.This probe involves releasing the rat from anuncued home base that is in a novel locationin the environment under light conditions.

For example, the home base is shifted 180°from the position experienced during cuedtraining (Fig. 38-1E). This probe produces aconflict between distal environmental cuesand self-movement cues. If the rat uses distalenvironmental cues to guide its movements,then it will attempt to return to the former lo-cation of the refuge. If the rat uses self-move-ment cues to organize its behavior, then therat will return directly to the "new" homebase.

After the rat leaves the home base in thenew location, it searches the table for the foodpellet (Fig. 38-3). After finding the food pel-

Figure 38-3. Solid black dots represent the searching seg-ment; gray dots represent the segment after finding the fooduntil the first hole was chosen; and white dots represent thesegment of the trip after the first hole was chosen. Top,Topography of a representative rat when released from anovel location (large black circle) after receiving trainingfrom a different location (large gray circle). Bottom, Kine-matics associated presented on top panel.

Page 421: 0195162854

404 COGNITION

let, the rat quickly orients to the old homebase location and makes a rapid direct path to-ward the old home base location. Thus, it ini-tially uses a place response. On finding thatthe refuge is not available at the old homebase location, the rat then orients to the newhome base location. As the rat returns to thenew home base location, it takes a direct path,and there is a symmetrical increase and de-crease in speed. Thus, it now appears to usedead reckoning.

This pattern of responses suggests thatrats have a hierarchy of navigational strategiesthey use as they are foraging for food items(Maaswinkel and Whishaw, 1999; Wallace etal., 2002b). First, they respond to familiar al-lothetic cues. If these cues are unreliable inpredicting the location of the refuge, then therats resort to the use of self-movement cuesto guide navigation.

That rats switch to idiothetic cues tosolve the "new" place problem can be con-firmed by repeating the test in the dark. Forexample, if a rat were using diffuse olfactorycues present in the testing room to organizeits spatial behavior, then one would anticipatea similar behavioral response as observedwhen the rat was released from a new posi-tion under light conditions. If, however, a ratwere using self-movement cues to guide nav-igation, then one would predict that the ratshould return directly back to the home baseindependent of its starting location. When ratsare released from a novel refuge location un-der dark conditions, they return directly to thestarting location and not to the previous lo-cation (Maaswinkel and Whishaw, 1999). Thisindicates that rats are not using nonvisual cuesin the environment to organize their spatialbehavior and are using dead reckoning.

NEURAL CONTROL OF DEAD RECKONING

Several studies have examined the neural basisof dead reckoning using the food-hoarding par-adigm. Considerable evidence suggests a rolefor the limbic system, including the hippocam-

pus, in the processing of self-movement cuesduring dead reckoning-based navigation. Re-search has demonstrated that although impair-ments in dead reckoning were observedafter hippocampal damage, other navigationalstrategies were intact (Maaswinkel et al., 1999).Whishaw et al. (ZOOlb) also demonstrated arole for the posterior cingulate cortex in deadreckoning during food hoarding. In examin-ing the role of the vestibular system in deadreckoning-based navigation Wallace et al.(2002b) also used the food-hoarding paradigm.Labyrinthectomies produced by intratympanicinjections of arsenic acid disrupted perform-ance on dead reckoning probes but did not im-pair performance on cued or place probes. Thisdemonstrates an important role of the vestibu-lar system and the hippocampus and other lim-bic structures in dead reckoning.

EXPLORATORY BEHAVIOR

Exploration is an obvious feature of the be-havior of many animals (O'Keefe and Nadel,1978). It is generally thought that this be-havior is useful in that once an animal hasexplored an environment, it is subsequentlyable to use the information it has acquired totravel through that environment again(Whishaw and Brooks, 1999). A rat faces twoproblems when engaged in exploratory be-havior. First, the information that it gatherson an outward trip may be of little value inguiding a return trip. Although it views andlearns about various cues on its outward trip,it does not see those cues, or move in rela-tion to them, from the vantage point of thejourney home. Second, an animal may wantto return directly home after a circuitous out-ward journey. How can it return home in astraight line after making a meandering out-ward trip? Rodents apparently solve thisproblem by using different strategies, one foroutward behavior and one for homeward be-havior. It is likely that a rat learns about theallothetic cues on the outward segment,

Page 422: 0195162854

Chapter 38. Dead Reckoning 405

whereas the homeward segment is guided bydead reckoning.

We have studied the exploratory be-havior of rats in a testing apparatus that inmany ways is designed to be a simplifiedanalogy of the animal's natural world. Weprovide the animals with a visible home base,the upstairs apartment described earlier.From their home base, the animals can ex-plore a large round table (on which there areno escape holes). There are no walls aroundthe table because walls might limit an animalto displaying thigmotaxic (wall-following)behavior. Although an animal's only homebase is the cage, its behavior is unconstrainedwith respect to opportunities to explore thetable, to examine the surrounding room, andto return to the home base. In a featurelessenvironment, rats set up virtual home baseswhere they turn, rear, and groom (Whishawet al., 1983; Golani et al., 1993); from whichthey make slow excursions (marked by anumber of pauses); and to which they returnat a higher movement speed than when theyleft (Tchernichovski et al., 1998; Drai et al.,2000). Thus, the home base and open fieldused here assist the rat in displaying orga-nized exploratory behavior.

THE OUTWARD TRIP

Our tests of rats' behavior on a tabletop con-firm that an animal's exploratory movementscan be divided into components, the mostsalient of which are excursions, stops, and re-turns. Initially, a rat makes excursions and re-turns to the vicinity of the home base en-trance. Then the rat makes longer circuitousexcursions on the table, and these excursionsare marked by head scans, pauses, or stops.Periodically, the long excursions are also in-terrupted by direct shortcut returns to thehome base. The elements of excursion (out-ward) and return (homeward) segments seemto be key components of the structure of ex-ploratory movements (Whishaw and Brooks,1999; Wallace et al., 2001c, 2002c).

We use the rat's final stop to behaviorallyfractionate the outward trip segment from thehomeward trip segment. A stop is defined as2 seconds or more of movements in whichspeed does not exceed 0.1 m/sec. Several as-pects of the exploratory data support this dis-tinction. First, work examining the distribu-tion of stops has demonstrated that thenumber of stops made during an exploratorytrip has an upper limit that is not dependenton the size of the arena (Golani et al., 1993).In addition, the kinematics associated withmovements before the last stop are qualita-tively different from movements observedafter the last stop. This latter point is consid-ered in greater detail next.

The outward trip segment is character-ized as a series of slow progressions punctu-ated by periods of stopping (Fig. 38-4). Speedsobserved on the outward trip segment rangefrom 0.2 to 0.6 m/sec. There is a high degreeof variability in the topography of the outwardtrip segment between exploratory trips. Eachoutward trip segment has a unique kinematic

Light

Figure 38-4. The kinematic and topographical representa-tion of a single exploratory trip tinder light and dark con-ditions are plotted (top and bottom, respectively). The out-ward trip segments are represented by the gray dots,whereas the homeward trip segments are represented bythe white dots. The solid black line in the kinematic plotsrepresents the criterion speed for a stop.

Page 423: 0195162854

406 COGNITION

profile (i.e., set of angular and linear speeds).Therefore, a rat gathers the maximal amountof information about an environment acrossexploratory trips.

THE HOMEWARD TRIP

In contrast to the outward trip segment, thehomeward trip is a rapid, noncircuitous paththat is directed toward the home base. Thespeeds observed on the homeward trip seg-ment vary from 0.2 to 1.6 m/sec. As the ratprogresses toward the home base, there is asymmetrical increase and decrease in speed inwhich the peak occurs at the midpoint inde-pendent of trip length (Fig. 38-4). The tem-poral pacing observed on the homeward tripsegment is reminiscent of work examining thekinematics of planar reaching movements inhumans that are executed independent ofvisual cues (Gordon et al., 1994). Such tem-poral pacing is consistent with the use of self-movement cues guiding behavior, rather thanexternal guidance. The homeward trip seg-ment originates from many different pointsin the environment across exploratory trips.This aspect of the homeward trip segment isfurther evidence that rats are not using a prox-imal cue such as olfactory guidance to organ-ize their spatial behavior (for a description ofodor tracking, see Wallace et al., 2002a).

EXPLORATION IN THE DARK

Spontaneous exploration has also been inves-tigated under completely dark conditions in anovel environment. Spontaneous exploratorytrips observed under dark conditions have asimilar topographical organization to that ob-served under light conditions (Fig. 38-4, bot-tom). The outward trip segment is circuitousand varies in location between exploratorytrips. The homeward trip segment is noncir-cuitous and is directed toward the home base.Although speeds observed on both trip seg-ments are slower under dark conditions rela-tive to light conditions; the temporal pacing

of the homeward trip segment is similar tothat observed under light testing conditions.This is consistent with rats using self-movement cues, or dead reckoning-basednavigation, to organize their homeward trips.

Because the light-versus-dark tests sug-gested that rats use dead reckoning in thedark, a parsimonious conclusion is that theanimals also use dead reckoning to returnhome in the light. We found support for thisidea with the following experiment. A rat wasplaced in a black box that was located on theedge of the table. This home base was clearlyvisible from all portions of the table. We hy-pothesized that if the rats were using allo-thetic cues to return to the starting locationin the light test, the visibility of the home basemight serve as a beacon. If the home was be-ing used as a beacon, then removing it afteran animal had left on an exploratory excur-sion should disrupt its homeward trip. Ini-tially, the rat emerged from the home basebox and explored its surface a number oftimes; therefore, it was clear that the rat wasinterested in the home base as an object. Af-ter this, it made some short excursions, fol-lowed by longer excursions and returns to thebox. On removal of the home base as the ratwas on a long outward excursion, thus elim-inating the visible cue that marked the loca-tion of the home base, the rat still returnedrapidly and directly to the previous locationof the home base. After first sniffing aroundthe vicinity of the previous location of thehome base, it then made a number of rapidexcursions and returned to that location,which seemed to indicate that it was at-tempting to find the missing home base.Thus, the behavior of the rats indicated thatthey could return accurately to the homebase, even though it was no longer visible,and they expected to find the home base atthat location. This result suggests that home-ward trips in the light are not dependent onthe ability to see the home base and thus theyare likely to be produced by dead reckoningjust as returns are in the dark.

Page 424: 0195162854

Chapter 38. Dead Reckoning

NEURAL CONTROL OF EXPLORATION

The observation that hippocampal lesionsdisrupt dead reckoning in the food-hoardingparadigm has prompted researchers to eval-uate the role of the hippocampus in the or-ganization of exploration. Wallace et al.(2002c) examined the organization of ex-ploratory trips subsequent to fimbria-fornixlesions, a major afferent/efferent pathwayfor the hippocampus proper. Although therewere no observable differences betweengroups on the outward trip segment, thegroups had different topographical and kine-matic profiles on the homeward trip seg-ment. Specifically, fimbria-fornix rats re-turned to the home base along a morecircuitous path without a distinct differencein kinematics observed on the outward seg-ment. Further evidence that the hippocam-pus is involved in dead reckoning comesfrom a similar study using rats with selectiveneurotoxic lesions of the hippocampus (Wal-lace and Whishaw, 2003). There is evidencefor a role of the hippocampus in the organi-zation of spontaneous exploration, but it islikely that other brain structures contributeto this naturally occurring behavior.

THEORETICAL FRAMEWORKFOR DEAD RECKONING

Dead reckoning observed during food hoardingand spontaneous exploration shares many com-ponents observed in nautical navigation (Fig.38-5). We suggest four components that arecritical for accuracy in dead reckoning-basednavigation: (1) home base, (2) measurement ofthe rate of linear and angular movement, (3)some index of the amount of time that haspassed, and (4) a central process that integrateslinear and angular rates of movement with theassociated temporal context to provide an in-ternal representation of current distance anddirection from the home base. Each of thesecomponents is discussed next.

407

Figure 38-5. (A) Example of dead reckoning course for aship. The ship starts from a known point of origin. Duringthe course of the trip, the ship changes direction and thenreturns to the point of origin. The ship leaves the point oforigin at 0800 hours, in an eastern direction (90°), and trav-els at 10 knots. The speed, direction, and current time arelogged at regular intervals to provide the necessary infor-mation to return to the point of origin independent of ex-ternal cues.

HOME BASE

The home base serves as a point of origin forrats as they explore or search for food. Thehome base is a critical component of a rat'senvironment. When a rat is placed in a novelenvironment, in the absence of a physicalhome base, it sets up one or more virtualhome bases (Eilam and Golani, 1989). Whena rat is provided with a physical shelter, ittreats that location as a home base. If the shel-ter is removed, then the rats treat the formerlocation as a home base (Whishaw et al.,200la). One possibility is that the rat uses thehome base as a discrete stimulus to reset orclear the dead reckoning process. Dead reck-oning-based navigation is subject to an ac-cumulation of errors (Maurer and Seguinot,1995). One way to reduce the accumulationof errors across trips is to clear the dead reck-oning process before the initiation of subse-

Page 425: 0195162854

408 COGNITION

quent trips. It is possible that the importanceof the home base to the rat reflects its func-tion to clear the dead reckoning process.

VELOCITY

Changes in linear and angular velocity occur asrats move through an environment, and thevestibular system processes this information.Examining the speeds at which a rat travelsthrough an environment, Drai et al. (2000) re-ported that animals typically travel at one ofthree speeds, or "gears." We observe that eachof the gears is associated with specific compo-nents of exploration. For example, a rat's out-ward trips reflect movement within secondgear, whereas third gear seems to be used forthe return trip. The use of a limited number ofspeeds simplifies the calculations of travel ve-locities as a variable in estimating a homewardtrip, thereby reducing the error present in esti-mating distance from the home base.

TEMPORAL CONTEXT

The temporal context in which the self-move-ment cues have occurred may play an impor-tant role in dead reckoning. Rats are able todiscriminate differences in temporal intervalson the scale of seconds (Church and Gibbon,1982), suggesting the rat has an adequate per-ception of time encountered during an ex-ploratory trip. Little work has been conductedto examine the role of time perception in spa-tial navigation. We suggest that the scalar in-variance property of interval time should bemanifested in a rat's judgment of distance anddirection from the home base. For example,a rat's ability to judge shorter distances will behighly reliable; however, as the distance in-creases, one will observe a systematic increasein the variability of distance judgments.

ONLINE PROCESSING

Dead reckoning involves online processing ofchanges in linear and angular velocities with

respect to the temporal context in which theyoccur. The resulting internal representationprovides information regarding the directionand distance of the home base from the rat'scurrent location. The temporal pacing of thehomeward trip is consistent with behaviorguided by an internal representation ratherthan external cues. The process of dead reck-oning may update the internal representationcontinually or just before returning to thehome base. It seems that the former wouldminimize the amount of error in direction anddistance estimates; however, the current datado not favor one method of updating overanother.

CONCLUSION

Dead reckoning-based navigation has all ofthe appearances of an innate action pattern ap-plied by rats to specific spatial problems. Therat processes online self-movement cues toreturn to the point of origin. The behavioraltechniques discussed here provide a founda-tion for observing dead reckoning in the rat.Dead-reckoning based navigation provides arich behavioral paradigm to evaluate the ef-fect of manipulating the nervous system. Inaddition, dead reckoning may provide a foun-dation that enables other forms of spatial nav-igation and may serve as a temporal and spa-tial marker for significant spatially relevantevents in a rat's life.

REFERENCES

Church RM and Gibbon J (1982) Temporal generaliza-tion. Journal of Experimental Psychology. AnimalBehavior Processes 8:165-186.

Eilam D and Golani I (1989) Home base behaviour ofrats (Rattus norvegicus) exploring a novel environ-ment. Behavioural Brain Research 34:199-211.

Drai D, Benjamini Y, Golani I (2000) Statistical discrim-ination of natural modes of motion in rat explor-atory behaviour. Journal of Neuroscience Methods96:119-131.

Page 426: 0195162854

Chapter 38. Dead Reckoning 409

Gallistel CR (1990) The organisation of learning. Cam-bridge, MA: The MIT Press.

Golani I, Benjamini Y, Eilam D (1993) Stopping be-haviour: constraints on exploration in rats (Rattusnorvegicus). Behavioural Brain Research 53:21-33.

Gordon J, Ghilardi MF, Cooper SE, Ghez C (1994) Ac-curacy of planar reaching movements. II. System-atic extent errors resulting from inertial anisotropy.Experimental Brain Research 99:112-130.

Maaswinkel H, Jarrard LE, Whishaw IQ (1999) Hip-pocampectomized rats are impaired in homing bypath integration. Hippocampus 9:553-561.

Maaswinkel H and Whishaw IQ (1999) Homing with lo-cale, taxon, and dead reckoning strategies by forag-ing rats: sensory hierarchy in spatial navigation. Be-havioural Brain Research 99:143-152.

Maurer R and Seguinot V (1995) What is modeling for?A critical review of the models of path integration.Journal of Theoretical Biology 175:457-475.

Techernichovski O, Benjamini Y, Golani I (1998) Thedynamics of long-term exploration in the rat. PartI. A phase-plane analysis of the relationship betweenlocation and velocity. Biological Cybernetics 78:423-432.

Wallace DG, Corny B, Whishaw IQ (2002) Rats cantrack odors, other rats, and themselves: implicationsfor the study of spatial behaviour. Behavioural BrainResearch 131:185-192.

Wallace DG, Hines DJ, Pellis SM, Whishaw IQ (2002)Vestibular information is required for dead reckon-ing in the rat. Journal of Neuroscience 22:10009-10017.

Wallace DG, Hines DJ, Whishaw IQ (2002) Quantifica-tion of a single exploratory trip reveals hippocam-

pal formation mediated dead reckoning. Journal ofNeuroscience Methods 113:131-145.

Wallace DG and Whishaw IQ (2003) NMDA lesions ofthe Ammon's horn and the dentate gyrus disruptthe direct and temporally paced homing displayedby rats exploring a novel environment: Evidence forthe role of the hippocampus in dead reckoning. Eu-ropean Journal of Neuroscience 18:513-523.

Whishaw IQ and Brooks BL (1999) Calibrating space:exploration is important for allothetic and idiotheticnavigation. Hippocampus 9:659-667.

Whishaw IQ, Coles BL, Bellerive CH (1995) Food car-rying: a new method for naturalistic studies of spon-taneous and forced alternation. Journal of Neuro-science Methods 61:139-143.

Whishaw IQ, Hines DJ, Wallace DG (2001) Dead reck-oning (path integration) requires the hippocampalformation: evidence from spontaneous explorationand spatial learning tasks in light (allothetic) anddark (idiothetic) tests. Behavioural Brain Research127:49-69.

Whishaw IQ, Maaswinkel H, Gonzalez CLR, Kolb B(2001) Deficits in allothetic and idiothetic spatial be-havior in rats with posterior cingulate cortex le-sions. Behavioural Brain Research 118:67-76.

Whishaw IQ, Kolb B, Sutherland RJ (1983) The analy-sis of behaviour in the laboratory rat. In: RobinsonTE, editor. Behavioural approaches to brain re-search, pp. 141-211. Oxford University Press: NewYork.

Whishaw IQ and Tomie JA (1997) Piloting and deadreckoning dissociated by fimbria-fomix lesions in arat food carrying task Behavioural Brain Research89:87-97.

Page 427: 0195162854

Fear

MATTHEW R. TINSLEY ANDMICHAEL S. FANSELOW

39

During the past 30 years, there have beengreat advances in the conceptualization of de-fensive behavior. Behavioral and motivationalapproaches have stressed the organization ofdefensive behaviors into a functional be-havior system tuned to the ecological andevolutionary requirements of the animals(Bolles, 1970; Fanselow, 1994). Advances inneurophysiological, neurochemical, and neu-roanatomical techniques have allowed themediating substrates of different defensive be-haviors to be mapped and information on thebehavioral and neural organization of defen-sive behavior to be integrated (Fanselow,1994; Fendt and Fanselow, 1999). This syn-thesis has resulted in perhaps the best under-standing of the processes involved in mediat-ing a complex and functionally significantsuite of behaviors in the rat. In the followingchapter we present defensive behavior as afunctional behavior system, describe the neu-ral substrates of many of the various defen-sive behaviors described, and present a modelthat integrates the behavioral and neural or-ganizations of defensive behavior.

THE ORGANIZATIONOF DEFENSIVE BEHAVIOR

Current theories on the organization of therat's defensive behavior are largely variationson Bolles' (1970) species-specific defensive re-action (SSDR) theory. This theory suggestedthat when confronted by a natural (e.g., apredator) or unnatural (e.g., a noxious stimu-

lus) threat, the rat's behavior is restricted to asmall number of innately determined defen-sive behaviors, the SSDRs. Determination ofwhich behavior is produced and why is wherethese theories of defensive behavioral organi-zation differ.

Bolles' (1970) original conception wasthat the behavior produced depended on priorexperience, with SSDRs that had been previ-ously unsuccesful being less likely than SSDRsthat had not. However, attempts to reducethe production of SSDRs through operantpunishment contingencies (Bolles, 1975) werenot successful. In any event, learning via pun-ishment requires the rat to be exposed to anumber of predatory encounters before learn-ing the correct behavioral strategy, which isunlikely in a natural situation (Fanselow et al.,1987).

A second theory suggests that the SSDRproduced depends on features of the envi-ronment (Blanchard et al., 1976). This theorysuggests that a rat will try to escape from asituation if escape is possible and will freeze ifit knows that escape is not possible. One find-ing that seems to support this suggestion isthat animals shocked immediately after beingplaced in a chamber do not show a freezingresponse. In contrast, animals shocked 3 min-utes after placement, who have perhapslearned there is no exit, do freeze (Blanchardet al., 1976). However, this immediate shockdeficit also occurs if other behaviors, such asdefecation or conditional analgesia, are ex-amined (Fanselow et al., 1994), suggestingthat this procedure causes reduced freezing

410

Page 428: 0195162854

Chapter 39. Fear 411

due to reduced fear conditioning. To furtherassess the influence of support stimuli,Fanselow and Lester (1988) presented a con-ditional fear stimulus in a variety of contextsthat differed with respect to the support stim-uli they contained. In each case, the stimuluselicited freezing, rather than some otherSSDR.

These results led to the suggestion thatwhat is important in determining the SSDR inwhich the animal is the level of fear it is ex-periencing (Fanselow and Lester, 1988).Fanselow and Lester (1998) and Fanselow(1994) suggested that defensive behavior is or-ganized along a continuum of perceivedthreat or predatory imminence (Blanchard etal., 1989). When the animal perceives thatpredatory imminence is low, that the threat istemporally or spatially distal, it will engage inpreencounter defensive behaviors such asmeal pattern reorganization to reduce the riskof predation (Helmstetter and Fanselow,1993) and stretched approach behavior to gaininformation (Blanchard et al., 1989). As preda-tory imminence increases, once a predator hasbeen detected, behavior shifts to posten-counter defensive behavior such as freezing.Finally, when contact with a predator is oc-curring or inevitable, the rat engages in circa-strike defensive behavior such as defensivefighting or jump attack. Conditional stimulipredictive of high predatory imminence, suchas those paired with predator presentation orshock, cause the animal to engage in posten-

counter, not circa-strike, behavior as the ratattempts to avoid actual contact with thethreat stimulus. This organization of defen-sive behavior along a predatory imminencecontinuum is illustrated in Figure 39-1.

BEHAVIORS DURINGLOW PREDATORY IMMINENCE:

FIELD STUDIES

The rat, being a heavily predated animal, islikely to spend most of its time in a state of atleast some perceived threat. This suggests thatmost, if not all, of its behaviors outside theburrow may be modified by tonically em-ployed, risk minimization strategies and over-lapping phasically employed, preencounterdefensive behaviors. Examples of the rat's ton-ically employed, risk minimization strategieshave been described from naturalistic studiesand include neophobia and trail making.

For a heavily predated animal, like therat, nonneutral changes in its physical envi-ronment are more likely to be negative thanpositive. Hence, the appropriate risk mini-mization strategy is to treat all changes as neg-ative until they are determined to be other-wise. The resulting avoidance of novelty hasbeen termed neophobia. This effect was notedby Calhoun (1963) following the introductionof activity recorders in a naturalistic environ-ment: "The rat would cautiously approach it,jump back, bypass it while keeping a foot or

Figure 39-1. Defensive behavior ishypothesized as being organizedalong a psychological dimension ofperceived predatory imminence. In-creases in predatory imminence,corresponding to changes in externalstimuli, shift the animal rightwardalong the continuum, resulting inqualitatively different defensive be-haviors at differing levels of per-ceived threat.

Page 429: 0195162854

412 COGNITION

two away" (Calhoun, 1963, p. 85). Impor-tantly for laboratory studies on preencounterdefensive behavior, exploratory behavior, andnot defensive behavior, is induced if either alaboratory rat or a wild rat is placed into acompletely novel environment, suggestingthat defensive behavior "occurs . . . whenthere is a change in an otherwise familiar situa-tion" (Barnett, 1963, p. 30 [author's italics]).

A second risk minimization strategy istrail making and use. Rats in both the wild anda controlled setting establish specific trails toresources such as sources of food or shelter(Calhoun, 1963). Trail following reduces therisk of predation by ensuring that the animalfinds resources as reliably and efficiently aspossible. These trails typically follow an ap-proximation to the shortest distance betweenthe two goals but deviate toward sources ofoverhead cover and vertical surfaces (thig-motaxis) (Calhoun, 1963). Trail-following be-havior has been demonstrated in the labora-tory by Timberlake and Roche (1998). In thesestudies, the radial-arm maze arrangement ofarms and food cups was placed on the floorof a large room, allowing the animals to ap-proach the food from all directions. Despitethis, the rats continued to move along or be-side the maze arms rather than using another,more efficient, search strategy.

BEHAVIORS DURINGLOW PREDATORY IMMINENCE:

LABORATORY STUDIES

At least two discrete behavioral responses toincreased but low levels of threat have beendescribed. Fanselow et al. (1988) describedstudies in which the foraging behavior of ratsin a closed economy was responsive to lowlevels of threat. Following establishment ofbaseline performance, low densities of ran-domly timed shock were introduced and theeffect on meal patterning and size was exam-ined. Rats responded by reducing the number,but increasing the size, of their meals, result-

ing in the animals avoiding 50% of the shocksthey would have experienced had they not al-tered their behavior. Subsequent studies havedetermined that rats will eat a given mealmore quickly after tail shock (Dess and Van-derweele, 1994) and will eat a particular sizeof pellet more quickly in lighter and more ex-posed environments (Whishaw et al., 1992).Importantly, the animal's meal patterning re-turns toward normal after the cessation ofshock (Fanselow et al., 1988), indicating thatthe change is a discrete behavior related to in-creases in perceived predatory imminence.

Light-enhanced startle is the second ex-ample in which exposure to conditions of mildthreat or low predatory imminence lead to aphasic increase in defensive behavior. Thestartle response is a fast muscle contraction,most noticeable around the head and neck,that follows an unpredicted, intense stimulus(Fendt and Fanselow, 1999). The magnitudeof the startle response is enhanced by manip-ulations that increase levels of fear and anxi-ety (Brown et al., 1951). In the light-enhancedstartle procedure, the rat is first tested for itsresponse to a number of short bursts of loudwhite noise in a darkened chamber before be-ing tested in the same chamber with the samenoise bursts under conditions of bright (700footlamberts) illumination (Walker and Davis,1997a). Startle amplitude is significantly en-hanced during the illuminated period of thetest (Walker and Davis, 1997a, 1997b).

BEHAVIORS DURING HIGHPREDATORY IMMINENCE:

RESPONSES ELICITEDBY PREDATOR CUES

Studies examining conditional defensive re-sponses reinforced by predator exposure arerare. We have examined whether cat expo-sure conditions context fear by exposing ratsto an inescapable cat for 5 minutes immedi-ately, 15 seconds, or 120 seconds after placingthe rats in a novel two-chamber context. Ro-

Page 430: 0195162854

Chapter 39. Fear 413

bust freezing, on the order of 40% to 50% forthe delayed cat groups, was observed duringcat exposure (Figure 39-2). Very little condi-tional freezing (5% to 10%) was seen duringlater context testing, indicating that cat expo-sure conditioned very little fear to the context.Adamec et al. (1998) found that cat odor-exposed rats later showed increased risk as-sessment behavior, whereas cat-exposed ratslater showed reduced risk assessment behav-ior. Blanchard et al. (2001) found defensivebehaviors conditioned with cat odor to acontext included more crouch/freeze withsniffing behavior and significantly less risk as-sessment behavior than when behaviors wereconditioned with cat odor on a discrete cue.

% Freezing Immediate vs. Delayed Cat: Trainining

Figure 39-2. The upper frame shows the percent of sam-ples during the 5 minute cat exposure during which the ratswere freezing. Exposure to the cat after 15 or 120 secondsin the context causes robust freezing in rats. The lowerframe shows the percent of samples during the 5 minutecontext reexposure 24 hours later. Despite the high levelsof freezing during the cat exposure, freezing to the contextis very low, indicating very little conditional fear of the con-text has been learned.

These results suggest that the conditional re-sponse to a context previously paired with apredator seems to be a suppression of behav-ior and an increase in freezing while the con-ditional response to a cue previously pairedwith a predator may be risk assessmentbehavior.

Studies have also examined uncondi-tional responses to predator cues. The mostcommon experimental method involves ex-posure to a cat (Blanchard and Blanchard,1971), although studies have also demon-strated defensive behavior after exposure tocat odor (Blanchard et al., 1975), odiferouscomponents of predator feces (Vernet-Mauryet al., 1992), or the experimenter (Blanchardet al., 1981).

The most naturalistic of these studieshave been those by Blanchard and Blanchard(1989) using their visible burrow system(VBS). Briefly placing a cat in the VBS leadsthe rats to flee into the burrows where theyremain for several hours. Other behaviors in-clude the production of 22 kHz ultrasonic vo-calizations during and for 30 minutes subse-quent to cat presentation (Blanchard et al.,1991) and the suppression of nondefensive be-haviors (Blanchard et al., 1989). The rats be-gin to emerge from their burrows about 4 to7 hours after the cat presentation and engagein risk assessment behaviors, although therats' behavior after cat presentation does notreturn to pre-cat presentation baseline valuesfor at least 24 hours (Blanchard et al., 1989).

Most studies of exposure to a predatoruse a procedure in which the rat and preda-tor are placed in a chamber and the rat's be-havior is assessed (Blanchard and Blanchard,1971). These procedures typically involve ashorter period of more intense exposure, asthe rat can continuously see, hear, and smellthe cat and is unable to escape from the situ-ation. Typical results show that rats react tothe presence of a predator by engaging in freez-ing: the cessation of all body movement ex-cept that necessary for respiration (Bolles andCollier, 1976). Other responses include inhi-

% Freezing Immediate vs. Delayed Cat: Context Testing

Page 431: 0195162854

414 COGNITION

bition of locomotion, inhibition of ex-ploratory behavior, and increased defecation(Satinder, 1976). Studies looking at the effectsof cat exposure on endogenous antinocicep-tion have also shown that exposure to a catleads to analgesia mediated by endogenousopioid systems (Lester and Fanselow, 1985).

Studies on the stimulus control of re-sponse to a predator have suggested that twomain classes of stimuli are important: the mo-tion of the predator and its odor. Blanchard etal. (1975) showed that rats do not freeze to thesound or the smell of a cat or to the sight ofa dead cat but that either a moving cat or dogor the abrupt and rapid movement of an inan-imate card causes freezing and inhibition ofapproach. In contrast, Griffith (1920) reportsthat presentation of a cat in the dark results inbehavior we would now describe as freezingbut that presentation of a cat in a glass jar doesnot. Presentation of a wooden block coveredin a cloth impregnated with cat odor in theVBS elicits similar, but less intense, behavioralchanges to those following cat presentation(Blanchard et al., 1989) and elicits risk assess-ment and crouch/freeze with sniffing behav-ior in a chamber (Blanchard et al., 2001). Ratsshow risk assessment behaviors when con-fined with a cat odor stimulus (Blanchard etal., 1991, 1993) but avoidance when the ani-mal can retreat from the stimulus (Dielenbergand McGregor, 1999). This avoidance re-sponse habituates over repeated presentationsof the stimulus (Dielenberg and McGregor,1999).

VALIDITY OF STUDIESUSING AVERSIVE STIMULI

The majority of studies of defensive behaviorin rats have not examined responses to pred-ators but have induced defensive behavior byusing stimuli that are unconditionally aver-sive, such as loud noise, bright light, and elec-tric shock. Just as studying laboratory rats toinfer the behavior of wild rats provokes ques-

tions of external validity, so does using a stim-ulus such as electric shock. The most influen-tial response to this criticism has come fromthe work of Robert Bolles (Bolles, 1970, 1975;Bolles and Fanselow, 1980). Bolles suggestedthat rats are equipped by evolution with pre-existing defensive behaviors and that "the im-mediate and inevitable effect of severe andaversive stimulation on a[n]. . . animal is [of]. .. restricting its response repertoire to a nar-row class of SSDRs" (Bolles, 1970, p. 34).Hence, using aversive stimulation to study de-fensive behavior is externally valid becausethe animal's reactions to both a predator andan aversive stimulus are governed by its SSDRorganization. This is currently the dominantview in ethoexperimental approaches to de-fensive behavior.

BEHAVIORS DURING HIGHPREDATORY IMMINENCE:

RESPONSES TO PREDICTORSOF AVERSIVE STIMULI

There are a number of conditional defensivebehaviors to cues or contexts paired with anaversive stimulus, including freezing (Bollesand Collier, 1976), various forms of crouching(Blanchard and Blanchard, 1969), ultrasonicvocalization (Kaltwasser, 1991), and defeca-tion (Fanselow, 1986), as well as autonomicmeasures such as hypertension (LeDoux et al.,1983), analgesia (Fanselow and Baackes,1982), and both increases and decreases inheart rate (LeDoux et al., 1984). In addition,cues paired with shock have been shown todisrupt ongoing behavior (Estes and Skinner,1941) and to potentiate unconditional startlebehavior (Brown et al., 1951).

Of this variety of conditional responsesto aversive stimulation, the most investigatedis freezing. It is important to note that al-though freezing is an unconditional responseto predator exposure (Blanchard and Blan-chard, 1969, 1971), it is a conditional responseto stimuli paired with aversive stimulation.

Page 432: 0195162854

Chapter 39. Fear 415

Fanselow (1980) compared freezing levels inthe training context or a novel context 30 sec-onds or 24 hours after two levels of foot shock.Rats froze more in the training context thanthe novel context regardless of interval, in-dicating that freezing was a conditional re-sponse that required the presence of shock-associated cues to be expressed. In addition,there was no difference in the amount of freez-ing between rats tested 30 seconds after shockand 24 hours after shock, indicating that theshock itself is not required to elicit freezing.

The immediate shock deficit providesfurther evidence that freezing is a conditionalresponse. Animals shocked immediately afterbeing placed in a chamber do not later showa conditional freezing response to that cham-ber. In contrast, animals shocked 3 minutes af-ter placement do freeze (Blanchard et al.,1976). This indicates that it is not merely theexperience of being shocked that elicits freez-ing behavior. Subsequent investigation of thisimmediate shock deficit in freezing, and othermeasures of conditional fear to the context(Fanselow et al., 1994), indicates that it is dueto the requirement for the animal to form acontext representation before the shock de-livery to be able to pair this representation,the to-be-conditioned stimulus, with theshock. Preexposure to the context alleviatesthe immediate shock deficit because it allowsthe animal to form the context representation(Fanselow, 1990).

BEHAVIORS DURING HIGHPREDATORY IMMINENCE:

RESPONSES TO AVERSIVE STIMULI

By far the most common aversive stimulusused in experiments examining defensive be-havior is electric shock. A rat's reaction toshock is a vigorous burst of activity that per-sists for a brief period after shock offset beforeit begins to engage in freezing behavior (Myer,1971). Anisman and Waller (1973) determinedthat the duration of the activity burst is posi-

tively correlated with the intensity of the shock,and Fanselow (1980, 1982) characterized thisreaction as the rat's unconditional response tothe shock stimulus. Fanselow (1982, p. 453)provides a description of the activity burst, stat-ing that it "was characterized by reflexive pawwithdrawal, jumping and squealing . . . the an-imal moved rapidly, though in an uncoordi-nated manner, about the chamber."

Unconditional analgesia has also beendemonstrated after footshock. Liebeskind andhis colleagues (e.g., Lewis et al., 1980) demon-strated that prolonged, strong shock (30-minute intermittent or 3-minute continuous3mA shock) induces opioid- and non-opioid-mediated analgesia in a tail-flick test. Subse-quent studies at more commonly used inten-sities and durations (Fanselow et al., 1994)suggest that the expression of unconditionalanalgesia may depend on this strong shock.

THE NEURAL SUBSTRATESOF DEFENSIVE BEHAVIOR

THE AMYGDALA

Any discussion of the neural substrates of de-fensive behavior must begin with the amyg-dala. Early studies of the effects of temporallobe lesions in monkeys (Brown and Schaffer,1888) described a substantial loss of defensivebehavior after extensive lesions of the medialtemporal lobe that included the amygdala.More selective lesions determined this loss ofdefensive behavior was due to damage tothe amygdala (Weiskrantz, 1956). Amygdaladamage has been shown to affect uncondi-tional freezing to a cat (Blanchard and Blan-chard, 1972) and unconditional analgesia(Bellgowan and Helmstetter, 1996) and to at-tenuate unconditional autonomic responses(Iwata et al., 1986; Young and Leaton, 1996).Amygdala lesions also affect context freezingbehavior conditioned by shock (Blanchardand Blanchard, 1972) and the acquisition ofconditional bar press suppression (Kellicuttand Schwartzbaum, 1963).

Page 433: 0195162854

416 COGNITION

Different regions within the amygdalaplay different roles in the coordination and ex-pression of fear behavior. These nuclei can beroughly delineated into two major subsys-tems (Maren and Fanselow, 1996; Maren,2001): the basolateral complex and the centralnucleus. The basolateral complex includes thelateral, basolateral, and basomedial nuclei andforms the sensory input into the amygdala.Lesions to these structures disrupt the acqui-sition and expression of conditional defensivebehaviors (LeDoux et al., 1990) by disruptingthis sensory input. Selective lesions of specificinput pathways from sensory processing re-gions to the basolateral complex affect condi-tioning of defensive behavior to CSs from theappropriate modality: lesions of the auditorycortex and auditory thalamus projections tothe basolateral complex affect conditioning toauditory CSs (Campeau and Davis, 1995a),whereas lesions of the perirhinal cortex affectconditioning to visual CSs (Campeau andDavis, 1995a).

Although the basolateral complex is in-volved in the formation of CS-US associa-tions, the central nucleus is involved in the ex-pression of defensive behavior (Fanselow andKim, 1994; Maren, 2001). Stimulation of thecentral nucleus produces autonomic re-sponses similar to those caused by presenta-tion of conditional fear stimuli (Kapp et al.,1982; Iwata et al., 1987). Although lesions ofthe central nucleus produce deficits in the ac-quisition and expression of fear conditioningthese deficits seem to be caused by reducedperformance of defensive behaviors ratherthan any inability to form the association be-tween CS and US (Fanselow and Kim, 1994).

Additional studies have implicated re-gions efferent to the central nucleus in theproduction of specific autonomic and behav-ioral defensive responses. Regions involved inautonomic responses include the paraventric-ular nucleus of the hypothalamus and bed nu-cleus of the stria terminalis (BNST), whichmediate glucocorticoid release; the lateral hy-pothalamus and dorsal motor nucleus of the

vagus (LeDoux et al., 1988; Kapp et al, 1991),which are involved in heart rate responses;and the parabrachial nucleus, which is in-volved in increased respiration. Regions in-volved in mediating behavioral responsesinclude the periaqueductal gray, which is in-volved in expression of freezing behavior(Liebman et al., 1970; Kim et al., 1993) (seelater), and endogenous opioid-mediated an-algesia (Helmstetter and Landeira-Fernandez,1990) and the nucleus reticularis pontis cau-dalis, which mediates the fear potentiation ofacoustic startle (Davis et al., 1982).

THE BED NUCLEUS OF THE

STRIA TERMINALIS

The BNST is a major output region of the ba-solateral amygdala, connecting it with the thehypothalamic-pituitary-adrenal axis (Alheid,1995). Despite these similarities, studies havedistinguished between the effects of lesions ofthe BNST and the central amygdala on de-fensive behavior. Pretraining and posttraininglesions of the BNST or infusions of the AMPAantagonist NBQX into the BNST do not affectfear-potentiated startle (Hitchcock and Davis,1991; Walker and Davis, 1997b) or passiveavoidance of a shock probe (Treit et al., 1998),whereas similar treatments directed at thecentral nucleus of the amygdala reduce ex-pression of conditional fear. In contrast, le-sions of the BNST, but not the central nucleus,reduce the expression of light-enhanced star-tle (Walker and Davis, 1997b) and CRH-enhanced startle (Lee and Davis, 1997). Morerecently. Fendt et al. (2003) demonstrated thatunconditional freezing to an odor derivedfrom predator feces was blocked by infusionof the GAB A agonist muscimol into the BNSTbut not the amygdala. This pattern of resultshas given rise to the suggestion that the BNSTis selectively involved in responses to uncon-ditional fear stimuli. In a recent review, how-ever, Walker et al. (2003) instead suggest thatthe BNST is involved in responding to long-duration aversive stimuli and can be charac-

Page 434: 0195162854

Chapter 39. Fear 417

terized as part of an anxiety-mediating system,distinct from the amygdala-mediated fearsystem.

THE PERIAQUEDUCTAL GRAY

The ventrolateral regions of the periaqueduc-tal gray (PAG) have long been implicated inthe production of defensive behaviors (Lieb-man et al., 1970). Lesions of this region, par-ticularly the caudalmost areas (LeDoux et al.,1988), reduce freezing (Kim et al., 1993) andsuppress opioid-mediated conditioned analge-sia (Helmstetter and Landeira-Fernandez,1990). These effects occur after the presenta-tion of shock, cues paired with shock, and af-ter presentation of a cat and occur whetherthe lesion is performed before or after (DeOcaet al., 1998). These results indicate that theventrolateral PAG is involved in mediatingthe freezing response to predators and to pre-dictors of aversive stimuli.

In contrast, lesions of the dorsolateral re-gions of the PAG have virtually no effect onfreezing in response to either a cat (DeOca etal., 1998) or stimuli associated with shock(Fanselow, 199la). Electrical stimulation ofthe dorsolateral PAG induces behaviors thatare similar to those of the postshock activityburst. Chemical stimulation of this regionshows similar effects, including the elicitation

of defensive postures and flight in response toconspecifics (Handler and Depaulis, 1988). Le-sions of this region also dramatically reducethe postshock activity burst. These results in-dicate a double dissociation of function withthe ventrolateral PAG involved in mediatingfreezing while the dorsolateral PAG is in-volved in mediating the postshock activityburst (Fanselow, 1991b).

INTEGRATING THE BEHAVIORALAND NEURAL ORGANIZATION

OF DEFENSIVE BEHAVIOR

Understanding the behavioral organization ofdefensive behavior and the neural substratesunderlying particular behaviors has allowedthe synthesis of this information into func-tional models of defensive behavior organiza-tion (Fanselow, 1994; Fendt and Fanselow,1999). While differing authors may focus onparticular parts of this organization—for ex-ample, Maren (2001) pays particular attentionto the role of the amygdala, whereas Walkeret al. (2003) review the role of the BNST, anarchetypal model could be expected to re-semble that presented in Figure 39-3.

A number of features of the organizationof defensive behavior are easily discerned inFigure 39-3. The most obvious of these is the

Figure 39-3. Sensory inputs and be-havioral outputs (in italics) are con-nected via a functional network ofbrain structures. Central to this or-ganization is the amygdala, whichacts as an interface between theseprocesses. Mapping the behavioralorganization onto the underlyingneural substrates allows a better ap-preciation for the relationship be-tween the various structures involvedin defensive behavior.

Page 435: 0195162854

418 COGNITION

centrality of the amygdala in defensive be-havior. Current conceptions of amygdalafunction in defensive behavior suggest it actsas an interface between sensory informationand defensive behavior output (e.g., Maren,2001) for both conditional and unconditionalresponses. Also clear is that the behavioraloutputs related to differing levels of perceivedpredatory imminence are effected by differingstructures. This is most clearly seen in thePAG, where circa-strike behaviors are medi-ated by the dorsolateral regions, whereas post-encounter freezing behavior is mediated bythe ventrolateral regions. This model alsohighlights the possible relationship betweenpreencounter defensive behaviors, which arehypothesized to be mediated by the BNST,and anxiety (Walker et al., 2003). Finally, thismodel also illustrates the distinction betweendefensive behavioral responses and auto-nomic responses such as bradycardia.

CONCLUSION

The understanding of defensive behavior atboth a behavioral and neural level has beengreatly enhanced over the past 30 years. Thefunctional mapping of behavioral organiza-tion onto a network of brain structures hasprovided an invaluable and productive frame-work for generating testable hypotheses of theeffects of experimental manipulations andmay be beginning to inform studies of the sub-strates of clinical disorders related to dysfunc-tion of defensive behavior in people (Walkeret al., 2003). This synthesis of brain and be-havior is much further advanced in the studyof defensive behavior as a functional systemand may provide a model for investigationsinto the neural substrates of behavior systems.

REFERENCES

Adamec R, Kent P, Anisman H, Shallow T, Merali Z(1998) Neural plasticity, neuropeptides and anxiety

in animals—implications for understanding andtreating affective disorder following traumatic stressin humans. Neuroscience and Biobehavioral Re-views 23:301-318.

Alheid GF, de Olmos JS, Beltramino CA (1995) Amyg-dala and extended amygdala. In: The rat nervoussystem, 2 ed. (Paxinos G, ed.), pp. 495-578. SanDiego: Academic Press.

Anisman H and Waller TG (1973) Effects of inescapableshock on subsequent avoidance performance: roleof response repertoire changes. Behavioral Biology9:331-355.

Bandler R and Depaulis A (1988) Elicitation of intraspe-cific defence reactions in the rat from midbrain peri-aqueductal grey by microinjection of kainic acid,without neurotoxic effects. Neuroscience Letters88:291-296.

Barnett SA (xxxx) A study in behaviour. London, UK:Methuen and Co.

Bellgowan PS and Helmstetter FJ (1996) Neural systemsfor the expression of hypoalgesia during nonasso-ciative fear. Behavioral Neuroscience 110:727-736.

Blanchard DC, Williams G, Lee EM, Blanchard RJ (1981)Taming of wild Rattus norvegicus by lesions of themesencephalic central gray. Physiological Psychol-ogy 9:157-163.

Blanchard RJ and Blanchard DC (1969) Crouching as anindex of fear. Journal of Comparative and Physio-logical Psychology 67(3):370-375.

Blanchard RJ and Blanchard DC (1971) Defensive reac-tions in the albino rat. Learning and Motivation:351-362.

Blanchard RJ and Blanchard DC (1972) Effects of hip-pocampal lesions on the rat's reaction to a cat. Jour-nal of Comparative and Physiological Psychology78(1): 77-82.

Blanchard RJ and Blanchard DC (1989) Antipredator de-fensive behaviors in a visible burrow system. Jour-nal of Comparative Psychology 103:70-82.

Blanchard RJ, Blanchard DC, Agullana R, Weiss SM(1991) Twenty-two kHz alarm cries to presentationof a predator, by laboratory rats living in visible bur-row systems. Physiology and Behavior 50:967-972.

Blanchard RJ, Blanchard DC, Hori K (1989) Ethoexper-imental approaches to the study of defensive be-havior. In: Ethoexperimental approaches to thestudy of behavior. (Blanchard RJEB and F Paul,eds.), pp. 114-136. New York: Kluwer Academic/Plenum Publishers.

Blanchard RJ, Flannelly KJ, Blanchard DC (1986) De-fensive behaviors of laboratory and wild Rattusnorvegicus. Journal of Comparative Psychology100:101-107.

Blanchard RJ, Fukunaga KK, Blanchard DC (1976) En-vironmental control of defensive reactions to foot-

Page 436: 0195162854

Chapter 39. Fear 419

shock. Bulletin of the Psychonomic Society 8:129-130.

Blanchard RJ, Mast M, Blanchard DC (1975) Stimulus con-trol of defensive reactions in the albino rat. Journal ofComparative Physiological Psychology 88:81-8.

Blanchard RJ, Yang M, Li CI, Gervacio A, Blanchard DC(2001) Cue and context conditioning of defensivebehaviors to cat odor stimuli. Neuroscience andBiobehavioral Reviews Special Issue: Defensive Be-havior and the Biology of Emotion 25:587-595.

Blanchard RJ, Yudko EB, Rodgers RJ, Blanchard DC(1993) Defense system psychopharmacology: anethological approach to the pharmacology of fear andanxiety. Behavioural Brain Research 58:155-165.

Bolles RC (1970) Species-specific defense reactions andavoidance learning. Psychological Review 77(l):32-48.

Bolles RC and Collier AC (1976) The effect of predictivecues on freezing in rats. Animal Learning and Be-havior 4:6-8.

Bolles RC, Uhl CN, Wolfe M, Chase PB (1975) Stimu-lus learning versus response learning in a discrimi-nated punishment situation. Learning and Motiva-tion 6:439-447.

Brown JS, Kalish HI, Farber IE (1951) Conditioned fearas revealed by magnitude of startle response to anauditory stimulus. Journal of Experimental Psy-chology 41:317-327.

Brown S and Schafer A (1888) An investigation into thefunctions of the occipetal and temporal lobes of themonkey's brain. Philosophical Transactions of theRoyal Society of London Series B 179:303-327.

Calhoun JB (1963) The ecology and sociology of theNorway rat. US Dept of Health, Education and Wel-fare (PHS Monograph 1008). Bethesda, MD.

Campeau S and Davis M (1995) Involvement of the cen-tral nucleus and basolateral complex of the amyg-dala in fear conditioning measured with fear-potentiated startle in rats trained concurrently withauditory and visual conditioned stimuli. Journal ofNeuroscience 15:2301-2311.

Davis M, Gendelman DS, Tischler MD, Gendelman PM(1982) A primary acoustic startle circuit: lesion andstimulation studies. Journal of Neuroscience 2:791-805.

De Oca BM, DeCola JP, Maren S, Fanselow MS (1998)Distinct regions of the periaqueductal gray are in-volved in the acquisition and expression of defen-sive responses. Journal of Neuroscience 18:3426-3432.

Dess NK and Vanderweele DA (1994) Lithium chlorideand inescapable, unsignaled tail shock differentiallyaffect meal patterns of rats. Physiology and Behav-ior 56:203-207.

Dielenberg RA and McGregor IS (1999) Habituation ofthe hiding response to cat odor in rats (Rattus

norvegicus). Journal of Comparative Psychology113:376-387.

Estes WK and Skinner BF (1941) Some quantitativeproperties of anxiety. Journal of Experimental Psy-chology 29:390-400.

Fanselow MS (1980) Conditioned and unconditionalcomponents of post-shock freezing. The PavlovianJournal of Biological Science 15:177-182.

Fanselow MS (1982) The postshock activity burst. Ani-mal Learning and Behavior 10:448-454.

Fanselow MS (1990) Factors governing one-trial con-textual conditioning. Animal Learning and Behav-ior 18:264-270.

Fanselow MS (199la) Analgesia as a response to aver-sive Pavlovian conditional stimuli: cognitive andemotional mediators. In: MR Denny (Ed.) Fear,avoidance, and phobias: a fundamental analysis.Hillsdale, NJ: Lawrence Erlbaum Associates, pp.61-86.

Fanselow MS (1991b) The midbrain periaqueductal grayas a coordinator of action in response to fear andanxiety. In: The midbrain periacqueductal gray mat-ter: Functional, anatomical and neuroschemical or-ganization (Depaulis A and Bandler R, eds.). NewYork: Plenum Press.

Fanselow MS (1994) Neural organization of the defen-sive behavior system responsible for fear. Psycho-nomic Bulletin and Review 1:429-438.

Fanselow MS and Baackes M (1982) Conditioned fear-induced opiate analgesia on the formalin test: Evi-dence for two aversive motivational systems. Learn-ing and Motivation 13:200-221.

Fanselow MS and Kim JJ (1994) Acquisition of contex-tual Pavlovian fear conditioning is blocked by ap-plication of an NMDA receptor antagonist D,L-2-amino-5-phosphonovaleric acid to the basolateralamygdala. Behavioural Neuroscience 108:210-212.

Fanselow MS and Kim JJ (1994) Acquisition of con-textual Pavlovian fear conditioning is blocked byapplication of an NMDA receptor antagonist D,L-2-amino-5-phosphonovaleric acid to the basolat-eral amygdala. Behavioral Neuroscience 108:210-212.

Fanselow MS, Landeira-Fernandez J, DeCola JP, Kim JJ(1994). The immediate-shock deficit and postshockanalgesia: Implications for the relationship betweenthe analgesic CR and UR. Animal Learning andBehavior 22:72-76.

Fanselow MS and Lester LS (1988) A functional behav-ioristic approach to aversively motivated behavior:Predatory imminence as a determinant of the to-pography of defensive behavior. In: RC Bolles andMD Beecher (Eds) Evolution and learning. Hills-dale, NJ: Lawrence Erlbaum Associates, pp.185-212.

Page 437: 0195162854

420 COGNITION

Fanselow MS, Lester LS, and Helmstetter FJ (1988)Changes in feeding and foraging patterns as an an-tipredator defensive strategy: a laboratory simula-tion using aversive stimulation in a closed economy.Journal of Experimental and Analytical Behavior50:361-374.

Fanselow MS, Sigmundi RA, and Williams JL (1987) Re-sponse selection and the hierarchical organizationof species-specific defense reactions: The relation-ship between freezing, flight, and defensive bury-ing. Psychological Record 37:381-386.

Fendt M, Endres T, Apfelbach R (2003) Temporary in-activation of the bed nucleus of the stria terminalisbut not of the amygdala blocks freezing inducedby trimethylthiazoline, a component of fox feces.Journal of Neuroscience 23:23-28.

Fendt M and Fanselow MS (1999) The neuroanatomicaland neurochemical basis of conditioned fear. Neu-roscience and Biobehavioral Reviews 23:743-760.

Griffith C (1920) The behavior of white rats in the pres-ence of cats. Psychobiology 2:19-28.

Helmstetter FJ and Fanselow MS (1993) Aversively mo-tivated changes in meal patterns of rats in a closedeconomy: The effects of shock density. AnimalLearning and Behavior 21:168-175.

Helmstetter FJ and Landeira-Femandez J (1990) Condi-tional hypoalgesia is attenuated by naltrexone ap-plied to the periaqueductal gray. Brain Research537:88-92.

Hitchcock JM and Davis M (1991) Efferent pathway ofthe amygdala involved in conditioned fear as mea-sured with the fear-potentiated startle paradigm. Be-havioral Neuroscience 105:826-842.

Iwata J, Chida K, LeDouxJE (1987) Cardiovascular re-sponses elicited by stimulation of neurons in thecentral amygdaloid nucleus in awake but not anes-thetized rats resemble conditioned emotional re-sponses. Brain Research 418:183-188.

Iwata J, LeDoux JE, Meeley MP, Arneric S, Reis DJ(1986) Intrinsic neurons in the amygdaloid field pro-jected to by the medial geniculate body mediateemotional responses conditioned to acoustic stim-uli. Brain Research 383:195-214.

Kaltwasser MT (1991) Acoustic startle induced ultra-sonic vocalization in the rat: a novel animal modelof anxiety? Behavioural Brain Research 43:133-137.

Kapp BS, Gallagher M, Underwood MD, McNall CL,Whitehorn D (1982) Cardiovascular responseselicited by electrical stimulation of the amygdalacentral nucleus in the rabbit. Brain Research234:251-262.

Kapp BS, Markgraf CG, Wilson A, Pascoe JP, Supple WF(1991) Contributions of the amygdala and anatomi-cally-related structures to the acquisition and expres-sion of aversively conditioned responses. In: L Da-

chowski and CF Flaherty (Eds) Current topics inanimal learning: Brain, emotion and cognition.Hillsdale, NJ: Lawrence Erlbaum Associates, pp.311-346.

Kellicutt MH and Schwartzbaum JS (1963) Formation ofa conditioned emotional response (CER) followinglesions of the amygdaloid complex in rats. Psycho-logical Reports 12(2):351-358.

Kim JJ, Rison RA, Fanselow MS (1993) Effects of amyg-dala, hippocampus, and periaqueductal gray lesionson short- and long-term contextual fear. BehavioralNeuroscience 107:1093-1098.

LeDoux JE, Cicchetti P, Xagoraris A, Romanski LM(1990) The lateral amygdaloid nucleus: sensory in-terface of the amygdala in fear conditioning. Jour-nal of Neuroscience 10:1062-1069.

LeDouxJE, Iwata J, Cicchetti P, Reis DJ (1988) Differentprojections of the central amygdaloid nucleus medi-ate autonomic and behavioral correlates of condi-tioned fear. Journal of Neuroscience 8:2517-2529.

LeDouxJE, Sakaguchi A, Reis DJ (1983) Alpha-methyl-DOPA dissociates hypertension, cardiovascular re-activity and emotional behavior in spontaneouslyhypertensive rats. Brain Research 259:69-76.

Lee Y and Davis M (1997) Role of the hippocampus, thebed nucleus of the stria terminalis, and the amyg-dala in the excitatory effect of corticotropin-releasinghormone on the acoustic startle reflex. Journal ofNeuroscience 17:6434-6446.

Lester LS and Fanselow MS (1985) Exposure to a catproduces opioid analgesia in rats. Behavioral Neu-roscience 99:756-759.

Lewis JW, Cannon JT, LiebeskindJC (1980) Opioid andnonopioid mechanisms of stress analgesia. Science208:623-625.

Liebman JM, Mayer DJ, Liebeskind JC (1970) Mesen-cephalic central gray lesions and fear-motivated be-havior in rats. Brain Research 23:353-370.

Maren S (2001) Neurobiology of Pavlovian fear condi-tioning. Annual Review of Neuroscience 24:897-931.

Maren S and Fanselow MS (1996) The amygdala and fearconditioning: has the nut been cracked? Neuron16:237-240.

Myer JS (1971) Some effects of non-contingent aversivestimulation. In: Aversive conditioning and learning(Bursh FR, ed.). New York: Academic Press.

Roche JP and Timberlake W (1998) The influence of ar-tificial paths and landmarks on the foraging behav-ior of Norway rats (Rattus norvegicus). AnimalLearning and Behavior 26:76-84.

Treit D, Aujla H, Menard J (1998) Does the bed nucleusof the stria terminalis mediate fear behaviors? Be-havioural Neuroscience 112:379-386.

Vernet-Maury E, Constant B, Chanel J (1992) Repellent

Page 438: 0195162854

Chapter 39. Fear 421

effects of trimethylthiazoline in the wild rat (Rattusnorvegicus Berkenout). In: Chemical signals in ver-tebrates (Doty R and Muller-Schwarze D, eds.), pp.305-310. New York: Plenum Press.

Walker DL and Davis M (1997) Anxiogenic effects ofhigh illumination levels assessed with the acousticstartle response in rats. Biological Psychiatry 42:461-471.

Walker DL and Davis M (1997) Double dissociation be-tween the involvement of the bed nucleus of thestria terminalis and the central nucleus of the amyg-dala in startle increases produced by conditionedversus unconditioned fear. Journal of Neuroscience17:9375-9383.

Walker DL, Toufexis DJ, Davis M (2003) Role of the bednucleus of the stria terminalis versus the amygdala

in fear, stress, and anxiety. European Journal ofPharmacology 463:199-216.

Weiskrantz L (1956) Behavioral changes associated withablation of the amygdaloid complex in monkeys.Journal of Comparative and Physiological Psychol-ogy 29:381-391.

Whishaw IQ, Dringenberg HC, Comery TA (1992) Rats(Rattus norvegicus) modulate eating speed and vig-ilance to optimize food consumption: Effects ofcover, circadian rhythm, food deprivation, and in-dividual differences. Journal of Comparative Psy-chology 106:411-419.

Young BJ and Leaton RN (1996) Amygdala central nu-cleus lesions attenuate acoustic startle stimulus-evoked heart rate changes in rats. Behavioural Neu-roscience 110:228-237.

Page 439: 0195162854

Cognitive Processes

ROBERT J. SUTHERLAND40

The goals in this brief overview are two-fold:to provide the reader with an exposure to thekind of thinking about behavioral processes inrats that is characteristic of a cognitive ap-proach and to provide some examples of ex-perimental paradigms that have been usedwith some success to probe the nature of therat's representations of its environment andits own actions. The examples are drawnmainly from recent work on attentional andmemory processes, two of the traditional sub-ject matters in cognitive science. At the out-set, however, it is important to identify whyan approach that is limited to studying just therat's processing of current stimuli and detailsof on-going movements is inadequate—whatmotivates the study of cognition?

WHAT IS A COGNITIVE PROCESS?

The study of cognition in the rat and othernonhuman species has often triggered majorcontroversies and misunderstandings. Themain reason for this is that at the core whatis "cognitive" about a cognitive process is con-trol over behavior by an event or a relation-ship between events that are not now present,that is, not given by the immediate stimuli.We are often surprised by an animal when itsbehavior seems to reflect knowledge about fu-ture outcomes or provides an inference aboutthe current situation or sensitivity to rela-tionships that are "emergent." In this context,emergent means sensitivity to a relationshipnot explicitly trained. The traditional account

of such behavior involves the idea thatthrough experience the rat's nervous systembuilds up a "representation" of certain, spe-cific events, or relationships in its environ-ments. The information contained in theserepresentations guides appropriate behavior.

Cognitive neuroscience work with ratsinvolves the study of how these representa-tions are built, how and where they are storedin the brain, and in what ways they influenceongoing behavior.

STIMULUS-RESPONSE VIEWSUNDERSPECIFY BEHAVIOR

A key to understanding why cognition in therat has generated heated controversy is an ap-preciation of the keen skepticism that animalbehaviourists historically have shown towardcomplex, inferred, inner causes of behavior, es-pecially when they involve conspicuous simi-larities to related human processes. There aremany good reasons to be suspicious ofanthropomorphic approaches. On the otherhand, there are many clear examples of howembracing a narrow stimulus-response per-spective has caused investigators to completelymiss interesting neurobiological processes. Inthis regard, the field of circadian rhythm re-search offers a telling parallel to research oncognition in animals. Two contrasting viewswere played out in developing an understand-ing of how the daily activities of animals wereorganized into a daily cycle. One view, akin tothe stimulus-response perspective, sought to

422

Page 440: 0195162854

Chapter 40. Cognitive Processes 423

discover the identity of the fluctuating cues inthe environment that were the basis for rhyth-mic behavior. The alternative view held thatthere were biologically and chemically com-plex internal clocks, endogenous oscillatingsystems, in the brains of animals that controlthe daily activity and rest cycles in behaviorand physiology. The former view yieldednothing of biological significance. This view al-ways underpredicts the existing organizationin the dynamics of behavior.

In contrast, the internal clock viewspawned an understanding of a rich set ofmechanisms involving "clock genes," hierar-chically nested rhythmic, interacting neuralsystems, and circuitry to synchronize or en-train internally generating rhythms to envi-ronmental cues, not to mention interestingclinical insights about certain psychopatholo-gies (Golombeck et al., 2003; Lowrey andTakahashi, 2000).

By analogy, in our subject domain,stimulus-response views always underpredictor underspecify complex behavioral processescontrolled by neural representations of eventrelationships. A simple thought experimentshould demonstrate this point. Consider a ratin a T-maze housed in a large room withabundant large visual cues around its perime-ter. Imagine that the stem of the T-maze (thestart arm) is always pointing southward dur-ing the rat's initial experiences. The rat is hun-gry, and on every trial it finds a small morselof tasty food at the end of the arm pointingeastward. The rewarded arm is white and theother (west) arm is black; after several trials,we observe the rat efficiently running downthe start arm and always making a right turninto the white goal arm. At this point, if wewere good stimulus-response experimenters,we could easily conduct a "probe" test to de-termine if the rat is navigating to the food byapproaching the white arm or by making aconstant right turn. The probe should involvepitting the two sources of response controlagainst each other, by having the white com-pete against the right turn. This can be done

by simply switching the white and black arms.When we conduct this "competitive test," wediscover that the rat continues to make a rightturn. We should be content that we have un-covered the simple stimulus-response rule,until some clever student carries out a secondprobe trial in which she rotates the T-maze sothat the stem (the start arm) is pointing north-ward, leaving the white arm in its westwardposition. Now the rat beginning the probetrial running southward in the stem, reachesthe point where it must turn right (alwaysrewarded) into the white arm (always re-warded) and to our dismay it turns left intothe black arm. We repeat this probe trial manytimes with other rats, always observing thesame outcome. We note that by turning leftinto the black arm, the rats have arrived at thesame location in the room at which they werealways rewarded during all earlier experiences.

So far we have demonstrated that neitherthe white arm nor the right turn is a constancyin the rat's choice, but location within theroom looks like it could be the key. One lastprobe trial should complete the picture. Wenotice that not only has the rat been going tothe same spot in the room, even during thecompetitive tests, but also it has been headingtoward the same direction in the room (east-ward) on every trial. By shifting the T-mazeeastward such that the end of the west arm isexactly where the end of the east arm hadbeen and the east arm projects much farthereastward than it ever had before, we can com-plete our competitive testing. If the rat headseastward on this trial to a point in the roomit had never experienced, then we can rule outlocation in the room as the constancy. By thesame token, if the rat turns into the westwardarm, we can rule out directional constancy.The fact of the matter is that with real ratsand real experiments it is likely that we couldarrange the complex factors influencing choicesuch that we could have some rats choosingwhite arms, others making only right turns,another set going to a specific location inthe room, and the rest always heading in a

Page 441: 0195162854

424 COGNITION

specific direction. A recent conceptually re-lated series of experiments by Skinner et al.(2003) demonstrates this point.

Our stimulus-response perspective is hereforlorn. The stimuli in the apparatus and inthe room are invariant, it is always the samefood reward, and no matter how fine-grainedare the measurements of the movements therat makes as it turns right, turns controlled bybrightness, direction, or locations are thesame turns.

IT IS THE REPRESENTATION

On a cognitive account of the rat in ourthought experiment, information about bright-ness, turning and going straight, direction, andlocation relative to room cues form the gristfor different types of representations. In addi-tion, these can all be simultaneously built upin different neural networks. To extend ouranalogy from circadian rhythms, in which dif-ferent fluctuating environmental cues such aslighting, temperature, noises, social activities,and feeding can synchronize or entrain vari-ous internal rhythms (they do not generatethem), so too, in our experiments, differentforms of cognitive representation are en-trained by the cues in the environments. Theproblem of understanding complex behaviordoes not have a determinate solution unlessit is known which of the various representa-tional systems is engaged as the rat experi-ences an environment. At this point it isimportant to note one meaning of "represen-tation" that is not intended here. If we say thata rat has built up representations of the re-wards that are available in an environment,we do not mean that there are internalizedcopies of the reward items that from time totime can be regurgitated to be retasted or lit-erally reexperienced. Here, reward representa-tion means that there is a neural network inwhich a pattern of activity that correspondsto the reward item in such a way that neuraloperations involving that pattern can lead to

valid conclusions or inferences about whatwould be adaptive behavior in an environ-ment that contains the specific reward item(see the section on reinforcer representations).An example from the spatial domain may bemore concrete. By saying that a rat has a rep-resentation of direction and distance from itscurrent location to its home, it is implied thatsome set of operations on the pattern ofactivity across the neural network re-presenting that information can lead to ahomeward trip with a direct trajectory havingadaptive acceleration profiles (as has beenshown by Wallace & Whishaw [Chapter 42],see also Gallistel, 1990). It is not meant thatthere is a literal Cartesian map that is calledup, inspected with the mind's eye, and, basedon mentally measured topographical coordi-nates on the map, a correct course plotted.

Cognitive processes in many ways re-semble basic sensory or motor processes andin important ways they are different. One re-semblance is that individual cognitive subsys-tems are specialized to enable specific types ofaction important in the rat's trafficking withits natural or social environment. Also, theprocesses are informed only by specific typesof information. This allows for highly spe-cialized and abstract processing, but this alsoimposes important constraints. Thus, the sortof cognitive systems that rats have do not per-mit all actions to be informed by all types ofinformation. The functional organization ofcognition confers the advantages of highlyspecialized processing that allows for flexiblebehavior in respect to cues and movements,as well as cognitive blind spots.

REPRESENTATIONS INTERACT

It was noted earlier that multiple different rep-resentations are being built up in differentneural networks as a rat experiences an envi-ronment. A wide area of investigation that isrelatively untapped involves the nature of theinteractions among representational systems.

Page 442: 0195162854

Chapter 40. Cognitive Processes 425

There are observations that suggest that un-der certain circumstances, different represen-tations interfere with or inhibit one anotherand obviously they can be synergistic or sup-portive of each other. An especially clear ex-ample of the latter interaction involves therepresentations of head direction in a networkincluding postsubiculum and anterior thala-mus with the place field representation in thehippocampus. When the head direction sys-tem is disrupted by damage to either post-subiculum or anterior thalamus, hippocampalplace fields are still intact, but aspects of theirinformation content and stability betweenepisodes in the same environment are de-graded (Calton et al., 2003). The representa-tion of head direction is clearly useful in build-ing up and maintaining a representation ofwhere the rat is in relation to visual and otherenvironmental cues.

We expect that representations of cues oractions could support the building of repre-sentations in downstream systems. There is,however, evidence for the counterintuitivenotion that acquiring one representation canblock learning in a separate representationalsystem and that this blocking is not due to thetwo systems merely driving competing move-ments. A nice example of this kind of interac-tion can be found in the work of McDonaldand White, who studied rats learning the re-lationship between food reward and a partic-ular cue in an arm of a simple maze. Therewas interference with this simple learning ifthe rats had had the opportunity to explorethat maze and environment. The interferenceeffect was shown to be due to the building upof a hippocampal system representation of theenvironment during initial exploration andthe conditioning to the cue was shown to de-pend on amygdala circuitry (McDonald andWhite, 1995; White and McDonald, 1993,2002).

In the rat, there are several nice examplesof synergistic and antagonistic interactions be-tween representations in different systems.What about different representations within

the same system? It is safe to say that we knowlittle about how representations interact be-tween systems, and we know almost nothingabout within-system representational interac-tions beyond that they occur. The recent ex-ample comes from studying contextual avoid-ance learning (Fenton et al., 1998). Rats areplaced on a circular tabletop with salient cuesaround the room. When they enter a specificregion on the table, they receive a mild footshock. They learn to quickly avoid enteringthat region. A moment's reflection revealsthat, as in our preceding T-maze example, theidentity of the region can be defined by morethan one kind of information. To consider justtwo, the region can be represented by its re-lationship to the available cues around theroom or by its relationship to the availablecues on the table (which would be supportedby self-motion information). We know fromother work measuring the place field proper-ties of hippocampal neurons that either ofthese types of information can serve as a frameof reference for the hippocampal representa-tion (Gothard et al., 1996). Which does the ratuse in this situation? The answer is both.

Fenton and co-workers (1998) demon-strated that both representations were "si-multaneously" active by rotating the table ina slow and continuous manner. The rat, in thesame episode, would avoid a region that ro-tated with the table frame and a region thatwas stable in the room frame. Furthermore,when one records from neurons in hippo-campus in this situation, some have placefields relative to the table frame and others tothe room frame (Zinyuk et al., 2000). Arethese two representations of the environmentsimultaneously active, or does the hippocam-pal network switch quickly and coherentlyfrom one representation frame to the other?We do not know. It is possible that these twoframeworks could be interleaved through thenetwork, but equally one can imagine that asthe rat attends to different features of the en-vironment, these two representations are suc-cessively recalled. This latter possibility sug-

Page 443: 0195162854

426 COGNITION

gests that attention could be a critical processin the rat differentially allocating processingresources to different parts of the environ-ment and hence to different representations.

We turn now to how it is that attentionis studied in the rat.

ATTENTION

I would like to describe a limited set of be-havioral tasks that have been used successfullyto study somewhat different aspects of atten-tion in the rat. Attention involves many sep-arable processes but we will consider onlytwo: (1} sustaining alert responsiveness in theface of occasional brief events and (2) selec-tive attention.

SUSTAINED ATTENTION

Robbins and his co-workers have made ex-tensive use of a serial reaction time task tomap several components of the rat's atten-tional networks (Muir et al., 1996). The task,the five-choice serial reaction time task, re-quires the rat to detect a brief (500 millisec-onds) visual stimulus. This visual target is pre-sented through one of five small holes in awall. Each of the five locations is used equally,often according to a random sequence. Therat's job is to quickly (within 5 seconds) pokeits nose into the hole that had been lit, where-upon it is rewarded with a food pellet at a sep-arate magazine. Five seconds later, the nexttrial begins. The task lends itself well to meas-uring several aspects of attentional perform-ance. The typical measures include reactiontime to nose-poke (from light target onset),response accuracy (nose-poking into the cor-rect hole), reaction time to move to the foodmagazine, and number of trials during whichno response is made. It is of importance that,in relation to attention, the rat must maintainan alert state to detect briefly presented visualtargets, and because the target can unpre-dictably appear at one of five locations, good

performance requires that the rat actively scanacross the relevant spatial layout. There arealso opportunities to detect other kinds of be-havioral effects such as nose-poking beforetarget presentation (premature responding),repeated poking into the same hole (perse-veration), and simply generally slow move-ment (Passetti et al., 2002).

SELECTIVE ATTENTION

Implicit in the five-choice serial reaction timetask is the idea the rats can divide attentionamong multiple spatial locations, presumablyby scanning across the layout, sampling suc-cessively, or simultaneously the target sites.The covert orienting task makes explicit thatrats have the capacity to selectively attend toa region of space, such that their processingof information from that region is selectivelyenhanced. Posner (1980) designed this simpleparadigm that permits measurement of themechanisms that specifically underlie the se-lective spatial attention process. In brief hisprocedure involves subjects responding asquickly and as accurately as they can whenthey detect a briefly presented visual target.The target appears equally often to the left orright of head direction. For example, in an ex-periment by Stewart et al. (2001), a trial be-gins with the rat inserting its snout a specifieddistance into a hole to interrupt a photobeam.Just before the target appears, a cue is pre-sented to the left or right. Usually the cue'slocation predicts where the target will appear,but sometimes the cue and target appear onopposite sides. The former case constitutescue valid trials, and the latter cue invalid tri-als. There also are trials during which the tar-get appears without prior cuing and still othertrials when two cues, one left and one right,appear before the target. The rat's task is todetect the visual target and to quickly with-draw its snout from the nose-poke hole.

It is thought that the appearance of thebrief cue on one side reflexively attracts at-tention to that region of space. If it did, then

Page 444: 0195162854

Chapter 40. Cognitive Processes 427

one should find that the reaction time (fromtarget onset to snout withdrawal) would befaster than if no cue or two cues had appeared.Further, attention should be allocated awayfrom target locations on the other side, lead-ing to longer reaction times when the cue isinvalid (i.e., when the cue and target appearon opposite sides). Each of these predictionshas been confirmed in work with rats, stronglysupporting the idea that rats do in fact havemechanisms for selectively and covertly direct-ing attention to specific locations. The fact thatreaction times are shorter on valid cue trialsthan on trials when two cues are presented rulesout an explanation of the valid cue effect that isbased on the cue simply alerting the rat to theimpending target appearance.

If rats have a mechanism for selectivelyshifting attention to different points in space,do they also have mechanisms that enable shift-ing attention between cues (or objects) or cuetypes, independent of where these might be lo-cated in space? One approach to the study ofnonspatial selective attention takes advantageof a cognitive phenomenon referred to as at-tentiowl set. Imagine that we present a subjectwith a variety of items that differ along severalstimulus dimensions, say color, shape, and size,and require initially that the subject discrimi-nate among the items based on color. Afterlearning, we will find relatively easy transfer ofdiscriminations to new colours that we intro-duce but more difficult transfer if we shift torequiring discriminations based on shape orsize. If the stimulus dimension stays the samefor a new discrimination, it is called an intradi-mensional shift, and if a different dimension isused, it is called an extradimensional shift. Thefinding that intradimensional shifts proceedmore readily than extradimensional shifts is thebasis for ascribing to the subject an attentionalset. It is believed that the initial training withone dimension induces a selective scanning ofobjects for relevant perceptual features. Thisselective attention to the relevant dimensioncomes at the cost of diminished processing ofvalues of irrelevant perceptual features.

Verity Brown and her colleagues de-signed a procedure for rats demonstrating thatthey can selectively allocate attention to per-ceptual features of objects, independent ofspatial location. They have designed an atten-tional set-shifting task in which rats learn to digin small bowls for a food reward. The bowlsdiffer in odor, digging medium, and surfacetexture. Rats learn to discriminate readily us-ing any of the three dimensions and readilylearn new discriminations if the perceptual di-mension of the new discrimination remainsthe same. They find it significantly more dif-ficult to learn a new discrimination if it in-volves a shift to a previously irrelevant di-mension. That is, rats have more difficultywith extradimensional compared with in-tradimensional shifts, a defining characteristicof an attentional set.

ATTENTIONAL NEURAL SYSTEMS

In several experiments, Sarter, Robbins, andothers have shown that forebrain cholinergicsystems projecting to neocortex are essentialin supporting accurate sustained attention inthe rat. The demonstrations of cholinergic in-volvement include reduced sustained atten-tion performance in rats after selective elimi-nation of forebrain cholinergic cells using theselective immunotoxin IgG-192 saporin aswell as extracellular unit recordings and invivo measurement of acetylcholine release incortex (Dalley et al., 2001; Everitt and Rob-bins, 1997; McGaughy and Sarter, 1998, 1999;Sarter and Bruno, 1997; Sarter et al., 2001).The same cholinergic systems have been im-plicated in maintaining spatially selective at-tention in the rat, but most of this work hasbeen conducted with systemically adminis-tered cholinergic drugs, limiting conclusionsabout locus of action (Phillips et al., 2000).

Medial prefrontal cortex damage in therat has more complex effects in sustained at-tention tasks (Passetti et al., 2002); nonethe-less, it is clear that circuits in this region are

Page 445: 0195162854

428 COGNITION

critical for sustained attention. Also in linewith work from primates, medial prefrontalcortex damage dramatically disrupts the rat'sability to shift attentional set (Birrell andBrown, 2000).

MEMORY

One of the most interesting controversiesbearing on cognitive processes in the rat is theextent to which memory representations en-able rats to predict the future and to revisitthe past.

CAN RATS PREDICT THE FUTURE?

Many cognitive representations, by their verynature, should confer on rats the ability to an-ticipate future outcomes of certain actions orevents without actually experiencing them.We have seen one example of this involvinga form of spatial navigation in which the rep-resentation of home base leads to the direc-tion and acceleration characteristics of a triphome that are consistent with the rat havingthe expectation of finding its home at a cer-tain spot—the rat begins to slow down at aparticular part of the trajectory in anticipationof arriving home (Wallace and Whishaw,Chapter 42). How general is this property ofmemories? Are there other clear examples inthe behavior of rats that show that they haverepresentations of the properties of antici-pated future events?

REINFORCER REPRESENTATIONS

In a typical Pavlovian conditioning experi-ment, a rat experiences some event (A) closelyfollowed in time by some significant event(US), say food delivery, that has an ability re-liably evoke a behavioral response (UR). Invirtue, of the relationship between A and thefood, A alone comes to evoke conditioned re-sponses (CR), say, approaches to the site offood delivery. There are very clear examples

of the fact that during conditioning not onlyis an association formed between A and theCR but also A comes to activate a represen-tation of the specific food. Further, that oper-ations involving this food representation canchange the way the rat subsequently respondsto A. This is shown by the fact that if afterconditioning we pair the food with illness (re-inforcer devaluation) without involving A, wediscover that when A subsequently occursconditioned responding is diminished (Hollandand Straub, 1979). This effect is specific to con-ditioned stimuli paired with the specific foodtype. Thus, without actually experiencing A to-gether with a devalued US, the rat treated Adifferently, because of its associations with amodified representation of the US.

We find an exactly parallel mechanism inthe case of instrumental learning. We have anenvironment that is structured such that if therat performs action X, then it obtains a rewardwith one flavor (Fl), and if it performs actionY, it obtains a reward with a second flavor (F2).Traditionally, it is thought that the rewards inthis situation function primarily to increase theprobability of occurrence of the two actions.More recent work, especially by Balleine andco-workers (Balleine and Dickinson, 1998a,1998b), has shown the rat forms a connectionbetween the actions and representations of thetwo outcomes. This is shown after initial train-ing is complete by feeding the rat to satiationwith one of the two flavors. The rat is thenplaced back into the training environment, andthe two actions are measured in extinction (nofurther reward presentation). If Fl was deval-ued by satiation, then the probability of actionX is selectively reduced; if F2 was devalued,then the probability of action Y is selectively re-duced. Many experiments on this mechanismconverge on the conclusions that the rat rep-resents the action-outcome relationship andthat the perceptual features of the outcome areincluded in this representation such that alter-ing the motivation associated with the out-come can come to alter motivation to performthe associated action.

Page 446: 0195162854

Chapter 40. Cognitive Processes 429

In both Pavlovian and instrumental con-ditioning, if the rat experiences a modificationof the value attached to an outcome, it willautomatically transfer a new value, withoutdirect experience, to previously associatedcues and actions.

CONFIGURAL/CONjUNCTIVEREPRESENTATIONS

Cues that rats encounter can occur alone or inconjunction with other cues. Sometimes thefact that cues occur together has an importantpredictive relationship to some outcome. Of-ten, regardless of any relationship to othercues, a specific cue is unambiguously predic-tive. Social transmission of food preferences(see Chapter 23) offers us an example of ameaningful conjunction between cues (foodflavor and essence of rat breath). There are of-ten more arbitrary cue conjunctions that areused by the rat. An especially good examplecan be found in context fear conditioning. Ratsthat experience a shock after a discrete cuelearn not only to fear the discrete cue but alsoto fear the constellation of cues comprising thecontext. In some interesting ways, the contextconditioning is different from the discrete cueconditioning (Rudy and O'Reilly, 1999). For ex-ample, placing the rat into the context for thefirst time and almost immediately deliveringthe shock produces no fear of the context. Therat must explore the context for an extendedinterval before pairing it with shock if it is toacquire reliable context fear (Fanselow, 1990,2000). If instead the rat explores individual fea-tures of the context to the same extent butnever together, then pairing of the completecontext with shock leads to no fear of the con-text (Rudy and O'Reilly, 1999). It appears to bethe case that during the period of preshock ex-ploration, the rat builds up a single configuralor conjunctive representation of the elementsof the context, a single reference frame. In prin-ciple rats could use just the individual, separateelements of the context to predict shock andgenerate fear—but they do not.

There are clear examples of experimen-tal procedures in which resolving a discrimi-nation requires that the animal represent therelationship among two or more cues. Thesimplest is the negative patterning discrimina-tion, in which either cue A or cue B is associ-ated with a reinforcer, but when A and B oc-cur together, the reinforcer is never delivered(A+, B+, AB—). Rats do learn to respondreadily to either cue but not to their co-occurrence. A simple account of this ability isthat the rat not only has representations of theindividual cues but also a configural or con-junctive representation of the two cuestogether. Each of these representations canenter into associations with outcomes. If therat had only representations of the individualcues, negative patterning could not be solved.A second example may make this point evenmore clearly. The transverse patterning problemhas the same formal structure as the "rock-paper-scissors" game. The rat experiencesthree cues, A, B, and C. They occur togetherin pairs. When A and B are together, choos-ing A is rewarded; when B and C are together,choosing B is rewarded; when C and A are to-gether, choosing C is rewarded. Again, if therat formed only representations of the indi-vidual cues, the problem cannot be solved. Itis only through building up representationsthat include the co-occurrence or conjunc-tions of the pairs of cues that the discrimina-tions can be resolved (Alvarado and Rudy,1992). It is important to note that the transi-tive inference problem used by Eichenbaum andco-workers (see, for example, Dusek andEichenbaum, 1997), unlike transverse pat-terning, does not require relational, config-ural, or conjunctive representations; instead,it can be solved by a series of simple or ele-mental representations (Frank et al., 2003;Van Elzakker et al., 2003). The transitive "in-ference" problem begins with training a rat ona series of discriminations problems, A+B—,B+C-, C+D-, D+E-. When presentedwith a novel combination BD, they choose B.Eichenbaum and co-workers assert that this

Page 447: 0195162854

430 COGNITION

must be due to formation of a complex or-dered relational representation from whichrelative value of new combinations can beinferred. Rudy and O'Reilly (1999) haveconvincingly demonstrated that this is notnecessary, that a solution exists based on thebuild up of associative strength of each cue'srepresentation.

HIERARCHICAL REPRESENTATIONS

In the real world, cues that co-occur do notalways appear simultaneously. When they doappear simultaneously, as described earlier,rats tend to come to represent them in a con-figural or conjunctive fashion. Configural rep-resentations likely participate in predictingoutcomes in the same way as elemental cuerepresentations (Rudy and O'Reilly, 1999;Sutherland and Rudy, 1989). This is often notthe case if they appear in a sequence. Holland(1992) and others (Miller and Oberling, 1998;Swartzentruber, 1995) have made a convinc-ing case that rats can represent cues in a spe-cial, hierarchical fashion, especially if theyoccur sequentially. The general name for ex-perimental procedures to study this form ofrepresentation is occasion-setting.

The simplest occasion-setting paradigm,and one studied with great success by Hollandand others, is the feature-negative discrimina-tion. The rat is exposed to a cue A that alwayspredicts a reinforcer, unless it is preceded bycue B (the occasion-setter). Thus the discrim-ination involves A+; B -» A— and the mean-ing of A depends on prior occurrence of B.Rats readily resolve such discriminationsresponding differently to A depending onwhether B preceded it. One could imaginethat the rat could simply learn that B is di-rectly inhibitory in respect to the reinforcer.However, it is clearly shown that the repre-sentation of B enters into a different kind ofbehavioral role. Instead of predicting a specificoutcome, B predicts that specific relationshipsexist between two or more other events. Thatis, the representation of B predicts which out-

comes other cue representations will predict.The specific cue predictions can thus benested inside a higher-level representationalstructure. It appears that occasion-setters es-tablish which of the possible event relation-ships that the rat has experienced are nowlikely to be operative. Interestingly, it hasbeen shown that at least some aspects of oc-casion-setting functions are disrupted by se-lective damage to the hippocampal system(Holland et al, 1999).

FUNCTIONAL EQUIVALENCE

Honey and colleagues (Coutureau et al., 2002)have recently observed interesting behavioralexamples showing that rats have even morecomplex, higher-level representations of eventrelationships. In their experiments, rats expe-rience two cues, X and Y, that have a predic-tive relationship to food delivery. For half ofthe occasions, X predicts food and Y predictsno food (X+; Y—); for the other halfof the occasions, the relationship is reversed(Y+; X—). When rats learned that a cue pre-dicted food, they approached the food welland pushed open a small covering flap, sofood-well inspection is the learned response.Honey et al. uses four different occasion-set-ters to signal for the rat, which of the tworelationships between cues and food is opera-tive. The occasion-setters never appear to-gether. Two of the occasion-setters, A and B,signal X+; Y—, and the other two occasion-setters, C and D, signal Y+; X-. Thus thefour component discrimination problemsare A: X+.Y-; B: X+,Y~; C: X+.Y-; D:X+, Y—. After training, the rats receive addi-tional experiences in the presence of A and C.With A (and not C), they receive delivery ofmany free food pellets. After this additional ex-perience, the conditioned responding during Band then during D is measured. The interest-ing finding is that increased responding to B andnot to D emerges during the test. The similar-ity between A and B and their distinctivenessfrom C and D does not reside in any percep-

Page 448: 0195162854

Chapter 40. Cognitive Processes 431

tual similarity but rather in the fact that theypredict the same relationships between otherevents. Thus, the new behavior to B emergeswithout any new experience with in virtue ofits functional equivalence to A. Honey et al.have also shown two additional facts. If the oc-casion-setters are either static context cues orcues with discrete onsets and offsets, the newbehavior transfers equally among them. Othermethods of revaluing the occasion-setters, forexample, shock in their presence, produce sim-ilar transfer through their equivalent signalingfunctions (Honey and Watt, 1999). Interest-ingly, the transfer does not occur in rats withselective entorhinal cortex damage (Coutureauet al., 2002).

This example demonstrates a further factabout the nature of the rat's representation ofthe elements of its environment. Not only dorats represent perceptual and motivationalproperties of events and outcomes and theirpredictive relationships, but also when anevent signals that a certain predictive relation-ship obtains between other events, that sig-naling function becomes part of that event'srepresentation. This permits interesting cog-nitive performance to emerge in behavioral re-sponses to functionally related signals.

DO RATS REVISIT THE PAST?

It is clear that rats have much more rich repre-sentational processes than would be predictedby traditional stimulus-response accounts oftheir behavior. There are numerous examples,only some of them outlined earlier, indicatingthat rats use these representations built upthrough their experiences with an environmentto predict the future—the outcomes and eventrelationships in that environment.

In terms of applying these cognitive par-adigms to attempts to model human condi-tions, we should be heartened by many im-portant convergences. Many of the sameneural systems that have been established inhumans as being critical for attentional

processes (e.g., forebrain cholinergic projec-tions, prefrontal and anterior cingulate cortexsystems) have been convincingly shown to beinvolved in very similar, possibly the same,processes in rats. Likewise in the memory do-mains, we know from work with brain-in-jured humans or in neuroimaging studies thata ventral prefrontal cortex and amygdala net-work is involved in representing the affectiveand motivational value of significant eventsand cues that predict them (Bechara et al.,1999). The same system in the rat has beenshown to be involved in the representationalprocesses involved in cue-reinforcer associa-tions and the behavioral effects of reinforcerrevaluation (Balleine et al., 2003; Gallagher etal., 1999; Hatfield et al., 1998). Furthermore,some spatial and configural memory tasks(e.g., Morris water task or transverse pattern-ing) that are affected by disrupting hip-pocampal system functions in rats are simi-larly affected by damage to the hippocampalsystem in humans (Astur et al., 2002; Reed andSquire, 1999; Rickard and Grafrnan, 1998) andcause reliable activations in this system dur-ing task performance in neuroimaging exper-iments (Ekstrom et al., 2003; Hanlon et al.,2002). Thus, the kinds of behavioral para-digms described earlier that are useful in clar-ifying cognitive processes in the rat are likelyto be very valuable in work applied to humancognitive disorders.

What about episodic memories? Manybelieve that it is deficits in an episodic mem-ory system that are at the core of human am-nesia, most memory disorders, and certain de-mentias. Thus, the topic is of considerableapplied importance. Tulving (1972) suggeststhat the key properties of episodic memoriesare that they represent when an episode oc-curred as well as what the event relationshipswere and where the episode took place. Fur-thermore, the reactivation of an episodicmemory representation can occur outside ofthe relevant context. Clayton et al. (2003) sug-gest three related criteria for establishingepisodic memory competence: (1) content, or

Page 449: 0195162854

432 COGNITION

what, where, and when; (2) an integrated rep-resentation of these three elements of content;and (3) flexibility in updating memory repre-sentations in light of new information gath-ered after the original episode. There is littledoubt that humans have such memories, butdo rats? Tulving and others doubt it, but theyclearly cannot prove that no such episodicmemory representations exist. It would bereasonable to conclude it is absent in the ratafter numerous clever attempts to demon-strate it have failed (assuming that there is rea-son in principle that it cannot be demon-strated in the rat). I assert that there haveprobably never been ambitious, clever at-tempts to demonstrate episodic memories inrats using a paradigm that covers all of the rel-evant criteria. The trick in such a demonstra-tion is to show that the rat's event memoryincludes an integrated representation of what,where, and when and that it can be updatedin light of new relevant events.

One way to arrange such a demonstra-tion has been devised for the experimentswith the Western scrub jay (Clayton et al.,2001). Clayton has taken advantage of the factthat these jays naturally cache perishable food.It is well established that they rememberwhere their caches are located and becausethey find some foods more palatable than oth-ers, their cache site preferences reveal thatthey remember what was cached. By varyingwhen more palatable, but perishable, foodswere cached in relation to the opportunitiesto retrieve cached foods, one can use thecache site retrieval preferences to learn if thejays remember when they cached the perish-able but more palatable, foods. Clayton has aclear example of episodic-like memory in abird species (Griffiths and Clayton, 2001). Nosimilar attempt has been reported in rats.

We have seen that the rat's event repre-sentations contain information about whereevents took place and, based on the reinforcerdevaluation experiments, it is clear that theyremember what outcomes are expected. (SeeDay et al., 2003, for a very nice demonstration

of this in a novel learning procedure involv-ing where-what paired associates.) What re-mains to be demonstrated is that the sameevent representation also contains informa-tion about when the event occurred.

I suggest that a simple way, given avail-able data, to accomplish the demonstration ofepisodic memory in the rat is to use the reval-uation procedure in Honey's functional equiv-alence experiments but with occasion-setterswhose functions depend not only on their per-ceptual properties but on when they occur(e.g., time of day). Each occasion setter canhave two time-stamps (say each cue occurs inthe morning and the afternoon). We arrangethe environment such that it is the conjunc-tion of the cue identity and the time-stampthat signals where and what to do. We thenrevalue cue A at time 1 and test for transferto cues that signal the same what-where re-lations as A when they are presented at thesame time of day as the revalued cue A. Ifthe rat "automatically" changes its behaviorin the presence of a different cue with thesame time-stamp and functional associationsas the revalued A without any further directexperience with the different cue, we willhave demonstrated that rats can show at leastone form of time travel through their memo-ries. Such a demonstration would bring us astep closer to measuring episodic memoryin rats.

Memory researchers working with ratshave only recently started to come to gripswith the difficult conceptual and method-ological challenges presented by demonstrat-ing true episodic memory processes usingnonlinguistic criteria.

CONCLUSION

The field of study of cognition in the rat hasborne recent fruit. We now have a rich vari-ety of behavioral tasks that enable reliable andvalid measurements of several components ofattention and of the characteristics of the rat's

Page 450: 0195162854

Chapter 40. Cognitive Processes 433

representations of important events and in-terevent relationships. We can be confidentthat there are striking similarities in the neu-ral systems that subserve attentional andmemory representational processes in ratsand humans. We should be equally confidentthat we have not yet probed the full extent ofthese similarities, nor have we come to gripswith what will likely be striking representa-tional differences between rats and humans.It is an exciting time to be a student of rat cog-nition, as we continue to develop and refinebehavioral tools for investigating the mind ofthe rat we will inevitably learn more aboutour own minds.

REFERENCES

Alvarado MC and Rudy JW (1992) Some propertiesof configural learning—An investigation of thetransverse-patterning problem. Journal of Experi-mental Psychology Animal Behavior Processes18:145-153.

Astur RS, Taylor LB, Mamelak AN, Philpott L, Suther-land RJ (2002) Humans with hippocampus damagedisplay severe spatial memory impairments in a vir-tual Morris water task. Behavioural Brain Research132:77-84.

Balleine BW and Dickinson A (1998a) The role of in-centive learning in instrumental outcome revalua-tion by sensory-specific satiety. Animal Learningand Behavior 26:46-59.

Balleine BW and Dickinson A (1998b) Goal-directed in-strumental action: contingency and incentive learn-ing and their cortical substrates. Neuropharmacol-ogy 37:407-419.

Balleine BW, Killcross AS, Dickinson A (2003) The ef-fect of lesions of the basolateral amygdala on in-strumental conditioning. Journal of Neuroscience23:666-675.

Bechara A, Damasio H, Damasio AR, Lee GP (1999) Dif-ferent contributions of the human amygdala andventromedial prefrontal cortex to decision-making.Journal of Neuroscience 19:5473-5481.

Birrell JM and Brown VJ (2000) Medial frontal cortexmediates perceptual attentional set shifting in therat. Journal of Neuroscience 20:4320-4324.

Calton JL, Stackman RW, Goodridge JP, Archey WB,Dudchenko PA, Taube JS (2003) Hippocampal pakecell instability after lesions of the head direction cellnetwork. Journal of Neuroscience 23:9719-9731.

Clayton NS, Bussey TJ, Dickinson A (2003) Can animalsrecall the past and plan for the future? Nature Re-views Neuroscience 4:685-691.

Clayton NS, Yu KS, Dickinson A (2001) Scrub jays(aphelocoma coerulescens) form integrated memo-ries of the multiple features of caching episodes.Journal of Experimental Psychology Animal Behav-ior Processes 27:17-29.

Coutureau E, Killcross AS, Good M, Marshall VJ, Ward-Robinson J, Honey RC (2002) Acquired equivalenceand distinctiveness of cues: II. Neural manipulationsand their implications. Journal of Experimental Psy-chology Animal Behavior Processes 28:388-396.

Dalley JW, McGaughy J, O'Connell MT, Cardinal RN,Levita L, Robbins TW (2001) Distinct changes incortical acetylcholine and noradrenaline efflux dur-ing contingent and noncontingent performance ofa visual attentional task. Journal of Neuroscience21:4908-4914.

Day M, Langston R, Morris RGM (2003) Glutamate-re-ceptor-mediated encoding and retrieval of paired-associate learning. Nature 424:205-209.

Ekstrom AD, Kahana MJ, Caplan JB, Fields TA, IshamEA, Newman EL, Fried I (2003) Cellular networksunderlying human spatial navigation. Nature 425:184-187.

Everitt BJ and Robbins TW (1997) Central cholinergicsystems and cognition. Annual Review of Psychol-ogy 48:649-684.

Fanselow MS (1990) Factors governing one-trial con-textual conditioning. Animal Learning and Behav-ior 18:264-270.

Fanselow MS (2000) Contextual fear gestalt memories,and the hippocampus. Behavioural Brain Research110:73-81.

Fenton AA, Wesierska M, Kaminsky Y, Bures J (1998)Both here and there: Simultaneous expression of au-tonomous spatial memories in rats. Proceedings ofthe National Academy of Sciences U S A 95:11493-11498.

Frank MJ, Rudy JW, O'Reilly RC (2003) Transitivity,flexibility, conjunctive representations and the hip-pocampus. II. An computational analysis. Hippo-campus 13:299-312.

Gallagher M, McMahan RW, Schoenbaum G (1999) Or-bitofrontal cortex and representation of incentivevalue in associative learning. Journal of Neuro-science 19:6610-6614.

Gallistel CR (1990) The organisation of learning. Cam-bridge, MA: The MIT Press, 1990.

Golombek DA, Ferreyra GA, Agostino PV, Murad AD,Rubio MF, Pizzio GA, Katz ME, Marpegan L, Bekin-schtein TA (2000) From light to genes: Moving thehands of the circadian dock. Frontiers in Bioscience8:S285-S293.

Page 451: 0195162854

434 COGNITION

Gothard KM, Skaggs WE, Moore KM, McNaughton BL(1996) Binding of hippocampal CAl neural activityto multiple reference frames in a landmark-basednavigation task. Journal of Neuroscience 16:823-835.

Griffiths DP and Clayton NS (2001) Testing episodicmemory in animals. Physiology and Behavior 73:755-762.

Hanlon FM, Weisend MP, Huang MX, Astur RS, MosesSN, Lee RR (2002) Neural activation during per-formance of transverse patterning using magneto-encephalography. Journal of Cognitive Neuro-science Suppl 8:163-163.

Hatfield T, Han J-S, Conley M, Gallagher M, Holland P(1996) Neurotoxic lesion of the basolateral, but notcentral, amygdala interfere with Pavlovian second-order conditioning and reinforcer-devaluation ef-fects. Journal of Neuroscience 16:5256-5265.

Holland PC (1992) Occasion setting in Pavlovian condi-tioning. In: Medin D, editor. The psychology oflearning and motivation. Vol. 28. San Diego: Aca-demic Press; p. 69-125.

Holland PC and Straub JJ (1979) Differential effects oftwo ways of devaluing the unconditioned stimulusafter Pavlovian appetitive conditioning. Journal ofExperimental Psychology Animal Behavior Processes5:65-78.

Honey RC and Watt A (1999) Acquired relational equiv-alence between contexts and features. Journal of Ex-perimental Psychology Animal Behavior Processes25:324-333.

Lowrey PL and Takahashi JS (2000) Genetics of themammalian circadian system: Photic entrainment,circadian pacemaker mechanisms, and posttransla-tional regulation. Annual Review of Genetics 34:533-562.

McDonald RJ and White NM (1995) Information ac-quired by the hippocampus interferes with acquisi-tion of amygdala-based conditioned cue preference(CCP) in the rat. Hippocampus 5:189-197.

McGaughy J and Sarter M (1998) Sustained attentionperformance in rats with intracortical infusions of192 IgG-saporin-induced cortical cholinergic deaf-ferentation: effects of physostigmine and FG 7142.Behavioral Neuroscience 112:1519-1525.

Miller RR and Oberling P (1998) Analogies between oc-casion setting and pavlovian conditioning. In:Schmajuk NA, Holland PC, editors. Occasion set-ting: Associative learning and cognition in animals.Washington, DC: American Psychological Associa-tion; p. 3-35.

Muir JL, Everitt BJ, Robbins TW (1996) The cerebralcortex of the rat and visual attentional function: dis-sociable effects of mediofrontal, cingulate.anterior,dorsolateral and parietal cortex lesions on a five

choice serial reaction time task. Cerebral Cortex6:470-481.

Passetti F, Chudasama Y, Robbins TW (2002) Thefrontal cortex of the rat and visual attentional per-formance: Dissociable functions of distinct medialprefrontal subregions. Cerebral Cortex 12:1254-1268.

Phillips JM, McAlonan K, Robb WGK, Brown VJ (2000)Cholinergic neurotransmission influences covertorientation of visuospatial attention in the rat. Psy-chopharmacology 150:112-116.

Posner MI (1980) Orienting of attention. Quarterly Jour-nal of Experimental Psychology 32:3-25.

Reed JM and Squire LR (1999) Impaired transverse pat-terning in human amnesia is a special case of im-paired memory for two-choice discrimination tasks.Behavioral Neuroscience 113:3-9.

Rickard TC and Grafman J (1998) Losing their config-ural mind: Amnesic patients fail on transverse pat-terning. Journal of Cognitive Neuroscience, 10:509-524.

Rudy JW and O'Reilly RC (1999) Contextual fear con-ditioning, conjunctive representations, pattern com-pletion, and the hippocampus. Behavioral Neuro-science 113:67-880.

Sarter M and Bruno JP (1997) Cognitive functions of cor-tical acetylcholine: toward a unifying hypothesis.Brain Research Reviews 23:28^46.

Sarter M, Givens B, Bruno JP (2001) The cognitive neu-roscience of sustained attention: where top-downmeets bottom-up. Brain Research Reviews 35:146-160.

Skinner DM, Etchegary CM, Ekert-Maret EC, Baker CJ,Harley CW, Evans JH, Martin GM (2003) An analy-sis of response, direction, and place learning in anopen field and T maze. Journal of Experimental Psy-chology Animal Behavior Processes 29:3-13.

Stewart C, Burke S, Marrocco R (2001) Cholinergic mod-ulation of covert attention in the rat. Psychophar-macology 155:210-218.

Sutherland RJ and Rudy JW (1989) Configural associa-tion theory: The role of the hippocampal formationin learning, memory, and amnesia. Psychobiology17:129-144.

Swartzentruber D (1995) Modulatory mechanisms inPavlovian conditioning. Animal Learning and Be-havior 23:123-143.

Tulving E (1972) Episodic and semantic memory. In:Tulving E, Donaldson W, editors, Organization ofmemory New York: Academic Press, p. 381-403.

Van Elzakker M, O'Reilly RC, Rudy JW (2003) Transi-tivity, flexibility, conjunctive representations andthe hippocampus. I. An empirical analysis. Hip-pocampus 13:292-298.

White NM and McDonald RJ (1993) Acquisition of a spa-

Page 452: 0195162854

Chapter 40. Cognitive Processes 435

tial conditioned place preference is impaired by Zinyuk L, Kubik S, Kaminsky Y, Fenton AA, Bures Jamygdala lesions and improved by fornix lesions. (2000) Understanding hippocampal activity by usingBehavioural Brain Research 55:269-281. purposeful behavior: Place navigation induces place

White NM and McDonald RJ (2002) Multiple parallel cell discharge in both task-relevant and task-irrele-memory systems in the brain of the rat. Neurobi- vant spatial reference frames. Proceedings of theology of Learning and Memory 77:125-184. National Academy of Sciences USA 97:3771-3776.

Page 453: 0195162854

Incentive Behavior

BERNARD W. BALLEINE41

An incentive is any source of motivation that isacquired through learning; that is not innate.Incentive behavior is therefore any responsethat is demonstrably controlled by an incen-tive learning process of one kind or another.For the purposes of this chapter, the incentivebehavior of rats is discussed in the context ofevaluative, Pavlovian, and instrumental con-ditioning procedures productive of what werefer to here as evaluative, Pavlovian, andinstrumental incentives, respectively (seeBalleine, 2001; Dickinson and Balleine, 2002;Dayan and Balleine, 2002, for recent reviews).As described in this chapter, these incentiveprocesses constitute a hierarchy: instrumentalincentives involve in part processes engagedby Pavlovian incentives that in part involveprocesses engaged by evaluative incentives.Whether these incentive processes can in factbe fully dissociated structurally is still a matterof debate, and some current issues are dis-cussed in the final section.

EVALUATIVE INCENTIVES

Since Pavlov (1927), it has become common-place for students of learning to divide per-ceptual events or stimuli into those that areconditioned and those that are unconditioned.However, for Pavlov, this distinction wasmade solely on the basis of the behavioral re-sponse that they evoke on first presentation.Although a physiological disturbance mayhave an immediate and unlearned motivatingeffect (i.e., elicit an unconditioned response),

the motivating effect of the perceptual fea-tures of the event productive of that distur-bance (e.g., the taste, smell, and visual, audi-tory, or textural features) is established onlyonce the animal has had some experience withthese events. Thus, for Pavlov, "the effect ofthe sight and smell of food is not due to an in-born reflex, but to a reflex which has been ac-quired in the course of the animal's own in-dividual existence" (1927; p. 23)—a process hereferred to as "signalization."

In contrast, contemporary theories oflearning acknowledge Pavlov's basic divisionbetween stimuli but obscure its basis by de-scribing events like food, water, or shock as"unconditioned stimuli" (USs). In this way,these theories tend to conflate what is un-conditioned about USs, the reflexive re-sponses (URs) they elicit on first contact, withwhat is not; that is, the association betweentheir sensory-perceptual features and, physio-logically based, motivational systems acti-vated by the detection of such things as calo-ries, fluids, pain, and so on (c.f, Chapters18-23). Though quite subtle, this distinctionbetween the representation of an event andthe response that it elicits is important. It iden-tifies a fundamental form of learning that haslargely been overlooked in contemporary ac-counts of learning and motivation in the ratand referred to here as evaluative incentivelearning (Fig. 41-1). (De Houwer et al., 2001.)

References to evaluative incentives inrats have, occasionally, surfaced in the past,although cloaked in quite different terms,such as in the work of P. T. Young in 1949

436

Page 454: 0195162854

Chapter 41. Incentive Behavior 437

Figure 41-1. A model of the representational and motiva-tional processes mediating evaluative incentive learning. Theacquisition of biological significance by the sensory-percep-tual features of events (i.e., olfactory [Ol], taste [Ta], so-matosensory [So], auditory [Au], and visual [Vi] features) isshown here to involve the formation of an acquired connec-tion (solid arrows) with physiologically based, motivationalsystems (M) modulated by visceral and humoral signals orig-inating in peripheral regulatory systems (D, solid circle).

and in analyses of research in the 1940s and1950s on what came to be called externalizedor acquired drive (Bolles, 1975). Moll (1964),for example, reports evidence consistent withPavlov's "signalization" process in young rats.On their first experience with food depriva-tion, Moll's rats ate substantially less than wasrequired to make up their deficit or even tomaintain them, although they rapidly learnedto increase consumption over time and overpresentations of the food. Similarly, Changizi,McGehee, and Hall (2002) observed that ratpups did not exhibit food-seeking behaviorwhen food deprived unless they had previousexperience with food deprivation and eating.Perhaps more surprisingly, Hall and colleaguesalso reported evidence that the same is true ofwater for thirsty rats (Hall et al., 2000; Changiziet al., 2002). Thus, for example, in prewean-ling rats or rats weaned onto a fluid diet, theinduction of a strong, extracellular thirst wasobserved to have no immediate effect on wa-ter consumption relative to rats not madethirsty. After experience with water in thethirsty state had been allowed, however, sub-

sequent induction of thirst produced an im-mediate increase in water consumption. Therepresentation of events such as food or wateras biologically significant for hungry or thirstrats appears, therefore, to be acquired.

The procedures used to assess the acqui-sition of evaluative incentives have obvioussimilarities to those used to generate condi-tioned taste preferences and aversions. Withregard to the former, rats are generally firstdeprived of some essential commodity orother (e.g., nutrients, fluids, or, more specifi-cally, sodium or calcium), after which a stim-ulus (usually a taste) is paired with the deliv-ery of the deprived commodity, presentedeither in solution with the taste or directly viaintragastric, intraduodenal, hepatic portal, orintravenous routes. Evidence for evaluativeincentive learning is established if, relative torats given the taste and the infusion of the de-prived commodity unpaired, the paired groupsignificantly increases their willingness to con-tact and consume the taste (Sclafani, 1999). Ithas also been reported that treatments such asthese increase the tendency of rats to show in-gestive, orofacial fixed action patterns (FAPs)when the paired taste is contacted (Forestelland Lolordo, 2003). Deprivation of one orother commodity appears to be necessary togenerate conditioned taste preferences of thiskind (Harris et al., 2000), suggesting that eval-uative conditioning is modulated by visceraland humoral signals originating in regulatoryprocesses such as those that control feeding,drinking, and so on (Fig. 41-1) (Sudakov, 1990).Indeed, studies that have specifically manipu-lated deprivation state report, for example,that the acquisition of taste preferences by nu-trient loads is strongly controlled by the de-gree of food deprivation (Harris et al., 2000).Studies of orofacial FAPs confirm that thesereactions to taste stimuli are also modulatedby motivational state. The taste reactivity pat-terns elicited by sugar solutions are aug-mented by hunger (e.g., Berridge, 1991) andthose elicited by saline are enhanced by asodium appetite (e.g., Berridge et al., 1984).

Page 455: 0195162854

438 COGNITION

Garcia (1989; Garcia et al., 1989) has ar-gued that conditioned taste aversions also arebest viewed as an example of evaluative in-centive learning (what he called "Darwinianconditioning"). This procedure involves thepairing of a (usually sweet) taste with the in-jection of an emetic agent, such as lithiumchloride. Subsequently, both orofacial FAPsshift from acceptance to those associated withrejection (Berridge, 2000), and the consump-tion of substances that contain that taste isstrongly and enduringly altered by this pair-ing. Garcia et al. (1989) argue that this effectreflects the formation of an association be-tween taste afferents and brain stem auto-nomic centers that subsequent work hasidentified as the parabrachial nucleus for con-ditioned aversions (Reilly, 1999) as well as,interestingly enough, for conditioned pref-erences (Sclafani et al., 2001). The site of in-tegration appears to differ for evaluative in-centive learning involving olfactory, visual,auditory, and somatosensory features andlikely involves the amygdala (Holland et al.,2002) along with its afferents in sensory cor-tex, brain stem, and hypothalamic nuclei.

PAVLOVIAN INCENTIVES

In contrast to evaluative conditioning, ac-counts of Pavlovian conditioning usually em-phasize the formation of associations betweenstimulus representations rather than betweenstimuli and the activity of intrinsic motiva-tional systems (Pearce and Bouton, 2001). Itis quite possible, in fact, for Pavlovian asso-ciative processes to operate without a pro-grammed motivational manipulation (e.g.,in sensory preconditioning). Usually, how-ever, evaluative incentives are heavily used inPavlovian conditioning in the pairing of stim-uli, usually called conditioned stimuli (CSs),that are relatively neutral with respect to aparticular motivational state, with events thatboth elicit biologically potent responses (URs)

and are represented as biologically significantevents (USs)—that have been established asevaluative incentives. It should come as nosurprise, therefore, that numerous authorshave suggested that Pavlovian CSs can acquireincentive properties (see Dickinson andBalleine, 2002, for a review).

One of the most sophisticated accountsof Pavlovian incentive learning is that devel-oped by Konorski (1967). In Konorski'saccount, Pavlovian conditioning comes intwo forms: consummatory and preparatory.Consummatory conditioning occurs whenthe form of the conditioned response reflectsthe specific sensory properties of the US, suchas when a signal, or CS, predicting a food USelicits salivation, licking, or chewing (DeBoldet al., 1965) or when predicting a shock tothe eye elicits a blink response (Schmajukand Christiansen, 1990). Hence, Konorski as-sumed that consummatory conditioning re-flects the formation of an association betweenthe representation of the CS and the sensoryand perceptual features of the US representa-tion and that it is CS-related activation of theUS representation via this connection thatelicits the consummatory CR.

In contrast, preparatory conditioning re-flects the acquisition of responses characteris-tic of the affective class to which that US be-longs rather than its specific properties. Theseresponses are quite general; for example, CSspaired with appetitive USs (e.g., food, water,etc.) often come to elicit approach, whereasthose paired with aversive USs (e.g., footshock or eye shock) elicit withdrawal.Konorski proposed that preparatory CRs areelicited by activation of affective processes bythe CS that may take either of two routes—via the representation of the sensory featuresof the US or through a direct association,thereby producing purely preparatory condi-tioning. Hence, preparatory and consumma-tory conditioning can be dissociated. Ginn etal. (1983) established that although a 0.5 sec-ond tone simultaneously presented with leg

Page 456: 0195162854

Chapter 41. Incentive Behavior 439

shock elicited a leg flexion CR, a longer, 4 sec-ond, tone did not, although both CSs eliciteda heart rate CR. A version of the associativestructure thought to underlie the acquisitionof Pavlovian incentives on this account is il-lustrated in Figure 41-2 (see Dickinson andDealing, 1978; Dickinson and Balleine, 2002).

The affective processes engaged by pre-paratory conditioning can facilitate the per-formance of consummately CRs based on theUS as well as the performance of orienting re-sponses (ORs) to the CS itself. Bombace et al.(1991) established a relatively long auditoryCS as a signal for a shock to the rear leg anda second, short visual stimulus as a CS for eyeshock. They found that the visual CS elicited

Figure 41-2. The Konorskian model of the representationaland motivational processes mediating Pavlovian incentivelearning (based on Dickinson and Dealing, 1979; Dickinsonand Balleine, 2002). Here the model is elaborated for ap-petitive conditioning involving sugar and aversive condi-tioning involving shock. Evaluative processes are includedin the dashed rectangle. Solid arrows represent acquiredconnections; open arrows, fixed connections; AP: appeti-tive; AV: aversive; ex: excitatory; inh: inhibitory; Nu: nu-tritive system; Fe: fear system.

a conditioned eye-blink response of greateramplitude when presented during the audi-tory stimulus. This rinding suggests that thesensory representation of the US both acti-vates and can be activated by the general af-fective state conditioned to the CS. Further-more, ORs to CSs (such as head jerking totones or rearing to lights) usually habituatequite quickly but when the CS is pairedwith a US, their incidence increases (Holland,1980), suggesting that the motivational sup-port for sensory-specific responses that is pro-vided by affective states conditioned to the CSmay be quite general.

In Figure 41-2, different USs from thesame affective class are proposed to activate acommon affective system. Further evidencefor this claim can be drawn from transrein-forcer blocking studies. Blocking refers to theobservation that pretraining one CS often re-duces conditioning to a second CS when acompound of these two stimuli is paired withthe US (Kamin, 1969); the pretrained CS is saidto block conditioning to the added CS. Al-though the standard blocking procedure usesthe same US during pretraining and compoundtraining, Bakal et al. (1974) observed that a CSpretrained with a footshock US blocked con-ditioning to an added CS when the compoundwas paired with loud auditory US, even thoughthe sensory properties of these USs differ sub-stantially. What they have in common, how-ever, is that they are aversive, and, as such,transreinforcer blocking is usually taken as ev-idence that the two USs activate a common af-fective process. Transreinforcer blocking alsoprovides the best evidence for a common ap-petitive affective process. Ganesen and Pearce(1988) pretrained a CS with a water US beforegiving compound training with the added CSwith a food US (or vice versa). Conditioned ap-proach to the site of food delivery during theadded CS was attenuated by pretraining withwater, indicating that pretraining with the wa-ter US blocked conditioning to the added CSwhen paired with food.

Page 457: 0195162854

440 COGNITION

APPETITIVE AND AVERSIVE INTERACTIONS

Although transreinforcer facilitation andblocking establish motivational commonali-ties, it is equally clear that there is a distinc-tion, at least between appetitive and aversiveprocesses. There is a wealth of evidence thatCSs of one affective class inhibit responsescontrolled by CSs of the other affective class(Dickinson and Pearce, 1977). This inhibitoryinterrelationship is most clearly illustrated bycounterconditioning experiments in which apreviously established predictor of an aversiveUS is subsequently paired with an appetitiveUS or vice versa. Generally, rather than en-hancing the performance of the previouslyconditioned CR, this treatment strongly at-tenuates it. This evidence has been reviewedextensively (Dickinson and Pearce, 1977;Dickinson and Dearing, 1979; Dickinson andBalleine, 2002) and favors the view that ap-petitive and aversive affective systems mutu-ally inhibit one another.

In addition to counterconditioning ef-fects, this opponent relationship also offersa straight-forward account of the propertiesof conditioned inhibitors. A conditioned in-hibitor is a stimulus that acquires the ca-pacity to inhibit the CR elicited by an exci-tatory CS as a result of being paired with theomission of an otherwise predicted US. Ithas long been known that an inhibitory CSof one affective class has properties in com-mon with an excitatory CS of the oppositeaffective class (Fig. 41-2). Thus, a CS pairedwith the omission of expected food is aver-sive; rats will learned to escape from it(Daly, 1974). Moreover, using a transrein-forcer blocking assay, conditioned excitorsand inhibitors of opposite affective classeshave been found to engage a common in-centive process; for example, a CS that pre-dicts the omission of a food US has been re-ported to block aversive conditioning witha shock US in rats (see Dickinson and Dear-ing, 1979; Dickinson and Balleine, 2002, forreviews).

MOTIVATIONAL CONTROL OFPAVLOVIAN CONDITIONING

Ramachandran and Pearce (1987) observedthat the asymptotic level of magazine ap-proach elicited by a CS paired with either foodor water was reduced by the presence of theirrelevant motivational state—thirst in thecase of the food reinforcer and hunger in thecase of the water reinforcer. The suppressionproduced by the irrelevant deprivation statewas motivational in origin because removalof the irrelevant state during an extinction testrestored performance to the level observed inrats trained under the relevant state alone. Ra-machandran and Pearce (1987) argue that thisinteraction between hunger and thirst doesnot occur at the level of the appetitive mo-tivational systems within the Konorskianmodel but reflects the mechanism by whichprimary motivational states modulate the ac-tivation of specific US representations. This isan important point that raises a critical dis-tinction between the interaction of motiva-tional systems and the interaction betweenthe mechanisms by which primary motiva-tional states, such as hunger and thirst, mod-ulate the capacity of relevant stimuli to acti-vate these systems.

That some form of motivational modu-lation of this kind is required is well recog-nized; performance of the CR in appetitivePavlovian conditioning has been found to de-pend directly on deprivation state (DeBold etal., 1965). Equally clearly, this modulation has,to some extent, to be US specific; it is hydra-tion that modulates fluid representations andnutritional need that modulates food repre-sentations. The implication of the Ramachan-dran and Pearce (1987) results for the model,therefore, is not only that hunger enhancesthe activation of the appetitive motivationalsystem by a CS paired with a nutritional USbut also that thirst counteracts this enhance-ment thereby reducing facilitation of the USrepresentation. Generally, therefore, and assuggested by Figure 41-2, motivational mod-

Page 458: 0195162854

Chapter 41. Incentive Behavior 441

ulation of evaluative incentive processesserves to gate the ability of CSs to activate theaffective system through the sensory repre-sentation of the US.

It is clear, however, that there should becases where shifts in primary motivation donot affect conditioned responding. On themodel presented in Figure 41-2, these casesshould be those in which the CR is entirelypreparatory, mediated by a direct connectionwith the affective system. A notable case ofthis kind is second-order conditioning in hun-gry rats, which is resistant to extinction of thefirst-order CS (Rizley and Rescorla, 1972) andis unaffected by devaluation of the US eitherby rotation or by satiation treatments (Hol-land and Rescorla, 1975).

INSTRUMENTAL INCENTIVES

Current evidence suggests that, in instru-mental conditioning, rats encode the rela-tionship between an action and its conse-quence or outcome and are extremelysensitive to the contingency between the per-formance of an action and the probability ofoutcome delivery (Balleine, 2001). Neverthe-less, it has long been recognized that refer-ence to the action-outcome association aloneis not sufficient to determine the performanceof an action; any learning that takes the form"action R leads to outcome O" can be usedboth to perform R and to avoid performingR. Of course, what is missing from this ac-count is mention of the role that the incen-tive value of the outcome plays in controllinginstrumental performance. It is now well es-tablished that the rats' experience of the af-fective and motivationally relevant propertiesof an outcome strongly controls the per-formance of actions instrumental to outcomedelivery (Dickinson and Balleine, 1994;Balleine, 2001, for reviews). Evidence for thisclaim has mainly come from studies assessingthe impact of shifts in primary motivation oninstrumental performance.

One of the most striking properties of in-strumental actions is that, in marked contrastto the Pavlovian CR, their performance is notdirectly sensitive to shifts in primary motiva-tion. This was first observed when rats trainedhungry to lever press and chain pull, with oneaction earning food pellets and the other aliquid sucrose solution, were subsequentlyshifted to water deprivation (Dickinson andDawson, 1989). In this situation, the rats didnot alter their performance of either actionand, in fact, only increased their performanceof the response trained with the liquid sucroseif they were first allowed to drink the sucrosewhen thirsty. The shift to water deprivationhad no direct impact on performance. In sub-sequent studies, the same pattern of resultshas been found after a number of other post-training shifts in motivation. For example, ratstrained to lever press for food when food de-prived do not immediately reduce their per-formance on the lever when they are suddenlyshifted to an undeprived state (Balleine, 1992).Nor do they increase their performance im-mediately if they are trained undeprived andthen given a test on the levers when food de-prived (Balleine, 1992; Balleine et al., 1994). Inboth of these situations, rats modify their in-strumental performance only after they havebeen allowed the opportunity to consume thespecific food outcome in the test motivationalstate.

The need for consummatory contactwith the instrumental outcome for a shift inprimary motivation to affect instrumental per-formance has been found to be quite generaland has been confirmed for a number of dif-ferent motivational states and using a numberof devaluation paradigms. For example, in ad-dition to shifts from hunger to thirst and be-tween hunger and satiety, incentive learninghas been found to play a roll in (1) taste aver-sion-induced devaluation effects (Balleine andDickinson, 1991, 1992), (2) specific satiety-induced devaluation (Balleine and Dickinson,1998a), (3) posttraining shifts in water depri-vation (Lopez et al., 1992), (4) changes in out-

Page 459: 0195162854

442 COGNITION

come value mediated by drug states (Balleineet al., 1994), and (J) changes in the value ofthermoregulatory (Hendersen and Graham,1979) and sexual (Everitt and Stacey, 1987) re-wards (see Balleine, 2001, for a review). In allof these cases, it is clear that, after a shift inprimary motivational state, rats have to learnabout the effect of this shift on the incentivevalue of an instrumental outcome throughconsummatory contact before the shift willact to affect instrumental performance. Thisform of learning is referred to as instrumentalincentive learning.

It is interesting to consider why Pavlov-ian CRs and goal-directed instrumental actionsshould differ in their sensitivity to the effectsof shifts in primary motivation. In a recent re-view of the literature, Balleine (2001) arguedthat the primary distinction between Pavlov-ian and instrumental conditioning may lie infact that, in instrumental conditioning, the rep-resentation of the outcome associated with anaction is not directly connected with the mo-tivational/affective structures typically di-rectly activated by the CS. Rather it is only in-directly related to these structures via aconnection with the emotional feedback in-duced by their activation. This account isbased on an elemental model of outcomerepresentation (Fig. 41-3), supposing that themost salient sensory features are directlyconnected with the motivational/affectiveprocesses whereas less salient features are not.It need only be assumed, on the basis of dif-ferential overshadowing, that CSs proximal tooutcome delivery become associated with themore salient features, whereas actions (beingmore distal to outcome delivery) are associ-ated with other, more diffuse, features. As aconsequence, the performance of actions is notaffected by shifts in motivational state until thediffuse elements of the outcome with whichthey are associated are revalued through con-summatory contact with the outcome in theprevailing motivational state (Balleine, 2001).

On this account, therefore, instrumentalincentive learning is mediated by an association

Figure 41-3. A model of instrumental incentive learning.Here the reward, or instrumental outcome, associated withan action such as lever pressing (i.e., LP) is illustrated ascomposed of several features (Si, S2, S3), the most salientof which is used in an evaluative connection with M thatgenerates emotional feedback (EM) via fixed connectionswith the affective systems (Af). Incentive learning reflectsthe formation of an association between the varioussensory-perceptual features of the outcome and the emo-tional response.

between the sensory features of the instru-mental outcome and emotional feedbackdriven by the same motivational/affectiveprocesses that are engaged by evaluative andPavlovian incentives. Instrumental incentivelearning, although dependent on these pro-cesses, involves a distinct associative connectionwith this emotional response. By establishingthe relative hedonic impact of the instrumentaloutcome, instrumental incentive learning al-lows rats (and, of course, other animals) to en-code the goal, or incentive, value of the conse-quences of specific actions and so plays a criticalrole in action selection (Balleine and Dickinson,1998b; Dickinson and Balleine, 2000).

CONCLUDING COMMENTS

The evaluative, Pavlovian, and instrumentalincentive learning processes around whichthis brief review of incentive behavior hasbeen organized can be readily distinguisheddescriptively and perhaps procedurally. Theycan also be distinguished functionally within

Page 460: 0195162854

Chapter 41. Incentive Behavior 443

a general analysis of the determinants of adap-tive behavior as anchoring exploratory be-havior, predictive learning, and behavioralchoice, respectively, although this issue liesmore in the province of evolutionary biologyand is beyond the scope of this chapter (seeBalleine and Dickinson, 1998b; Dickinson andBalleine, 2000, for further discussion). It re-mains to be established, however, whetherthese learning processes are mediated by dis-tinct mechanisms, that is, whether their con-trol over behavior can be dissociated one fromthe other. Unfortunately, to date, there hasbeen relatively little systematic research ofthis issue, and what there has been has not al-ways been focused on rats.

For example, arguably the best evidencethat evaluative and Pavlovian incentive learn-ing are mediated by distinct processes hascome from studies of human learning whereit has been reported that changes in the in-centive value of a perceptual stimulus (like facesor flavors) can be produced by pairing themwith noxious consequences in situations wheresubjects appear to be unaware of the contin-gency between these events (De Houwer et al.,2001). For obvious reasons, experiments ofthis kind are difficult to conduct in rats. Nev-ertheless, some dissociations have been re-ported that are suggestive of a similar effect.For example, lesions of the basolateral amyg-dala have been reported to have no effect oneither sensory preconditioning or first-orderPavlovian conditioning using food (Holland etal., 2001; Blundell et al., 2003) or, in a condi-tioned suppression paradigm, shock USs (Kill-cross et al., 1997). These lesions do, however,strongly affect conditioned flavor aversions,aspects of fear conditioning (notably freezingbehavior), and conditioned eating responses(Holland et al., 2001). In contrast, lesions ofdorsal hippocampus can influence the associ-ation between stimuli in sensory precondi-tioning (Talk et al., 2002) but do not affectevaluative incentives produced either by con-ditioned taste preference or taste aversion pro-cedures (Reilly et al., 1993).

Furthermore, Wyvell and Berridge (2000)reported that microinjection of a dopamine ag-onist into the nucleus accumbens of rats en-hanced appetitive Pavlovian conditioning with-out affecting ingestion and rejection-relatedorofacial FAPs elicited by the intraoral infu-sion of a bitter-sweet sucrose-quinine solution.Conversely, dopamine antagonists attentuatePavlovian conditioning for appetitive rewards(Berridge and Robinson, 1998) but do not af-fect ingestive orofacial FAPs (Pecina et al.,1997). Rather than affecting evaluative in-centive learning, therefore, dopamine ago-nists and antagonists appear to act on thePavlovian form of incentive learning. Indeed,Berridge and Robinson (1998; Robinson andBerridge, 1993) have consistently argued thatthe dopamine system mediates the motiva-tional properties of CSs, something they referto as its "incentive salience."

With regard to Pavlovian and instru-mental incentive learning, numerous studiessuggest that Pavlovian incentives can exertquite selective excitatory effects on instru-mental performance. Thus, for example, thereis good evidence that superimposing a Pavlov-ian excitor on an instrumental baseline can af-fect the performance of the instrumental re-sponse (Balleine, 1994), an effect referred toas Pavlovian-instrumental transfer. More im-pressive still is evidence of selective transfereffects; CSs paired with the outcome earnedby an instrumental action have a greater ex-citatory effect on performance of that actionthan CSs paired with a different outcome(Colwill and Rescorla, 1988). Thus, and incontrast to the differential effects of shifts inprimary motivation on Pavlovian and instru-mental performance, this evidence appears toindicate that Pavlovian and instrumental con-ditioning share a common incentive process.

There is, however, good evidence againstthis claim. For example, a number of treat-ments have been found to affect transfer with-out affecting instrumental devaluation effectsand vice versa. First, peripheral administrationof the dopamine antagonists pimozide or

Page 461: 0195162854

444 COGNITION

a-flupenthixol has been observed to attenuatethe excitatory effects of a Pavlovian CS on in-strumental performance without affecting in-strumental outcome devaluation produced bya shift from a food deprived to a nondeprivedstate (Dickinson et al, 2000).

Second, Corbit and Balleine (2001) reportthat cell body lesions of the shell subregion ofthe nucleus accumbens profoundly attenuateselective transfer effects produced when a CSis paired with the same outcome as thatearned by the instrumental action but have noeffect on the sensitivity of rats to selective de-valuation of the instrumental outcome in-duced by a specific satiety treatment. Con-versely, lesions of the core subregion ofthe accumbens had no effect on the selectivetransfer effect abolished by the shell lesionsbut had a profound effect on the sensitivity ofrats to the selective devaluation of the in-strumental outcome. This study presents,then, a double dissociation between the neu-ral processes that mediate Pavlovian and in-strumental incentive learning processes.

Corbit and Balleine (2003) also reportthat Pavlovian CSs and instrumental incentivelearning have dissociable effects on the per-formance of components of a heterogeneouschain of instrumental actions. Hungry, ratswere trained to press first one lever (Rl)and then a second lever (R2) to earn a foodoutcome (Rl —» R2 —> O). They were thengiven Pavlovian training in which a CS waspaired with that same outcome. In a test ofPavlovian-instrumental transfer, the CS wasfound to elevate performance but only of theresponse proximal to food delivery (R2). Therats were then retrained on the chain afterwhich they were shifted to an undeprivedstate and given an extinction test on the twoactions. Half of the rats were allowed to eatthe food outcome when sated before this test,whereas the remainder were not. On the test,performance on the chain was reduced but,importantly, only in the rats reexposed to theoutcome when undeprived and then only onthe distal response in the chain (Rl). Perfor-

mance on R2 was not differentially affectedby the instrumental incentive learning treat-ment.

The unavoidable conclusion from thesestudies is that Pavlovian and instrumental in-centive processes not only are mediated byanatomically and neurochemically distinctsystems but also are functionally independent.It is tempting to conclude that the same is trueof evaluative incentive processes, although atpresent, and particularly in rats, there is toolittle in the way of systematic data on this is-sue to do so with any confidence.

ACKNOWLEGMENT

The preparation of this chapter was supported by a grant fromthe National Institute of Mental Health #MH56446 to theauthor.

REFERENCES

Bakal CW, Johnson RD, Rescorla RA (1974) The effectof change in US quality on the blocking effect.Pavlovian Journal of Biological Sciences 9:97-103.

Balleine B (1992) Instrumental performance following ashift in primary motivation depends upon incentivelearning. Journal of Experimental Psychology: Ani-mal Behavior Processes 18:236-250.

Balleine B (1994) Asymmetrical interactions betweenthirst and hunger in Pavlovian-instrumental trans-fer. Quarterly Journal of Experimental Psychology47B:211-231.

Balleine BW (2001) Incentive processes in instrumentalconditioning. In: Handbook of contemporary learn-ing theories (Mowrer R and Klein S, eds.), pp.307-366. Hillsdale, NJ: Erlbaum.

Balleine B, Ball J, Dickinson A (1994) Benzodiazepine-induced outcome revaluation and the motivationalcontrol of instrumental action in rats. BehavioralNeuroscience 108:573-589.

Balleine B and Dickinson A (1991) Instrumental per-formance following reinforcer devaluation dependsupon incentive learning. Quarterly Journal of Ex-perimental Psychology 436:279-296.

Balleine B and Dickinson A (1992) Signalling and incen-tive processes in instrumental reinforcer devalua-tion. Quarterly Journal of Experimental Psychology456:285-301.

Balleine B and Dickinson A (1998a) The role of incen-tive learning in instrumental outcome revaluation

Page 462: 0195162854

Chapter 41. Incentive Behavior 445

by sensory-specific satiety. Animal Learning andBehavior 26:46-59.

Balleine BW and Dickinson A (I998b) Consciousness:the interface between affect and cognition. In: Con-sciousness and human identity (Cornwell J, ed.), pp.57-85. Oxford: Oxford University Press.

Berridge, KC (1991) Modulation of taste affect byhunger, caloric satiety, and sensory-specific satietyin the rat. Appetite 16:103-120.

Berridge KC (2000) Reward learning: Reinforcement, in-centives, and expectations. In: The psychology oflearning and motivation (Medin DL, ed.), Vol. 40,pp. 223-278. New York: Academic Press.

Berridge KC, Flynn FW, Schulkin J, Grill HJ (1984)Sodium depletion enhances salt palatability in rats.Behavioral Neuroscience 98:652-660.

Berridge KC and Robinson TE (1998) What is the roleof dopamine in reward: hedonic impact, rewardlearning, or incentive salience? Brain Research Re-views 28:309-369.

Blundell P, Hall G, Killcross S (2003) Preserved sensi-tivity to outcome value after lesions of the basolat-eral amygdala. Journal of Neuroscience 23:7702-7709.

Bolles RC (1975) Theory of motivation. New York:Harper & Row.

Bombace JC, Brandon SE, Wagner AR (1991) Modula-tion of a conditioned eyeblink response by a puta-tive emotive stimulus conditioned with a hindlegshock. Journal of Experimental Psychology: AnimalBehavior Processes 17:323-333.

Changizi MA, McGehee RM, Hall WG (2002) Evidencethat appetitive responses for dehydration and food-deprivation are learned. Physiology and Behavior75:295-304.

Colwill RM and Rescorla RA (1988) Associations be-tween the discriminative stimulus and the rein-forcer in instrumental learning. Journal of Experi-mental Psychology: Animal Behavior Processes14:155-164.

Corbit L, Muir J, Balleine BW (2001) The role of the nu-cleus accumbens in instrumental conditioning: evi-dence for a functional dissociation between ac-cumbens core and shell. Journal of Neuroscience21:3251-3260.

Corbit LH and Balleine BW (2003) Pavlovian and in-strumental incentive processes have dissociable ef-fects on components of a heterogeneous instru-mental chain. Journal of Experimental Psychology:Animal Behavior Processes 29:99-106.

Daly HB (1974) Reinforcing properties of escape fromfrustration aroused in various learning situations.In: The psychology of learning and motivation(Bower GH, ed.), Vol. 8, pp. 187-231. New York:Academic Press.

Dayan P and Balleine BW (2002) Reward, motivationand reinforcement learning. Neuron 36:285-298.

DeBold RC, Miller NE, Jensen DD (1965) Effect ofstrength of drive determined by a new techniquefor appetitive classical conditioning of rats. Journalof Comparative and Physiological Psychology 59:102-108.

De Houwer J, Thomas S, Baeyens F (2001) Associationlearning of likes and dislikes: A review of 25 yearsof research on human evaluative conditioning. Psy-chological Bulletin 127:853-869.

Dickinson A and Balleine BW (2002) The role of learn-ing in the operation of motivational systems. In:Learning, motivation & emotion, Volume 3 ofSteven's handbook of experimental psychology, 3rded. (Gallistel CR, ed.), pp. 497-533. New York JohnWiley & Sons.

Dickinson A and Balleine BW (2000) Causal cognitionand goal-directed action. In: The evolution of cog-nition (Heyes C and Huber L, eds.), pp. 185-204.Cambridge, MA: MIT Press.

Dickinson A and Dawson GR (1989) Incentive learningand the motivational control of instrumental per-formance. Quarterly Journal of Experimental Psy-chology 41B:99-112.

Dickinson A and Dearing MF (1979) Appetitive-aversiveinteractions and inhibitory processes. In: Mecha-nism of learning and motivation (Dickinson A andBoakes RA, eds.), pp. 203-231. Hillsdale, NJ:Lawrence Erlbaum Associates.

Dickinson A and Pearce JM (1977) Inhibitory interac-tions between appetitive and aversive stimuli. Psy-chological Bulletin 84:690-711.

Dickinson A, Smith J, Mirenowicz J (2000) Dissociationof Pavlovian and instrumental incentive learningunder dopamine antagonists. Behavioral Neuro-science 114:468^83.

Everitt BJ and Stacey P (1987) Studies of instrumentalbehavior with sexual reinforcement in male rats(Rattus novegicus): II. Effects of preoptic area lesions,castration, and testosterone. Journal of Compara-tive Psychology 101:407-419.

Forestell CA and Lolordo VM (2003) Palatability shiftsin taste and flavour preference conditioning. Quar-terly Journal of experimental Psychology 56B:140-160.

Ganesen R and Pearce JM (1988) Effects of changing theunconditioned stimulus on appetitive blocking.Journal of Experimental Psychology: Animal Be-havior Processes 14:280-291.

Garcia J (1989) Food for Tolman: Cognition and cathexisin concert. In: Aversion, avoidance, and anxiety:Perspectives on aversively motivated behavior(Archer T and Nilsson L-G, eds.), pp. 45-85. Hills-dale, NJ: Lawrence Erlbaum Associates.

Page 463: 0195162854

446 COGNITION

Garcia J, Brett L, Rusiniak KW (1989) Limits of Dar-winian conditioning. In: Contemporary learningtheories: Instrumental conditioning theory and theimpact of biological constraints on learning (KleinSB and Mowrer RR, eds.), pp. 181-203. Hillsdale,NJ: Lawrence Erlbaum Associates.

Ginn SR, Valentine JD, Powell DA (1983) ConcomitantPavlovian conditioning of heart rate and leg flexionresponses in the rat. Pavlov Journal of Biological Sci-ence 18:154-160.

Hall WG, Arnold HM, Myers KP (2000) The acquisitionof an appetite. Psychological Science 11:101-105.

Harris JA, Gorissen MC, Bailey GK, Westbrook RF(2000) Motivational state regulates the content oflearned flavor preferences. Journal of Experimen-tal Psychology: Animal Behavior Processes26:15-30.

Hendersen RW and Graham J (1979) Avoidance of heatby rats: Effects of thermal context on the rapidity ofextinction. Learning and Motivation 10:351-363.

Holland PC (1980) CS-US interval as a determinant ofthe form of Pavlovian appetitive conditioned re-sponses. Journal of Experimental Psychology: Ani-mal Behavior Processes 6:155-174.

Holland PC and Rescorla RA (1975) The effect oftwo-ways of devaluing the unconditioned stimulusafter first- and second-order appetitive conditioning.Journal of Experimental Psychology: Animal Be-havior Processes 1:355-363.

Holland PC, Hatfield T, Gallagher M (2001) Rats withbasolateral amygdala lesions show normal increasesin conditioned stimulus processing but reduced con-ditioned potentiation of eating. Behavioral Neuro-science 115:945-950.

Holland PC, Petrovich GD, Gallagher M (2002) The ef-fects of amygdala lesions on conditioned stimulus-potentiated eating in rats. Physiology & Behavior76:117-129.

Kamin LJ (1969) Selective association and conditioning.In: Fundamental issues in associative learning(Mackintosh NJ and Honig WK, eds.), pp. 42-64.Halifax: Dalhousie University Press.

Killcross S, Robbins TW, Everitt BJ (1997) Different typesof fear-conditioned behaviour mediated by separatenuclei within amygdala. Nature 388:377-380.

Konorski J (1967) Integrative activity of the brain: Aninterdisciplinary approach. Chicago: University ofChicago Press.

Lopez M, Balleine B, Dickinson A (1992) Incentive learn-ing and the motivational control of instrumentalperformance by thirst. Animal Learning and Be-havior 20:322-328.

Moll RP (1964) Drive and maturation effects in the de-velopment of consummatory behavior. Psycholog-ical Reports 15:295-302.

Pavlov IP (1927) Conditioned reflexes. Oxford Univer-sity Press.

Pearce JM and Bouton ME (2001) Theories of associa-tive learning in animals. Annual Review of Psy-chology 52:111-139.

Pecina S, Berridge KG, Parker LA (1997) Pimozide doesnot shift palatability: separation of anhedonia fromsensorimotor effects. Pharmacology, Biochemistryand Behavior 58:801-811.

Ramachandran R and Pearce JM (1987) Pavlovian analy-sis of interactions between hunger and thirst. Jour-nal of Experimental Psychology: Animal BehaviorProcesses 13:182-192.

Reilly S (1999) The parabrachial nucleus and conditionedtaste aversion. Brain Research Bulletin 48:239-254.

Reilly S, Harley C, Revusky S (1993) Ibotenate lesionsof the hippocampus enhance latent inhibition inconditioned taste aversion and increase resistanceto extinction in conditioned taste preference. Be-havioral Neuroscience 107:996-1004.

Rizley RC and Rescorla RA (1972) Associations insecond-order conditioning and sensory precondi-tioning. Journal of Comparative and PhysiologicalPsychology 81:1-11.

Schmajuk NA and Christiansen BA (1990) Eyeblink con-ditioning in rats. Physiology & Behavior 48(5):755-758.

Sclafani A (1999) Macronutrient-conditioned flavor pref-erences. In: Neural control of macronutrient selec-tion (Bertoud H-R and Seeley RJ, eds.), pp. 93-106.Boca Raton, FL: CRC Press.

Sclafani A, Azzara AV, Touzani K, Grigson PS, NorgrenR (2001) Parabrachial nucleus lesions block taste andattenuate flavor preference and aversion condition-ing in rats. Behavioral Neuroscience 115:920-933.

Sudakov KV (1990) Oligopeptides in the organization offeeding motivation: a systemic approach. Biomed-ical Science 1:354-358.

Talk AC, Gandhi CC, Matzel LD (2002) Hippocampalfunction during behaviorally silent associative learn-ing: dissociation of memory storage and expression.Hippocampus 12:648-656.

Wyvell CL and Berridge KG (2000) Intra-accumbensamphetamine increases the conditioned incentivesalience of sucrose reward: Enhancement of reward"wanting" without enhanced liking or response rein-forcement. Journal of Neuroscience 20:8122-8130.

Young PT (1949) Food seeking drive, affective processand learning. Psychological Review 56:98-121.

Page 464: 0195162854

Models and Tests VIII

Page 465: 0195162854

This page intentionally left blank

Page 466: 0195162854

Neurological Models

BRYAN KOLB42

The junction of any region of the brain is toproduce behavior. It follows that if a brain re-gion is dysfimctioning, then behavior will be al-tered in some way. The general presumptionin neurology is that it ought to be possible torestore at least some normal functioning bypharmacological, behavioral, or surgical in-tervention. Consider Parkinson's disease asan example. Parkinson's patients have a widerange of symptoms, including two obviousones: tremor and akinesia (i.e., the absence orpoverty of movement). Although the cause ofParkinson's disease is the death of dopamin-ergic cells in the brain stem, the loss of thosecells has a cascading effect on neurologicalfunctioning so that forebrain regions such asthe basal ganglia and thalamus do not func-tion properly, which in turn produces theobserved behavioral symptoms. The simplestway to treat the disease would be to use drugs(such as L-DOPA) to increase the productionof dopamine, but to date this treatment canonly partially restore function. Other treat-ments thus are necessary, an example beingthe transplantation of embryonic dopaminer-gic cells into the affected brain. Again, al-though once believed to be a promisingtreatment, it has become clear that transplan-tation also has limitations. New treatmentstherefore must be developed.

The major problem in developing treat-ments for any neurological disorder, how-ever, is that like most new treatments in med-icine, they must be developed in nonhumansubjects first. The nervous system provides aunique problem for medical science, how-

ever, because unlike other body organs, suchas the heart or pancreas, which appear to func-tion similarly across a wide swath of animalspecies, the brain is different. The most obvi-ous difference is in the relative size of thebrain. The brain is more than twice a big rel-ative to body size than that of our closest rel-atives, the chimpanzee, and about 15 timesbigger than that of our most common labo-ratory animal, the rat. Thus, a fundamentalproblem for neurological science is the issueof whether nonhuman brains are similarenough to human brains to be useful insearching for cures to human neurological dis-orders. This problem is compounded furtherwhen we consider the issue of whether labo-ratory animals actually contract the same dis-orders as we do; further, there is the issue ofhow we would collect a sufficient number ofanimals with particular diseases to use as sub-jects in experiments.

The solution to these problems is actu-ally quite simple. First, although species suchas rats have brains that are much smaller thanhuman brains, the overwhelming evidence isthat the fundamental organization of the ratbrain is not much different from the humanbrain. (For an extensive discussion of this is-sue, see Kolb and Whishaw, 1983, 2003.) Cer-tainly rats do not have as complex a cognitivelife as humans and obviously do not talk, butthey do have sensory and motor systems thatare sufficiently similar to ours to allow us tomake generalizations across the species di-vide. Furthermore, rats have neurological sys-tems controlling cognitive, emotional, and

449

Page 467: 0195162854

450 MODELS AND TESTS

attentional processes, which are remarkablysimilar in general organization to human sys-tems. Second, it is not necessary to wait untilrats show symptoms of diseases. Rather, wecan devise ways to induce different types ofneurological disorders in otherwise healthyanimals. The question remains, however,whether "artificially induced" disorders suchas Parkinson's disease actually are goodenough models of the "real," naturally occur-ring, disorder observed in humans. This is anempirical issue that requires careful analysisof behavior for each disorder under study.Of course, a disorder such as attention-deficit/hyperactivity disorder (ADHD) is go-ing to provide special problems for animalmodelers because the most obvious problemin children with hyperactivity syndromes isthat they have problems in school. Rats obvi-ously do not go to school! Nonetheless, itremains possible to study effectively othersymptoms of such disorders, as we shall see.

The goal of this chapter is to examinesome of the most well-developed models ofhuman neurological disorders that involve thecerebral hemispheres. First, however, becauseso many neurological disorders involve corti-cal functioning, we must examine the useful-ness of the rodent cerebral cortex as a modelof human cortical organization.

CORTICAL ORGANIZATION IN RATS

One of the major obstacles in comparing thebehavior of different species of mammals isthat each species has a unique behavioralrepertoire that permits the animal to survivein its particular environmental niche. Thereis, therefore, the danger that neocortical or-ganization is uniquely patterned in differentspecies in a way that reflects the unique be-havioral adaptation of different species. Oneway to address this problem is to recognizethat although the details of behavior may dif-fer somewhat, mammals share many behav-ioral traits and capacities (e.g., Kolb and

Whishaw, 1983). For example, all mammalsmust detect and interpret sensory stimuli, re-late this information to past experience, andact appropriately. Similarly, all mammals ap-pear to be capable of learning complex tasksunder various schedules of reinforcement.The details and complexity of these behaviorsclearly vary, but the general capacities arecommon to all mammals. Warren and Kolb(1978) proposed that behaviors and behavioralcapacities demonstrable in all mammals couldbe designated as class-common behaviors. Incontrast, behaviors that are unique to aspecies and that have presumably been se-lected to promote survival in a particularniche are designated as species-typical behav-iors. This distinction is important because ithas implications for the organization of thecerebral cortex. We note that just becausemammals have class-common behaviors doesnot prove that they have not independentlyevolved solutions to the class-common prob-lems. There is little evidence in support of thisnotion, however. Neurophysiological, anatom-ical, and lesion studies reveal a similar topog-raphy in the motor, somatosensory, visual,and auditory cortices of the mammals, a to-pography that provides the basis for class-common neural organization of fundamentalcapacities of mammals.

Kaas (1987) has argued, for example, thatall mammalian species have similar regionsdevoted to the analysis of basic sensory infor-mation (e.g., areas VI, Al, Si), the control ofmovement (Ml), and a frontal region in-volved in the integration of sensory and mo-tor information. We can extend Kaas's idea bysuggesting that these regions have class-com-mon functions. To be sure, there are largespecies differences in the details of the class-common behaviors. Monkeys (and humans)have chromatic vision and fine visual acuitycompared with the largely achromatic visionand reduced acuity of cats or rats. Neverthe-less, in all mammalian species studied, re-moval of visual cortex severely disrupts objectrecognition. Similarly, rats and cats have a

Page 468: 0195162854

Chapter 42. Neurological Models 451

large somatosensory representation of thewhiskers, whereas monkeys and humans haveno such representation, but in all species, thesomatosensory cortex functions to representskin-related receptors for tactile sensations. Fi-nally, we have seen (see Chapter 15) that themotor systems of rats and primates are re-markably similar in general organization, al-though some details such as the use of an op-posable thumb are obviously different. Thus,the basic operation of the sensory and motorsystems represents class-common functions,even though the details of this recognitionmay vary in a species-typical manner.

But what about the so-called associativefunctions of the cerebral cortex? That is, arethere parallel posterior parietal, anterior tem-poral, and prefrontal cortical systems in ratsand primates? Pandya and Yeterian (1985)have noted that as the number of sensory rep-resentations increase in evolution, there is acorresponding increase in the size and num-ber of associative regions that function to in-tegrate this sensory information with motoroutput. It follows that although the complex-ity of each of the associative regions will varyacross species with the nature of the basic sen-sory representations, there should be parallelassociative regions across mammalian species.Here I consider each briefly.

As a gross generalization, we can arguethat the posterior parietal region functions touse sensory information, and especially visualand tactile information, to guide movementsin space. For primates much of the expansionin posterior parietal functioning is thus di-rected to visual and tactile guidance of limbmovements to grasp and manipulate objectsas well as to navigation in space. As notedin Chapter 15, the guidance of skilled limbmovements in rats is largely under olfactorycontrol, so it is hardly surprising that the pos-terior parietal region does not play a majorrole in guiding limbs of rats. The evidencefrom both lesion and electrophysiologicalstudies is clear, however, in showing that asin primates, rats with posterior parietal lesions

have deficits in spatial navigation (Kolb et al.,1994).

The primary function of the temporal as-sociative regions in humans is in the recogni-tion of complex visual and auditory infor-mation, and especially in the recognition ofmeaningful patterns of such information suchas in the recognition of faces, objects, words,and music. The temporal associative regionsof rats have a much simpler organization thanthose of primates, but there now is little doubtthat these regions are involved in complexpattern recognition of visual (e.g., objects) andauditory (e.g., species-typical vocalizations)inputs (Dean, 1990; Kolb et al., 1994).

Finally, the prefrontal region, which ispresumed to control the somewhat mysteri-ous "executive functions" in primates also hasa parallel organization in rats. Recent studieshave shown a striking parallel in anatomicalorganization and behavioral functioning inrats and primates (Uylings et al., 2003). Thus,for example, both rats and primates have atleast two major subdivisions of the prefrontalcortex (a dorsal and medial division and an or-bital division), each of which can be furthersubdivided into many subregions. Further-more, there are a wide range of parallel symp-toms after injury to the two regions in rats andmonkeys (see a review, Uylings et al., 2003).

In sum, although obviously simpler, thecerebral cortical organization of rats provides agood approximation of what is observed in pri-mates. This parallel provides a nice startingpoint for investigating the effects of behavioraldisorders that involve cerebral dysfunctioning.

MODELS OFNEUROLOGICAL DISORDERS

A discussion of all neurological disorders oreven a detailed discussion of a representativesubset of disorders is beyond the scope of ashort chapter. I therefore have selected themost common disorders, which I discuss inonly general terms. (For a more detailed dis-

Page 469: 0195162854

452 MODELS AND TESTS

cussion of such models, see the two volumesedited by Boulton, Baker, and Butterworth,1992.) One advantage of studying neurologi-cal, rather than psychiatric, disorders (seeSzechtman and Eilam, Chapter 42) is thatthe former disorders are presumed to have apurely neurological cause, and thus it shouldbe possible to induce the disease without par-ticular concern about experiential factors. Inparallel to psychiatric disorders, however, isthe difficulty that many neurological disor-ders are characterized by disordered cognitivefunctioning. Thus, as discussed in detail bySzechtman and Eilam (Chapter 42) the analy-sis of behavior becomes critical as a means tomaking inferences about the nature of the dis-eases and the potential treatments that maybe used to ameliorate the symptoms. Thus,depending on the location of the neurologicaldisorder, there must be appropriate behav-ioral assays such as those described in the restof this volume. I shall therefore not focus onthe behavioral analysis of the different mod-els so much as on the nature of neurologicalmodel itself.

In principle, it is possible to distinguishbetween disorders that are largely found inthe adult brain and those that occur develop-mentally. These disorders present differentchallenges in developing models and are con-sidered separately.

MODELS OF DISORDERSOF THE ADULT BRAIN

StrokeHuman ischemic stroke is diverse in its causes,location, and symptoms. A major advantageof rat models of stroke is that the injuries canbe controlled to be reproducible, which al-lows for a systematic study of the behavioralsequelae and treatments of stroke. There aremarked variations in details of cerebral bloodflow in different strains so care must be takenin selecting a strain that can be compared withother literature (Ginsberg and Busto, 1989).Some strains show considerable variability

within the strain as well. For example, theSprague-Dawley rat has at least six differentbranching patterns, which means that therewill be considerable variance across strokes indifferent animals.

There are two general categories ofstroke models: focal ischemia and global is-chemia (Table 42-1) (Seta et al, 1992). Wehave found the Long-Evans rat to be a goodmodel, particularly for behavioral studies offocal stroke, although they are not so com-monly used for studies of ischemic stroke. Thenature of the behavioral analysis is different inthe two models because the pattern of injuryis very different. Focal models that involve ei-ther the MCA or pial stripping involve motorcortex and variable amounts of striatum. Theadvantage is that there are excellent tech-niques for assessing motor recovery and com-pensation (see Chapter 15). One disadvantageof the MCA models is that large lesions mayproduce extreme motor deficits that are verydifficult to assess, except with rather grossmeasures.

Ischemic models have been used exten-sively in studies of neuroprotective agents. Theadvantage of these models is that there is hip-pocampal cell death, thus allowing studies ofcognitive behavior (see Chapter 39). The dis-advantage is that the lesions tend to be variableand large group samples are often needed.

Cerebral InjuryPeople acquire cerebral injury in a variety ofways in addition to stroke, including traumatichead injury and surgically induced injury, suchas when there is tumor removal, vascular sur-gery, or surgery for drug-intractable epilepsy.Cerebral injuries in rats are most commonlyproduced using either suction ablation or in-jections of selective neurotoxins or with a fluidpercussion model of head trauma. The formermethods produce focal injuries, whereas thehead trauma model produces a much more dif-fuse injury. The behavior of animals with bothtypes of injuries can be assessed using behav-ioral tasks sensitive to the functions of the dis-

Page 470: 0195162854

Chapter 42. Neurological Models

Table 42-1. Models of Stroke in Rats

453

Model Preparation

FOCAL MODELS

Embolism models

Photothrombolitic model

Endothelin-1

Middle cerebral artery (MCA)occlusion

Pial stripping

Injection of blot clot fragments or microspheres into the carotid artery

Systemic injection of a chemical (e.g., Rose Bengal). Laser illumination offocal region of skull induces photochemical reaction, causing platelets anderythrocytes aggregation.

Local injection of endothelin-1 causes local degradation of blood vessels,causing cell death.

Permanent occlusion by thermocoagulation of all, or part of, the MCAReversible occlusion by using a snare ligature around the MCA

The pia and blood vessels are stripped off of a localized region.

GLOBAL MODELS

Bilateral carotid occlusionTwo-vessel occlusion

Four-vessel occlusion

Levine hypoxia-ischemia

Both common carotids are occluded, and blood pressure is reduced to50 mm Hg.

Vertebral arteries are cauterized; carotids are temporarily occluded with aclasp or snare.

Unilateral carotid occlusion followed by exposure to anoxic environment forabout 45 minutes 24 hours later.

crete cortical areas injured (see Chapters 5-17,39, 43), particularly in animals with focal in-juries. Studies of head trauma typically also in-volve examination of cognitive functions us-ing tests similar to those used for dementia (seeChapter 39).

Historically, the suction model has beenmost extensively used, but the major disad-vantage of the suction model is that theremust be a craniotomy to expose the tissue tobe removed, although the advantage is thatthe tissue is visualized and by using stereotaxiccoordinates it is possible to make consistentlesions. A second disadvantage is that the tis-sue is removed from the brain and thus doesnot slowly die as in a natural head injury.There are differences in the neuroanatomicaland neurochemical reaction to injuries that door do not have dying tissue, so this may beimportant in studies examining treatments forcerebral injury. The excitotoxic models onlyrequire that small bur holes be made to ac-commodate cannulae for infusion of the toxin,

but in the case of large lesions, such as re-moval of the entire motor cortex, there mustbe many bur holes and there is considerablevariability in the lesion extent. One advantageof the excitotoxic lesions is that the lesions canbe restricted to gray matter, thus reducing thediffuse effects of the injury. This may be a dis-advantage in studies of recovery of function,however, because it would be atypical in hu-mans to have lesions that spared the whitematter.

The fluid percussion model of headtrauma requires a craniotomy to expose thebrain. A plunger strikes the brain with a spe-cific amount of force, causing cell death andshearing and tearing of white matter. The le-sions are quite variable in extent but do mimicclosed head injuries in people.

Parkinson's DiseaseParkinson's disease has a complex and vari-able etiology and neuropathology, but theconsistent common feature is a loss of

Page 471: 0195162854

454 MODELS AND TESTS

dopamine neurons in the substantia nigra thatleads to a variety of motor impairments. Al-though it was once believed that only primatemodels are useful in studying Parkinson's dis-ease, there is no consensus nor current basisto suggest that one species is more predictivethan another in the transition from researchto clinical practice (Schallert and Tillerson,2002). The most common rat model of Parkin-son's disease is produced by injecting the neu-rotoxin 6-hydroxydopamine into either themedial forebrain bundle or the rostral stria-turn in one hemisphere. Such procedures nor-mally produce a range of dopamine depletion(as measured by neurochemical assays suchas high-performance liquid chromatography),and this variation can be correlated to behav-ioral impairments. For example, Tillerson etal. (1998) found a correlation of .92 betweenstriatal dopamine depletion and the degree offorelimb motor impairment. The toxin injec-tions are normally unilateral both because an-imals with extensive bilateral depletions willnot eat and because unilateral lesions allowthe possibility of comparisons between themotor performance of the two sides of thebody.

Although the original behavioral meas-ures used to reflect the extent of motorimpairments were based on dopaminergicdrug-induced asymmetries in behavior, theseasymmetries do not predict the extent ofmotor impairment and may not be the bestmeasure. A variety of tests have been devisedthat do not require drugs, including skilledforelimb reaching behavior (Miklyaeva andWhishaw, 1996), forelimb asymmetry duringvertical exploration (Schallert and Lindner,1990), tactile extinction (Schallert et al., 1982),and a variety of simple motor reflex tests de-scribed by Schallert and Tillerson (2002).

Huntington's DiseaseHuntington's disease is an inherited, progres-sive, neurodegenerative disorder that is char-acterized by bizarre uncontrollable movementsand postures. Anatomically, Huntington's dis-

ease is associated with gross generalized atro-phy of the cerebral hemispheres and extensivecell death in the striatum. The striatal celldeath is confined to a loss of medium spinyneurons that are the primary output neuronsof the striatum. This cell death was originallybelieved to be the major cause of the behav-ioral abnormalities, but it has become clearthat neural degeneration associated with thedisease is quite widespread, involving thecerebral hemispheres, brain stem, and cere-bellum (Emerich and Sanberg, 1992).

Rodent models of Huntington's diseasehave focused on the striatal pathology, but itmust be remembered that the human disor-der includes much broader pathology. Therehave been two general types of models de-veloped: dyskinesia models based on striatalneurotransmitter imbalances and excitotoxinlesion models. The neurochemical modelshave been based on the assumption that it isnot so much the striatal cell death but ratherthe effect of the striatal cell death on thebalance of dopaminergic, GABAergic, andcholinergic systems in the striatum. Thus, var-ious dopaminergic agonists and GABAergicand cholinergic antagonists have been infusedinto the striatum, and in each case dyskineticmovement patterns can be produced. The ad-vantage of these models is that they are easyto prepare and the motor abnormalities do re-semble those of the disease, but the major dis-advantages are that the effects are not longlasting and it is difficult to see how such mod-els would prove useful in developing clinicallyrelevant treatments.

The excitotoxic models use selectivetoxic compounds (kainic acid, quinolinic acid)to kill striatal neurons. Both toxins producepathology and behavioral effects that are rem-iniscent of Huntington's disease. The rat mod-els have lead to the hypothesis that a funda-mental deficit in Huntington's disease is adysfunction of glutamatergic transmission,which results in the slow progressive celldeath characteristic of the disease (Emerichand Sanberg, 1992). However, because the dis-

Page 472: 0195162854

Chapter 42. Neurological Models

ease is genetic, these rat studies cannot pro-vide information about the genetic defect (butsee Viral Vector-Mediated Neurodegenera-tion). The models do provide a viable tech-nique for examining clinical treatments.

Dementia of the Alzheimer's TypeAlthough there are many forms of dementiain humans, the most common and the mostintensively studied is Alzheimer's disease.Like most dementias, Alzheimer's disease is aneurodegenerative disorder that can be reli-ably diagnosed only with postmortem exam-ination of the cerebral pathology. The path-ology in Alzheimer's is rather widespread,including death of magnocellular cholinergicneurons in the basal forebrain, loss of neuronsin the brain stem monoaminergic projectionsfrom the raphe and locus ceruleus, and senileplaques within the amygdala, hippocampus,and cerebral cortex. The degeneration of thebasal forebrain cholinergic neurons that in-nervate the cortex and hippocampus are be-lieved to underlie some of the cognitiveimpairments associated with Alzheimer's dis-ease; thus, animal models have tended to fo-cus on the cholinergic loss. Like the modelsof Huntington's disease, although the pathol-ogy in Alzheimer's disease is rather wide-spread, the models have focused on a singleaspect of the disease, which in Alzheimer's dis-ease is the cholinergic aspect.

The three most common models ofAlzheimer's are (1) aged rats, (2) lesions ofthe basal forebrain, and (3) pharmacologicalagents. The age-related models are based onthe observation that many of the cognitive im-pairments in Alzheimer's disease also occur,although to a lesser degree, in normal aging.Thus, there is mild amnesia that progressesgradually over a period of years as well as aslow decline in general intellectual functions.Rats show age-related cell loss in the basalforebrain that are correlated with behavioralimpairments in cognitive tasks. Surprisingly,however, there is no age-related loss in corti-cal markers of cholinergic activity, which may

455

limit the usefulness of the aged rat model. Fur-thermore, most aging studies in rats use agenetically homogeneous strain of rats, theFisher 344, which is healthier in old age thanother strains such as the Long-Evans, butthere is some question over how representa-tive this strain's spatial memory abilitiesmight be.

Young animals do not develop the bio-chemical changes seen in Alzheimer's disease,but it is possible experimentally to producesuch changes by making discrete basal fore-brain lesions. Normally, studies use neuro-toxins related to glutamate including kainicacid, ibotinic acid, and quinolinic acid. All arepotent toxins that affect a slightly differentpopulation of neurons within the injectionsite. All produce mild, but consistent, deficitsin memory in rats, and all produce a depletionin cortical and hippocampal markers foracetylcholine. One difficulty is that the mem-ory impairments tend to recovery over timeor with extensive postoperative training (Bar-tus et al., 1985), which means that the base-line performance will shift over time, thusmaking studies of therapeutic agents moredifficult.

Finally, anticholinergic agents can beused to impair behavioral performance, thelogic being that muscarinic antagonists canproduce cognitive deficits in both humans andrats that resemble the impairments inAlzheimer's disease (but see Whishaw, 1985).The anticholinergic models have been usedwith some success to examine the effectsof putative cognitive-enhancing drugs. Theproblem remains, however, that the effects ofthe anticholinergics are transient and are spe-cific to brain cholinergic systems yet the dis-ease is progressive and far more widespread.

Viral Vector-Mediated NeurodegenerationThe defective handling of proteins is a centralfeature of most major neurodegenerative dis-eases. Because the production of proteins iscoded by genes, it is not surprising that it wasrecently discovered that neurodegenerative

Page 473: 0195162854

456 MODELS AND TESTS

diseases can be caused by mutations in singlecellular proteins. The genetic mutations canbe induced in rats through the generation oftransgenic animals carrying the disease-caus-ing gene, but this model has the drawbacksthat the mutated gene affects the entire brainand that the animal has the abnormal genethroughout its lifetime, rather than just inadulthood. A new rodent model has been de-veloped that is based on advances in recom-binant viral vector technology. Thus, a viruswith a specific affinity for certain cell types canbe injected into specific regions of the brain,such as the substantia nigra. Once inside thecell, the virus induces the expression of the ab-normal gene (Kirik and Bjorklund, 2003). Al-though this technique is still in its infancy, vec-tor systems have been used successfully toexpress either the mutated human huntingtinprotein in striatal neurons as a model of Hunt-ington's disease or mutated human alpha-synuclein in nigral dopamine neurons as amodel of Parkinson's disease (de Almeida etal, 2002; Kirik et al., 2002). Viral vector-mediated gene transfer is expected to proveuseful for producing other disease-causingproteins, the most likely candidate being amy-loid precursor protein involved in Alzheimer'sdisease pathology. There are few reports ofbehavioral changes in animals infected withthe viral vectors as the studies to date havebeen anatomical.

Seizure DisordersOne of the most common neurological con-ditions is the occurrence of seizures, eitheracutely in some disorders such as traumatichead injury or encephalitis or chronically as inepilepsy. The main reason for developing an-imal models is to understand the molecularmechanisms producing the seizures, as well asto screen potential anticonvulsant agents.

Most seizures are caused by an irritant ofsome sort in the nervous system, such as aninjury, that results in abnormal electrical dis-charges. Seizures can be acute, however, re-sulting from electroshock or a chemical im-

balance. Thus, there are two general types ofseizure models in rats, one based on the de-velopment of an irritant in the brain and theother being chemically induced.

The most common focal model of sei-zures is known as kindling. Kindling was ametaphor proposed by Goddard to describethe observation that, in a manner similar tothat in which burning small twigs ultimatelyproduces a large fire, repeated subconvulsivestimulation of the brain by electrical currentor pharmacological agents produces seizuresthat gradually increase in intensity. Kindlinghas been the object of intense study over thepast 25 years, leading to the development ofan enormous body of literature (e.g., Corco-ran and Moshe, 1998; Teskey, 2001). One ad-vantage of the kindling model is that the de-velopment of the seizures can be objectivelymeasured both electrophysiologically and be-haviorally, and thus it is possible to initiateanticonvulsant medications at many differenttime points in the development of the seizuredisorder.

The other principal method of inducingseizures is to systemically or locally injectdifferent drugs or other compounds, includ-ing penicillin, strychnine, tetanus toxin, alu-minum, and bicucullin (see McCandless andFineSmith, 1992, for a review). These modelshave the advantage of ease of induction andreliability in the seizure disorder produced.The disadvantage is that chemically inducedseizures are rare in humans and the mecha-nisms underlying the seizures are likely to bequite different in the different models.

Drug AddictionMany people commonly take stimulant drugslike nicotine, cocaine, or heroin, all of whichaffect behavior and thus are said to be psy-choactive. The long-term consequences ofabusing psychoactive drugs are now well doc-umented, but less is known about the howthese drugs can chronically alter the nervoussystem. One experimental demonstration ofdrug-induced changes in the rat brain is

Page 474: 0195162854

Chapter 42. Neurological Models

known as drug-induced behavioral sensitization,often referred to just as behavioral sensitization(Robinson and Berridge, 2003). Behavioral sen-sitization is the progressive increase in the be-havioral actions of a drug that occur after re-peated administration of a constant dose ofthe drug. Behavioral sensitization occurswith most psychoactive drugs and is especiallystrong with psychomotor stimulants such asamphetamine. For example, when a rat isgiven a small dose of amphetamine, it mayshow a small increase in activity. When therat is given the same dose on subsequent oc-casions, the increase in activity is progres-sively larger, showing behavioral sensitiza-tion. This drug-induced behavioral changepersists for weeks or months so that if the drugis given in the same dose as before, the be-havioral sensitization is still present. In asense, the brain has some memory of the ef-fects of the drug. Thus, addiction now is in-creasingly being viewed as a pathologicalprocess of learning (Berke and Hyman, 2000;Nestler, 2001).

Like the changes associated with learningdifferent tasks (Kolb and Whishaw, 1999),behavioral sensitization is associated withchanges in dendritic and spine morphologyin both prefrontal cortex and the striatum(Robinson and Kolb, 1998), changes in striatalphysiology (Gerdeman et al, 2003), andchanges in the production of immediate-earlygenes and neurotrophic factors (Flores andStewart, 2001). These various changes in brainorganization and function are believed to un-derlie much of the pathological behaviorof human drug addicts, and there is consider-able interest in trying to relate changes incognitive behaviors with morphologicalabnormalities.

MODELS OF DEVELOPMENTAL DISORDERS

One major advantage of using the rat as amodel of developmental disorders is that therat is born embryologically younger thanlaboratory primates or carnivores and thus it

457

is possible to perform many manipulations,which would normally be prenatal in otherlaboratory species, postnatally in the rat. Therat's gestation period is about 22 days, withcerebral neurogenesis beginning around em-bryonic day 11 and finishing by birth (seeBayer and Altman, 1990, for details). Cell mi-gration continues throughout the first post-natal week, at which time synaptogenesis be-gins and continues at a high rate until aboutpostnatal day 25. The developing rat can thusbe exposed to a variety of perturbations (e.g.,injury, behavioral and pharmacological treat-ments) postnatally, at a time that would cor-respond to the third trimester of human de-velopment. Because the third trimester is sosensitive in human development, it is there-fore possible to examine the effects of differ-ent brain manipulations and to explore possi-ble ameliorative treatments in the developingrat model. I consider here two manipulations(injury, stress), although there is also exten-sive literature on the effects of alcohol andother drugs in both the prenatal and postna-tal infant rat.

Perinatal Brain InjuryThe perinatal rat brain can be disturbed usinga variety of manipulations, including suctionremoval of discrete cortical areas, excitotoxicinjuries, and hypoxia/ischemia. In general,these different manipulations show one con-sistent result. Thus, damage in the first weekof life in infant rats produces more severefunctional dysfunctions than similar damagein adulthood. In contrast, damage in the sec-ond week of life produces much less severefunctional dysfunctions than similar injury inadulthood (for a review, see Kolb et al., 2001).These time points correspond roughly to in-jury in the third trimester and the first fewpostnatal months of human development, re-spectively. It is known that injury during thethird trimester, including ischemia in prema-ture infants, is particularly deleterious to thedeveloping human brain whereas childrenwith cerebral injuries in the first postnatal year

Page 475: 0195162854

458 MODELS AND TESTS

show better outcomes, especially when thelanguage zones are damaged (Kolb andWhishaw, 2003).

The markedly different behavioral out-comes from week 1 and week 2 injuries in thedeveloping rat are correlated with distinctlydifferent anatomical responses as well. For ex-ample, damage in the second week inducesneurogenesis, gliogenesis, and dendritic hy-pertrophy, whereas damage in the first weekgenerally has little effect on neurogenesis orgliogenesis and leads to dendritic atrophy(Kolb et al., 2001). In addition, my colleaguesand I have shown recently that injury at thetwo timpanist produce differential effects onprotein expression, changes that can be seenin some cases into adulthood. One example isthe expression of a neurotrophic factor, basicfibroblast growth factor (bFGF). bFGF ex-pression is markedly increased after injury inthe second postnatal week but not after injuryin the first postnatal week. In sum, the exten-sive literature on the effects of perinatal cor-tical injury in the rat provide an excellentmodel for understanding the effects of injuryto the developing human brain. More impor-tant, however, the rat models have proved tobe especially useful in understanding the ef-fects of different treatments on stimulatingfunctional compensation after early injuries(Kolb et al., 2001). Thus, the rat with cerebralinjury in the first few days of life can show sig-nificant functional benefit from a variety oftreatments, including tactile stimulation, di-etary supplements, complex housing, andbFGF.

Perinatal ExperienceExtensive literature shows that the develop-ing rat is especially sensitive to postnatal ma-nipulations. Two examples are neonatalhandling and maternal separation. In moststudies, handling involves maternal separa-tion for about 15 minutes a day over the first2 weeks of life (Levine, 1961; Caldji et al.,2000). This treatment enhances maternal be-haviors such as licking and grooming of the

infants, which in turn appears to alter the feed-back regulation of stress-related systems in thenervous system. These changes have long-lasting benefits on cognitive functioning andreactivity to stress in adulthood. In contrast,maternal separation for longer periods, usually3 to 4 hours per day, produces animals that arehyperresponsive to stress, including a hyperre-sponsive hypothalamic-pituitary response (Liuet al., 1997). Finally, other forms of perinatalexperience, such as tactile stimulation duringinfancy or complex housing during the juvenileperiod, are also known to produce long-lastingchanges in cortical organization (Kolb et al.,2003).

Taken together, these studies of perina-tal experience are believed to provide modelsof how perinatal experience might chronicallyinfluence behavioral development in children.For example, it is believed that the studies ofhandling and maternal separation will provideinsight into different developmental behav-ioral conditions (e.g., ADHD) and will providemodels to investigate ways to treat such be-havioral pathologies.

Attention-Deficit/Hyperactivity DisorderNumerous recent reviews have describedthe behavioral and cognitive characteristics ofADHD (Barkley, 1997). It is now generally be-lieved that ADHD results from a frontal-striatal abnormality, possibly lateralized to theright hemisphere (Heilman et al., 1991). Ab-normalities in the dopaminergic systems pro-jecting to the prefrontal cortex and basalganglia are believed to underlie much ofthe frontal-striatal abnormality, and the mostcommon treatment for ADHD is methyl-phenidate (Ritalin), which blocks dopaminereuptake. Correlated with the dopaminergicabnormality is an impairment in prefrontalfunctions, including in particular workingmemory and attentional functions. The causeof the dopaminergic abnormality has been dif-ficult to determine, but predisposing factors,including genetics, premature birth, prenatalstress, and hypoxia/ischemia, among others,

Page 476: 0195162854

Chapter 42. Neurological Models 459

have been proposed (Sullivan and Brake,2003).

ADHD has not proved to be easy to treatclinically and thus has led to considerable in-terest in developing an animal model. Oneway to proceed has been to take advantage ofthe normal variance in the performance of ratson various tests of working memory and cog-nitive functioning. Many studies have nowshown that treating rats with methylphenidatecan actually improve performance of poorlyperforming rats on tests of attentional pro-cesses. One strain of rats, the Kyoto SHR rat,has been proposed to be an especially goodmodel, largely because there are known abnor-malities in prefrontal dopaminergic innervationand this is correlated with behavioral abnor-malities such as hyperactivity. The behavioralabnormalities can be reversed by dopaminergicagonists such as methylphenidate (Sullivan andBrake, 2003).

Other models of ADHD have focused onmanipulating prefrontal development by peri-natal anoxia (Brake et al., 1997). Interestingly,this treatment leads to prefrontal abnormali-ties that are lateralized to the right hemi-sphere, as is seen in humans (Brake et al.,2000). Anoxia is not the only manipulation toproduce dopaminergic abnormalities; manip-ulations of the postnatal social environmentbetween mothers and infants have shown ab-normalities in dopaminergic systems in rats aswell (Sullivan and Gratton, 2003).

In sum, there are now several differentmodels in which dopaminergic innervationof the prefrontal/striatal circuitry in the ratis abnormal and is correlated with behavioralabnormalities including increased activity,poor working memory, and attentionaldeficits.

CONCLUSION

It is possible to produce brain dysfunctions inrats that can be shown to mimic a wide rangeof human neurological conditions. Although

it would be ideal to study neurological dis-eases in animals with brains very similar toours, such as chimpanzees, this is impracticalfor both ethical and financial reasons. Ratsprovide an reasonable alternative, althoughtheir smaller brain and less complex cognitiveprocesses place some constraints on the na-ture of the behavioral questions that can beaddressed. There remains, however, an ani-mal ethics issue that must be considered.There is no question that inducing neurolog-ical disorders in any animal means that theremust be special consideration of animal care.Particular consideration must be given to an-imals used in models that might produce dis-tress. Humans with neurological disorders of-ten experience irritability, fear, anxiety, andpain. If the rat models successfully mimic thehuman condition, then we might expect thatsome of these symptoms would also be man-ifest in the rats. As Olfert (1992) has empha-sized, any pain, suffering, distress, or deficitsin function that negatively affect the animal'swell-being and are not scientifically "neces-sary" for the study should be alleviated orminimized, regardless of the cost or conven-ience to the experimenter. Presumably the ex-periences of humans with disorders such asParkinson's and Huntington's can serve as aguide to the distress that rats in models ofthese conditions might be enduring.

REFERENCES

de Almeida LP, et al (2002) Lentiviral-mediated deliveryof mutant huntingtin in the striatum of rats inducesa selective neuropathology modulated by polyglut-amine repeat size, huntingtin expression levels, andprotein length. Journal of Neuroscience 20:219-229.

Barkley RA (1997) Behavioral inhibition, sustained at-tention, and executive functions: constructing aunifying theory of ADHD. Psychological Bulletin121:65-94.

Bartus RT, Flicker C, Dean RL, Pontecorvo M, FiguerdoJC, Fisher SK (1985) Selective memory loss followingnucleus basalis lesions: long term behavioral recoverydespite persistent cholinergic deficiencies. Pharma-cology, Biochemistry, and Behavior 23:125-135.

Page 477: 0195162854

460 MODELS AND TESTS

Berke JD and Hyman SE (2000) Addiction, dopamine,and the molecular mechanisms of memory. Neuron25:515-532.

Boulton AA, Baker GF, Butterworth RF (eds.) (1992) An-imal models of neurological disease. Totowa, NJ:Human Press.

Brake WG, Noel MB, Boksa P, Gratton A (1997) Influ-ence of perinatal factors on the nucleus accumbensdopamine response to repeated stress during adult-hood: an electrochemical study in the rat. Neuro-science 77:1067-1076.

Brake WG, Sullivan RM, Gratton A (2000) Perinatal dis-tress leads to lateralized medial prefrontal corticaldopamine hypofunction in adult rats. Journal ofNeuroscience 20:5538-5543.

Caldji C, Diorio J, Meany MJ (2000) Variations in ma-ternal care in infancy regulate the development ofstress reactivity. Biological Psychiatry 48:1164-1174.

Corcoran ME and Moshe SL (1998) Kindling 5. NewYork: Plenum Press.

Dean P (1990) Sensory cortex: Visual perceptual factors.In: The cerebral cortex of the rat (Kolb B and TeesRC, eds.), pp. 275-308. Cambridge, MA: MIT Press.

Emerich DF and Sanberg PR (1992) Animal models ofHuntington's disease. In: Animal models of neuro-logical disease (Boulton AA, Baker GB, ButterworthRF, eds.), pp. 65-134. Totowa, NJ: Human Press.

Gerdeman GL, Partridge JG, Lupica CR, Lovinger DM(2003) It could be habit forming: drugs of abuse andstriatal synaptic plasticity. Trends in Neuroscience26:184-192.

Ginsberg MD and Busto R (1989) Rodent models of cere-bral ischemia. Stroke 20:1627-1642.

Heilman KM, Voeller KK, Nadeau SE (1991) A possiblepathophysiologic substrate of attention deficit hy-peractivity disorder. Journal of Child Neurology6(suppl):S76-S81.

Kaas JH (1987) The organization of neocortex in mam-mals: Implications for theories of brain function. An-nual Review of Psychology 38:129-151.

Kirik D and Bjorklund A (2003) Modeling CNS neuro-degeneration by overexpression of disease-causingproteins using viral vectors. Trends in Neuro-sciences 26:386-392.

Kirik D, et al (2002) Parkinson-like degeneration inducedby targeted overexpression of alpha-synuclein inthe nigrostriatal system. Journal of Neuroscience22:2780-2791.

Kolb B, Gibb R, Gonzalez C (2001) Cortical injury andneuroplasticity during brain development. In: To-ward a theory of neuroplasticity (Shaw CA andMcEachern JC, eds.), pp. 223-243. New York:Elsevier.

Kolb B, Gibb R, Corny G (2003) Experience-dependentchanges in dendritic arbor and spine density in

neocortex vary with age and sex. Neurobiology ofLearning and Memory 79:1-10.

Kolb B, Buhrmann K, MacDonald R, Sutherland RJ(1994) Dissociation of the medial prefrontal, poste-rior parietal, and posterior temporal cortex for spa-tial navigation and recognition memory in the rat.Cerebral Cortex 4:15-34.

Kolb B and Whishaw IQ (1983) Generalizing in neu-ropsychology: problems and principles underlyingcross-species comparisons. In: Behavioral contribu-tions to brain research (Robinson TE, ed.). NewYork: Oxford University Press.

Kolb B and Whishaw IQ (2003) Fundamentals of humanneuropsychology, 5th ed. New York: Worth.

Levine S (1961) Infantile stimulation and adaptation tostress. Research Publication of the Association ofNervous and Mental Disorders 43:280-291.

Liu D, Diorio J, Tannenbaum B, Caldji C, Francis D,Freedman A, et al (1997) Maternal care, hippocam-pal glucocorticoid receptors, and hypothalamic-pituitary-adrenal responses to stress. Science 277:1659-1662.

McCandless DW and FineSmith RB (1992) Chemicallyinduced models of seizures. In: Animal models ofneurological disease, II (Boulton AA, Baker GB, But-terworth RF, eds.), pp. 133-151. Totowa, NJ: Hu-man Press.

Miklyaeva El and Whishaw IQ (1996) Hemiparkinsonanalogue rats display active support in good limbsversus passive support in bad limbs on a skilledreaching task of variable height. Behavioral Neuro-science 110:117-125.

Nestler EJ (2001) Molecular basis of long-term plasticityunderlying addiction. Nature Neuroscience Re-views 2:119-128.

Olfert ED (1992) Ethics of animal models of neurologi-cal diseases. In: Animal models of neurological dis-ease, I (Boulton AA, Baker GB, Butterworth RF,eds.), pp. 1-29. Totowa, NJ: Human Press.

Pandya D and Yeterian EH (1985) Architecture and con-nections of cortical association areas. In: Cerebralcortex, 4: Association and auditory cortices (PetersA and Jones EG, eds.), pp. 3-61. New York: Plenum.

Robinson TE and Berridge KC (2003) Addiction. AnnualReview of Psychology 54:25-53.

Robinson TE and Kolb B (1999) Alterations in the mor-phology of dendrites and dendritic spines in thenucleus accumbens and prefrontal cortex follow-ing repeated treatment with amphetamine orcocaine. European Journal of Neuroscience11:1598-1604.

Schallert T and Lindner MD (1990) Rescuing neuronsfrom trans-synaptic degeneration after brain dam-age: helpful, harmful or neutral in recovery of func-tion? Canadian Journal of Psychology 44:276-292.

Page 478: 0195162854

Chapter 42. Neurological Models 461

Schallert T and Tillerson JL (2002) Intervention strate-gies for degeneration of dopamine neurons inparkinsonism. In: Central nervous system diseases(Emerich DF, Dean RL, Sanberg PR, eds.), pp.131-151. Totowa, NJ: Human Press.

Schallert T, Upchurch M, Lobaugh N, Farrar SB, SpirusoWW, Gilliam P, Vaughn D, Wilcox RE (1982) Tac-tile extinction: distinguishing between sensorimotorand motor asymmetries in rats with unilateral ni-grostrial damage. Pharmacology, Biochemistry, andBehavior 18:753-759.

Seta KA, Crumrine RC, Whittingham TS, Lust WD, Mc-Candless DW (1992) Experimental models of hu-man stroke. In: Animal models of neurological dis-ease, II (Boulton AA, Baker GB, Butterworth RF,eds.), pp. 1-50. Totowa, NJ: Human Press.

Sullivan RM and Brake WG (2003) What the rodentprefrontal cortex can teach us about attention-deficit/hyperactivity disorder: The critical role ofearly developmental events on prefrontal function.Behavioural Brain Research 146:43-55.

Sullivan RM and Gratton A (2003) Behavioral and neu-roendocrine correlates of hemispheric asymmetries

in benzodiazepine receptor binding induced bypostnatal handling in the rat. Brain and Cognition51:218-220.

Teskey GC (2001) Using kindling to model the neuro-plastic changes associated with learning and memory,neuropsychiatric disorders, and epilepsy. In: Towarda theory of neuroplasticity (Shaw CA and McEachernJC, eds.), pp. 347-358. Philadelphia: Taylor & Francis.

Uylings HBM, Groenewegen HJ, Kolb B (2003) Do ratshave a prefrontal cortex? Behavioural Brain Re-search 146:3-17.

Warren JM and Kolb B (1978) Generalizations in neu-ropsychology. In: Brain damage, behavior and theconcept of recovery of function (Finger S, ed.). NewYork: Plenum Press.

Whishaw IQ (1985) Cholinergic receptor blockade in therat impairs locale but not taxon strategies for placenavigation in a swimming pool. Behavioral Neuro-science 99:979-1005.

Wenk GL (1992) Animal models of Alzheimer's disease.In: Animal models of neurological disease, I (Boul-ton AA, Baker GB, Butterworth RF, eds.), pp. 29-63.Totowa, NJ: Human Press.

Page 479: 0195162854

Psychiatric Models

HENRY SZECHTMAN AND DAVID EILAM

Animal models have always played an integralpart in the advancement of medical research,but their acceptance in psychiatry is both re-cent and instructive. The change in attitudefrom a past anathema to the present em-bracement of animal models reflects the pro-found shift of how today psychiatry viewsmental disorders and the successful efforts ofbasic science in showing the utility and thelimits of animal studies of psychopathology.We first discuss the conflicts associated witha rise of a biological psychiatry perspective onmental illness and suggesting that the currentpreeminence of this viewpoint has alignedclinical psychiatry with behavioral neuro-science for the source of much of its basic sci-ence knowledge. We then elaborate on themethods of behavioral neuroscience andshow how this methodology is in fact justwhat is needed to investigate mechanisms ofmental disorders of interest to psychiatry.Next, we attempt to clear up some confusionrevolving around the distinction betweentests, models, and theories. Finally, we illus-trate some of the discussed principles by ex-amining a rat model of a psychiatric illness—obsessive-compulsive disorder.

PSYCHIATRY AND MENTAL ILLNESS

To understand what is asked of animal mod-els in psychiatry, one must have an apprecia-tion of the concerns and practices of psychia-try. Psychiatry is a medical specialty chargedwith the diagnosis, treatment, and prevention

of mental disorders. This simple and straight-forward textbook definition of psychiatry be-lies its true complexity. Unlike other medicalspecialties that deal with visible pathology ofpalpable bodily parts, the purviews of psychi-atry are not physical entities but an intangiblemental life. Thus, at the very heart of psychi-atry lie theories about subjective experiencesand human normality, theories that frame thepsychiatric constructs of a mental disorder. In-evitably, notions what constitutes psychiatricillnesses intersect our most profound concep-tions about the nature of man. A glimpse ofthis true essence of psychiatry can be obtainedby considering the emergence of psychiatry inits historical context. As described by Berriosand Markova (2002):

Born as a by-product of the nineteenth-centurymovement to organize society on a scientificbasis, psychiatry was charged with the con-struction and enactment of normative views ofmadness. . .. Under the protection of medicineand the economic practices following the In-dustrial Revolution, psychiatrists developedrepresentations of mental disease together withthe professional and institutional apparatus toenjoin them. (p. 3)

Not surprisingly, a diversity of psychiatricconstructions of mental disease arose over theyears, attuned to major philosophical and so-cial norms of humans as well as to the tech-nological advances of the day. However,while with time clinical psychiatry had pro-gressed in terms of successful patient man-agement, no school of psychiatry can claimsuccess in terms of cure or prevention of men-

462

43

Page 480: 0195162854

Chapter 43. Psychiatric Models 463

tal disease. Therefore, using this yardstick, anappropriate conceptualization of mental dis-ease does not yet exist. Nevertheless, of rele-vance to the present discussion, we contrastthe likelihood that a successful clinical scienceof mental illness will emerge from a biologicalversus a nonbiological psychiatry perspective.

BIOLOGICAL PSYCHIATRYPERSPECTIVE

Taking schizophrenia as an example of men-tal illness, we highlight two extreme positionson this disorder because the dichotomy thatthey represent illuminates the range of distinctconceptual approaches and attitudes in psy-chiatry. At one end of the pole are viewpointsthat characterize schizophrenia not as an ill-ness but rather as something extrinsic to theperson—"a social fact and a political event"(Laing, 1967). At the other end is the per-spective that schizophrenia is like any othermedical disease caused by a distinct patho-physiology and as such is correctable by theappropriate medical intervention.

A reader of this book may be incredulousthat schizophrenia could be conceived as any-thing but an illness. However, not only isthere a long history of such viewpoints butone should expect that they will legitimatelycontinue into the future until such time whenthe etiology of schizophrenia is ultimatelyidentified. Before the advent of psychiatry, itwas the philosophers who denied the merepossibility of mental disease "because the soul(mind) was a marker of the divine, [and there-fore] it could become neither divided or dis-eased" (Berrios and Markova, 2002, p. 4). Inrecent times, psychiatrists such as Szasz (1961)and Laing (1967) similarly promulgated theposition that schizophrenia is not a disease ofthe person but instead a societal constructionand thus extrinsic to the individual.

In contrast to this, there is the perspec-tive that the mental life and behaviors char-acteristic of schizophrenia are nothing but a

disorder of brain function caused by a physi-cal insult to normal brain biology. While thereader may resonate with the latter perspec-tive, it too is an extreme position. This purely"medical" outlook seeks to explain all of thesymptoms and all of the causes of schizo-phrenia solely in terms of biology, without al-lowing for any significant contribution fromnonbiological factors. Ironically, probably inreaction to the extreme medical viewpoint,many scientists and mental health profession-als fail to fully endorse, in turn, the role of bi-ological factors in disorders such as schizo-phrenia.

A lack of consideration by the medicalperspective of environmental and experientialfactors in the causes, presentation, and treat-ment of schizophrenia is probably not theonly reason for the skepticism by many vis-a-vis a biology of mental illness. For many sci-entists the opposition is no doubt also deeplyphilosophical as those scientists hold that bio-logical explanations of schizophrenia will nec-essarily invalidate and eliminate explanationsat a psychological level. This follows, they ar-gue, from the supposition that the philosophyof biological reductionism that characterizesthe scientific efforts of today's disciplines suchas behavioral neuroscience or biological psy-chiatry leads to more fundamental accountsof mental life and disease by virtue of "re-ducing" psychological events to elements ofbiology. Consequently, the thinking goes, ex-planations of normal and abnormal behaviorusing a nonbiological language are at best ten-tative, awaiting replacement by an appropri-ate biological account in the future. Indeed, aspointed out by Teitelbaum and Pellis (1992),some have argued that a scientist would bewasting time by not directly proceeding to abiological account of behavior.

Biological reductionism is in fact the sci-entific philosophy that underlies the search inpsychology and psychiatry for explanations ofbehavior and psychopathology, but it is alsoonly an instance of a more general philosophyof reductionism that pervades all fields of sci-

Page 481: 0195162854

464 MODELS AND TESTS

ence. Unfortunately, reductionism is muchmisinterpreted, as one can gleam from theopening sentence in The Oxford Companion toPhilosophy entry for reductionism: "One of themost used and abused terms in the philo-sophical lexicon" (http://www.xrefer.com/entry/553365). Clark (1980) reviews many ofthe misconceptions associated with the termreductionism, especially as it relates to issues ofexplaining psychological phenomena in thelanguage of biology (i.e., the brain). The mainthesis advanced by Clark (1980) is that reduc-tionism is simply a rational methodology withobjective rules of procedure and evidence formapping a theory in one field of science ontoa theory in another field. Thus, with regard tophenomena at the level of behavior or men-tal life, biological reductionism refers to themethodology by which scientists attempt tofind the correspondence (mapping) betweencomponents of a psychological theory of be-havior or mental life and entities employed ina conceptual framework of another discipline,namely neuroscience (Clark used the olderterm, physiology, from which present-dayneuroscience emerged). According to Clark(1980), psychological theories aim to providean account for the observed relationships be-tween environmental stimuli and behaviorand are usually framed in terms of answers totwo general questions: "What sort of ma-chinery is required to produce the input-out-put relationships we observe? What jobs mustparts of the system fulfill if the system as awhole is to manifest these relationships?"(Clark, 1980; p. 72). By specifying the "jobs"of the parts, psychological theory identifiesthe junction of each component, that is, the na-ture of the transformation that each part per-forms on the inputs to it and the nature of theconsequences that the output of each part hason the other components and on the systemas a whole. In other words, psychological the-ories specify what different parts do but do notidentify how it is physically done in the or-ganism—how the function is implemented inthe nervous system. Clark argues that reduc-

tionism of psychology to neuroscience is inessence the localization of function—the de-termination of the neural (structural) arrange-ment by which the psychologically identifiedfunction is implemented—and that functionaland structural endeavors are complimentary:

Psychological and physiological concepts canbe related to one another in a unified model ofthe functioning of the nervous system. The var-ious hypothetical states and processes invokedin psychological accounts are not abandoned inreduction; instead they are given a physiologi-cal identification. The presupposition of thepsychological terms are not contradicted by in-creasing physiological knowledge; instead thatknowledge allows us to explain how it is possi-ble that neural states can be psychologicalstates.

(Clark, 1980, p. 184).

Before closing the present section, we con-sider the reason for our previous suggestionthat biological psychiatry is more likely toyield a successful clinical science of mental ill-ness than a nonbiological framework.

It should be first noted that a biologicalpsychiatry perspective is not the extreme medi-cal viewpoint described earlier but is confinedonly to the assertion that a mental disorder re-flects some fault in the workings of the brain.In other words, using the language of biolog-ical reductionism, normal mental functionsare carried out by particular arrangements ofneural entities, and therefore mental diseaseor malfunction must correspond to a break ina required neural component(s) or to a break-down in communication amongst the neuralparts. How the neural failure is produced orcan be repaired is an open question and bio-logical psychiatry does not predetermine thenature of the answer. Thus, the conceptualframework of biological psychiatry allows forthe possibility that neural parts can break be-cause of psychological factors or because ofphysical insults or faulty genetics and that con-versely both psychological and physical fac-tors may yield neural repair effects. All that abiological psychiatry perspective specifies is

Page 482: 0195162854

Chapter 43. Psychiatric Models 465

that the roles of nonbiological and biologicalvariables must be assessed at the proper lev-els of discourse. Hence, we suggest, the per-spective of biological psychiatry is more likelyto provide a successful clinical science of men-tal illness because it is a more comprehensiveframework than either of the extreme posi-tions described previously and because it in-cludes the possible impact of psychologicalfactors on the causes, presentation, and treat-ment of mental diseases. Moreover, by iden-tifying mental disorders with a malfunction ofthe brain, the biological psychiatry perspec-tive makes explicit that an understanding ofthese disorders can come about only throughan understanding of the neural machinery un-derlying normal psychological functions. Thepursuit of such basic knowledge is thepurview of behavioral neuroscience, to whichwe now turn.

FROM PSYCHIATRY TOBEHAVIORAL NEUROSCIENCE:

NEURAL MACHINERY OFMENTAL FUNCTIONS

Because the pathogenic mechanisms for mostmental illnesses are unknown, psychiatristsmake their diagnosis of mental disorder fromevidence comprised primarily of behavioraldata. Such data consist of (1) reports by pa-tients on their subjective experiences and ownbehavior and (2) records by an external ob-server of patients' behavior and physical pres-entation. As mental disorders are classified onthe basis of specific profiles in the behavioraldata set, psychiatrists arrive at a diagnosis ofa particular mental illness by matching the ob-tained behavioral findings to an identified datapattern of a disease. While in practice theprocess of diagnosis is more complicated, itremains the case that still today the primarysource of psychiatric data comes from meas-ures of behavior. This point is crucial—eventhough biological psychiatry holds that a faultin the workings of the brain underlies the pre-

sentation of a mental disorder, the classifica-tion of such a disease is not based on any iden-tified brain pathology but on behavior. Thisis because no diagnostic brain pathology forpsychiatric disorders has been yet identified.As is noted later, the demonstration of faultybrain workings in mental disorders is notlikely to come from a clinical field but from abasic science discipline such as behavioralneuroscience.

Behavioral neuroscience is the disciplinethat seeks to map normal psychological func-tions onto specific arrangements of neurons.This task is doubly complex for it requires, onthe one hand, that the entity to be localized isindeed a proper unit of psychological functionand, on the other hand, that the mapping con-forms to proper units of nervous system or-ganization. Thus, the discipline requires ex-pertise in the methods of both behavioral andneural sciences. This in principle is no differ-ent from requirements of biological psychia-try, where expertise in the evaluation of be-havior and understanding of brain function isalso needed. However, biological psychiatry,as an experimental science, can examine onlycorrelations between mental life and brain func-tion. In contrast, because its methods includemanipulation of brain tissue, behavioral neu-roscience can demonstrate the cause-effectmechanism by which a malfunction in neuralmachinery may result in abnormal mental ex-periences and behavior (for a discussion of be-havioral neuroscience methods, see TeitelbaumandPellis, 1992; Teitelbaum and Strieker, 1994).In other words, in the pursuit of localizationof function, behavioral neuroscience simulta-neously addresses two questions: What is theneural machinery of normal psychologicalfunctions? What is the nature of the behavioralabnormality that ensues from a perturbationof a particular set of neurons? In this respect,behavioral neuroscience is closely aligned withbiological psychiatry by providing the basic sci-ence tools and knowledge required to identifywhat brain malfunction yields which type ofmental disorder.

Page 483: 0195162854

466 MODELS AND TESTS

ON METHODS, TESTS,MODELS, AND THEORIES

Evidence that a specific neural dysfunction hasa mechanistic role in a particular psychiatricdisorder requires the experimental demon-stration that the induction of this brain path-ology yields the behavioral symptoms of thedisease. Such experiments are ethically feasi-ble using animals as subjects. Consequently,most laboratory studies use animal subjects,with rats and mice being the most commonsubjects. Of course, given the many obviousdifferences between rodents and humans, ex-periments on animals to understand phe-nomena in the human raise issues pertainingto the conduct and interpretation of such stud-ies. The problem is especially vexing in rela-tion to phenomena of interest to psychiatrywith its focus on the mental life of the indi-vidual: How can one study and relate to thehuman the normal and abnormal mental lifeof so different an animal as a rat? Answers tosuch questions can be expected to be formal-ized in the use of animal models of psychiatricdisorders.

Animal models of psychopathology are arelatively recent scientific phenomenon andthe rationale underlying their use is still un-der debate. Nevertheless, several influentialexpositions have been published (McKinneyand Bunney, 1969; Willner, 1984; Willner etal., 1992; Geyer and Markou, 1995). McKin-ney (1988), in Models of Mental Disorders: A NewComparative Psychiatry, reviewed the history ofanimal modeling in psychiatry and noted sev-eral pitfalls of early modeling in psychiatry.First, it was conceptually unreasonable ofearly workers to seek comprehensive animalmodels of psychiatric disorders because nomodel can be a miniature replica of the hu-man illness. Rather, McKinney argued, one"should focus on the development of specificexperimental systems to investigate selectedaspects of the human syndromes" (p. 143).Second, and related to this point, there hadbeen in the past a tendency to overgeneralize,

jumping too quickly to clinical conclusionsfrom a given set of animal behavior experi-ments. Any generalization must consider notonly the problem of cross-species compar-isons but also the fact that animal experimentswill not model the entire spectrum, but onlyaspects, of the disease. Third, early studies re-lied on descriptions, rather than on a quanti-tative analysis of animal behavior. Moreover,they did not consider the full richness of ani-mal behavior, its evolution, and social struc-ture. Consequently, comparisons to humansseemed strained considering the obvious com-plexity of human behavior and of psychiatricdisorders. Finally, modeling of some disorderssuch as schizophrenia had been (and contin-ues to be) especially problematic because di-agnostic criteria keep changing, making it dif-ficult to determine what should be modeledin animals.

A rationale for modeling psychopath-ology in the laboratory (which still seems ap-propriate today) had been proposed byAbramson and Seligman (1977). In accordancewith the more general framework of whatare models in psychology (Chapanis, 1961),Abramson and Seligman (1977) defined mod-eling as "the production, under controlled con-ditions, of phenomena analogous to naturallyoccurring mental disorders. Its goal is to un-derstand the disorders" (p. 1). Furthermore,the authors viewed modeling as the attemptto bring the study of psychopathology undera "controlled setting in which particular symp-tom or constellation of symptoms is producedin miniature to test hypotheses about causeand cure. Confirmed hypotheses [would] thenbe further tested in situations outside the lab-oratory" (p. 1). The authors envisioned twotypes of paradigms. In one type, a standardtest is administered under controlled condi-tions to subjects with mental disorder andcompared with performance of normal con-trols. The purpose is to examine some a pri-ori hypothesis about the determinants of themental disorder and consequently the chosentest is one that measures the aspect of psy-

Page 484: 0195162854

Chapter 43. Psychiatric Models 467

chological function which the investigatorshypothesize is disordered in the disease. Forexample, to examine hypotheses of disorderedinformation processing in schizophrenia, in-vestigators may compare performances ofsubjects with and without the disease on testsof sensorimotor gating such as prepulse inhi-bition of the startle response (Geyer et al.,1999). Using appropriate tests, other hy-potheses of the disorder may be examined inthe same manner.

The alternative paradigm is potentiallymuch more powerful because it involves theuse of the laboratory experimental modelmethod. As argued by Abramson and Selig-man (1977, p. 3),

In this method, psychopathology is not merelybrought into the laboratory for study; it is mod-eled and reproduced there. The advantage ofstudying a model over observing and compar-ing two groups is that because the disorder isproduced in the lab, we can specify what causesit. The model is rarely an exact replica of thespontaneous psychopathology, and it is usefulonly to the degree that it mimics the phenom-enon in question or suggests interestingexperiments. . . .

The authors suggested that an ideal modelwould show similarity with the real-word dis-order in terms of cause, symptoms, preven-tion and cures; they proposed the following 4criteria in evaluating how good is a laboratorymodel of psychopathology (Abramson andSeligman, 1977; p. 3):

1. Is the experimental analysis of the laboratoryphenomenon thorough enough to describe theessential features of its causes as well as its pre-ventives and cures?

2. Is the similarity of symptoms between themodel and naturally occurring psychopathol-ogy convincingly demonstrated?

3. To what extent is similarity of physiology,cause, cure and prevention found?

4. Does the laboratory model describe in all in-stances a naturally occurring psychopathologyor only a subgroup? Is the laboratory phenom-enon a model of a specific psychopathology,

or does it model general features of all psy-chopathologies?

While the meaning of the middle two criteriais self-evident, the authors elaborated thatthe first criterion regarding the "essential fea-tures" of the model refers to establishing howrelevant is the procedure of inducing the dis-order in the model to the real-world mannerof precipitating the disease in the human (andlikewise for the procedures of preventing andcuring the disease). For instance, in a modelof depression (Seligman, 1972; Miller andSeligman, 1973), the inducing procedure is theadministration of electric shock but the au-thors show that the critical feature of this pro-cedure for the induction of depression is notthe shock itself but rather the production ofan inescapable and uncontrollable situation,which the authors hold resembles the real-world inductor of depression. With regard tothe last criterion, it refers to the need of es-tablishing that the model is of a specific dis-ease (or subtype) rather than of a general re-sponse to being ill.

It is important to note that, as outlinedby Abramson and Seligman (1977), an animalmodel of a psychiatric disorder represents atheory of a naturally occurring psychopathol-ogy and as such is subject to the usual rulesfor confirmation of scientific theories. How-ever, this terminology is often applied—especially in the psychopharmacology litera-ture—to animal preparations that do not havethe attributes of a theory (that is, the inten-tion of describing in all its facets the truemechanism of the real-world psychopathol-ogy). Rather, the term "animal model" of amental disorder is often used with referenceto a preparation that is best described as a "be-havioral test" that is sensitive, either directlyor indirectly, to the presence of a particularpsychopathology. For instance, apomorphine-induced vomiting is a behavior that respondsto antipsychotic drugs (and was in fact one ofthe first "animal models" used in identifyingpotential drugs with antipsychotic activity

Page 485: 0195162854

468 MODELS AND TESTS

Qanssen et al., 1966]), but its status is moreappropriately labeled as a "behavioral assay"of stimulation of dopaminergic receptors(Szechtman et al., 1988) and, for this reason,as a test that is indirectly sensitive to any psy-chopathology associated with dopaminergichyperactivity. In essence, such "behavioral as-says" conform to "standard tests" of psycho-logical or neurobiological function as used inthe first type of paradigm described earlierbut do not constitute an animal preparation(model) of the psychopathology.

TOWARD AN ANIMAL MODEL OFOBSESSIVE-COMPULSIVE DISORDER

In this section we use a specific example ofan animal model of obsessive-compulsivedisorder (OCD) to evaluate its characteristicswith reference to the criteria of Abramsonand Seligman (1977) and to highlight themethods and principles relevant to the de-velopment of an animal model of a psychi-atric disorder. The particular rat modelseems strongest in addressing the similarityof symptoms and cure between the modeland human OCD (criteria 2 and 3 of Abram-son and Seligman, 1977; see earlier). Conse-quently, we discuss this model with refer-ence to these criteria first.

The model of OCD (Szechtman et al.,1998) is produced by chronic treatment of ratswith the D2/D3 dopamine agonist quinpirole.Rats receive repeated injections of the drug(0.5 mg/kg twice weekly for 10 doses) in alarge, open field with four small objects(boxes) in it. On the last injection, the animals'behavior is evaluated and compared withthe behavior of rats treated chronically withsaline. The authors used the results oftheir behavioral analysis to stake the claimthat quinpirole-treated rats show compulsivechecking and that this is a symptom like theone shown by patients with OCD. The in-vestigators provided two lines of evidence toshow this similarity.

The first line of evidence came from acomparison of the structure of compulsivechecking in the rat with that of the human.It was reasoned that whatever is the natureof the mental aberration in OCD, it must findexpression in a characteristic pattern of be-havior that is readily observable. Unfortu-nately, no ethological descriptions of humanOCD compulsions were available. Conse-quently, the authors had to surmise what thespatiotemporal structure of compulsivechecking is likely to be. For this they drew onthe type of information solicited from pa-tients by psychiatrists to identify OCD symp-toms and on attempts in the psychiatric lit-erature to characterize the essential featuresof clinical compulsions (Reed, 1985). Basedon these sources, it became apparent thatsalient features of compulsions include (1)preoccupation with the performance of com-pulsions and a reluctance/resistance to en-gage in them, (2) a "ritual-like" quality in theperformance of the compulsions, and (3)close coupling between OCD motor ritualsand environmental stimuli/context. Accord-ingly, the authors suggested that a checkingcompulsion would reflect itself as (1) a pre-occupation with and an exaggerated hesi-tancy to leave the item(s) of interest in one'sterritory, (2) a ritual-like motor activity pat-tern and (3) dependence of checking behav-ior on environmental context. Consequently,the authors proposed that the following fiveperformance criteria identify the spatiotem-poral structure of compulsive checking(Szechtman et al., 1998, p. 1477):

1. In the subject's territory, there would be one ortwo places/objects to which the subject returnsexcessively more often than to other places/objects in the environment.

2. The time to return to these preferred places/objects would be excessively shorter than thetime to other places/objects.

3. Excessively fewer places would be visited be-tween returns to the preferred places/objects.

4. A characteristic set of acts would be performedat the preferred place/object, which would dif-

Page 486: 0195162854

Chapter 43. Psychiatric Models 469

fer from the acts performed at other loca-tions/objects.

5. Activity would be altered when the en-vironmental properties of the places/objectsare changed.

Quinpirole rats met all of the five criteriafor compulsive checking. As shown in Figure43-1, their behavior indicated a preoccupationwith and a reluctance to leave a particularlocation/object in the open field (the homebase), as measured by the first three ofthe proposed criteria. Moreover, their be-havior had a ritualistic quality in the preferredplaces/objects (Table 43-1). Finally, when anobject was moved to a new location, theirchecking activity did shift to the new locationas well as producing other changes in behav-ior (Szechtman et al., 1998).

The second line of evidence for similar-ity between quinpirole-induced compulsivechecking and the symptom of OCD checkinginvolved comparison of the possible motiva-tional basis of this activity in the rat model andthe human condition. In the human, compul-sive checking appears motivated by concernsof safety and security and is seen as an exag-gerated form of normal checking regardingone's well-being and security (Reed, 1985).The characteristics of the checking behaviorin the rat model were consistent with havinga similar motivational basis. In particular, ratsnormally do engage in checking behaviors("risk assessment"; Blanchard, 1997) and thebehavior of quinpirole-treated rats was di-rected at a likely stimulus for checking activ-ity—the home base (Eilam and Golani, 1989).Thus, quinpirole checking was attached tostimuli with a plausible relationship to safetyand security and in this regard could be calledan exaggerated form of normal checking inthe rat, similar to the human condition.

The similarity of symptoms in the ratmodel and human OCD was strengthened bythe additional finding that quinpirole-treatedrats, like OCD patients, may temporarilydesist from engaging in compulsive check-ing (Szechtman et al., 2001). Criterion 3 of

Abramson and Seligman (1977)—similarity of"cure"—was shown by the effectiveness ofclomipramine, a drug used in the treatmentof human OCD, to partially attenuate quinpi-role-induced checking behavior (Szechtmanet al., 1998). Thus, in several respects, there isa strong similarity in the symptoms displayedby the quinpirole model and those foundin human OCD psychopathology. However,this similarity would be strengthened byshowing directly that the spatiotemporalstructure of human OCD compulsions doeshave the same form as observed in the rat. Wehave begun to conduct such "ethological"studies and show the results of analyzing acompulsive ritual of one OCD patient (Table43-2). As is evident, the structure of that rit-ual, even though it is of cleaning the noserather than being a checking ritual, is re-markably similar in its form to the compul-sive checking behavior of the quinpirole-treated rat (compare with Table 43-1).

With reference to criterion 1 of Abram-son and Seligman (1977) regarding the "es-sential features" of the quinpirole model, thisis a question that is currently not resolved.Quinpirole is a dopamine agonist, so clearlyan important aspect of the "cause" of OCDsymptoms in the model is chronic activationof dopamine receptors. However, it must beexamined whether activation by only the drugproduces compulsive checking (in which casethe quinpirole preparation would be only amodel of drug-induced OCD) or whether useof the drug represents a more general class ofparticular environmental conditions that hy-peractivate the dopamine system(s). Likewise,how important is the fact that rats experiencethe drug injections in a large open field—whatdoes this indicate about the induction of com-pulsive checking? From other studies in whichthe locomotor response to quinpirole was ex-amined (Szechtman et al., 1993; Einat andSzechtman, 1993; Einat et al., 1996; Szumlin-ski et al., 1997), it is clear that the nature ofthe environmental context has a profound in-fluence on the degree of locomotor sensitiza-

Page 487: 0195162854

470 MODELS AND TESTS

Checking of Home Base

Figure 43-1. Induction of compulsive checking as identi-fied by formal performance criteria. Performance measuresare in reference to the home base established by each raton the tenth open field test shown here and recognized asthe locale with the longest total duration of stops. Gener-ally, it is also the most frequently visited place in the openfield. The open field is 160 X 160 cm without walls and sub-divided into 25 locales (places). Locomotor activity is meas-ured in relation to these 25 places. Quinpirole (QNP)-treated rats (gray bars) satisfied the first three criteriafor compulsive checking because, compared with saline-treated rats (open bars), they showed (A) more frequent re-turns to the home base, (B) a higher-than-expected rate ofreturning to the home base, (C) more rapid returns to thehome base, and (D) fewer visits to other places in the openfield on trips from the home base. Values are given as meanand SEM. All differences between quinpirole- and saline-treated rats are significant. Adapted from Szechtman et al.(2001).

tion to quinpirole, with the induction ofminimal sensitization when the drug is ad-ministered in the rats' home cage (Szechtmanet al., 1993; Szumlinski et al., 1997). There-fore, it is likely that the type of environmentin which the rat experiences chronic injectionsof quinpirole plays also an important role inwhether the rat does or does not developcompulsive checking. We suspect that to de-velop OCD rituals, the environment must besuch as to direct the rat's attention to eventsrelated to its safety and security, but this mustbe shown in formal studies.

As to whether the quinpirole model de-scribes all instances or only a subtype of OCD(criterion 4 of Abramson and Seligman, 1977),it seems more likely that it is only a subtypeof OCD that is being modeled. This opinion

is based on the observation that the effects ofclomipramine on quinpirole checking werenot lasting, as found for a subpopulation ofOCD patients (Szechtman et al., 1998). Thus,there seem to be different populations ofOCD patients, and the quinpirole model mayrepresent the subtype that is less sensitive tothe effects of clomipramine. However, this isanother aspect of the model that needs inves-tigation. Still other major aspects of the modelthat need examination concern how the neu-robiology of checking in the quinpirole modelrelates to the neurobiology of the human dis-order. In this endeavor, because knowledgeabout the neurobiology of the human psy-chopathology is relatively rudimentary, it isequally likely that the rat model will be usedto generate hypotheses to be examined in hu-man OCD. Indeed, the usefulness of an ani-mal model lies not only in its convenience ofproviding a physical preparation to investigatethe mechanisms of a disorder but also in itsability to generate novel hypotheses that canbe examined in the clinic.

CONCLUDING REMARKS

Until quite recently, there was great skepti-cism in psychiatry regarding the usefulness ofanimal models of human psychiatric disor-ders. Now, however, the stage is set and wideopen for the development of animal modelsfor virtually any psychiatric disorder. Thisprofound shift in attitude reflects largely theacceptance of the viewpoint that malfunctionof mental life may be embodied in a break-down of normal brain function, an attitudeembraced by the rising field of biological psy-chiatry. Acceptance of animal models followsas a necessary corollary of this viewpoint be-cause such models not only provide the phys-ical preparation to investigate the mecha-nisms of normal and disordered mental lifebut also acknowledge the evolutionary rootsof humans in the continuity of biological life.Of course, development of animal models of

Page 488: 0195162854

Table 43-1. A Record of 2 Minutes of Activity of a Rat in a Large, Open Field, 40 Minutes After the Tenth Injection of Quinpirole

Location in the open field

Trip

12345

6

78

9101112

Home Base (Corner Locale With Box)

A3

A3

A3

A3

A3

A3

A3

A3

A3

A2

A3

A3

Vu

Vu

vuVu

Vu

Vu

Vu

Vu

Vu

Vu

Vu

Vu

T4

T4

T4

T2

T2

T2

T2

To

To

To

T2

T2

vuvdvdVd Rc6 S Rc2 T2

Rc5 S Ra4 Vd T2

Ra2 Rc6 Vu Rc3

Vu T0 Rc4 Vd

Ra2 Vu T2

Ra2 Rc6 Vu

Ra4 Ra2 Vu T2

vdvu

Rc5 Ra3 Vd Vu Vd

Vu T4 Vd Rc4 Vd

Rc6

Rc6

Rc6

Rc6

Rc6

Rc7

Rc7

Rc7

Rc6

Rc6

Rc6

Rc6

SSSSS

S

SS

SSSS

Trot6Trot6

Trot6

Trot6

Trot6

Trot7

Trot7

Trot7

Trot6

Trot6

Trot6

Trot6

A2

A2

A2

A2

A2

A2

A2

A2

A2

Edge Locale

RajRa0

Ra0

Ra0

Rax

Ra2

RajRc2

Ra!

TrotjTrot0

Trot0

Trot0

Troti

TrotjTrot2

Trot,

A5

A4

A4

A4

As

A3

A3

A3

A5

As

Center Locale With Box

Rc3

Rc3

T2Rc3

Rc3

Vu TO Rc3

Rc7 Vu Ra6 Vu

T0Rc2

Vu TO Ra3

Rcj Vu TO Rag T4

Vu Vu T4 Rc3

Vu T4 Rc3

Vu T4 Rc3

Trot3Trot3

Trot3

Trot3

Trot3

Trot3

Trot3

Trot3

Trot2

Trot3

Trot3

A indicates arrival, with the direction of arrival indicated by the subscript numeral and specified in units of 1 = 45°; V, vertical movement of the forequarters (u, upward; d, downward); T, establishing contact with thebox (4, climbing on top of it with all four legs; 2, leaning on it with the forelegs; 0, only the snout makes contacts with the box); Re, rotation (turning) clockwise to one of eight specified directions in intervals of 1 = 45°; Ra, ro-tation (turning) anticlockwise to one of eight specified directions in intervals of 1 = 45°; S, stepping down from the box; Trot, running in a trot gait to the specified direction, 1 = 45°.Note: The record describes the sequence of movements performed by the rat during 12 successive round trips from the home base, which was a corner locale of the open field that contained an object (small box). At this object,the rat displayed a typical set of acts starting with arrival (A3), vertical movement (Vu), and touching the box (T). The sequence of movements at this locale always ended with turning clockwise to direction 6 or 7 (Rc6 or Rc7),stepping down from the box (S), and trotting in that direction. On trips 1 to 5 and 9 to 12, the rat then stopped at an edge locale, where it made an anticlockwise turn (Ra; except for trip 11, when a clockwise turn occurred). Itthen ran to a center locale with an object in it and moved around the object, often making a vertical movement (Vu) and touching the object (T). It then trotted back to the home base, where it begun another trip. On trips 6 to8, the rat did not stop at the edge locale but ran directly from the home base to the center locale, where it displayed the usual ritual and then returned to the home base. Thus, although there is a certain amount of flexibilitywithin and across round trips, the behavior of this rat is a highly structured and repetitive ritual.

Page 489: 0195162854

472 MODELS AND TESTS

Table 43-2. Four Repetitions of a Ritual of Blowing and Wiping the Nose in anObsessive-Compulsive Disorder Patient

Repeat Act

1

2

3

4

BlowWipe

Blow

Wipe

BlowWipe

BlowWipe

Blowing or Wiping

the Nose

Taking a Tissue Paper and /or Mouth

H C Pr F Ra PiH C Pr F RcRa Pi

PiPI

H C Pr F Ra Pi

H C Pr F Re Ra PiPIPI

Missing on tapeH C Pr F RaRc Pi

PIPI

H C Pr F Ra PiH C Pr F RaRc Pi

PIPI

Wi X 4W2

W3 X 4W4 X 4

Wj X 4

W 2 X 2W3 X 4W4 X 4

W 2 X 4W3 X 4W4 X 4

W!

W2 X 2W3 X 4W4 X 2

SSSS

S

SSS

SSS

SSSS

Handling the Used Tissue Paper

ReFo

RaRc FRe F

Re F Ra F

FoRe Re F

F

FoRe Re F

F

RcRa FFo

RcRa FF

Pu

Pu Cleaning shirtand hands

Pu Cleaning shirtand hands

Pu Cleaning shirtand hands

Pu Cleaning shirtand hands

Pu

Pu Cleaning shirtand hands

Blow indicates the entire first row in each repetition and is a label for taking a piece of tissue paper, blowing the nose, and putting away thetissue; Wipe, the remaining rows in each repetition and a label for taking a piece of tissue paper, wiping the nose, upper lips, and under the nose,and putting away the tissue; H, holding horizontally in the left hand the roll of tissue paper; C, turning the roll so that two pieces of tissue hangdown, holding them between the thumb and the other four fingers, and detaching the two pieces—the act takes place with what appears to be highconcentration, with the eyes following every movement being performed; Pr, placing the tissue roll back on the dresser, with the left hand; F, hold-ing vertically the tissue piece and folding it along a line perpendicular to the ground, and then smoothing the fold with two fingers from top to bot-tom; Fo, flipping over the folded piece of tissue; R, rotating the piece of the tissue by 90° clockwise (Re) or anticlockwise (Ra); PI, placing the pieceof tissue on the body part to be wiped. The tissue covers the nose and the mouth, is held first with only one finger and then by four ringers at eachside; WL blowing the nose strongly after a deep breadth and blowing the air intensely so that two holes are formed in the tissue—piping movementsoccur in the fingers starting with the median and ending with the lateral fingers; Wa, wiping the upper lip while moving the head to the right andopening the mouth, and then wiping the lower lip while moving the head to the left and closing the mouth; W3, wiping the nose while moving thehead up and down; W4, wiping under the nose while moving the head up and down with closed eyes; S, staring at the tissue with high concentra-tion for several seconds; Pu, placing the used piece of tissue, with the left hand, on top of the previously used tissues; X 2 or X 4, two or four rep-etitions, respectively.

Note: The record shows four repetitions of a ritual of blowing and wiping the nose observed during a 1 hour videotaped interview in the pa-tient's home. Each repetition of the ritual took approximately 4 minutes. Although there is a certain amount of flexibility within and across repeti-tions of the ritual, behavior is highly structured and repetitive, as found in the quinpirole-treated rat in Table 43-1.

psychopathology is a complex task that atevery stage must be evaluated with rigor, butthe usefulness of such models is undisputableby providing both precise control over hered-ity and experience and the opportunity to con-stantly monitor the disorder (Abramson andSeligman, 1977). In developing the models,the challenge for behavioral neuroscientists

lies, on the one hand, in applying their knowl-edge of psychology and skills in behavioralanalysis to construct tests and methods ofmeasurement that capture the expression ofnormal and disordered mental function. As il-lustrated in the rat model of OCD, an etho-logical approach that focuses on recording thespatiotemporal structure of behavior provides

Page 490: 0195162854

Chapter 43. Psychiatric Models 473

a natural window on mental life that lends it-self readily to comparisons of the form of be-havior across species as diverse as the humanand the rat. The challenge is equally de-manding, on the other hand, in using theavailable neuroscience tools to manipulate thenervous system in a meaningful fashion andthereby identify which perturbations of thebrain yield disorders of psychological functionand thereby psychiatric disorders.

ACKNOWLEDGMENTS

This work was supported by the Ontario Mental Health Foun-dation, the Canadian Institutes of Health Research, and the Is-rael Science Foundation. H.S. is a Senior Research Fellow ofthe Ontario Mental Health Foundation.

REFERENCES

Berrios GE and Markova IS (2002) Conceptual issues. In:Biological psychiatry (D'haenen HAH, Boer JA,Willner P, eds.), pp. 3-24. Chichester: Wiley.

Blanchard DC (1997) Stimulus and environmental con-trol of defensive behaviors. In: The functional be-haviorism of Robert C. Bolles: Learning, motivationand cognition (Bouton M and Fanselow MS, eds.),pp. 283-305. Washington, DC: American Psycho-logical Association.

Chapanis A (1961) Men, machines, and models. Ameri-can Psychologist 16:113-131.

Clark A (1980) Psychological models and neural mech-anisms: An examination of reductionism in psy-chology. Oxford: Clarendon Press.

Eilam D and Golani I (1989) Home base behavior of rats(Rattus norvegicus) exploring a novel environment.Behavioural Brain Research 34:199-211.

Einat H, Einat D, Allan M, Talangbayan H, Tsafhat T,Szechtman H (1996) Associational and nonassocia-tional mechanisms in locomotor sensitization to thedopamine agonist quinpirole. Psychopharmacology127:95-101.

Einat H and Szechtman H (1993) Environmental mod-ulation of both locomotor response and locomotorsensitization to the dopamine agonist quinpirole.Behavioural Pharmacology 4:399-403.

Geyer MA, Braff DL, Swerdlow NR (1999) Startle-response measures of information processingin animals: Relevance to schizophrenia. In: Ani-mal models of human emotion and cognition(Haug M and Whalen RE, eds.), pp. 103-142.

Washington, DC: American Psychological Asso-ciation.

Geyer MA and Markou A (1995) Animal models of psy-chiatric disorders. In: Psychopharmacology: Thefourth generation of progress (Bloom FE andKupfer DJ, eds.), pp. 787-798. New York: RavenPress.

Janssen PA, Niemegeers CJ, Schellekens KH (1966) Is itpossible to predict the clinical effects of neurolep-tic drugs (major tranquillizers) from animal data?Arzneimittel-Forschung 16:339-346.

Laing RD (1967) The politics of experience. Harmonds-worth: Penguin.

McKinney WT Jr and Bunney WE Jr (1969) Animalmodel of depression. I. Review of evidence: Impli-cations for research. Archives of General Psychiatry21:240-248.

McKinney WT (1988) Models of mental disorders: Anew comparative psychiatry. New York: PlenumMedical Book Co.

Miller WR and Seligman ME (1973) Depression and theperception of reinforcement. Journal of AbnormalPsychology 82:62-73.

Reed GF (1985) Obsessional experience and compul-sive behaviour: A cognitive-structural approach.Orlando, FL: Academic Press, Inc.

Seligman ME (1972) Learned helplessness. Annual Re-view of Medicine 23:407-412.

Szasz TS (1961) The myth of mental illness: Foundationsof a theory of personal conduct. New York:Hoeber-Harper.

Szechtman H, Eckert MJ, Tse WS, Boersma JT, BonuraCA, McClelland JZ, Culver KE, Eilam D (2001)Compulsive checking behavior of quinpirole-sensi-tized rats as an animal model of obsessive-compul-sive disorder (OCD): Form and control. BMC Neu-roscience 2:4.

Szechtman H, Eilam D, Ornstein K, Teitelbaum P,Golani I (1988) A different look at measurement andinterpretation of drug-induced behavior. Psychobi-ology 16:164-173.

Szechtman H, Sulis W, Eilam D (1998) Quinpirole in-duces compulsive checking behavior in rats: A po-tential animal model of obsessive-compulsive dis-order (OCD). Behavioral Neuroscience 112:1475-1485.

Szechtman H, Talangbayan H, Eilam D (1993) Envi-ronmental and behavioral components of sensitiza-tion induced by the dopamine agonist quinpirole.Behavioural Pharmacology 4:405-410.

Szumlinski KK, Allan M, Talangbayan H, Tracey A,Szechtman H (1997) Locomotor sensitization toquinpirole: Environment-modulated increase in ef-ficacy and context-dependent increase in potency.Psychopharmacology 134:193-200.

Page 491: 0195162854

474 MODELS AND TESTS

Teitelbaum P and Pellis SM (1992) Toward a synthetic Willner P (1984) The validity of animal models of de-physiological psychology. Psychological Science pression. Psychopharmacology 83:1-16.3:4-20. Willner P, Muscat R, Papp M (1992) Chronic mild stress-

Teitelbaum P and Strieker EM (1994) Compound com- induced anhedonia: A realistic animal model of de-plementarities in the study of motivated behavior. pression. Neuroscience and Biobehavioral ReviewsPsychological Review 101:312-317. 16:525-534.

Page 492: 0195162854

Neuropsychological Tests

GERLINDE A. METZ, BRYAN KOLB,AND IAN Q. WHISHAW

44

Having examined the behavior of the rat over43 chapters focused on a wide range of be-haviors, we must now try to present a simple,yet comprehensive battery of tests that couldbe recommended to provide a starting pointfor the description of the behavior of the rat.Obviously, particular chapters on specific top-ics such as sex or eating would be the place tostart if there were a reason to suspect changesin such behaviors, but often that is not thecase. For example, if one were interested inthe effect of a genetic manipulation or perhapsin the effect of a drug to be used for the treat-ment of bowel cancer, the place to begin a be-havioral analysis would be with a broad bat-tery of tests that would serve to point theinvestigator to further examination with morespecific tests. The behavioral repertoire of ananimal, or its ethogram, is a summary of allof its innate and learned behaviors. Such a be-havioral assessment combines principles ofthe neurological examination, developed forthe assessment of the motor behavior ofhumans, and the neuropsychological assess-ment, developed for the assessment of humanpsychological functions. These assessmentprocedures have been tailored for the rat.

A neuropsychological examination of ratbegins with a description of general appear-ance and sensorimotor responsiveness andcontinues by examining posture and immo-bility, locomotion, skilled movements, species-specific behavior, and learning (Table 44-1).It may not always be necessary to evaluate allof the behavioral categories, or to use all of

the tests, the guiding principle being the an-swer to the question, "What do you want toknow?" Thus, for example, studies of spinalcord function are unlikely to include measuresof visual acuity. Nevertheless, one cannot al-ways be certain about what one needs toknow. Accordingly, it is often necessary touse the principles of differential diagnosis, inwhich both unimpaired and impaired func-tions are described. This chapter provides anoverview of these seven main categories of be-havioral assessment and describes some indi-vidual tests of behaviors in each category. Thechapter is broken into seven sections that pro-vide overviews on a selection of behavioraltests useful in assessing that behavioral cate-gory. Representative references are providedfor each test, but the reader is directed to spe-cific chapters earlier in the book for more ex-tensive reference lists. The test descriptionsare intended to present a summary of a giventest procedure rather than presenting the ap-paratus and procedural details; these can beobtained in cited methods references.

It is important to note that there is nostatic test or test battery. Existing tests areconstantly being revised and adapted to thedemands and purpose of an experiment andto the type and severity of deficit predicted af-ter a neurological insult (Whishaw et al., 1983,1999). Furthermore, these tests are also beingrefined as understanding of the behavior ofthe rat and its neural substrate grows. The in-dividual tests and assessment parameters mayneed to be modified to fit the specific research

475

Page 493: 0195162854

476 MODELS AND TESTS

Table 44-1. Categories of Behavior andNeuropsychogical Assessment

Category Tests Described

Appearance andresponsiveness

Sensory andsensorimotorbehavior

Posture andimmobility

Locomotion

Skilled movements

Species-specificbehavior

Learning

General examination of healthTests of orientingFormalin pain testHot and cold temperature testvon Frey hair testSticky dot testLimb-use asymmetry

(cylinder) testPostureRighting responsesSwimmingWalking and runningCircadian activityBBB scaleFood handlingTray-reaching taskPellet-reaching taskRung and beam walkingGroomingFood hoardingExplorationPlay behaviorSexual behaviorWater taskRadial arm mazeBarnes mazeElevated-plus mazeContext conditioningOne-way avoidanceSpontaneous alternationTwo-choice association

protocol of the investigator to reveal param-eters that best predict chronic outcome orlong-term consequences of a manipulation.Furthermore, an optimal assessment woulduse a number of measurements from the testbattery, each describing specific aspects of be-havior. Therefore, a combination of differenttests is used to gain a detailed and broad func-tional profile of an animal.

In addition to the selection of appropri-ate tests, the method of analysis determinesthe sensitivity of the individual measures. Pa-rameters can include end point measures toquantify the frequency of occurrence of a spe-cific event. Kinematic measures reconstruct a

movement to obtain information about its di-rection, velocity, or angle. Descriptive analy-sis of movement, which can include use of aformal language, describes the movement ofbody segments and the intersegmental rela-tions. These methods of analysis will be men-tioned in the description of the individual testprocedures. It is important to remember thatdifferent measures may not necessarily corre-late. A rat may perform a task in the correctway but have a terrible end point score or, al-ternatively, its performance may be terriblebut its score excellent. Thus, single measuresseldom capture function. As a rule of thumb,in initial screening, we favor observational de-scriptions along with end point measures.Obviously, because a test battery procedureinvolves the necessary use of multiple com-parisons, some differences may occur bychance. Thus, we prefer to identify deficitsonly after we obtain a number of convergingmeasures.

One of the problems in assessing behav-ior is that if a test procedure is laborious ortime consuming, it will torpedo the usefulnessof a test battery. Consequently, most of thetests described here are relatively simple andcan be administered fairly quickly. Further-more, the tests can be given concurrently sothat an extensive battery of tests can be givenin a relatively short time. In addition, there isa growing interest on the part of test design-ers to develop automated systems of analysis.So, for example, a photocell automated openfield test can be administered in 10 minutesand can provide a good range of measures, in-cluding overall activity, habituation, rearingactivity, thigmotactic tendencies, and turningbiases. Such a test is sensitive to motor ab-normality or the laterality of a motor impair-ment. Because rats are typically more activein the dark than in the light, an open field ac-tivity test repeated in the light and dark mightalso reveal something about a rat's visual sen-sitivity. Video-based automated monitoringsystems can also reveal the organization of ex-ploratory behavior in single animals or in an-

Page 494: 0195162854

Chapter 44. Neuropsychological Tests 477

imal pairs, thus providing a measure of socialbehavior.

Finally, we reiterate that the tests outlinedin the following pages represent a general setof tests that are useful as a tool to generate abehavioral phenotype. Specific chapters earlierin this volume provide more detailed and spe-cific measures of specific behavioral functions.

APPEARANCE AND RESPONSIVENESS

Much can be learned by simply examining thegeneral appearance and responsiveness of an-imals (Table 44-2). This can be achieved byexamining them in their home cage, by plac-ing them in a novel environment, and by han-dling them. The analysis of appearance andresponsiveness provides important clues onwhat other types of measures may be appro-priate. For example, long claws may reflect adifficulty in making fine movements of themouth that are necessary for claw trimming,and weeping eyes may indicate the accumu-lation of Hardarian material and thus poorgrooming (Whishaw et al., 1983). The extentof a physical abnormality can be quantified byusing a small ruler, such as to measure clawlength. In the course of examining strain dif-ferences in spatial behavior of rats, we ob-served that male Dark Agouti rats, an inbredpigmented rat strain, were smaller than Long-Evans rats, tended not to grow as quickly, andhad somewhat delicate heads and long snouts.Their appearance suggested that the process of

inbreeding had feminized the animals (Harkerand Whishaw, 2002). Feminization may havecontributed to their poorer spatial perform-ance than that obtained in Long-Evans rats.Thus, the observation of physical appearancelead to hypotheses concerning function. Thisin turn may lead to experimentation in whichmales are given additional testosterone to de-termine if both general morphology and be-havior might become more like that of themale Long-Evans rat.

It is useful to make up a checklist of ap-pearance features and behaviors as a way ofstandardizing the appearance examination. Wethen recommend a five-step sequence for theassessment of appearance and responsiveness.

1. The animal is observed in its home environ-ment. It is useful to have a test animal housedwith, or beside, an experimental animal so di-rect comparisons can be made. Observationsare made on the animals' fur, eyes, and feet todetermine whether they are clean, a featurethat indicates whether the animal is grooming.

2. The animal is given some simple tests of re-sponsiveness in its home environment. Objectscan be introduced into an animal's cage to ob-serve its responsiveness to novel stimuli, and asubstantial battery of sensorimotor tests can begiven while the animal is in its home cage (seenext section).

3. The animal is removed from its home cage andobserved in a novel environment. Again it isuseful to have a yoked control animal as a stan-dard against which to compare an experimen-tal subject. An animal's behavior can be quitedifferent in a safe versus novel environment.

Table 44-2. Examination of Appearance and Responsiveness

Appearance

Cage side examination

Handling

Body measurements

Inspect body shape, eyes, vibrissae, limbs, fur, tail, and coloring.

Examine the animal's cage, including bedding material, nest, food storage, anddroppings.

Remove the animal from its cage, and evaluate its response to handling, includingmovements and body tone, and vocalization. Lift the lips to examine teeth,especially the incisors, and inspect the digits and toenails. Inspect the genitalsand rectum.

Weigh the animal and measure its body proportions (e.g., head, trunk, limbs, andtail). Measure core temperature with a rectal or aural thermometer.

Page 495: 0195162854

478 MODELS AND TESTS

4. The animal is given some simple tests in thenovel environment (see next section). These in-clude picking up the animal, assessing its mus-cle tone, examining its placing responses, dan-gling it by the base of the tail to observe itsorientation in relation to gravity, and its re-sponse to the experimenter.

5. Specific morphological measures are made.The animal is weighed, and its eyes, ears, vib-rissae, head shape, body shape, and tail shapeare examined. Using a small plastic ruler, it isuseful to directly measure physical features.

SENSORY ANDSENSORIMOTOR BEHAVIOR

The object of sensory tests is to assess the gen-eral functional status of the sensory systems,including somatosensory, visual, and olfac-tory functions. Such measures may include aquick assessment both in the home cage andin an open space using the procedures out-lined in Table 44-3, or they may involve spe-cific tests outlined later, or more detailed testsprovided in specific chapters elsewhere in thevolume. It is important to note that informalobservations, during which an animal is sim-ply observed along with its control partnerseveral times a day, can provide good obser-vations that may later be refined into moreformal tests. Many skilled behaviors can alsobe observed in the home cage. By presentingstrips of paper each day, nest building behav-ior can be observed, especially in female rats.By presenting sunflower seeds, food handlingcan be examined, and by presenting tasty food

on a spatula, tongue use can be examined. Ex-amination of the home cage can also provideclues about cognitive and motor status. Forexample, one can determine if food is chewedand eaten efficiently or if there are excessivefood fragments and crumbs suggesting motorimpairments in eating. Similarly, normal spa-tial behavior is suggested by compartmental-ization of sites for elimination, eating, andsleeping. More formal tests involve removingthe animal from its home cage; these includeCO somatosensory and pain testing using thevon Frey hair test, or the formaline test (2) theformalin pain test, (3) thermal sensitivityusing the hot and cold plate tests, (4) so-matosensory detection and skilled movementusing the sticky dot test, and (5) placing re-sponses using the cylinder test.

THE VON FREY HAIR TEST

A vital modality in rats is their haptic sense,essential to their nocturnal life style. Rats usetactile receptors on their body, including theirvibrissae, special receptors on their paws, andthe many large "sinus" hairs on their body.When combined together, the informationfrom these haptic receptors provide the ani-mals with an acute sense of their surround-ings. The responsiveness to tactile stimulationcan be evaluated by mechanically stimulatingthe skin of the animal with various objects. Acalibrated monofilament, the von Frey hair, canbe used for threshold sensitivity evaluation.

In laboratory rodents, von Frey hair testshave been used to evaluate the effects of le-

Table 44-3. Examination of Sensory and Sensorimotor Behavior

Home cage Response to auditory, olfactory, somatosensory, gustatory, vestibular, and visual stimuli. Thehome cage should provide easy viewing of an animal. Holes in the sides and bottom of the cageprovide entry for probes to touch the animal or to present objects to the animal or to presentfood items. Animals are extremely responsive to inserted objects and treat capturing the objectsas a "game." Slightly opening an animal's cage can attenuate its response to introduced stimuli,showing that it notices the change.

Openfield Response to auditory, olfactory, somatosensory, taste, vestibular, and visual stimuli. The sametests are administered. Generally, animals taken out of their home cage are more interested inexploring and so ignore objects that they responded to when in the home cage.

Page 496: 0195162854

Chapter 44. Neuropsychological Tests 479

sions and maladaptive plasticity on sensoryfunction as well as tactile sensitivity and painsensitivity. Before testing, the animal is habit-uated to the testing environment. A good en-vironment is a wire mesh cage or an elevatedplatform, either of which allows access to allparts of the animal's body. When the animalis still, specific locations on the skin canprobed with a monofilament. von Frey mono-filaments are calibrated to exert a force rang-ing from 5 to approximately 178 g/mm2. Thethreshold force for each monofilament is thepoint at which it bends. The animal's re-sponses of withdrawal or orienting toward thestimulus are recorded. Several scoring sys-tems have been developed that allow catego-rizing the animals' reaction (e.g., see Marshallet al., 1971).

There are two ways to use von Frey hairs.One way is to present different strengths ofmonofilaments to one particular location ofthe skin to evaluate the threshold at which theanimal starts responding to the stimulation ofa part of the body. The other way is to pres-ent one monofilament to multiple regions ofthe body to topographically map the segmentsof the body (dermatomes) to tactile stimula-tion. This approach is useful when determin-ing the severity and level of spinal lesions(Takahashi et al., 1995). The von Frey hairpinch test has been used in rat models of spinalcord injury, Parkinson's disease, and strokeand is sensitive to both spinal and supraspinalfunctions.

THE FORMALIN PAIN TEST

To assess analgesia (absence of pain percep-tion) or hyperalgesia (increased pain percep-tion) to mechanical, thermal, and chemicalstimulation, several standard pain tests havebeen developed. Chemically induced pain iscommonly assessed using formalin (Dubuis-son and Dennis, 1977). In the formalin paintest, animals are injected with a 3% or 5% for-malin solution subcutaneously into the dorsalsurface of the hindpaw. Behavior is monitored

for a time interval of up to 60 minutes afterthe injection. There are two phases in the an-imals' response to the injection: an early phaseand a late phase. The number of paw flinchesper time interval and paw licking are easilyquantifiable responses.

There are several confounding factorsthat can modify the response to formalin,including ambient temperature, stress, andnumber of treatments. Different strains of an-imals have a different stress response, thus re-stricting the validity of comparisons betweengroups of animals (Ramos et al., 2002). It isnoteworthy that repeated exposure to forma-lin might influence the animals' responsivityto other behavioral manipulations and subse-quent tests (Sorg et al., 2001).

HOT AND COLD TEMPERATURE TEST

Hot or cold plate tests are assessments of bothloss of sensory perception and plasticity-induced changes in pain thresholds. Temper-ature and pain senses are mainly mediatedvia the spinothalamic, spinoreticular, andspinomesencephalic tracts, also known as theanterolateral system. These tracts decussate atthe spinal level and ascend in the dorsal col-umn to the contralateral thalamus and the so-matic sensory cortex. Hot and cold tests makeuse of a range of temperatures to induce ther-mal sensations and induce a withdrawal re-sponse. A preferred apparatus is a heavycopper plate with heating or cooling coils dis-tributed equally below its surface, so that theentire plate is at a constant temperature.

One form of the test involves placing ananimal on a heated plate that is maintained ata temperature between 45° to 55° C or a coldplate of 4° C. Another form of the test in-volves briefly exposing the animal to a plateset at specific temperatures (e.g., an ascend-ing or a descending series) to determine the"threshold" at which an animal makes a re-sponse. Responses include orientation to thesite of the stimulation, flexion withdrawal re-flexes, licking of the paw, and a generalized

Page 497: 0195162854

480 MODELS AND TESTS

escape or attack response. These responsesare transient and occur only during applica-tion of the stimulus. The latency for appear-ance of heat avoidance behavior (orientation,lifting and licking the paw, withdrawal) andthe number of responses per time interval canbe recorded. To avoid injury, the trial in thehot plate test should be terminated quickly af-ter an animal responds.

Some commercial heat tests use infraredbeams that can be selectively pointed to a spe-cific part of the body, e.g., plantar foot or tail.Sensitivity thresholds to the heat are obtainedby gradually increasing the temperature (Al-masi et al., 2003).

The measures of withdrawal responseson a hot or cold plate can be used to screenanalgesic effects of drugs and substances ofabuse. Furthermore, they are part of standardtest batteries used to evaluate the extent ofspinal cord injury and supraspinal lesions.Structural rearrangements after a localizedtissue injury or drug treatment might alterthe pain threshold and excitability and thuschange the latency of the initial withdrawalresponse. Using the hot or cold plate tests,malfunctional plasticity and abnormal pro-cessing that contribute to the development ofneurogenic pain can be detected (e.g., Woolfet al., 1992).

STICKY DOT TEST

When adhesive-backed labels are attached tothe paw of a normal rat, the animal orients to-ward the stimulus and removes it using itsteeth. The latency and asymmetry in thebehavioral response can be used to assess so-matosensory function (Schallert and Whishaw,1984). To determine the presence of a so-matosensory asymmetry, an animal receivesadhesive stimuli (usually round sticky dots ofapproximately 12 mm diameter) attached tothe distal-radial aspect of each forelimb. Theanimal then is returned to a familiar cage andobserved. After being replaced in its cage, a ratquickly contacts and removes each label from

the forelimbs by using the teeth. The dots areremoved one at a time, and the latencies to con-tact and remove each stimulus, and the orderof removal (left or right first), are recorded overseveral trials (Schallert et al., 2000). The orderof contact and removal reflects whether the an-imal shows a bias. After a unilateral lesion af-fecting a forelimb, the animal may first contactand remove the dot from the good limb (thelimb unaffected by the injury). The animal laterorients toward the dot on the bad forelimb andmay take more time to remove that dot. If ananimal shows a somatosensory asymmetry inthe majority of the trials, the magnitude of theasymmetry can be assessed in a follow-up test(Schallert et al., 2000). In this test, the size ofthe dot on the bad limb is progressively in-creased while the size of the dot on the goodlimb is decreased by an equal amount. If thestimulus on the bad limb becomes more salientthan the stimulus on the good limb, the animalstarts contacting and removing the stimulusfrom the bad limb first. Studies using thismethod have shown that the ratio sufficient toreverse the initial bias is related to the severityof brain damage (Earth et al., 1990).

The sticky dot test has been shown to re-flect sensory impairments after unilateral le-sions involving the sensorimotor cortex, lateralhypothalamus, and subcortical lesions such aspyramidal tract transection and striatal lesions(Schallert et al., 2000). A critical issue in lesionstudies is that the damage of sensorimotorareas also results in impairments of motorfunction. Motor impairments are likely to leadto longer latency to remove the dot due todeficits in using the mouth and forelimb to re-move the sticky dot. This criticism might be re-solved by recording the contact time in addi-tion to the removal time. The contact does notrequire the skill required to remove the dot.

LIMB USE ASYMMETRY (CYLINDER) TEST

Rats explore both horizontal and vertical su-faces (Gharbawie et al., 2004). When explor-ing a vertical surface, the animal lifts its fore-

Page 498: 0195162854

Chapter 44. Neuropsychological Tests 481

quarters to support itself as it investigates.When bracing against the wall, either one orboth forelimbs are used. A normal animal typ-ically uses both forelimbs equally often forsupport, but a unilateral lesion may bias theanimal to prefer one limb (Schallert et al.,2000).

The cylinder apparatus can be placed ei-ther on a transparent bottom to allow video-recording from underneath the cylinder or ona table surface with a mirror arranged at anangle behind the cylinder to allow observa-tion of the behavior from all directions. Threemain categories of forelimb use are analyzed:lifting, moving along the wall, and landing.The forelimb used is recorded as the inde-pendent use of the left versus the right limb,or as the simultaneous use of both forelimbs.The standard measure is the number of leftand right forelimb uses calculated as per-centage of the total number of contacts(Schallert et al., 2000). Unilateral lesions that

affect limb use reduce the use of the impairedforelimb. Interrater reliability for the test isfound to be high (Schallert et al., 2000). Nev-ertheless, the sensitivity of this test to discretemotor disturbances might be reduced by thedevelopment of compensatory strategies,which could be misleadingly interpreted asrecovery.

POSTURE AND IMMOBILITY

Animals spend a great deal of waking timepartly or completely immobile. The postureof immobility is not always the same, how-ever, and certain postures can be consideredpathological. The initiation of movementfrom an immobile state can also reveal ab-normalities in posture and body support.Table 44-4 summarizes several gross meas-ures of posture and immobility. One test, thelighting response, is described here.

Table 44—4. Examination Strategies of Posture and Immobility

Immobility and movementwith posture

Immobility and movementwithout posture

Movement and immobility ofbody parts

Restraint-induced immobility

Righting responses

Environmental influences onimmobility

Animals usually have postural support when they move about, and theymaintain posture when they stand still and remain still while rearing. Postureand movement can be dissociated: in states of catalepsy, postural support is re-tained while movement is lost.

An animal has posture only with limb movement. When a limb is still, theanimal collapses, unable to maintain posture when still. When still, the animalremains alert but has no posture, a condition termed cataplexy.

Mobility and immobility of body parts can be examined by placing a limb inan awkward posture or placing it on an object such as a bottle stopper and tim-ing how long it takes an animal to move it.

Restraint-induced immobility, also called tonic immobility or hypnosis, isinduced by placing an animal in an awkward position, such as on its back. Thetime it remains in such a position is typically measured. Animals will maintainawkward positions while maintaining body tone or when body tone is absent.During tonic immobility, animals are usually awake.

Supporting, righting, placing, and hopping reactions are used to maintain aquadrupedal posture. When placed on side or back or dropped in a supine orprone position, adjustments are made to regain a quadrupedal position. Right-ing responses are mediated by tactile, proprioceptive, vestibular, and visualreflexes.

Feeding fatigue potentiates immobility. Warning induces heat loss postures,such as sprawling, and thus potentiates immobility without tone. Coolinginduces heat gain posture with shivering and thus potentiates immobility withmuscle tone.

Page 499: 0195162854

482

RIGHTING RESPONSES

Postural control is necessary for all types ofmotor performance. Postural adjustments de-pend on the position of the center of gravityin the body. In rats, the mechanics and con-trol mechanisms for postural maintenancewhen standing or walking are relatively sim-ple as compared to bipeds. When placed in aposition of unstable equilibrium or whenresponding to a passive displacement oftheir limbs, rats attempt to maintain a pronequadrupedal body position in relation to thecenter of gravity of the body.

The magnitude and ability of righting re-sponses in animals reflect their ability for sen-sorimotor integration. Sensory disruption ofproprioceptive, vestibular, tactile, or visual in-put can disturb the position of the body andbody parts in space. In turn, difficulties in re-sponding to the sensory stimulus of limb orbody displacement may be caused by the in-ability to recruit the appropriate musculature.

A number of tests for righting responsesmeasure specific aspects of sensorimotor in-tegration. The righting response most com-monly tested requires that an animal quicklyregain its quadrupedal position when placedon its side or back (stationary placing re-sponse). A normal animal first adjusts its headposition by dorsiflexing the head and neck andthen turns its forequarters of the body andlater the hindquarters to right itself. Whendropped facing upside down from the heightof less than 1 meter onto a cushion, the ani-mal will quickly right itself and adjust its pos-ture to land on all four feet (acceleratoryplacing response; Pellis and Pellis, 1994). Thisrighting response is mainly dependent onvestibular and visual cues. The sensory andmotor aspects of the righting response arequantifiable by the use of limbs and axial mus-culature, completeness of righting, and the la-tency to regain a prone body position.

Because righting responses are easy toassess in laboratory rodents and do not re-quire pretraining, they are often part of stan-

MODELS AND TESTS

dard test batteries in neurotoxicology andpharmacology. Furthermore, it has beenfound that righting in the air occurs at a de-fined time point during development (Hardand Larsson, 1975), and so righting responsescan be reliable indicators of maturation ofthe vestibular system. Righting responsesalso provide insight into processing mecha-nisms of motor control. For instance, an es-tablished phenomenon is an abnormality inregaining a prone body position after lesionsof the dopaminergic system (Martens et al.,1996). Although righting responses are usu-ally fast and require high-speed videorecord-ing, they provide reliable measures of sen-sory and motor system function.

LOCOMOTION

Analysis of locomotion includes observationof rats as they walk, run, jump, turn, and swim(Table 44-5). Virtually any manipulation ofthe brain will affect at least one of these meas-ures, although the effects are often subtle andmay require analysis of slow motion video-tapes. Locomotion can be measured with pho-tocells, running wheels, or automated videotracking systems and by measures of groundreaction forces. Ultimately, a refined descrip-tion can be obtained by using a movement no-tation system to document the movement ofevery joint and limb. One example is a systemdevised by Eshkol and Wachmann (1958) forclassical ballet and later adapted for describ-ing animal movements by Golani (1976). Inaddition, there are standardized tests of swim-ming, circadian rhythms, and there are com-posite tests such as the BBB scale, widely usedfor spinal cord injury.

SWIMMING

The swimming test takes advantage of the factthat rats are semiaquatic and thus are profi-cient swimmers. When rats swim, onlythe hind limbs are used for propulsion, while

Page 500: 0195162854

Chapter 44. Neuropsychological Tests 483

Table 44-5. Tests of Locomotion

General activity

Movement initiation

Turning and climbing

Walking and swimming

Exploratory activity

Circadian activity

Included are videorecording, movement sensors, activity wheels, and open field tests.

The warm-up effect: Movements are initiated in a rostral-caudal sequence, smallmovements precede large movements, and lateral movements precede forwardmovements, which precede vertical movements.

Components of movements can be captured by placing animals in cages, alleys,tunnels, etc.

Rodents have distinctive walking and swimming patterns. Rats and mice walk bymoving limbs in diagonal couplets with a forelimb leading a contralateral hindlimb.They swim using the hindlimbs with the forelimbs held beneath the chin to assist insteering.

Rodents select a home base as their center of exploration, where they turn andgroom, and make excursions of increasing distance from the home base. Outwardtrips are slow and involve numerous pauses and rears while return trips are morerapid.

Most rodents are nocturnal and are more active in the night portion of their cycle.Peak activity typically occurs at the beginning and the end of the night portion of thecycle. Embedded within the circadian cycle are more rapid cycles of eating and drink-ing, especially during the night portion of the circadian cycle.

the forelimbs are held immobile underneaththe chin or are tilted for steering. Althoughthe ability to swim is rarely abolished by ex-perimental manipulations, the performance ofswimming movements changes during devel-opment and aging, after certain types of braindamage, and after some drug treatments.Most of these manipulations lead to disinhi-bition of the forelimb use when swimming inthat animals make rhythmic strokes with theirforepaws. Such changes in swimming havebeen described after lesions of the cortex, pos-terior hypothalamus, and cerebellum (Kolband Whishaw, 1983; Whishaw et al, 1983).

The swim test requires a rectangular,transparent swimming pool with a water levelof approximately 30 cm. A visible platform isplaced on one end of the pool. The rat is in-troduced on the other end of the pool andquickly learns to swim to the platform (Stoltzet al., 1999). A few pretraining sessions beforetesting might be helpful to habituate the ani-mal to the procedure. To analyze limb use, ananimal's performance is videorecorded froma side view. The tapes can be analyzed on aframe-by-frame basis to count the number offorelimb strokes per limb. If the effects of uni-

lateral lesions are being assessed, the intactside serves as a control for the lesion side. Bi-lateral lesions, aging, or drug studies mightimpair both sides, so that data from controlanimals or preoperative test sessions will benecessary as a comparison.

CIRCADIAN ACTIVITY

The periodic changes of light intensity duringnight and day serve as a salient natural Zeit-geber in mammals. The light cycle regulates avariety of body functions, including cyclicbehavior such as sleep and wakefulness, andphysiological and biochemical processes, in-cluding body temperature and hormonalchanges. The circadian clock assists in organ-izing the time course of these processes, thusoptimizing the organism's performance inanticipating the rhythmic change of environ-mental conditions (Holzberg and Albrecht,2003).

Circadian activity is usually tested byrecording general activity in the home cageover one day and night cycle, but it can alsobe tested in constant light or in constant dark.Normal rats are active during the night por-

Page 501: 0195162854

484 MODELS AND TESTS

tion of a cycle even if the lights are on or offduring that portion of the cycle. Testing cir-cadian activity requires a separate room withlighting regulated by a timer. The activity ofanimals over a 24-hour time interval is usu-ally recorded by the interruption of photo-beams mounted on the walls of the home cageor by the number of turns of a running wheelthat is provided in the cage. The number ofinterruptions of photobeams or number ofturns of the running wheel are summarizedby a computer system that analyzes changesin the activity. Furthermore, the use of pho-tobeams allows the distinction between loco-motion and stereotyped movements. Succes-sive interruptions of different beams reflectthe animal's locomotion from one point to an-other, and successive interruptions of a spe-cific beam indicate stereotyped movementssuch as grooming or circling.

The circadian cycle of rodents usuallyshows a burst of activity when the lights turnoff followed by bouts of activity throughoutthe dark period, and another burst just beforelight turns on. During the light period, the an-imals are generally inactive. Animals that areplaced in a new environment also show in-creased activity during the first hour when ex-ploring the new environment. Detailed analy-ses of activity can reveal differences inexperimental groups depending on environ-mental conditions such as external light cyclesand feeding schedules, physiological conditionssuch as stress, aging, drugs, or brain dysfunc-tions (Weinert, 2000). Moreover, the comput-erized system makes it easy to record overall

changes in activity after various manipulationsand lesions to the brain.

BBB LOCOMOTOR RANKING SCALE

The BBB rating system, developed by Basso,Beattie, and Bresnahan (1995, 1996b), is usedto rate the degree of paralysis in rat spinal cordinjury models. The BBB scale is a widely usedmeasure of spinal cord integrity. The integrityof spinal function is reflected in interlimb co-ordination, alternating activity of limb mus-cles and rotational paw position, and posture.Lesions of the spinal cord usually not only af-fect movement but also interrupt descendingand ascending connections necessary to walkon difficult territory.

The 21-point BBB scale is based on a five-category score originally developed by Tarlov(1954) and has been refined to display discreteimpairments of locomotion. The BBB scalerates the magnitude of limb movement, posi-tion of trunk and abdomen, paw placementand stepping positions, coordination, toe clear-ance, paw rotation, trunk stability, and tail po-sition (Table 44-6). The 21 categories of therating scale allow conclusions about trunk sta-bility, weight support, and stepping and thusis tailored to the aim of preclinical studies inspinal cord injury.

Animals spontaneously explore a novelenvironment, and so no pretraining or othermotivation is ordinarily required to test ani-mals. There are, however, a few cautionarynotes regarding the testing. For example, fre-quent repetition of test sessions leads to ha-

Table 44-6,

Limb

Movement

None

Slight

Extensive

. Categories and Gait Features Assessed by the BBB Scale

Trunk

Position Abdomen

Side Dragor mid

Parallel

High

Paw

Placement

Sweeporsupported

Stepping

Dorsalorplantar

Coordination

Never

OccasionalFrequentConsistent

Toe

Clearance

Never

OccasionalFrequentConsistent

Paw

Position

Initialcontact

Liftoff

Trunk

Stability

AbsentorPresent

Tail

Up

Down

Page 502: 0195162854

Chapter 44. Neuropsychological Tests 485

bituation of the animals to the testing envi-ronment and can reduce the exploratory ac-tivity. Stepping patterns of paralyzed animalsalso can be modified by textured surfaces, likelybecause of reflexive limb movements.

The BBB scale is a well-standardized scor-ing system that has been validated for spinalcord injury in rats in large-sample studies(Basso et al., 1996a); the results have beenshown to correlate closely with results fromother motor tasks (Metz et al., 2000). The be-havioral abnormalities are directly related tothe loss of spinal tissue (Basso et al., 1996b),and the scale was recently adapted for theuse in mouse models (Ma et al., 2001), thusstrengthening the applicability of the BBBscale as a multipurpose tool.

SKILLED MOVEMENTS

Skilled movements are voluntary movementsthat require irregular motor patterns, rotatorymovement and movements that consist of acomplex sequence of movements, or move-ments that counteract movements that nor-mally move the body against gravity. Thesemovements are likely mediated by neuralmechanisms that are at least partly independ-ent of those that support locomotion. There-fore the quantification of skilled movementscan provide insights into neural function thatare at least independent of those obtained fromthe study of locomotion. At its simplest, theuse of skilled movements can be observed as

an animal eats laboratory chow, specialty fooditems such as sunflower seeds, which need tobe shelled, and prey items, such as crickets,which have to be caught and prepared for eat-ing. The movements used in these activitiesmay be difficult to score objectively, and somany researchers prefer more formal tests(Table 44-7). Formal tests have the advantagein that the same movement is performed re-peatedly and therefore can be scored with endpoint measures or qualitative measures. Trayreaching, single pellet reaching, and rung walk-ing are examples of such skilled movementtasks that assess forelimb function. Each testassesses specific forelimb abilities, such as ex-tending the forelimb, aiming and grasping, orapplying force to retrieve an object (Whishawet al., 1986; Montoya et al., 1991; Ballermannet al., 2001). The rung-walking task assesseshindlimb function as well as forelimb function.

TRAY-REACHING TASK

The tray-reaching task is a simple test of fore-limb use (Whishaw et al., 1986). The reachingbox consists of three solid walls and a frontwall made of thin vertical metal bars to allowthe animals to extend the forelimb throughthe full width of the front wall. Mounted onthe outside of the front wall is a tray that ismatched to the width of the wall. The traycan be filled with food, such as chicken feedor small food pellets, so animals can reachthrough the bars and retrieve food from anyangle and position in the box. The box is

Table 44-7. Skilled Movements and Test Methods

Limb movements

Climbing and jumping

Oral movements

Examples include bar pressing, reaching and retrieving food through a slot,spontaneous food handling of objects or nesting material, and limb movements usedin fur grooming and social behavior. Rodents use limb movements that areorder-typical and species-typical.

Examples include movements of climbing up a screen, rope, ladder, etc. and jumpingfrom one base of support to another.

Examples include mouth and tongue movements in acceptance or rejection of foodsuch as spitting food out or grasping and ingesting food, as well as movements used ingrooming, cleaning pups, nest building, and teeth and claw cutting.

Page 503: 0195162854

486 MODELS AND TESTS

mounted on a grid floor, so that lost anddropped food items fall through.

Training requires moderate food depri-vation. The training procedure is relativelysimple because there is no need to replacefood items or handle the animals while theyare being tested. Training sessions last about30 minutes, but test sessions can be as brief as5 minutes, during which performance can bevideorecorded for analysis. Animals can alsolive in the test apparatus, should that berequired. There are two main methods foranalysis of limb use and reaching perform-ance. First, animals can be allowed to use ei-ther forelimb and the preference of one limbover the other can be evaluated. Thus, whenunilateral lesions are assessed, the ratio of theuse of the good limb versus the bad limb re-veals the degree of limb use asymmetry. Sec-ond, to measure success rates of one limb ex-clusively, the use of the other limb can beprevented with a cuff made from adhesivefabric tape wrapped around the distal aspectof the limb (Whishaw and Miklyaeva, 1996).

For end point measures of success, a reachis defined as the insertion of a forepawthrough the bars of the cage, and a successfulreach is defined as a movement that obtainsfood and brings it to the mouth for con-sumption. Performance is then described ashit percent: the number total reaches dividedby successful reaches. Skilled forelimb move-ments in this task have been shown to bechronically affected by motor cortex lesion(Kolb et al., 1997) or corticospinal tract tran-section (Whishaw and Metz, 2002).

PELLET REACHING TASK

A sensitive test for fine motor control andbodily supporting adjustments in rats is thesingle pellet reaching task (Whishaw et al.,1991). The task is designed so that not onlycan measures of success be recorded but alsobehavior can be filmed and scored using stan-dard scoring procedures. The pellet-reachingtask requires that the animal extends its fore-

limb through a slit to aim the forelimb to asingle target, grasp a food pellet located on ashelf in front of the slit, for eating. Animals re-quire 1 or 2 weeks of training to habituate tothe apparatus and optimize the success ofreaching movements. From videorecordedperformance, a large variety of measures canbe obtained from a single test session such asend point measures of reaching success anddescriptive movement analysis of the limb andbody movement components (Whishaw etal., 1991). The sequence and performance ofthe components are relatively fixed so that theability to modify the elements of the reachingmovement to adapt to context is limited. An-imals with brain damage might return to base-line levels in their success rates, but the qual-itative analysis of reaching movements willreveal whether the movement strategy is per-manently changed.

In addition to the ample informationabout the organization of the motor systemprovided by the pellet reaching task, its rele-vance for investigations in animal models isestablished. Comparisons of rats with humans(Whishaw et al., 1992) demonstrates a ho-mology between the reaching movements inthe two species. Moreover, reaching move-ment abnormalities are similar between ro-dent models of human neurodegenerative dis-ease and human patients (Whishaw et al.,2002). Performance of reaching movementscan be compromised by even subtle braindamage that would spare coordinated move-ments such as locomotion, swimming, orgrooming. Discrete lesions of the motor sys-tem, including the motor cortex and its effer-ent corticospinal tract, basal ganglia, and rednucleus cause recognizable disturbances ofreaching movements (Whishaw et al., 1986;Metz and Whishaw, 2002a).

RUNG WALKING

Recent investigations also revealed that ratsuse skilled hindlimb movements to adapt theirgait pattern to a difficult territory. A simple

Page 504: 0195162854

Chapter 44. Neuropsychological Tests 487

and sensitive test is the rung-walking task(Metz and Whishaw, 2002b). The rung-walk-ing task resembles a horizontal ladder withrungs that can be adjusted individually. Therungs can be arranged in a regular pattern thatallows the animals to anticipate the locationof a rung or learn a specific sequence of pat-terns or in an irregular configuration to pre-vent animals from learning a specific pattern.Therefore, changing the pattern allows re-peated testing sessions.

Animals can easily be trained to cross thehorizontal ladder to reach their home cage.Performance is videorecorded for furtheranalysis of end point measures and qualitativemovement analysis. End point measures in-clude the number of errors in placing a limbwhen crossing the beam (number of errors perstep). The number of limb-placement errorsalso increases in the intact limbs, thus reveal-ing compensatory adjustments. The errorcounts can be supplemented by describing thetype of an error using a scoring system thatallows a distinction between severe and milderrors and limb positioning on the rung (Table44-8). Furthermore, the behavioral analysiscan be supplemented by other techniques,such as electrophysiological recordings ofmuscle activity (Merkler et al., 2000).

The rung-walking test is sensitive tochronic movement deficits after adult andneonate lesions to the motor system and tophysiological variables such as aging or stress(Metz et al., 2001).

SPECIES-TYPICAL BEHAVIORS

One of the primary functions of the brain isto produce behaviors that allow animals toadapt to their environment with a minimumof learning and its associated error. A numberof behaviors that are mainly innate, relativelystereotyped, and characteristic of a species arecalled species-typical behaviors. Table 44-9 pre-sents a catalogue of rat species-typical behav-iors and their description. The following sec-tion provides some examples of tests ofspecies-typical behaviors that can be used in ageneral analysis of behavioral phenotype.

FOOD HOARDING

Rats are in danger of predation whenever theyleave their home base to search for food. Abehavior that they use to minimize risk is toeat smaller pieces of food immediately and tocarry larger pieces of food (i.e., food items thattake more than a few seconds to consume) toa safe location for later consumption. Usually,this location is the nest or a home box in thelaboratory. The rule that they use is that iffood takes longer to consume than the roundtrip to the refuge, the exposure can be mini-mized by food carrying. Rats will not onlycarry food to a shelter to eat; they also carryand store food when they are not hungry. Thisbehavior indicates that they place an incentivevalue on food.

Table 44-8. The Seven Categories of Limb Placement That Can Be Rated for Limb Placement

Category Type of Foot Placement

0 Total miss1 Deep slip

2 Slight slip3 Replacement4 Correction

5 Wrist or digits6 Correct placement

Characteristics

Deep fall after foot missed the rungDeep fall after foot missed the rung OR slight fall after foot missed

the rungSlight fall after foot slipped off the rungFoot replaced from one rung to anotherFoot aimed for one rung but was placed on another OR foot

position on same rung was correctedFoot placed on rung with either digits /toes or wrist /heelFoot placed on rung in plantar fashion with full weight support

Page 505: 0195162854

488 MODELS AND TESTS

Table 44-9. Species-Typical Behaviors and Their Assessment

Grooming

Food foraging/hoarding

Eating

Exploration / neophobia

Foraging and diet selection

Sleep

Nest-building

Maternal behavior

Social behavior

Sexual behavior

Play behavior

Grooming movements are species-distinctive and are used for cleaning andtemperature regulation. Begin with movements of paw cleaning and proceedthrough face washing, body cleaning, and limb and tail cleaning.

Food carrying movements are species-distinctive and used for transporting food toshelter for eating, scattering food throughout a territory, or storing food in depots.Size of food, time required to eat, difficulty of terrain, and presence of predatorsinfluence carrying behavior. Both mouth-carrying or cheek-pouching are used bydifferent species. Rodents also engage in food wrenching, in which food is stolenfrom a conspecific, and dodging, in which the victim protects food by evading therobber.

Incisors are used for grasping and biting, rear teeth are used for chewing, tongueis used for food manipulation and drinking.

Species vary in responses to novel territories and objects. Objects are exploredvisually or with olfaction, avoided, or buried. Spaces are explored by slow excur-sions into space and quick returns to a starting point. Spaces are subdivided intohome bases, familiar territories, and boundaries.

Food preferences are based on size and eating time of food, nutritive value, taste,and familiarity. For colony species the colony is an information source with ac-ceptable foods identified by smelling and licking snout of conspecifics.

Rodents display all typical aspects of sleep including resting, napping, quiet sleepand rapid eye movement sleep. Most rodents are nocturnal, thus sleeping duringthe day with major activity periods occurring at sun up and sun down. Cycles innatural habitats vary widely with seasons.

Different species are nestbuilders, tunnel builders, and build nests for small familygroups or large colonies. All kinds of objects are carried, manipulated, and shred-ded for nesting material.

Laboratory rodents typically have large litters that are immature when born. Pupsare fed for the first 2 to 3 weeks of life and thereafter become independent.

Colony or family rodents have rich social relations including territorial defense,social hierarchies, family groupings, and greeting behaviors. Solitary rodents mayhave simplified social patterns. Defensive and attack behavior in males and fe-males is distinctive.

Characteristic sexual behavior displayed by males and females. Males displayterritorial control or territory invasion, and engage in courtship and often groupsexual behavior. Sexual behavior is often long-lasting with many bouts of chasing,mounting and intromission, and incidents of ejaculation. Mounting is followed bygenital grooming and intromission is followed by immobility and high frequencyvocalizations. Females engage in soliciting including approaches and darting,pauses and ear wiggling, and dodging and lordosis to facilitate male mounting.

Many rodents have rich play behavior with the highest incidence in the juvenileperiod. Play typically consists of attack in which snout-to-neck contact is theobjective and defense in which the neck is protected.

Food-hoarding behavior can be tested inany apparatus in which an animal has a refugeand there is an open area that contains food. Forexample, a two-compartment box with one com-partment illuminated more than the other or a

home base attached to an alley provides a sim-ple test enclosure. Food hoarding has also beenused in structured test situations, including thosein which an animal has to solve a maze puzzleor learn a spatial problem to find food.

Page 506: 0195162854

Chapter 44. Neuropsychological Tests 489

The neural basis of food hoarding is notfully understood, but it is disrupted by frontaldecortication, even though individual com-ponents of hoarding behavior may still be in-tact (Kolb, 1974). Food-hoarding behavior isdisrupted after mesolimbic dopamine deple-tion (Stam et al, 1989), indicating that thedopaminergic projection to the prefrontal cor-tex plays a central role in structuring this be-havior. Furthermore, rats with limbic systemlesion, including hippocampal damage, showa disruption of food-carrying behavior un-less stimulated by external signals (Whishaw,1993). Other deficits found in food-hoardingtasks point toward a deficit in spatial naviga-tion and acquiring problem-solving strategies.Thus, food-hoarding tasks are versatile testsin revealing sensory and cognitive aspects ofnatural behavior.

EXPLORATION

When rats are removed from their home cageand placed in an unfamiliar environment, theydisplay structure in their exploratory behav-ior (Eilam and Golani, 1989). The rat will es-tablish a home base at the location at whichit was first placed or at a shelter provided. Itwill start exploring its environment and returnto the home base to pause, rear, and groom.As it begins to explore, it will stretch its fore-quarter and head to examine the area sur-rounding the home base. It will then start toexplore more remote areas by making tripsaway from the home base. The animal willtravel mostly along a wall or the edge of theopen field, and it will frequently return di-rectly and quickly to the home base. Returntrips are usually shorter than the outward-bound trips. The outward trips will becomegradually longer until the entire area has beenexplored, and the animal might also adopta new home base over the course of itsexploration.

Although scoring the structure in ex-ploratory behavior is laborious, computer-based programs are being developed. In addi-

tion, an interval of 5 or 10 minutes can pro-vide ample data including the number of tripsand stops, number of rears, the frequency ofgrooming, duration of trips, and kinematicmeasures of speed and path trajectories. Openfield tests are used in test batteries because thebehavior is easy to elicit and is reliable becausepretraining or habituation is not necessary.Exploratory tests have proved useful in stud-ies of the effects of drugs, brain damage, orphysiological changes. Common neurologicaltest batteries include the assessment of anxi-ety in the open field, which is indicated by ex-ploratory activity, the time spent in peripheralcompartments versus time spent in the centerof the open field, the latency to enter a novelhome base, and the habituation to the novelaspects of the field.

PLAY BEHAVIOR

Play is the most prominent form of socialinteraction during the juvenile stage of life;however, rats show play behavior at any stageof their development and in adulthood. Playconsists of a ritualized sequence of move-ments. The individual components of play inrodents are often part of other behaviors suchas aggression, predation, and sex. The com-plex movement sequences of play can be sep-arated into the individual movement compo-nents, and each sequence is characteristic fora specific type of play. For instance, play fight-ing is marked by the attempt to repeatedly at-tack the nape area of a recipient with thesnout, while the recipient attempts to avoidthe contact. Because play fighting is the mostreported form of social play and persists intoadulthood, it is suitable as a measure of socialinteraction in rats.

Play fighting movements in juvenile andadult rats offer a variety of options for be-havioral analysis. Although the defensive tac-tics undergo changes during development, thesequence of individual acts in play fighting isrelatively fixed in both the initiator and the re-cipient. The major categories of play fighting

Page 507: 0195162854

490

are chasing, dodging, wrestling, and tumbling,all of which occur in play fighting in rats.The individual movement components ofplay fighting are easily recognizable. The re-lationship between the two interacting indi-viduals or individual motor acts can be ana-lyzed by the means of detailed descriptiveanalysis systems (Pellis and Pellis, 1983).

Play behavior has a cognitive compo-nent, and it is thought to serve gathering so-cial and emotional information about con-specifics. In line with this theory is thefinding that the complex sequences of playbehavior can be disturbed or lost by neuro-logic conditions (Pellis et al., 1993). Further-more, play behavior has been described asbeing a measure of the ability to initiate orrespond to social contacts (Daenen et al.,2002). In summary, the analysis of play be-havior represents a valuable measure of mo-tor and cognitive function involved in socialinteraction.

MODELS AND TESTS

LEARNING

Psychologists have been studying learning inrats for at least 100 years, and there are liter-ally hundreds of different learning tests avail-able in the literature (Table 44-10). Thechoice of learning tests will depend on the na-ture of the question being asked by the in-vestigator, but a few simple tasks can provideconsiderable information about the cognitivestatus of an animal. Learning tests reveal atleast four insights into behavior: (1) Can ananimal master the procedures that the task re-quires? (2) What neural system is an animalusing in performing/learning a task? (3) Islearning normal? (4) What is the structure oflearning behavior? Investigators who screenanimals for the effects of pharmacological ma-nipulations or genetic influences will mainlybe interested in the first three questions,whereas students of learning are likely to beinterested in the fourth question as well.

Table 44-10. Learning and Measures of Learning

Classic conditioning

Instrumental conditioning

Avoidance learning

Object recognition

Spatial learning

Memory

Unconditioned stimuli are paired with conditioned stimuli and the strength of anunconditioned response to the conditioned stimuli is measured. Almost anyarrangement of stimuli, environments, treatment, or behavior can be used.

Animals are reinforced for performing motor acts such as running, jumping, sittingstill, lever pressing, or opening puzzle latches.

Passive responses including avoiding preferred places or objects which have beenassociated with noxious stimuli such as electric shock. Active responses includingmoving away from noxious items or burying noxious items.

Including simple and recognition of one or more objects, matching to sample, andnonmatching to sample in any sensory modality. Tasks are formal in which an ani-mal makes an instrumental response of inferential in which recognition is inferredfrom exploratory behavior.

Dry land- and water-based tasks are used. Spatial tasks can be solved usingallothetic cues, which are external and relatively independent to movements, or idio-thetic cues, which include cues from vestibular or proprioceptive systems, reaffer-ence from movement commands, or sensory flow produced by movements them-selves. Animals are required to move to/away from locations. Cue tasks requireresponding to a detectable cue. Place tasks require moving by using the relation-ships between a number of cues, no one of which is essential. Matching tasks re-quire learning a response based on a single information trial.

Memory includes procedural memory in which response and cues remain constantfrom trial to trial. Tasks are constructed to measure one or both types of learning.Memory is typically divided into object, emotional, and spatial, and each categorycan be further subdivided into sensory and motor memory.

Page 508: 0195162854

Chapter 44. Neuropsychological Tests

SWIMMING POOL TASKS

Some of the most popular tasks that provideinformation about place learning ability aswell as procedural and working memory inanimal studies are swimming pool spatialtasks (Morris et al., 1982). These tasks are es-pecially useful in rats because they are semi-aquatic. Numerous modifications of this testhave been developed, however, and the fol-lowing will focus on the basic procedure thatapplies to all versions of the water task. Atraining or test session begins with introduc-ing an animal to a round swimming pool thatcontains a hidden platform. The goal is for therat to discover and to localize the platform andescape from the water. By using skim milkpowder, the water is made opaque so that theanimal is unable to see the platform sub-merged about 1 cm underneath the water sur-face. Over consecutive trials, the time to findthe platform decreases as the animal learns toswim directly to the invisible platform withrespect to distal cues surrounding the swim-ming pool. The animal's performance is meas-ured as the time it takes to find the platform(escape latency), the distance traveled (swimdistance), and the accuracy in targeting theplatform over consecutive trials.

The procedural simplicity of the watertask is opposed to the complex underlyingprocesses that determine performance of thistask. The task requires a variety of behavioralprocesses, including navigation strategy for-mation, place learning, memory, and the per-formance of visually guided behavior (Cainand Saucier, 1996; D'Hooge and De Deyn,2001). Various aspects of spatial learning canbe assessed using the following procedures:

1. Procedural learning. Procedural learning involvesevaluating whether an animal can acquire theskills necessary to escape from the water. Thewater is tepid, and a platform is hidden at a fixedlocation with its surface 1 cm below the surfaceof the water. The cardinal compass points serveas starting points, and a rat swims until it findsthe platform or until 60 seconds has elapsed, at

491

which pint it is removed from the water. Twotrials are administered each day for a total of 5days. Latency to reach the platform, distanceswum, and head direction can all be measuredusing commercial tracking systems.

2. Matching-to-place learning. If an animal can ac-quire the procedural aspects of the task, it canbe tested for spatial learning ability using amatching-to-place version of the task. Each daythe platform is moved to a new location, for atotal of 5 to 7 days. Each day, the animal re-ceives two swims from the same location—asample swim in which it has to locate the plat-form at its new location and a matching swim—in which it demonstrates that it has learned thenew location on the sample trial. Rats quicklybecome adept in learning new locations in a sin-gle trial. Typically, animals have long latencieson the first trial, because they display a win-staybehavior and search for the platform at its oldlocation, and a short second trial latency, be-cause they can learn the new location in a sin-gle trial.

3. Cue learning. If an animal cannot acquire theprocedural aspects of the task, it can be pre-sented with trials on which the platform is vis-ible and so serves as a cue or beacon. A cue trialprocedure is used to demonstrate that an ani-mal can see, swim, and escape.

Many variations on these basic procedures areused, with major variations being the watertemperature, size of the apparatus, number ofswimming trials per day, and so on. The ma-jor advantages of the task are that no specialdeprivation is required to motivate the ani-mals, so testing can proceed quickly. Conse-quently, a large number of studies haveconfirmed that acquisition, retention, and re-versal of navigational strategies in the watermaze involve a number of brain regions andneurochemical systems, each affecting specificparameters of performance in the water maze.Spatial learning has been related to the inte-gration of hippocampal CAl and CA3 regions,and hippocampal connectivity to nucleus ac-cumbens and raphe nuclei via the fimbriafornix (Whishaw, 1987). Thus, interruption ofhippocampal projections leads to deficits in

Page 509: 0195162854

492 MODELS AMD TESTS

the acquisition of new platform positions ornew distal cues and to impairments in reten-tion of previously learned information. Otherbrain structures influencing water task per-formance include the prefrontal cortex, stria-turn, cerebellum, and various neurochemicaland neuromodulator systems (for reviews, seeMcNamara and Skelton, 1993; D'Hooge andDe Deyn, 2001). It is worth noting that thenumber of structures sensitive to the taskrequire caution in interpreting results. Ani-mals may be impaired at the task for manyreasons, only one of which is an impairmentin learning or memory, per se.

Finally, it is noteworthy that a virtual wa-ter task for use in humans has been developedthat could replicate the basic findings from an-imal studies (Hamilton et al., 2002). Thus, notonly is the water maze a flexible test for ani-mal studies but also the methodology is di-rectly applicable to human spatial behavior.

RADIAL ARM MAZE

The radial arm maze analyzes spatial naviga-tion strategies in rodents in that the animal isrequired to learn the location of food at oneor more locations. The radial arm maze con-sists of a central platform from which a num-ber of arms originate (Olton et al., 1979). Mostcommonly, eight arms are arranged aroundan octagonal central platform in equal spaces.Access to the arms can be controlled by guil-lotine doors that can be opened or closed foreach trial. The maze is usually located in aroom with salient visual cues such as posterson the walls, counters, cupboards, etc. The lo-cation of the arms is either fixed or flexibleand can be marked by a cue on the arm suchas a light or color. Food is located in an in-dentation at the farthest end of one or morearms. Before the start of an experiment, theanimals must be food deprived and habituatedto the food reward provided in the test. Thetask of the animal is to learn the location ofthe food over a number of trials. When usingmultiple locations of food, the animal has to

learn a sequential response strategy to pick upthe food. By using doors to control access tothe arms, the retention of a learned strategycan be tested by leaving the doors closed fora certain period of time before a new trialstarts.

There are a number of standard trainingand testing procedure protocols available,with each focusing on specific aspects of pro-cedural memory and retention (Jarrard, 1983).A common test strategy in the radial armmaze is to expose a rat to a habituation phasein which all arms are baited. In consecutivetrials, the animal learns a strategy of enteringthe arms to collect all food items. After a fewdays, the training phase begins in which onlyfour arms are baited. When the animal is re-leased from the center platform, it is requiredto obtain all four food rewards within a cer-tain time period. Over a number of test ses-sions, the same arms remain baited and thenumber of visits of every arm are recorded. Inthis arrangement, working memory is re-flected by an animal visiting each of the fourbaited arms only once. Reference memory isreflected by an animal not visiting any of thefour arms that has never been baited before.

The radial arm maze is especially suitedto evaluate the ability to form a proceduralmemory of the task that might be interruptedby treatments or brain lesions. Like the waterbased tasks, the radial arm task is sensitive todamage of many brain structures. What dis-tinguishes the swimming pool task and the ra-dial arm task is that the former evaluates win-stay behavior, return to the previous escapelocation, whereas the latter evaluates win-shift behavior, because reward has beenremoved from that location, search at a newlocation.

BARNES TASK

The Barnes task is essentially a dry maze ver-sion of the swimming pool task, and it allowsan animal to escape from an open area to arefuge. The Barnes task is a circular platform

Page 510: 0195162854

Chapter 44. Neuropsychological Tests

made of wood or stainless steel. In its originalversion, there are eight holes cut in the tablethat are spaced equidistant and arranged alongthe perimeter of the table (Barnes, 1979). Acage similar to the animal's home cage isplaced underneath one of the holes and servesas a refuge. The common test procedure usesa fixed location for the refuge. The rat is re-leased in the center of the table, and its latencyand accuracy in rinding the refuge are de-pendent measures. Because the eight poten-tial refuge holes all appear the same to aviewer on the table, the correct hole can bequickly reached only by learning its locationin relation to surrounding room cues.

The Barnes maze has been modified toevaluate search behavior. A food-deprived ratis placed in the refuge and can leave the refugeto find food on the table. When a rat is placedinto its refuge, it will soon leave the refugeand start exploring the platform. Its outward-bound trips will become increasingly longeruntil the rat explored the entire surface of theplatform. This version of the test allows forstudying the structure of exploratory behav-ior (Whishaw et al., 2001). In addition, largefood pellets can be placed on various or fixedlocations on the table. The rat will then for-age for food, pick up the food item, and carryit to its refuge for consumption. This behaviorrequires that the animal memorize the locationof its home base relative to the distant cues inthe room and relative to self-movement cues.Interestingly, it was found that rats not only usethese cues for path integration but also trackolfactory information (Whishaw and Gorny,1999). In their studies, Whishaw and Gorny(1999) described the use of strings that carry anew odor. Rats can be trained to track the odorand follow a string to locate a food reward onthe platform. The animals also are able to fol-low their own scent or the scent of a conspe-cific to locate food or refuge locations.

To record the data, a camera can bemounted on the ceiling above the center ofthe platform to videorecord the behavior ofthe animals. When analyzing the video tapes,

493

automated tracking systems might be usefulto trace the path of the animal as it exploresthe platform.

ELEVATED-PLUS MAZE

When a rat is placed in a novel environmentor it experiences a disturbance in its homecage, it can show an emotional response withsigns of anxiety. Anxiety refers to an internalemotional state related to threat. The meas-ures of anxiety that have been established inanimal research are based on observable be-havior that reflects the emotional state of fearand apprehension. One of these measures isthe tendency of a rat to avoid lit open spaces,and the preference to move along walls whenexploring a new environment.

The elevated-plus maze takes advantageof the fact that rats prefer to remain in en-closed compartments when they are placed ina novel environment. The elevated-plus mazeis elevated above the floor. It has a smallcenter area and four arms of equal lengtharranged in a plus-sign shape. Two of the armsthat are facing each other are enclosed withside walls, and the other two arms are openwithout walls. When a rat is introduced intothis test environment, it will spend more timeexploring the enclosed arms than in the openarms. The typical measures of an animal's per-formance are the time spent in the enclosedarms versus the open arms and the numberof entries of the two types of arms (Fellow etal., 1985).

The elevated-plus maze is a common testprocedure to determine the effectiveness ofdrug treatments. For instance, it can be ex-pected that an anxiolytic drug would increasethe number of entries in the open arms andalso elevate the time spend in the open arms(Fellow and File, 1986). In contrast, an anxio-genic treatment or procedure would lead tothe opposite effects by reducing the numberof entries in the open arms and by promotingthe animal's preference to stay in the enclosedarea. Similar consequences can be drawn by

Page 511: 0195162854

494 MODELS AND TESTS

deletion of genes that are involved in the con-trol of these emotional responses. Thus, theelevated-plus maze is widely used not only inpharmacological research but also as a tool totest emotional behavior in genetic mousemodels (Belzung and Griebel, 2001).

There are number of factors that caninfluence the behavior of an animal in theelevated-plus maze. First, manipulations thatmight arouse the animals and lead to stress-induced anxiety can be a disturbing confoundthat makes the interpretation of data difficult.Furthermore, the test apparatus itself influ-ences the animals' responses. The size of theapparatus, the elevation above the ground,and the shape of the open arms might mod-ify the fear to explore the open arms. Inter-estingly, it has been described that when theopen arms are confined by a small ledge, thebehavior changes over repeated test sessionsin a different manner from trials that use noledge on the open arms (Fernandez and File,1996). These findings led to the conclusionthat the elevated-plus maze in fact measurestwo distinct types of anxiety: the fear of theopen space and the fear of elevation.

CONTEXT CONDITIONING

The rats' preference for some stimulus objectsover others, its tendency to explore novelstimuli, and its avoidance to stimuli that sig-nify threat or harm form the bases for "con-textual" or "context conditioning" tasks (Ottoand Giardino, 2001). The term context refersto the test situation prepared by the experi-ment, and the term conditioning refers to thefact that some previous experience in the sit-uation will influence the rat's choice or pref-erence behavior. In most of these tests, thecontextual cues to which the rat responds areunspecified and are unexamined, as it is theanimal's choice behavior that is of primary in-terest to the experimenter. Nevertheless, it iscertainly the case that a rat's behavior in thetask contains structure and that some cues areprepotent, although the measures of "prefer-

ence" made by the experimenter may not in-clude the structure- or address-specific cues.The strengths of context conditioning tests arethat they are simple to administer and requirelittle pretraining of the animal and that endpoint scores of choice and preference are sim-ple to collect. Many of the tests are also basedon the facts that rats have a strong tendencyto explore novel options and, when presentedwith a number of food sources, rats tend toalternate choices on successive trials. Amongthe many context conditioning tests that havebeen designed, those in greatest use are spon-taneous alternation, forced alternation, andconditioned place preference.

Spontaneous AlternationSpontaneous alternation tests are typicallyconducted in a T-maze or a Y-maze. The an-imal is placed in one arm of the maze to be-gin a trial and is allowed to move around inthe maze until it has completely entered oneof the other arms. At a specified time of min-utes to as long as a day, it is given a secondtrial in which it is placed at the previously usedstarting location. If the animal chooses thepreviously unentered arm, it is scored ashaving performed an alternation. In principle,there is no limit to the number of additionaltrials that are administered, although for pur-poses of rapid screening, two trials may be suf-ficient. In some versions of the task, roomcues are visible and it is assumed that the an-imal is making choices based on room cues.In other versions of the task, the alley is cov-ered and it is assumed that the animal is mak-ing its choice using local cues, such as odorsin the alley including those odors that the ratitself has left, or the animal is using self-move-ment cues, which is a record of its previousmovements that is derived from vestibular orproprioceptive cues.

Forced Choice AlternationA forced choice alternation test is a variationof the spontaneous alternation test in whichon the first trial one of the two potential

Page 512: 0195162854

Chapter 44. Neuropsychological Tests 495

choice alleys is blocked so that the animal isforced to chose the only open alley. On thesubsequent choice test, both choice alleysare open. If the animal chooses the previouslyblocked alley it is scored as having made analternation.

Rewarded Alternation TasksIn rewarded alternation tasks, animals arefood or water deprived and reinforcement canbe obtained by entering one of the choices.An animal may also be negatively reinforcedwith shock for entering an alley. The advan-tage of using reward is that many trials can beadministered and the schedule and delay be-tween choices can be controlled. In sponta-neous-rewarded alternation tasks, reinforce-ment is always present in both choice alleysand the spontaneous alternations over suc-cessive trials are counted. In forced-choice re-warded alternation tasks, one of the alleysis blocked on a forced choice trial, and on achoice trial, both alleys are open. Reward maybe removed from the forced choice alley forthe choice trial to increase the probability ofalternation behavior or reward may be pres-ent in both choice arms.

Conditioned Place PreferenceAnimals become conditioned to the locationof an object or event that has been experi-enced as pleasant or noxious in a previousencounter (Jodogne et al., 1994). As a result oftheir previous experience, they may seek outor avoid places in which they have been rein-forced previously. The apparatus for condi-tioned place preference is typically a box withtwo compartments, and the dependent meas-ure is the time spent in either compartment.The box may allow the animal to view sur-rounding room cues and the compartments ofthe box may be distinctively marked with lo-cal cues (i.e., visual cues, odor cues, or bed-ding material). In the designs of the experi-ments, an animal is given some experience inone of the compartments of the box, the con-ditioning trial, and then later given an oppor-

tunity to chose which compartment it entersand subsequently spends its time, the placepreference trial. Dependent measures arecompartment choice and/or time spent in thecompartments.

A typical experimental paradigm for con-ditioned place preference is conditioned re-ward trials using drugs. An animal is placedin one of the compartments after having re-ceived a drug, or the drug is administeredwhile the animal is in that compartment. Ona subsequent trial, the rat is allowed to choosecompartments. If it chooses the previouslyrewarded compartment, it can be concludedthat the drug is positively rewarding, whereasif it chooses the other compartment, it is con-cluded that the drug is negatively rewarding.

Shock-Induced AvoidanceA widely used variation of conditioned placepreference test is a shock-induced avoidanceor "passive-avoidance" test. Two compart-ments are used, and often the compartmentsvary in their reinforcing properties. The floorof the box is made of metal grids that can beelectrified to present a mild shock to the feetof the animal. One compartment of the boxmay be painted black and may be in the dark,whereas the other arm is painted white and isilluminated. After one or two familiarizationexperiences with the apparatus, rats typicallyquickly enter the darkened compartment andremain there. On a subsequent test, the ani-mal is placed in the nonpreferred compart-ment and then receives a foot shock as soonas it enters the preferred compartment. Afterthis, the animal is then removed from the ap-paratus. In a subsequent avoidance phase ofthe test, the animal is replaced in the nonpre-ferred compartment, and its latency to enterthe preferred compartment is measured.

The escape response can be used for clas-sic conditioning experiments. In active avoid-ance conditioning, the animal can escape fromone compartment to avoid a noxious stimu-lus such as gentle food shock in the other. Aneutral stimulus (conditioned stimulus) such

Page 513: 0195162854

496 MODELS AND TESTS

as a light flash or tone is presented before de-livery of a footshock event (unconditionedstimulus). The animal will first learn to escapethe footshock, but it will eventually associ-ate the preceding neutral stimulus with thepainful stimulus. When conditioning is com-plete, the escape response will occur to theonset of the conditioned stimulus and beforethe arrival of the unconditioned stimulus. De-pendent measures are the number of trials forconditioning to occur and response latency.

CONCLUSION

In this chapter, we described a general batteryof tests that can be used to describe the be-havioral phenotype of the rat given a particu-lar treatment. It was not possible to describeevery behavior in detail, and many behaviorswere not examined at all; we made no attemptto be inclusive in citing the relevant literature,only to provide examples. We noted in the in-troduction that the general test battery wouldnormally be followed by more detailed analy-sis such as that described in the other chaptersof the volume. The key point is that a greatdeal can be inferred about the functional stateof the brain by careful behavioral analysis.Thus, although the analysis of molecular,chemical, anatomical, and physiological fac-tors are critical to understanding brain func-tion, in the end it is behavior that the brain isdesigned to produce. It thus could be arguedthat until you understand how the brain con-trols behavior, you really do not understandmuch at all about brain function.

REFERENCES

Almasi R, Petho G, Bolskei K, Szolcsanyi J (2003) Effectof resiniferatoxin on the noxious heat thresholdtemperature in the rat: a novel heat allodynia modelsensitive to analgesics. British Journal of Pharma-cology 139:49-58.

Ballermann M, Metz GA, McKenna JE, Klassen F,Whishaw IQ (2001) The pasta matrix reaching task:

a simple test for measuring skilled reaching distance,direction, and dexterity in rats. Journal of Neuro-science Methods 106:39-45.

Barnes CA (1979) Memory deficits associated with senes-cence: a neurophysiological and behavioral study inthe rat. Journal of Comparative and PhysiologicalPsychology 93:74-104.

Earth TM, Jones TA, Schallert T (1990) Functional sub-divisions of the rat somatic sensorimotor cortex. Be-havioural Brain Research 39:73-95.

Basso DM, Beattie MS, Bresnahan JC (1995) A sensitiveand reliable locomotor rating scale for open fieldtesting in rats. Journal of Neurotrauma 12:1-21.

Basso DM, Beattie MS, Bresnahan DK, Anderson DK,Faden AI, Gruner JA, Holford TR, Hsu CY, NobleLJ, Nockels R, Perot PL, Salzman SK, Young W(1996a) MASCIS evaluation of open field locomotorscores: effects of experience and teamwork on reli-ability. Journal of Neurotrauma 13: 343-359.

Basso DM, Beattie MS, Bresnahan JC (1996b) Graded his-tological and locomotor outcomes after spinal cordcontusion using the NYU weight-drop device versustransection. Experimental Neurology 139:244-256.

Belzung C and Griebel G (2001) Measuring normal andpathological anxiety-like behaviour in mice: a re-view. Behavioural Brain Research 125:141-149.

Daenen EW, Wolterink G, Gerrits MA, Van Ree JM(2002) The effects of neonatal lesions in the amyg-dala or ventral hippocampus on social behaviourlater in life. Behavioral Brain Research 136:571-582.

D'Hooge R and De Deyn PP (2001) Applications of theMorris water maze in the study of learning andmemory. Brain Research and Brain Research Re-views 36:60-90.

Dubuisson D and Dennis SG (1977) The formalin test:a quantitative study of the analgesic effects of mor-phine, meperidine, and brain stern stimulation inrats and cats. Pain 4:161-174.

Eilam D and Golani I (1989) Home base behaviour ofrats (Rattus norvegicus) exploring a novel environ-ment. Behavioural Brain Research 34:199-211.

Eskhol N and Wachmann A (1958) Movement notation.London: Weidenfeld and Nicolson.

Fernandez C and File SE (1996) The influence of openarm ledges and maze experience in the elevatedplusmaze. Pharmacology Biochemistry and Behav-ior 54:31-40.

Gharbawie O, Whishaw PA, Whishaw IQ (2004) Thetopography of three-dimensional exploration: a newquantification of vertical and horizontal explo-ration, postural support, and exploratory bouts inthe cylinder test. Behavioural Brain Research,151:125-35.

Golani I (1976) Homeostatic motor processes in mam-malian interactions: a choreography of display. In

Page 514: 0195162854

Chapter 44. Neuropsychological Tests 497

Perspectives in ethology, Vol II (Bateson PPG andKlopfer PH, eds.). New York: Plenum Press.

Hamilton DA, Driscoll I, Sutherland RJ (2002) Humanplace learning in a virtual Morris water task: someimportant constraints on the flexibility of place nav-igation. Behavioural Brain Research 129:159-170.

Hard E and Larsson K (1975) Development of air right-ing in rats. Brain Behavior and Evolution 11:53-59.

Harker TK and Whishaw IQ (2002) Place and matching-to-place spatial learning affected by rat inbreeding(Dark-Agouti, Fischer 344) and albinism (Wistar,Sprague-Dawley) but not domestication (wild rat vs.Long-Evans, Fischer-Norway). Behavioural Brain Re-search 134:467-477.

Holzberg D and Albrecht U (2003) The circadian dock:a manager of biochemical processes within the or-ganism. Journal of Neuroendocrinology 15:339-343.

Jarrard LE (1983) Selective hippocampal lesions and be-havior: effects of kainic acid lesions on performanceof place and cue tasks. Behavioral Neuroscience97:873-889.

Jodogne C, Marinelli M, Le Moal M, Piazza PV (1994)Animals predisposed to develop amphetamine self-administration show higher susceptibility to de-velop contextual conditioning of both ampheta-mine-induced hyperlocomotion and sensitization.Brain Research 657:236-244.

Kolb B (1974) Prefrontal lesions alter eating and hoard-ing behavior in rats. Physiology and Behavior12:507-511.

Kolb B, Cote S, Ribeiro-da-Silva A, Cuello AC (1997)Nerve growth factor treatment prevents dendriticatrophy and promotes recovery of function aftercortical injury. Neuroscience 76:1139-1151.

Kolb B and Whishaw IQ (1983) Dissociation of the con-tributions of the prefrontal, motor and parietal cor-tex to the control of movement in the rat. CanadianJournal of Psychology 37:211-232.

Ma M, Basso DM, Walters P, Stokes BT, Jakeman LB(2001) Behavioral and histological outcomes fol-lowing graded spinal cord contusion injury in theC57B1/6 mouse. Experimental Neurology 169:239-254.

Marshall JF, Turner BH, Teitelbaum P (1971) Sensoryneglect produced by lateral hypothalamic damage.Science 221:389-391.

Martens DJ, Whishaw IQ, Miklyaeva El, Pellis SM (1996)Spatio-temporal impairments in limb and bodymovements during righting in an hemiparkinsonianrat analogue: relevance to axial apraxia in humans.Brain Research 733:253-262.

McNamara RK and Skelton RW (1993) The neurophar-macological and neurochemical basis of place learn-ing in the Morris water maze. Brain Research Re-views 18:33-49.

Merkler D, Metz GA, Raineteau O, Dietz V, SchwabME, Fouad K (2001) Locomotor recovery in spinalcord-injured rats treated with an antibody neutral-izing the myelin-associated neurite growth in-hibitor Nogo-A. Journal of Neuroscience 21:3665-3673.

Metz GA and Whishaw IQ (2002a) Drug-induced rota-tion intensity in unilateral dopamine-depleted ratsis not correlated with end point or qualitative meas-ures of forelimb or hindlimb motor performance.Neuroscience 111:325-336.

Metz GA and Whishaw IQ (2002b) Cortical and sub-cortical lesions impair skilled walking in the variablyspaced ladder rung walking task. Journal of Neuro-science Methods 115:169-179.

Metz GA, Dietz V, Schwab ME, van de Meent H (1998)The effects of unilateral pyramidal tract section onhindlimb motor performance in the rat. BehaviouralBrain Research 96:37-46.

Metz GAS, Merkler D, Dietz V, Schwab ME, Fouad K(2000) Efficient testing of motor function in spinalcord injured rats. Brain Research 883:165-177.

Metz GA, Schwab ME, Welzl H (2001) The effects ofacute and chronic stress on motor and sensory per-formance in male Lewis rats. Physiology and Be-havior 72:29-35.

Montoya CP, Campbell-Hope LJ, Pemberton KD, Dun-nett SB (1991) The "staircase test": a measure of in-dependent forelimb reaching and grasping abilitiesin rats. Journal of Neuroscience Methods 36:219-228.

Morris RGM, Garrud P, Rawlins JN, O'Keefe J (1982)Place navigation impaired in rats with hippocampallesions. Nature 297:681-683.

Olton DS, Becker JT, Handlemann GE (1979) Hip-pocampus, space and memory. Behavioral andBrain Sciences 2:313-365.

Otto T and Giardino ND (2001) Pavlovian conditioningof emotional responses to olfactory and contextualstimuli: a potential model for the development andexpression of chemical intolerance. Annals of theNew York Academy of Science 933:291-309.

Pellis SM and Pellis VC (1983) Locomotor-rotationalmovements in the ontogeny and play of the labo-ratory rat Rattus norvegicus. Developmental Psy-chobiology 16:269-286.

Pellis SM and Pellis VC (1994) Development of rightingwhen falling from a bipedal standing posture: evi-dence for the dissociation of dynamic and staticrighting reflexes in rats. Physiology and Behavior56:659-663.

Pellis SM, Castaneda E, McKenna MM, Tran-NguyenLT, Whishaw IQ (1993) The role of the striatumin organizing sequences of play fighting in neona-tally dopamine-depleted rats. Neuroscience Letters158:13-15.

Page 515: 0195162854

498 MODELS AND TESTS

Fellow S and File SE (1986) Anxiolytic and anxiogenicdrug effects on exploratory activity in an elevatedplus-maze: a novel test of anxiety in the rat. Phar-macology Biochemistry and Behavior 24:525-529.

Pellow S, Chopin P, File SE, Briley M (1985) Validationof open:closed arm entries in an elevated plus-mazeas a measure of anxiety in the rat. Journal of Neu-roscience Methods 14:149-167.

Ramos A, Kangerski AL, Basso PF, Da Silva Santos JE,Assreuy J, Vendruscolo LF, Takahashi RN (2002)Evaluation of Lewis and SHR rat strains as a geneticmodel for the study of anxiety and pain. BehaviouralBrain Research 129:113-123.

Schallert T and Whishaw IQ (1984) Bilateral cutaneousstimulation of the somatosensory system in hemi-decorticate rats. Behavioural Neuroscience 98:518-540.

Schallert T, Fleming SM, Leasure JL, Tillerson JL, BlandST (2000) CNS plasticity and assessment of forelimbsensorimotor outcome in unilateral rat models ofstroke, cortical ablation, parkinsonism and spinalcord injury. Neuropharmacology 39:777-787.

Sorg BA, Tschirgi ML, Swindell S, Chen L, Fang J (2001)Repeated formaldehyde effects in an animal modelfor multiple chemical sensitivity. Annals of the NewYork Academy of Science 933:57-67.

Stam CJ, de Bruin JP, van Haelst AM, van der GugtenJ, Kalsbeek A (1989) Influence of the mesocorticaldopaminergic system on activity, food hoarding,social-agonistic behavior, and spatial delayed alter-nation in male rats. Behavioral Neuroscience 103:24-35.

Stoltz S, Humm JL, Schallert T (1999) Cortical injuryimpairs contralateral forelimb immobility duringswimming: a simple test for loss of inhibitory mo-tor control. Behavioural Brain Research 106:127-132.

Takahashi Y, Takahashi K, Moriya H (1995) Mapping ofdermatomes of the lower extremities based on ananimal model. Journal of Neurosurgery 82:1030-1034.

Tarlov IM (1954) Spinal cord compression studies. III.Time limits for recovery after gradual compressionin dogs. Archives of Neurology and Psychiatry 71:588-597.

Weinert D (2000) Age-dependent changes of the circa-dian system. Chronobiology International 17:261-283.

Whishaw IQ (1987) Hippocampal, granule cell and CA3-4 lesions impair formation of a place learning-set inthe rat and induce reflex epilepsy. Behavioural BrainResearch 24:59-72.

Whishaw IQ (1993) Activation, travel distance, and en-vironmental change influence food carrying in ratswith hippocampal, medial thalamic and septal le-

sions: implications for studies on hoarding and the-ories of hippocampal function. Hippocampus 3:373-385.

Whishaw IQ and Gorny B (1999) Path integration ab-sent in scent-tracking fimbria-fornix rats: evidencefor hippocampal involvement in "sense of direc-tion" and "sense of distance" using self-movementcues. Journal of Neuroscience 19:4662-4673.

Whishaw IQ, Haun F, Kolb B (1999) Analysis of behav-ior in laboratory rodents. In: Modern techniques inneuroscience research (Widhorst U, ed.). Heidel-berg: Springer.

Whishaw IQ, Hines DJ, Wallace DG (2001) Dead reck-oning (path integration) requires the hippocampalformation: evidence from spontaneous explorationand spatial learning tasks in light (allothetic) anddark (idiothetic) tests. Behavioural Brain Research127:49-69.

Whishaw IQ, Kolb B, Sutherland RJ (1983) A neuropsy-chological study of behavior of the rat. In: Behav-ioral contributions to brain research (Robinson TE,ed.). New York: Oxford University Press.

Whishaw IQ and Metz GA (2002) Absence of impair-ments or recovery mediated by the uncrossed py-ramidal tract in the rat versus enduring deficits pro-duced by the crossed pyramidal tract. BehaviouralBrain Research 134:323-336.

Whishaw IQ, Miklyaeva El (1996) A rat's reach shouldexceed its grasp: analysis of independent limb anddigit use in the laboratory rat. In: Measuring move-ment and locomotion: From invertebrates to hu-mans (Ossenkopp KP, Kavaliers M, Sanberg PR,ed.), pp. 135-169. Austin, TX: RG Landes.

Whishaw IQ, O'Connor RB, Dunnett SB (1986) Thecontributions of motor cortex, nigrostriatal dopamineand caudate-putamen to skilled forelimb use in therat. Brain 109:805-843.

Whishaw IQ, Pellis SM, Gorny BP (1992) Skilled reach-ing in rats and humans: evidence for parallel devel-opment or homology. Behavioural Brain Research47:59-70.

Whishaw IQ, Pellis SM, Gorny BP, Pellis VC (1991) Theimpairments in reaching and the movements ofcompensation in rats with motor cortex lesions: anendpoint, videorecording, and movement notationanalysis. Behavioural Brain Research 42:77-91.

Whishaw IQ, Suchowersky O, Davis L, Sarna J, MetzGA, Pellis SM (2002) Impairment of pronation,supination, and body co-ordination in reach-to-grasp tasks in human Parkinson's disease (PD) re-veals homology to deficits in animal models. Be-havioural Brain Research 133:165-176.

Woolf CJ, Shortland P, Coggeshall RE (1992) Peripheralnerve injury triggers central sprouting of myeli-nated afferents. Nature 355:75-78.

Page 516: 0195162854

Index

Action pattern, 260-3262Activity, 40Acuity, 51Ad libitum, 324ADHD, 458-459Adolescence, 278-286

behavioral development, 278-281Adrenal gland, 21Aging, 228Aggression, 22-23. See also Fear

conditioning, 412-414deviant forms, 350initiation of, 346-347lateral attack, 299supine defense, 299, 301tactics, 298termination of, 348

Alarm call, 374Alarm cries, 335Albino, 53Allodynia, 69. See also PainAllogrooming, 346Allothetic cues, 392AUothetic navigation, 393, 401. See also Spatial learningAlpha rat, 344Alzheimer's models, 454Ambient thermoneutrality, 227Ambulation, 278Amphetamine, 41-44

stereotypy, 39Amygdala, 202, 256,285,359, 415. See also FearAndrogens, 256. See also Sexual behaviorAngiotensin, 209-210Animal models, 449-461,

neurological, 449-461psychiatric, 464-474psychopathology, 446-468

Animal psychophysics, 113Anogenital, 288Anosmia, 99-100, 269. Seeabo OdorAnterior hypothalamus, 229Anthropomorphism, 18Antigen-presenting cells, 246Anxiety, neural mechanisms, 358-360Anxiety/defense, 340Anxiolytic tests, 358. Seeabo StressApparent motion, 54Appearance, 477-^78

Appetitive behavior, 200-3201Arousal, 41Arpeggio movement, 166-167Associative learning, 263-264Attack bite, 347Attack jump, 347Attention, 426-428Attention, neural, 427-428Attentional set, 427Attentional-hyperactivity disorder, 458-459. See also ADHDAuditory function, 268

Bait shyness, 18Barrel field, 61Barrelettes, 82Barrels, 83. See abo VibrissaeBasal ganglia, 144BBB scale, 156, 484-^(85Biological psychiatry, 463-465Biological reductionism, 463Blocking, 439Body temperature, 232-242

heat stress, 232-233measurement of, 231-232neural control, 229-231

Boxing, 20, 280Bracing, 122-123Bradycardia, 262Brain derived neurotropic factor, 324Brain stimulation, 38-39Breeding, 23. See abo Sexual behaviorBrown Norway, 30

Cage, 294Carbohydrate, 203Carbon disulfide, 367Cat, 413Cat odor, 414Catalepsy, 122-123Categories, 476Caudate nucleus, 168Cerebellum, 64Cerebral injury, 452-453. See abo Neurological modelsChemical senses, 267Cholinergic system, 427, 454Chorda tympani, 107Cinnamon, 366Orcadian activity, 483-484

499

Page 517: 0195162854

500 INDEX

Circadian rhythm, 227-228Circular tract, 390Class-common, 450Cocoa, 366Coefficient of inbreeding, 26Cognitive development, 281-294Cognitive processes, 422-435Cold stress, 226-235Colony, 365Commensal, 15Conditioned. See Configural, learning; Incentive behaviorConditioned place preference, 200, 292, 437Conditioned taste aversion, 281-283, 438Cones, 51Configural, 429

learning, 393Conjunctive, 429Consummatory, 201-203, 438Copulation, 373,Cortical organization, 450-3451Corticosterone, 237Cosegration analysis, 32Crawl under, 19,346Cross maze, 395Crouching, 289Crowding, 23Cue navigation, 283, 393. See also Spatial learningCutaneous, 61Cylinder, 480-3481Cytokines, 245

Dark agouti, 55Dead reckoning, 401-409

measurement of, 401-406theories of, 407-408

Death, 21Decerebrate, 144Defeat/defense, 340Defensive behavior, 335-341, 410-3421

conditioned, 414-418defensive attack, 335laboratory models, 339-341neural substrates of, 415-418organization of, 410-3411

Defensive burying, 239, 335, 353-362conditioned, 356-357paradigms of, 353-356sex and age, 356unconditioned, 354-356

Defensive threat, 335Dehydration, 207-212. See also FoodDelayed matching to sample, 383-387Dendritic spines, 324Development 257-321

disorders, 457-458motor, 284thermoregulation, 229-230

Diabetes, 324Diabetes mellitus, 34Diestrus, 310Diet, 323-324, 366Digits, 162-163

Diversity, 23-24Dodging, 280Domestication, 54-56Dopamine, 39, 147-148, 469Drinking, 38, 207-216Drive, 22-23Drug addiction, 456-457Drugs of abuse, 37Dysaesthesia, 69. See also PainDyskinesia model, 454

Eat, 203Eating, 38, 197-216. See also Food

time, 217-218Ecology, 15-25Ectotherm, 226Ejaculation, 372Electromyograph, 158Emotion, 376Emotionality, 38, 341-342Endotherm, 226. See also TemperatureEnergy intake, 17. See also FoodEnrichment, 327Environment, 212-213

enrichment, 130Episodic memory, 431—432Escape, 336. See also Social behaviorEstrogen, 308Estrous cycle, 307Ethogram, 269Evolution, 3-14

fitness, 29Exercise, 327Exploration task, 397, 404-407Extracellular fluid, 207-209Extradimentional shift, 427Extrastriate cortex, 54Eye, 50-351

Fear, 410-3421Feature-negative, 430Fecundity, 23. See also Maternal behaviorFeeding. See FoodFetal action patterns, 260-3262

behavior, 257-266development 257-258

Fever, 229Fibroblast growth factor, 458Fimbria-fornix, 407Fischer 344, 30, 55,Fisher-Norway, 53Flight, 335Fluid, 207-209Fluid balance, 207-210Fluid percussion model, 453Fluoxetine, 202Focal ischemia, 452Follical-stimulating hormone, 308Food, 197-203

avoidance, 363-364carrying, 217, 220-3221deprivation, 213, 385

Page 518: 0195162854

Index 501

dodging, 218hoarding, 401intake and diet choice, 203-204intake and patterns, 198-200intake and time, 222-223restriction, 197sampling, 16selection, 17wrenching, 218

Foraging, 217-225, 412Forced swimming, 241Forelimb, 162-163Formaline pain test, 75, 479Free radicals, 324Freezing, 335Frontal lobe, 325Future, 428

GABA, 454Genetics, 25Genome, 30Glabrous skin, 266. See also Sensorimotor behaviorGlobal ischemia, 452-453Glomerulonephritis, 21Glucocorticoids, 21Grooming 141-149, 202

bilateral, 142chain, 142-144nonchain, 142-144sampling, 16syntactical, 144-147

Ground reaction forces, 153Group housing, 323, 325, 328Guidance navigation, 393Gustatory system, 105Gustometer, 111

Haloperidol, 121-122Handling, 294Hardy-Weinberg principle, 25Head trauma, 452-453Heritability, 28, 32Heterozygosity, 27. See also DifferenceHind limb function, 136-138Hind limb stepping, 138Hippocampus, 223, 285, 325, 327, 359, 398, 404, 407Hoarding, 15, 223-224Home base, 221, 407Homeward trip, 406Hormonal priming, 293Hormones, 229Hot plate test, 73Housing, 321Hovering, 289Huddling, 270. See also DevelopmentHumidity, 323Huntington's disease, 148, 454—453Husbandry, 3266-Hydroxydopamine, 34Hyperalgesia, 69. See also PainHypertension, 236, 250Hypothalamus. 38-39

Hypothermia, 226-227Hypovolemia, 209Hypoxia, 228, 262

Idiothetic, 401-409. See also Spatial learningcues, 392navigation, 40-341, 392

Immobility, 119-128,121-122, 481Immune system, 245-254

brain, 248-249cellular interaction, 246function, 323, 326stress, 212-213

Immunoglobulin, 246Inbred strains, 25-36Inbreeding, 26, 54-56Incentive behavior, 436—446

instrumental, 441-442learning, 442modulation of, 440-3441motivation, 437-438Pavlovian, 438-441

Indian mole rat, 23Individual differences, 37-46, 237-238Infant behavior, 266-277

amnesia, 282-283development, 266-268movement 268-269

Insular cortex, 285Intracellular fluid, 207-209Intradimensional shift, 427Ischemia, 452-453Isolation, 325-326

Jumping, 138Jumping stand, 49

Kinematics, 157Kinetic measurement, 157-158

Landmark, 393. See also Spatial learning, navigationLateral geniculate, 54L-dopa, 262Leutal phase, 308Leuteinizing hormone, 308Lick rate, 111Life span, 324, 327Lighting, 321Limb use asymmetry, 132-134Litter, 294Locomotion, 140-3161, 269,482-485

activity, 269galloping, 153-154measurement, 156-159mechanics, 150-3154neural control, 154-156step cycle, 150-3151subterranian, 4tasks, 158-159trotting, 152-153walking, 151-152

Long-Evans, 38, 41. See also Strain difference

Page 519: 0195162854

502 INDEX

Lordosis, 310Lympocytes, 246

Ml, 61Macrophages, 246Macrosomatic, 90Macrovibrissae, 82Map, 60Maternal behavior, 287-297

care, 323, 325, 326measurement, 290-295memory, 292methodology, 294-295

Maudsley reactive, 30Medial frontal cortex, 223Medial prefrontal cortex, 427Medial preoptic area, 295, 313Memory, 384, 428-432Methylphenidate, 459Microvibrissae, 82Midazolam, 359Milk letdown, 270Milk production, 273Milk teeth, 266Monocytes, 246Morphine, 122, 260Morris water task, 283, 328, 396Motivation, 376Motor cortex, 168Motor learning, 62-63Multiparous, 295Murinae, 6,7Muroidea, 6Muscimol, 359Music, 326Myelin, 326

N:NIH rat, 30Nasogenital contact, 345Nasonasal contact, 345Neophobia, 16, 204, 363-366Nest building, 289-290Nests, 15Neurological, 449-461Neurological models, 449—461Neuronal morphology, 325Neuropathetic pain, 69Neuroplasticity, 324Neuropsychological tests, 475-498Neurotropic factor, 324Nociceptors, 69. See also PainNorway rat, 363Novel object preference, 387-390Nucleus accumbens, 223Nursing, 288-289

Obesity, 324Object. See Configural learning; Spatial learningObject recognition, 383-391Object recognition memory, 383Observational learning, 363-370

Obsessive compulsive disorder, 468Occasion-setting, 430Odor, 16, 64, 90-3104, 164-165

aversion, 95diffusion, 92discrimination, 97genes, 91habituation, 92-93learning, 93-95lesion, 99-101matching-to-sample, 97non-matching-to-sample, 97preference, 93psychophysics, 97-98quality perception, 98-99stimuli, 91-92

Olfactory bulb, 90Olfactory epithelium, 90Olfactory measurement, 95-96Open field test, 324, 327Operant conditioning, 201Optimal foraging, 217, 223Oral grasp, 261Orienting, 129-149

asymmetric, 131-132cylinder test, 132-134sticky dot test, 130-3132tactile, 130-3132

Origins, 11Orofacial fixed action patterns, 437Oromotor, 105Oromotor response to taste, 109Osmoreceptors, 207Outward trip, 405Oxygen capacity, 35

Pacing behavior, 311-313Pain, 69-80, 373Pain modules, 70-373Pain pressure, 74Parabrachial, 417Parabrachial nucleus, 105, 285Paraesthesia, 69Parkinson's disease, 248, 168, 449, 453-454Parturition, 269Pavlovian conditioning, 200, 428, 436-437Paw preference, 165Paw withdrawal, 73. See also PainPelagic hair, 81Periadeductal gray, 373, 417Phagocytosis, 246Phantom pain, 70Phenotype, 28Pheromones, 326Phylogenic hypothesis, 5Phylogeny, 10-311Physostigmine, 359Piloerection, 347Piloting, 394-399. See also Spatial learningPivoting, 138Place navigation, 283, 394

Page 520: 0195162854

Index 503

Place preference task, 396Place probe, 402Placing, 125, 129-149

forelimb, 135wheelbarrow, 136

Play, 11,298-306. See also Developmentaggression, 304behavior, 301-305fighting, 301-305origins, 302-303tactics, 301-302

Poison, 18Polymorphonudear phagocytes, 246Posterior parietal cortex, 451Postural development, 268-270Posture, 121-128,481Pouncing, 280Prandial drinking, 212Predators, 412-414Prefrontal cortex, 451Pregnancy, 325, 326Prenatal, 258-260Prenatal behavior, 257-266Prenatal learning, 364-365Preoptic, 229Preovulatory, 308Preparatory, 438Primiparous, 295. See also Maternal behaviorProcedural memory, 394Preceptive behavior, 310-3311. See also Sexual behaviorProestrus, 310. See also Sexual behaviorProgesterone, 308Proprioceptive response, 125-126, 262-263Protein, 203Provocative signals, 345Psychiatric models, 462-475Psychiatric, 462-474Psychiatry, 462Psychopathology, 466-468

obsessive-compulsive, 468Pup licking, 288Pup retrieval, 288Pyramidal tract, 168

Quinpirole, 469. See also Obsessive compulsive disorder

Radial arm maze, 395Rat breath, 366-367. See also Social behaviorReaching, 279, 279Red nucleus, 65,168Releaser, 22Representations, 424-425. See also Incentive behaviorResident-intruder test, 38, 349-350Responsivness, 477-478. See also Sensorimotor behaviorRetrograde, 389Righting, 123-124

body tactile, 125-126tests of, 124-127trigeminal, 124-125vestibular, 126visual, 127

Risk assessment, 335Robbing, 280. See also FoodRodentia, 3-6Rotorod, 31. See also Neuropsychological testsRoute learning, 393. See also Spatial learningRung walking, 486-487Running, 30

Scent marking, 20Schedule-induced polydipsia, 38-39Schizophrenia, 463Selected lines, 29Selection, 30Selective attention, 426-427Selective breeding, 30Self-administration, 40Self-mutilation, 70Sensitization, 293Sensorimotor behavior, 478—481. See also SomatosensorySensorimotor capacity, 31Septum, 358, 407Serotonin, 202Set point, 227Sex difference and dodging, 220Sex differences, 197Sexual behavior, 307-318

female, 307-315male, 315-118measurement, 315-316neural substrates of, 313-314, 316-317

Sexual development, 281Sexual motivation, 314-315, 317Shivering, 232Siatic nerve, 71. See also PainSingle pellet reaching, 486Sinus hairs, 60, 65Skilled movement, 162-169,485-486

food handling, 163-164grasping, 162-163limb guidance, 164-165limb structure, 162-163reaching, 165-166strain differences, 167

Social behavior, 236, 238, 241-242ethology, 22interactions, 18-19learning, 363-370play, 280posture, 346-348relationships, 19-20signals, 344-346status, 21stress, 21-22

Sodium appetite, 214-215Solitary tract, 105Somatosensory, 60-368Somatosensory receptive field, 61Sonogram, 345Sparague-Dawley, 38. See also Strain differenceSpatial acuity, 51Spatial learning, 56-57, 392-400

Page 521: 0195162854

504 INDEX

memory, 283navigation, 394-399tests of, 394-394theories of, 392-393

Species-typical, 450Spinal cord, 157-158Startle, 412Stereotyped behavior, 202Sticky dot, 480Stimulus response theories, 422-424Strain difference, 25, 36, 295

defensive burying, 356Stress, 236-244, 325-327

acute, 237chronic, 237coping strategy, 237-238immune system, 249-251

Stress and models of, 238-242Stretch response, 261,270Stria terminalis, 416Striate cortex, 54Striatum, 145, 454Stroke, 34, 452-A53Suckling, 269-270Sustained attention, 426Swimming, 138, 482^83Syntax, 143-144

Tactile hair, SI. See also Barrels; VibrissaeTactile hair function, 267Tactile stimulation, 259Tail-flick, 73Tapered beam test, 137Targets, 298Tarlov scale, 156Taste, 105-118,267

aversion, 108conditioning, 112reactivity, 108-110responsiveness, 107-108stimuli, 106-107

Taste-olfactory interactions, 114Taxon system, 393Taxonomy, 4Teeth, 3Teeth-chattering, 347Temperature, 267, 322, 479-480Temporal cortex, 451Territoriality, 334-352Thalamus, 62Thermoregulation, 226-235Threat posture, 19, 20Threat stimuli, 336-341

cat odor, 341intensity, 338

Thumb, 163. See also Skilled movementTime sampling, 290-3291T-maze, 394Tongue protrusion, 109Tourette syndrome, 148Trait, 33

Transitive inference, 429Transport reflex, 272Transportation, 326Transverse patterning, 429Tray reaching, 485-486Trigeminal nerve, 82-83Trimethylthiazoline, 240Trotting, 151

Ulcers, 236Ultrasonic emissions, 76Ultrasonic vocalization, 323, 371-380Ultrasound, 271Uterus, 258

Vaginal cytology, 309-310Variation, 26-27Ventilation, 323Ventromedial hypothalamus, 313Vernier acuity, 53Vestibular cues, 131,135Vestibular functions, 267Vibrissae, 16, 61, 81-89, 135. See also Barrels

cortex, 82distance detection, 86follicle, 82growth, 83kinematics, 84motor, 84object position, 83object size, 86-87organization, 82straddlers, 83

Viral vector mediated neurodegeneration, 455-456Vision, 54-59

critical period, 57deprivation, 56discrimination thresholds, 50function, 268plasticity, 57water task, 50

Vocalization, 271, 371-378anatomy of, 373characteristics of, 371environmental contexts of, 371-373functions of, 373-376measurement of, 377models of disorders, 376-377pain, 76

Volumetric thirst, 209Von Frey hair, 478-479. See also Somatosensory

Water deprivation, 212Water intake, 207-216Water measurement, 212-214Water task, 49Weaning, 273, 326Wild pigmented, 53Wistar, 30. See also Strain differenceWorking memory, 394Wrestling, 280. See also Development; Play