psychological review 73, no. 1,...

Psychological Review1966, Vol. 73, No. 1, 16-43

HABITUAT1ON:

A MODEL PHENOMENON FOR THE STUDY OFNEURONAL SUBSTRATES OF BEHAVIOR *

RICHARD F. THOMPSON AND WILLIAM ALDEN SPENCER2

University of Oregon Medical School

The recent habituation literature is reviewed with emphasis on neuro-physiological studies. The hindlimb flexion reflex of the acute spinalcat is used as a model system for analysis of the neuronal mechanismsinvolved in habituation and sensitization (i.e., dishabituation). Ha-bituation of this response is demonstrated to follow the same 9parametric relations for stimulus and training variables characteristicof behavioral response habituation in the intact organism. Habituationand sensitization appear to be central neural processes and probably donot involve presynaptic or postsynaptic inhibition. It is suggested thatthey may result from the interaction of neural processes resembling"polysynaptic low-frequency depression," and "facilitatory afterdis-charge." "Membrane desensitization" may play a role in long-lastinghabituation.

Perhaps the most fundamental prob-lem in physiological psychology is theelucidation of neuronal mechanismsunderlying alterations in behavior.This implies prediction of behavioralchanges from descriptions in terms ofpatterns of synaptic interactions amongidentifiable populations of neurons.Two general strategies in vogue todayare (a) correlational studies relatingchanges in various neuronal responsesof the intact organism to changes inbehavior (cf. Morrell, 1961), and (&)analytic studies of relatively long-termchanges induced in simplified neuronalsystems such as the monosynaptic re-flex (cf. Eccles, 1964). At present,phenomena measured in the intact or-

1 Supported in part by NIH ResearchGrants NB-03494 and B-2161 and by NIHResearch Career Awards l-K-NB-19485 andK3-MH-6650. The authors are indebted toJohn M. Brookhart, Judson S. Brown, SethSharpless, and Eric Kandel for helpful com-ments and suggestions.

2 W. A. Spencer's current address: De-partment of Physiology, New York Univer-sity School of Medicine.

ganism cannot be analyzed in termsof neuronal mechanisms. Analyticstudies, while utilizing simpler systems,generally deal with changes whose rela-tionships to alterations in the behaviorof the intact organism are not alwaysclear.

The solution to this impasse wouldseem to be a type of relatively long-lasting and nontrivial alteration in be-havior as a result of training which hasidentical characteristics both in the in-tact organism and in simplified prep-arations, where some degree ofneuronal analysis is possible. Therehas been a resurgence of interest inrecent years in the phenomena ofhabituation, due in part to studies ofneural response habituation (e.g.,Hernandez-Peon, Jouvet, & Scherrer,1957; Sharpless & Jasper, 1956). Be-havioral response habituation appearsto satisfy both the requirements ofbehavioral significance and occurrencein simplified preparations. A widevariety of responses of the intact or-ganism habituate (Harris, 1943), as

16

HABITUATION : A MODEL PHENOMENON 17

do reflex responses of the chronicspinal animal (Prosser & Hunter,1936). Sufficient information has beendeveloped about the neural organiza-tion of the spinal cord to permit somedegree of analysis of reflex mechanisms(e.g., Eccles, 1964). This paper willdiscuss recent studies of habituationwith emphasis on neurophysiologicalfindings. We will attempt to demon-strate that habituation of the spinalflexion reflex is identical in characteris-tics to habituation of more complexresponses of the intact organism and isat least partially amenable to analysisin terms of neuronal mechanisms.Original data to be reviewed were allobtained using reflex activity of a"model" mammalian system: the acutespinal cat.

DEFINITIONS

Harris (1943), in his careful andcomprehensive review of the earlierliterature, defined habituation opera-tionally to be "response decrement as aresult of repeated stimulation [p.385]." Humphrey's (1933) earlierdefinition, although expressed in less-behavioral terms, is substantially thesame. Response decrement resultingfrom very rapid stimulation (e.g., 500/second) which might be due to invasionof the neuronal relative refractoryperiod has different characteristics andis best excluded, as are decrementsresulting from trauma, growth, aging,etc. Habituation can usually be dis-tinguished from the latter phenomenabecause it is reversible: Habituatedresponses exhibit spontaneous recov-ery. The only authority to divergefrom Harris' characterization of habit-uation is Thorpe (1956), who regardsit as a "relatively permanent waningof a response as a result of repeatedstimulation [p. 54; italics ours]." Itwill be seen that the time course forspontaneous recovery of an habituated

response depends on a great manyvariables; hence distinctions in termsof recovery time would seem somewhatarbitrary. Certainly Harris' definitionhas gained widespread acceptance.Several authors have suggested a dis-tinction between "short-term" and"long-term" habituation processes(Shapless & Jasper, 1956; Sokolov,1955).

Two peripheral mechanisms haveoften been assumed either explicitly orimplicitly to account for response de-crement : alterations in receptors or ineffectors. These may be given morespecific operational definitions whichpermit empirical evaluation. In thispaper, a decrease in receptor activitywill be called receptor adaptation, anda decrease limited to the effector re-sponse (including both effector andneuroeffector junction) will be termedeffector fatigue. Following Harris(1943), response decrements that canbe accounted for entirely by thesemechanisms will not be consideredhabituation.

If a common-denominator definitionof learning such as "change in behaviorunder conditions of practice" is adopted,habituation must be included as anaspect of learning. Habituation can bedistinguished from many phenomena oflearning in that the change in behavioris a decrease in response strength. Itis customary, however, to assume thatthe response exhibiting habituation isitself unlearned (Harris, 1943).James (1890) and Konorski (1948)have proposed a more inclusive term,plasticity, to include habituation, condi-tioning, and other aspects of behavioralchange. Our general definition ofhabituation may also apply to a varietyof other terms: adaptation, accommo-dation, fatigue, inhibition, negativelearning, extinction, stimulus satiation,etc. Habituation will be used here tostand for whichever term the reader

18 RICHARD F. THOMPSON AND WILLIAM ALDEN SPENCER

prefers. Martin (1964) has recentlyproposed that such diverse constructsas response habituation, evolutionaryadaptation, and photochemical darkadaptation have a common conceptualbasis. The utility of this approach isunclear.

PARAMETRIC CHARACTERISTICS OFHABITUATION

The extensive research on behavioral re-sponse habituation in the intact vertebratehas demonstrated a number of parametricrelations for stimulus and training variables.Although not previously enumerated as such,examples of all of these may be found in theearlier literature for various reflex responses.Some of the more recent observations, par-ticularly those dealing with types of re-sponses not covered in Harris' (1943) re-view, will be noted in the brief summarythat follows.

1. Given that a particular stimuluselicits a response, repeated applicationsoj the stimulus result in decreased re-sponse (habituation}. The decrease isusually a negative exponential functionof the number of stimulus presentations.

Examples of response habituation canprobably be found in essentially all behav-ioral studies where a stimulus is regularlypresented. In earlier experiments devotedto habituation, per se, parametric character-istics were studied for a variety of responses(cf. Harris, 1943) ranging from postrotatorynystagmus (Griffith, 1920; Wendt, 1951) tostartle (Prosser & Hunter, 1936) and gal-vanic skin response (GSR—Davis, 1934).With the possible exception of the "kneejerk" reflex (Lombard, 1887; Prosser &Hunter, 1936) habituation was a consistentfinding, usually exhibiting an exponentialcourse. The recent habituation literaturecontains many investigations of various re-sponses which can be subsumed under thegeneral heading of response to novel stimu-lation, ranging in complexity from simplestartle through variously defined orientingreflexes to curiosity, stimulus satiation, andplay behavior (Berlyne, 1950, 1960; Bindra,1959; Butler & Harlow, 1954; Davis, Buch-wald, & Frankmann, 1955; Dember, 1961;Dykman, Reese, Galbrecht, & Thomasson,1959; Engen & Lipsitt, 1965; Glanzer, 1953;

Melzak, 1961; Montgomery, 1953; Robinson& Gantt, 1947; Rodgers, Melzak, & Segal,1963; Scholander, 1960; Sokolov, 1960;Welker, 1956, 1961). Interestingly, if asingle novel stimulus is repeated regularly,habituation of approach behavior follows anexponential course (Glanzer, 1953). Un-learned behavior sequences of birds elicitedby certain categories of stimuli have alsobeen found to habituate under some condi-tions (Hinde, 1954; Thorpe, 1956). Finally,much of the extensive research on responsedecrement or fatigue under conditions ofwork (Bartley & Chute, 1947; Robinson,1934; Solomon, 1948) can be consideredwithin the framework of habituation.

2. // the stimulus is withheld, theresponse tends to recover over time(spontaneous recovery).

Spontaneous recovery is reported in mostof the studies noted above and has come tobe the most common method of demonstrat-ing that a given response decrement is anexample of habituation (Harris, 1943). Thetime course of spontaneous recovery ismarkedly influenced by many variables andis not necessarily characteristic of a givenresponse. Thus the habituated startle re-sponse to sound in the intact rat may recoverin 10 minutes (Prosser & Hunter, 1936) orfail to recover in 24 hours,8 depending upondetails of testing. Consequently any cate-gorization of types of habituation basedsolely on recovery time is likely to be some-what artificial.

3. // repeated series of habituationtraining and spontaneous recovery aregiven, habituation becomes successivelymore rapid (this might be called poten-tiation of habituation).

Humphrey (1933) noted this effect in hisstudies on turtle leg withdrawal to shell tap.Konorski (1948) describes it for the orient-ing response, and it has been described inmany studies where repeated habituationseries were given (e.g., Davis, 1934).

4. Other things being equal, themore rapid the frequency of stimula-tion, the more rapid and/or morepronounced is habituation.

3 J. S. Brown, personal communication,1964.


Numerous examples of this were noted inthe earlier reflex studies (Harris, 1943) aswell as in more recent work on stimulussatiation and curiosity (Glanzer, 1953;Welker, 1961). The effect occurs in termsof real time course and occurs within certainlimits in terms of number of trials as well.

5. The weaker the stimulus, themore rapid and/or more pronounced ishabituation. Strong stimuli may yieldno significant habituation.

This relationship is characteristic of mosttypes of responses ranging from simple re-flexes (Harris, 1943) to complex explora-tory behavior (Welker, 1961).*

6. The effects of habituation trainingmay proceed beyond the zero or asymp-totic response level.

Additional habituation training given afterthe response has disappeared or reached astable habituated level will result in slowerrecovery. Although relatively few experi-ments have studied "below-zero" habituationas such (Humphrey, 1933; Prosser &Hunter, 1936; Wendt, 1931), the observa-tions may be viewed as an extension of therelationship between number of stimuluspresentations and degree of habituation.Zero response level is of course to somedegree dependent upon the particular re-sponse measures used.

7. Habituation of response to agiven stimulus exhibits stimulus gener-alisation to other stimuli.

Coombs (1938) demonstrated generaliza-tion of GSR habituation to different typesof auditory stimulation, and Porter (1938)demonstrated cross-modal generalization ofthe habituated GSR for light and tonestimuli. Mowrer (1934) showed some gen-eralization of postrotatory nystagmus habitu-ation in the pigeon. In a recent study,Crampton and Schwam (1961) reportedgeneralization of optic nystagmus habituationin the cat to different degrees of angularacceleration.

4 Postrotary optic nystagmus may be anexception in that under some conditions thedegree of habituation is directly related tovelocity of rotation (G. Crampton, personalcommunication, 1964).

8. Presentation of another (usuallystrong} stimulus results in recovery ofthe habituated response (dishabitua-tion).

This phenomenon appears to be as ubiqui-tous as habituation itself and is commonlyused to demonstrate that habituation has oc-curred. Pavlov (1927) was perhaps the firstto describe this process (i.e., disinhibition)in relation to an extinguished conditioned re-sponse (CR), but also applied it to thehabituated orienting response. Humphrey(1933) studied dishabituation extensively inlower vertebrates. Essentially all responsesof mammals that can be habituated can alsobe dishabituated (Harris, 1943). It is notalways necessary for the dishabituatory stim-ulus to be strong. In fact Sokolov (1960)and Voronin and Sokolov (1960) reportedthat a decrease in the intensity of an auditorystimulus results in dishabituation of the ha-bituated orienting response in humans. Dis-habituation, viewed as neutralization of theprocess of habituation (Humphrey, 1933),has been perhaps the most important methodof distinguishing between habituation and"fatigue."

9. Upon repeated application of thedishabituatory stimulus, the amount ofdishabituation produced habituates(this might be called habituation ofdishabituation).

Most studies of dishabituation (see above)have noted its habituation. Lehner (1941)has done the most careful parametric studies,showing that habituation of dishabituationfollows a negative exponential course for thestartle response in the rat and the abdominalreflex in man. More recently, Hagbarth andKugelberg (1958) and Hagbarth and Finer(1963) verified and extended Lehner's find-ings for the abdominal and leg flexion re-flexes in humans. Crampton and Schwam(1961) have shown that dishabituation ofpostrotatory nystagmus in the cat by audi-tory or cutaneous stimuli habituates in asimilar fashion.

In reviewing the behavioral habitua-tion literature, it is striking to findvirtually complete agreement on theparametric characteristics of the phe-nomenon in such a wide variety ofanimals and responses. These nine


common characteristics may conse-quently serve as the detailed opera-tional definition of habituation, replac-ing the more general definition givenabove. The extent to which any otherresponse decrements satisfy these char-acteristics will thus determine whetherthey can be called habituation.

NEUROPHYSIOLOGICAL STUDIES OFHABITUATION

Electroencephalogram (EEC) Arousal

The EEC arousal response—a rapidtransition from slower waves to fastlow-voltage activity, usually followingsudden stimulation—was observed todecrease upon repeated stimulation inearly studies both in animals (e.g.,Rhineberger & Jasper, 1937) and man(Knott & Henry, 1941). The firstinvestigation of this effect, per se, wasthe now classic study by Sharpless andJasper (1956). Using repeated pres-entations of brief tones they found thatcortical EEG arousal of the sleeping catbecomes shorter and finally disappears.After cessation of stimulation, thearousal response exhibited spontaneousrecovery over a period of minutes tohours. Although not studied para-metrically, examples of all nine rela-tions listed above for behavioral re-sponses are to be found here and insubsequent confirmatory studies (Cas-pers, Lerche, & Greuter, 1958; Roig &Sommer-Smith, 1959; Segundo, Roig,& Sommer-Smith, 1959). Apelbaum,Silva, Frick, and Segundo (1960)noted a more restricted generalizationgradient when using waking cats thanSharpless and Jasper found with sleep-ing animals. The human "alpha block-ing" response has been shown to habit-uate to tactile, auditory, and visualstimulation (Sokolov, 1960; Voronin &Sokolov, 1960). Cortical EEG arousalalso exhibits habituation, spontaneousrecovery, and dishabituation when

elicited by electrical stimulation of themesencephalic reticular formation(Glickman, 1958; Glickman & Feld-man, 1961; Roig & Sommer-Smith,1959). Habituation of electrographicarousal has also been reported for thal-amus and reticular formation (Sharp-less & Jasper, 1956), hippocampus(Green & Arduini, 1954), and olfac-tory bulb (Adrian, 1950; Hernandez-Peon, Lavin, Alcocer-Cuaron, & Mar-celin, 1960).

Sensory Evoked Potentials

A number of investigators have re-ported habituation of click- or tone-evoked responses in various regionsof the central nervous system (CNS)ranging from the cochlear nucleus tothe auditory cortex and mesencephalicreticular formation in the waking ani-mal (Galambos, Sheatz, & Vernier,1956; Gershuni, Kozhevnikov, Maru-seva, Avakyan, Radionova, Altman, &Soroko, 1960; Hernandez-Peon et al.,1957; Hernandez-Peon & Scherrer,1955; Jouvet & Hernandez-Peon,1957). Worden and Marsh (1963)have emphasized the importance ofcontrolling stimulus intensity, demon-strating statistically reliable changes inresponse, and conducting other neces-sary control procedures which were notalways observed in the studies listedabove. Experiments satisfying allthese methodological criteria have not,for the most part, found any consistentor significant evidence of habituationof auditory evoked responses in thewaking animal (Huttenlocher, 1960;Marsh, McCarthy, Sheatz, & Galam-bos, 1961; Sharpless & Jasper, 1956;Worden & Marsh, 1963). HoweverMarsh and Worden (1964) have re-cently reported decreases in click-evoked response amplitudes from theauditory cortex when the EEG was de-synchronized, even in the absence ofany consistent response changes at the


cochlear nucleus. In a recent and care-ful study, Webster, Dunlop, Simons,and Aitkin (1965) reported a rapidinitial decrease in the amplitude ofsound-burst-evoked responses of thecochlear nucleus that occurred duringthe first 30 seconds of stimulation, ifthe sound (20-millisecond duration)was presented 1/second or faster. Nodecreases occurred with slower ratesof stimulation. (The experiments de-scribed above generally presented stim-uli slower than 1/second and examinedresponses over periods of many min-utes or hours.) Auditory evoked re-sponses of the reticular formation doappear to habituate in the sleeping ani-mal (Huttenlocher, 1960). Auditoryevoked cortical association responsesdo not habituate (Shaw & Thompson,1964) ; in fact the amplitude of this re-sponse increases during the first fewpresentations of a click (Thompson &Shaw, 1965). Head position in thesound field is a powerful and consistentdeterminer of evoked-response ampli-tude (Marsh, Worden, & Hicks,1962). Contractions of middle-earmuscles have been implicated in somestudies of auditory evoked responsehabituation (Guzman-Flores, Alcaraz,& Harmony, 1960; Hugelin, Dumont,& Paillas, 1960), but denied by otherworkers (Altman, 1960; Moushegan,Rupert, Marsh, & Galambos, 1961).Further discussion of methodologicalproblems is given in the careful studiesby Marsh et al. (1962) and Wordenand Marsh (1963). (See also the dis-cussion by Deutsch, 1962.)

Many reports have noted habituationof evoked responses from various re-gions of the visual system to repeatedflash stimulation (Cavaggioni, Gian-nelli, & Santibafiez, 1959; Hernandez-Peon, Guzman-Flores, Alcaraz, &Fernandez-Guardiola, 1958; Mancia,Meulders, & Santibanez, 1959; Pales-tini, Davidovich, & Hernandez-Peon,

1959). Pupil diameter apparently wasnot controlled in many of these earlystudies. When the eye is atropinizedand an artificial pupil used, the evi-dence for visual evoked response habit-uation is inconsistent. Fernandez-Guardiola, Roldan, Fanjul, and Cas-tells (1961) found no habituation ofresponses to flash under these condi-tions recording from the optic tract andlateral geniculate body and inconsistentchanges from the visual cortex: Somepreparations showed partial decreases,particularly of the later components ofthe evoked response, and others showedno decreases. Naquet, Regis, Fischer-Williams, and Fernandez-Guardiola(1960) noted that flash-evoked re-sponses of the visual cortex decreasedwhen pupil diameter increased (con-comitant with EEG desynchroniza-tion), but did not change when an arti-ficial pupil was used. Affanni, Man-cia, and Marchiafava (1962) reportedno reduction in responses of the visualcortex to a flash repeatedly presentedfor hours when an artificial pupil wasused. On the other hand, Garcia-Austt (1963) reported decreases inscalp averaged evoked responses toflash in humans with artificial pupils,and decreases of averaged visualevoked responses from the rat after re-moval of the iris. Responses evokedby somatic sensory stimulation havebeen reported to habituate (Hernan-dez-Peon, 1960; Hernandez-Peon,Scherrer, & Velasco, 1956), and toexhibit no habituation (Santibanez,Trouche, & Albe-Fessard, 1963).

Conclusion

The evidence for EEG arousal habit-uation is clear and consistent; it ex-hibits essentially all parametric rela-tions characteristic of behavioral re-sponse habituation. Studies of evokedresponses, on the other hand, permitonly one safe generalization at present;


There is as yet no consistent agreementregarding the occurrence of habituationin sensory systems of the waking brain.Where habituation has been reported,it may be the result of decreases in ef-fective stimulus intensity due to altera-tions in head position, middle-earmuscle contractions, pupilary constric-tion, etc., or an indirect result ofchanges in state of EEG arousal. Theissue can only be settled by furtherresearch.

EFFECTS OF NEURAL LESIONS ONBEHAVIORAL RESPONSE HABITUATION

Although the majority of studies de-scribed here were not planned as para-metric investigations of the effects oflesions on behavioral response habitua-tion, per se, many were so designedthat such effects could be noted. Theorienting response exhibits markedlyreduced habituation following chronictotal decortication in dogs (Konorski,1948; Lebidinskaia & Rosenthal, 1935)and chronic decerebration in cats(Bard & Macht, 1958). Lesions ofauditory cortex in the cat reduce theextent of intersession habituation oforientation to sound (Thompson &Welker, 1963). Activity level of mon-keys with Cortical Area 9 lesions ex-hibits less habituation than that of nor-mal animals (French, 1959; French &Harlow, 1955). Frontal or temporallesions in monkeys reduce the amountof visual exploration and decrease theextent of habituation in repeated test-ing sessions (Butler & Harlow, 1954;Symmes, 1959, 1963). Frontal lesionshave a comparable effect in the cat(Hagamen, Lance, & Ungewitter,1959). Damage to frontal cortex de-creases the extent of habituation of theheart-rate response to the cold-pressortest in man and the rat (Glaser & Grif-fin, 1962; Griffin, 1963). Schwartz-baum, Wilson, and Morrissette (1961)report less-rapid habituation to novel

stimuli following bilateral amygdalec-tomy. Pribram (1963) reports unpub-lished results by Kimble and Bagshawshowing comparable effects of amygda-lectomy on GSR response to novelstimulation and emphasizes the import-ance of this structure in habituation.

Response alternation or variability inmaze behavior can be related to habit-uation of response to novel stimuli (seebelow; also Glanzer, 1953; Welker,1961). Partial cortical lesions in therat reduce the tendency to vary move-ment through a simple maze (Krechev-sky, 1937a, 1937b, 1937c). Frontallesions are more effective than parietalor occipital lesions (Morgan & Wood,1943). Roberts, Dember, and Brod-wick (1962) found that bilateral abla-tion of the hippocampus resulted in lesshabituation of T-maze responses andless alternation behavior.

Results of these studies, involvingvarious responses and many differentneural structures, are in essentiallycomplete agreement. Brain damagetends to produce a reduction in theextent of behavioral response habitua-tion. There is some degree of spe-cificity in that damage to structuresrelated to particular responses tends tohave a greater effect on habituation.It would seem that habituation maybe mediated, at least in part, by what-ever neural structures are most in-volved in a particular response, sug-gesting that common neuronal mechan-isms may be involved.

SPINAL REFLEXES

Sherrington noted a decrement ofthe digital flexion of the monkey in1898, but his experiments on the acutespinal dog (1906) were the first toanalyze habituation of spinal reflexesin detail. Although he used the termfatigue, he was careful not to implyany necessary meaning other than re-sponse decrement by the term. He


found that continuous or repeatedmechanical or electrical stimulation ofthe appropriate skin region resulted ina decrement of the scratch reflex andthe flexion reflex, the former decreas-ing more rapidly. Both exhibitedspontaneous recovery. Several lines ofevidence suggested that both receptoradaptation and muscle fatigue could beruled out. Indirect evidence from avariety of reflex-interaction studies ledSherrington to conclude that the"final common path," that is, the motorneurons, were not fatigued. Prosserand Hunter (1936) completed a care-ful and extensive study of habituationand dishabituation of the flexion reflexin the chronic spinal rat, using elec-trical or mechanical stimulation of thelegs, tail, and hindlimb nerves. Theydemonstrated habituation, spontaneousrecovery, and dishabituation by strongextraneous stimulation. Weak stimuliled to more rapid habituation thanstrong stimuli. In one experimentthey showed that no habituation oc-curred in the knee-jerk reflex of aspinal cat. Lehner (1941) demon-strated that the duration of dishabitua-tion of the flexion reflex in the spinalrat habituated.

In conditioning studies of the acutespinal dog, Shurrager and Culler(1941) and Shurrager and Shurrager(1941) demonstrated habituation ofthe leg flexion reflex to tail stimula-tion. More recently, Nesmeianova(1957) has demonstrated habituation,spontaneous recovery, and dishabitua-tion of the scratch reflex and theflexion reflex in chronic spinal dogs.Habituation below zero was marked;the response sometimes required daysor weeks to recover after long periodsof training. Kozak, McFarlane, andWesterman (1962) have reported sim-ilarly long-lasting habituation of flexionand scratch reflexes in the chronicspinal kitten, Habituation of the

scratch reflex has also been reportedfor the chronic spinal frog (Afelt,1963). Lethlean (1965) has presentedevidence suggesting that the flexion re-flex may continue to habituate for manyminutes after the stimulus has beenwithdrawn in the chronic spinal rat.Very recent experiments by Buchwald,Halas, and Schramm (1964, 1965) re-ported habituation of single-unit dis-charges of motor neurons recordedfrom ventral root fibers in the acutespinal cat, and a preliminary summaryof our own work has already appeared(Spencer, Thompson, & Neilson, 1964).

REFLEX HABITUATION IN THEACUTE SPINAL CAT

Studies reviewed above indicated thathabituation of the flexion reflex in the acutemammalian spinal preparation appears toexhibit some of the parametric relationscharacteristic of behavioral response habitu-ation in the intact organism. Since a gooddeal is known about the synaptic organiza-tion of the spinal cord, this preparation waschosen as a model system in which to studydetailed characteristics of habituation andthe possible neuronal mechanisms involved.

The preparation used in these experimentswas the acute unanesthetized decerebrate catwith spinal transection at T-12. Proceduresare described in. detail in papers (in prepara-tion) devoted to analyses of neuronal mech-anisms. Data given here are all in terms ofisometric "twitch" contraction strength ofthe tibialis anterior muscle, a hindlimb flexorparticipating in the generalized limb flexionreflex to afferent stimulation. Other muscleswere also studied to insure generality ofobservations. The flexion reflex is poly-synaptic, at least one interneuron interveningbetween the incoming afferent fibers andthe output motor neurons. Dishabituatingstimuli were either electrical or natural, thatis, pinch. The basic phenomena of habitu-ation and spontaneous recovery were ob-served on all of the more than 100 catsstudied. Spontaneous recovery was used asthe criterion that habituation had occurredrather than response decrement due to de-terioration or various artifacts. Althoughnot shown in many of the figures, spon-taneous recovery was always obtained fol-lowing every habituation series.


100

50

J U

8 \e\ 24 32 40 48 56 64 72 80 88 96 104 HZ 120 128

0 3 6 9 12 5 18 21 24

FIG. 1. A: Habituation (zero minutes to arrow) and spontaneous recovery (arrow to128 minutes) of the hindlimb flexion reflex of the spinal cat in response to repeated skinshocks. (Stimuli were brief trains of shocks— 5 in SO milliseconds— delivered every 10 sec-onds during habituation and every 3 minutes during spontaneous recovery, except for a 12-minute period of no stimuli at about 100 minutes. In this and all subsequent figures theresponse measured is tension developed by contraction of the tibialis anterior muscle, ex-pressed as a percentage of mean initial-control response amplitude. B : Effect of repeated-habittiation and spontaneous-recovery series on degree of habituation. Response recoveredto control level following first habituation series and was then rehabituated— second series.Conditions as in A. Data are averages of 10-trial blocks. C : Effect of stimulus frequencyon habituation. Single shocks given 1/3.2 seconds in one habituation series and 1 /secondin the other to the saphenous nerve. Data are averages of 10-trial blocks. D : Effect ofstimulus intensity on habituation. Brief trains of shocks, as in A, were delivered every10 seconds to the saphenous nerve with spontaneous recovery allowed after each series.Voltages refer to output of stimulator and were attenuated, but in the same ratios, whendelivered to the nerve. Data averaged over 3-trial blocks. E: Stimulus generalization ofhabituation. Single shocks to two separate branches of the saphenous nerve. The habituatingstimulus to one branch was given I/second, and the test stimulus to the other branch wasgiven I/minute. Data are averages over 3-trial blocks for response to the test stimulusand averages over the same periods of time for response to the habituating stimulus.)


In all experiments, initial-control responseamplitudes were first determined by teststimuli delivered no more frequently than1/minute. Subsequent response amplitudesare plotted in terms of percentage of meancontrol level. Habituating stimuli wereeither single shocks (.01-.03-millisecond du-ration) or brief trains of five such shocksdelivered over a 50-millisecond period. Al-though trains generally gave somewhat morerapid and less variable habituation, the para-metric relations of all other variables werecomparable for both. Stimuli were deliveredto skin or to afferent nerves.

Characteristic habituation and spontaneousrecovery of the spinal flexion reflex areshown in Figure 1A. Brief trains of shockswere delivered 1/10 seconds to skin duringhabituation (0 minutes to arrow) and 1/3minute during recovery. Individual re-sponses are given to illustrate normal vari-ability. Note the exponential form of ha-bituation and the long and variable timecourse of spontaneous recovery. Althoughthe time courses of habituation and spon-taneous recovery are dependent on the manyvariables described below, 40 separate seriesinvolving a variety of conditions were se-lected at random from our experiments todetermine average values. The mean habitu-ation time was 9.4 minutes (range 1.2-33minutes), and the mean recovery time was25.7 minutes (range .5 minute-108 minutes).

The effect of repeated habituation-spon-taneous recovery series on habituation is il-lustrated in Figure IB. Habituation andrecovery conditions are the same as in 1A,and the data shown are averages for 10-trialblocks. The second series exhibits signifi-cantly more habituation than the first (p<.01; sign test).

Figure 1C shows the effect of stimulusfrequency. Here single shocks were de-livered either at I/second or 1/3.2 seconds tothe saphenous nerve. Data are averagesover 16-second blocks of time. Habituationis significantly greater with the more rapidstimulus (p <.01; sign test).

The effect of stimulus intensity on habitu-ation is indicated in Figure ID. Stimuliwere brief trains of shocks to the saphenousnerve every 10 seconds at the indicatedvoltages. Data are averages of 3-trialblocks. There is no habituation for the 45-volt stimulus, significant habituation for the25-volt stimulus, and significantly greaterhabituation for the 7-volt stimulus (p <.01for each comparison; sign test). Thus rate

and degree of habituation are inversely re-lated to stimulus strength.

The development of stimulus generaliza-tion of habituation is shown in Figure IE.Single shocks were delivered to two separatebranches of the saphenous nerve. The ha-bituating stimulus was given I/second, andthe generalization test stimulus to the othernerve branch was given I/minute. Dataare averages over 3-trial segments at com-parable times for the habituating stimulus.Response to the training stimulus habituatedto 10% of control level, and the test-stim-ulus response decreased to about 60% ofcontrol level. Amplitudes of the last 10habituated, generalized, and control responsesall differed significantly from each other(habituated versus control, t=l8.1, dj = 18,p <.01; generalized versus control, t — 2.43,df = 18, p <.05 ; habituated versus general-ized, f = 7.58, df= 18, p <01). Thus thereis significant but not complete generalizationof habituation from one input channel to an-other.

The effect of "below-zero" habituation onspontaneous recovery is shown in Figure 2A.Stimuli were brief trains to the sural nerveevery 5 seconds. In the recovery curves fol-lowing habituation to "zero," recovery beganwhen the response had reached a stablelevel. For the recovery curves followinghabituation below zero, habituation stimuliwere continued for 10-20 minutes after theresponse had reached a stable level. As seen,spontaneous recovery has a longer timecourse following below-zero habituation.The habituated levels (i.e., zero recoverytime) are comparable for both. The twoconditions were alternated 22 times in twoseparate experiments. The mean spon-taneous recovery time following habituationto zero was 2.1 minutes, significantly lessthan the 7-minute mean recovery time fol-lowing below-zero habituation (t — 4.45, df —20,/><.01).

Two examples of dishabituation are shownin Figure 2B. In both cases the response wasfirst habituated to single-shock stimulationof the tibial nerve at 1/3.2 seconds. Weakdishabituation was achieved by increasing thestimulus frequency to 3/second and strongdishabituation by a comparable increase infrequency plus an intensity increase. Ha-bituating stimuli were continued in eachseries. Data are averages of 8-trial blocks.The weak dishabituating stimulus yielded in-complete recovery, and the strong stimulusresulted in a response of greater than con-trol amplitude. In both cases the response


100

50 h

IHabituationp] 1ozero

Habituation-below zero

= " 0 2 3 4 5 6 7 8 9 10

.IOO

•O3

"|-50o<DV)cQ- 0(Da:

Strong dishabituation

2 3Time in minutes

C-

-

iMiS

IS

•t|!fcs»;S!E:

<j|%:

ill

•

SvvIS

;.;•%

•K

::?:fe:

:.:.:':!!:'•

||

i;?is

?:.W.K;'"•"•:•:•"-ISiiiil

•Xv.'ivX-

.piS!

''"•Iv"

;V~C

'::•&::

3'fa

V ., ,. •

'c"c

-20Cg

o3

15o-5•5 10'o

|B

^ n

D

-

ai;

l-f«...

2S

;>5

___

R 1

100 h

50 h

1st 2nd 3rd 4th ~ 1st 2nd 3rd 4thSuccessive dishabituations

FIG. 2. A: Effect of below-zero habituation on the time course of spontaneous recovery.(Stimuli were brief trains to the sural nerve 1/5 seconds during habituation and 1/minuteduring the recovery curves shown. Habituation in the to-zero condition was habituationtraining until the response reached a stable amplitude, and below zero was 10 minutes ofadditional habituation training after this point had been reached. The to-zero and below-zero conditions were alternated. B: Effects of weak and strong dishabituation on anhabituated response. Habituating stimulus was a single shock to the tibial nerve 1/3.2seconds. Weak dishabituation, bar, was an increase in stimulus frequency to 3/second for10 seconds, and strong dishabituation was an increase in both frequency and intensity for 10seconds. Data are averages of 8-trial blocks. C: Habituation of dishabituation. Bars repre-sent the percentage of increase in response amplitude following successive dishabituationsdelivered when the response was at a stable habituated level. Habituation conditions as in1A. Dishabituation was a strong pinch of the digits for S seconds. D: Same as 2Conly duration of the dishabituation effects are plotted.)


returned to habituated levels in about 4minutes.

Figures 2C and D indicate habituation ofdishabituation. Here four successive dis-habituating stimuli were given with recoveryallowed after each. The habituating stimuluswas a brief train of shocks to skin every 10seconds. Dishabituation was "natural" inthis experiment—strong pinching of thedigits. There is a reduction both in durationof dishabituation and in amplitude of dis-habituated responses following repeateddishabituation.

The data given above illustrate allnine parametric relations characteristicof response habituation in the intactanimal. Consequently the flexion re-flex of the acute spinal cat exhibitshabituation as genuine as that for anyresponse of the intact animal. Thetime relations for habituation and spon-taneous recovery here are within therange of those for responses of the in-tact animal. It is perhaps somewhatremarkable that a weak stimulus de-livered every few seconds for a fewminutes can produce a relatively long-lasting but reversible alteration in thebehavior of this very much "simplified"nervous system.

Independence of Habituation andDishabituation

As noted earlier, dishabituation by ex-traneous stimulation is customarily believeddue to disruption of the habituation process,whatever the latter may be. If this were thecase, dishabituation ought not to increase thehabituated response above its initial controllevel. However in Figure 2B a strong dis-habituatory stimulus was seen to increase theresponse well above the control level. Thisobservation suggests an alternative hypothe-sis, namely that dishabituation is a separatefacilitatory process superimposed upon thehabituated system and probably many otherresponse systems as well. Several testabledeductions may be made from this hy-pothesis. First, nonhabituated responsesought to exhibit dishabituation as well ashabituated responses. Figure 3A shows thisto be the case; the nonhabituated control-level response is increased markedly by adishabituatory stimulus.

10 12 14 16 16 20 22 24 26

FIG. 3. A: Effect of a strong dishabitu-ating stimulus on an unhabituated (i.e., con-trol-level) response. (Test stimulus was abrief train of shocks to skin every 2 minutes,and dishabituation was a strong shock traindelivered elsewhere on the limb for IS sec-onds. B: Independence of habituation anddishabituation. Habituation stimulus, arrowto bar, was brief skin-shock train given I/second and test stimulus given I/minute.Following the strong, pinch, dishabituatorystimulus, bar, the habituating stimulus wasdiscontinued and a test stimulus deliveredI/minute. Note that the response dishabitu-ates, then rehabilitates and spontaneously re-covers, even though there is no furtherhabituation training.)

A more crucial deduction concerns the roleof the habituating stimulus in the decay ofthe dishabituation effect. If dishabituation ispurely a disruption of habituation, a responsewhich is dishabituated following habituationought not to rehabituate unless additionalhabituation stimuli are given. Thus in Fig-ure 2B if the habituating stimulus wasturned off after dishabituation, the responseshould remain at the high dishabituated orcontrol level. Alternatively, if dishabituationis a superimposed facilitatory process havingan intrinsic decay time, the dishabituated re-sponse should gradually decrease toward thehabituated level, even if no habituation train-ing is given subsequent to dishabituation.This cruical test of the two alternative hy-potheses is illustrated in Figure 3B. Afterhabituation to brief trains at 1/second (ar-row to bar), the response is dishabituated(bar), and the habituating stimulus with-


held and introduced only I/minute to testresponse amplitude. As seen, the responseamplitude declines (or "rehabituates") al-most to the habituated level following dis-habituation, even though it is no longerbeing habituated. Subsequently it recoversspontaneously to the control level. Thisfinding has been consistently obtained inmore than 20 separate experiments. Todemonstrate it clearly, conditions must besuch that spontaneous recovery has a longtime course relative to dishabituation, other-wise the habituated response will partiallyrecover spontaneously by the time dishabitu-ation effects are dissipated. Since dishabitu-ation is not a disruption of habituation, butrather a separate process having all thecharacteristics of response sensitization (Hil-gard & Marquis, 1940), it will hereafter betermed sensitisation.

Although the independence of habituationand sensitization has not previously beendemonstrated as such, several investigatorshave suggested the possibility. Humphrey(1933), in his studies of habituation in thesnail and turtle, noted that a strong dis-habituatory stimulus caused a response ofgreater amplitude than prehabituated controlresponses and would also increase control-level responses prior to habituation. Hiscomment deserves quoting: "These resultsat once suggest . . . that what has beentermed 'dishabituation' is simply an increasein excitability, which, under conditions notknown, may summate with repeated stimu-lation [p. 143]." Sharpless and Jasper(1956) suggested that dishabituation ofhabituated EEG arousal was probably atransient superimposed facilitation. Pavlov(1927), it should be noted, used the sameparadigm we employed in Figure 3B todemonstrate that the disinhibitory effect ofan extraneous stimulus on a previously ex-tinguished CR was also transient and had thesame decay time whether or not the un-reinforced conditioned stimulus (CS) wasrepeatedly applied.

Sensitization thus emerges as a sep-arate and significant process in its ownright. A strong stimulus of very shortduration can cause an increase in theexcitability of the spinal cord lastingfor many minutes. Since sensitizationis independent of habituation, at leastfor the spinal flexion reflex, parametricrelations Numbers 8 and 9 are notreally relevant to habituation as such,

but rather describe a separate process.In maintaining the comparability ofhabituation characteristics for re-sponses of the spinal cord and the in-tact animal, we are forced to theprediction that the independence ofhabituation and sensitization will alsohold for responses of the intact animal.Future research will determine the ex-tent to which this is true. The habit-uation of sensitization (characteristicNumber 9), incidentally, now becomessimply an additional example of habit-uation, the response being increasedexcitability of spinal reflexes by thesensitizing stimulus.

Neuronal Mechanisms of Habituationand Sensitisation

Analytic studies of possible neuronalmechanisms involved in habituationand sensitization of the spinal flexionreflex are presented elsewhere (Spen-cer et al., 1964; also papers in prepara-tion). Results of these experimentswill be summarized briefly here. Re-ceptor adaptation was ruled out byelectrical stimulation of afferent nerveswith monitored neurograms indicatingno change in input volleys. Effectorfatigue was eliminated as a significantvariable by recording activity from ef-ferent nerves and ventral roots ratherthan muscles. The possible role offeedback from muscle activity waseliminated by nerve and root sectionand by drug paralysis of neuromuscularjunctions, including those betweengamma motor neurons and intrafusalmuscle fillers. These results indicatethat habituation and sensitization occurwithin the spinal cord. In numerousexperiments testing possible habitua-tion of the monosynaptic reflex wehave found no evidence of habituationin the sense of progressive decrementin response having a time course com-parable to flexion reflex habituation.

The excitability of motor neurons in-


volved in a habituated response can betested by interposing monosynaptic vol-leys activating them. In addition,using both ventral-root electrotonus asa population measure, and intracellularmicroelectrode recording of identifiedmotor-neuron membrane potentials forbehavior of neuronal elements, thesubliminal responses of motor neuronsduring habituation and dishabituationwere also studied. There was no back-ground alteration in the monosynapticreflex excitability of motor neuronsduring habituation, but there was aclear-cut decrease in motor-neuronpolysynaptic potentials. Thus what-ever processes are responsible forhabituation occur prior to motorneurons (i.e., in interneuron path-ways) and cannot be due to inhibitionacting on the motor neurons. Duringsensitization there was often, althoughnot invariably, an increase in theexcitability of motor neurons. Theseobservations further substantiate thebehavioral demonstration of the inde-pendence of habituation and sensitiza-tion.

The possibility that habituation isclue solely to altered conduction in af-ferent terminals of nerve fibers stimu-lated during training is ruled out bythe generalization experiment of Fig-ure IE. If habituation were due en-tirely to changes in afferent terminalsof the habituated channel, there wouldbe no generalization of habituation tothe other channel since completely sep-arate input channels were used. Gen-eralization of habituation did occur, sothe decrement must have involved cen-tral processes shared by both inputs.

To summarize, it has been provenbeyond reasonable doubt that habitua-tion of the spinal flexion reflex is notdependent upon receptor adaptation oreffector fatigue and occurs after theafferent input terminals but prior tothe output motor neurons in the spinal

cord. Consequently the process ofhabituation must occur in the courseof transmission through interneurons.Prosser and Hunter (1936), incident-ally, suggested some time ago that thiswas probably the case. Sensitization,on the other hand, is reflected by in-creased excitability of both motor andother neurons. Hypotheses and fur-ther evidence concerning the processesresponsible for habituation and sensiti-zation are developed below.

THEORETICAL ISSUES—BEHAVIORAL

Extinction

The similarity of habituation andexperimental extinction was perhapsfirst emphasized by Humphrey (1930),although Pavlov (1927) had earlier in-sisted on the basic comparability of un-conditioned and conditioned responses.Humphrey (1930, 1933) noted thatclose parallels exist between the twoprocesses in terms of spontaneous re-covery, below-zero effect, and dishabit-uation. Konorski (1948), in discus-sing the general properties of plasticity,notes several other parallels includingthe effect of stimulus frequency andpotentiation of habituation upon re-peated testing.

The experimental literature on ex-tinction of conditioned responses ac-tually contains many illustrations ofthe nine parametric relations listedabove for habituation (cf. Brogden,1951; Hull, 1943; Kimble, 1961; Pav-lov, 1927). In fact many of these areamong the commonly cited characteris-tics of experimental extinction. To as-sert that habituation is really extinctiondoes not of course constitute any kindof explanation for either process. Bydefinition extinction presupposes priorconditioning, and the variables of train-ing have a potent influence on thecharacteristics of extinction. Althoughunspecified prior conditions of learning


may modify or shape many "reflex" ac-tivities of the intact organism, habitua-tion does imply an unlearned response.However the fact that habituation andextinction appear to show similar para-metric features does raise the possibil-ity that they may share common under-lying mechanisms. Although manydifferent pieces of the nervous systemare clearly involved in extinction andhabituation of various responses, wewould like to suggest (in agreementwith Sharpless, 1964) that commonneuronal mechanisms exist. Responsedecrements would then be due to a gen-eral property of neuronal activity com-mon to all such systems. At leastpartial analysis of such neuronalmechanisms seems possible in thespinal animal. Hypothetical neuralmechanisms consistent with currentknowledge of neurophysiology are de-veloped below.

Stimulus Satiation and RelatedConstructs

A variety of complex behavioral re-sponses such as spontaneous alterna-tion, exploration, and curiosity havebeen subsumed under the construct of.stimulus satiation (Glanzer, 1953).Glanzer's ingenious theory begins witha five-part postulate describing theproperties of stimulus satiation and de-duces a number of testable predictionsabout maze running and exploratorybehavior. The majority of such pre-dictions have been verified, particularlywhere deductions from Hull's postu-lated reactive inhibition lead to oppo-site predictions (Dember, 1961; Glan-zer, 19S3). Glanzer's postulate can beviewed as a formalised restatement ofsome of the parametric characteristicsof response habituation listed above.All of Glanzer's deductions may alsobe made from the earlier establishedproperties of response habituation.His general postulate reads: "Each

moment an organism perceives a stim-ulus object or stimulus objects, A,there develops a quantity of stimulussatiation to A [p. 259]." We mayrephrase it as: Each moment an or-ganism responds to a stimulus objector stimulus objects, A, there develops aquantity of response habituation to A.Glanzer's deductions follow equallywell from this restatement of stimulussatiation in terms of response habitua-tion.

Glanzer's approach has been cate-gorized as a peripheral rather than acentral theory (Dember, 1961; Walker,1958). The above review of neuro-physiological studies indicated thatthere is as yet no entirely satisfactoryor consistent evidence for habituationor decrement of evoked responses inprimary sensory systems at any levelfrom receptor to cortex. In conse-quence it would seem misleading tolabel as stimulus satiation a decre-mental process that does not appear toexist at any level of neural activitywhich can be identified with the im-mediate responses to sensory activa-tion. Insofar as behavioral responsesare concerned, the formal properties ofstimulus satiation and response habit-uation are the same.

Walker (1958) suggested that theterm "action decrement," having cen-tral locus, be substituted for conceptslike stimulus satiation or reactive inhi-bition. Although he does not relate itto the literature on behavioral habitua-tion, its properties appear to be thesame. Dember (1956, 1961) andDember and Earle (1957) have devel-oped still another theoretical frame-work for this category of behaviorwhich they term "stimulus change."The organism seeks to obtain the"ideal" amount of "stimulus complex-ity." Thus in an alternation experi-ment, the animal alternates choices notbecause he is satiated or habituated to


one stimulus, but because the other at-tracts him more (Dember, 1961).Dember (1956) reported an experi-ment designed as a critical test of stim-ulus-satiation versus stimulus-changehypotheses. Rats were given experi-ence in a T maze (one arm white, oneblack) and then tested with both armsthe same. He argued that satiationtheory would predict no difference intest-choice behavior, but stimuluschange would predict more choices ofarm different in experience and testsessions. Dember's results verified hisprediction. This argument would seemjustified if stimulus satiation is limitedonly to alley brightness. However, ifthe total response pattern of running,orienting, turning, and response tobrightness is considered, then generali-zation of response habituation will begreater to arm same than arm different.Thus analysis in terms of responsehabituation would also predict morechoices of arm different than arm same.As yet there seems to be no clear wayof differentiating experimentally be-tween the alternative hypotheses of"avoiding" or habituating to one stim-ulus complex versus "approaching"another stimulus complex. It has beensuggested (Brown, 1961) that this dis-tinction may be fundamentally untest-able. Again, the formal behavioralproperties of Dember's and Earle's con-structs do not appear to differ greatlyfrom those listed above.

To summarize, the formal behavioralproperties of stimulus satiation, actiondecrement, and search for ideal stim-ulus complexity appear similar to theproperties of response habituation.They may well be indistinguishable atthe behavioral level. The earlier andmore neutral term, habituation, viewedas response decrement, might betterserve the search for central neural cor-relates.

Reactive Inhibition

The construct of reactive inhibition(Ir), developed by Hull (1943) fromPavlov's (1927) more general conceptof "inhibitory state," has been shown tohave considerable predictive utility,particularly in experiments relating towork performance (Hull, 1943; Solo-mon, 1948). It is unquestionably themost carefully and extensively analyzedconstruct dealing with response decre-ment.

The parametric characteristics ofhabituation listed above are consistentwith deductions from reactive inhibi-tion, except for Number 5, indicatingthat amount of habituation is inverselyrelated to stimulus strength. At firstglance this may seem contradictory, inthe sense that response strength is oftenrelated to stimulus strength. How-ever, Ir has been developed within theframework of performance variables,where the stimulus events cannot easilybe specified, and no close relation be-tween stimulus strength and responseamplitude holds. In habituation, par-ticularly of spinal reflexes, initial re-sponse strength is directly related tostimulus strength. The increased ac-tivation of the response by a strongerstimulus could overcome the I, buildup.Sherrington's analysis of reflex fatigue(see below) elaborates this point.

Glanzer (1953) showed that deduc-tions from Ir in T-maze alternation be-havior appeared to lead to the em-pirically contracted prediction that sub-jects will alternate turns even if thediscriminable stimuli on the two sidesare switched each trial. This has beenconsidered a telling argument againstIr (Dember, 1961; Glanzer, 1953). Itseems to us that it may be a ratherartificial contradiction resulting fromthe definition of response which waschosen. From the actual responsesequence of running down the alley,


orienting toward and approaching oneo\ the stimulus complexes, and henceturning, Glanzer abstracts turning asthe response. If we instead emphasizestimulus orientation and approach asthe more relevant aspect of the re-sponse sequence, then application ofHull's 7r analysis leads to exactly thesame predictions as does Glanzer'ssatiation hypothesis. This characteri-zation of the response, incidentally, iscomparable to that used by Spence(1960, p. 374) in his analysis of spatialdiscrimination learning.

To summarize, 7r is not inconsistentwith the phenomena of habituation, andmay even be viewed as a molar analysisof the role of response habituation inbehavior. Experimentally contradictedpredictions made from Ir about alterna-tion behavior may be ascribed to theparticular response definition used.The hypothetical neuronal model ofhabituation developed below might wellbe considered as a possible neurophysi-ological basis for certain aspects of Ir.

Fatigue

There is no agreed upon definition offatigue, either in psychology or physi-ology. We have defined effector fa-tigue as response decrement localizedto neuroeffector and effector processes(cf. Merton, 1954) and shown that itcannot account for habituation, at leastof the flexion reflex. It is extremelyunlikely that effector fatigue has anygreat importance for habituation gen-erally since all responses that showhabituation can be elicited morestrongly, more rapidly, and withoutdecrement under other conditions. Asfar as central fatigue is concerned, ar-bitrary definitions specifying certainsynaptic states could also be used, buthave not come into general parlance.Sherrington's (1906, p. 220) molardefinition of reflex fatigue is perhaps themost commonly used: " . . . a spinal

reflex under continuous excitation orfrequent repetition becomes weaker,and may cease altogether." He wasvery careful to hold firmly to thispurely operational definition, which isof course identical to the definition ofhabituation given above. In like man-ner, Hull (1943, p. 278) identifies re-active inhibition with fatigue, the latterbeing defined as ". . . decrement inaction evocation potentiality ratherthan an exhaustion oi the energy . . .[italics ours]."

The inverse relationship betweenstimulus strength and rate and degreeof fatigue or habituation was analyzedvery carefully by Sherrmgton (1906,p. 220) :

In my experience, these spinal reflexes fadeout sooner under a weak stimulus than un-der a strong one. This seeming paradox in-dicated that even under feeble intensities ofstimulation the threshold of the reactiongradually rises, and that it rises above thethreshold value of the weaker stimulus be-fore it reaches that of the stronger stimulus.

Many students of habituation haveargued against fatigue as the mechan-ism. Most of these have implied effec-tor fatigue (e.g., Harris, 1943 ; Prosser& Hunter, 1936). However the moregeneral concept of fatigue has also beenquestioned on the grounds that dis-habituation could not occur if a simplefatigue process were responsible fordecrement- Such arguments of coursemake an implicit assumption about thenature of central fatigue; namely, thatit is a process which cannot be brokenup rapidly by dishabituation. Howeverwhen dishabituation is viewed as anindependent superimposed process(i.e., sensitization), the argument losesits force. Thus in the flexion reflex,response decrement occurs in inter-neuronal pathways without any back-ground change in motor-neuron excit-ability, but sensitization is commonly


reflected in transient depolarization ofmotor neurons as well.

In summary, operational definitionsof central fatigue do not differ fromdefinitions of habituation and reactiveinhibition. Even if this central processis characterized as having a slow re-covery time, dishabituation is not acontradiction if viewed as an independ-ent superimposed sensitization process.

THEORETICAL ISSUES—NEURAL

Stimulus Model Comparison

On the basis of a wide variety ofingenious experiments concerning ha-bituation of the human orienting reflex,Sokolov (1960) has developed a theo-retical neural schema. While thistheory is somewhat specific to theorienting response, it has occasionedwidespread interest (cf. Magoun, 1963;Pribam, 1963) and can be generalizedto other types of response habituation.In essence it is a feedback comparatoranalogue. One system (cerebral cor-tex) is concerned with the formationof a model of the stimulus-input pat-tern. This interacts with an amplify-ing system (reticular formation) whichserves to alter response tendency. Asthe model develops in the cortex, retic-ular amplification is inhibited. If adishabituating stimulus is given, theamplification is no longer inhibited, andan increased response occurs. Theessential point in the theory is that amodel of the habituating stimulus ac-tivity develops, against which any otherstimulus pattern is compared. Re-sponse strength is directly proportionalto the degree of discordance betweenany stimulus input and the model.

The crucial experimental observa-tions forming the basis of this theoryconcern dishabituation by altered stim-ulus conditions where intensity is notincreased. Thus after habituation toa given tone, a weaker tone or a shorter

tone produces dishabituation5 of theorienting reflex (Sokolov, 1960).Hence a comparison model is neces-sary. There are few direct tests ofSokolov's theory. Habituation of theorienting reflex does occur, albeit moreslowly, in the totally decorticated ani-mal (Lebidinskaia & Rosenthal, 1935),and auditory orientation habituationoccurs in animals without auditory cor-tex (Thompson & Welker, 1963).These observations tend to raise diffi-culties for the role of the cortex as themodel formation system, but the gen-eral construct could still be used byinvoking primary sensory relays belowthe thalamocortical level for the modelcomparater. In like manner it couldbe generalized to spinal reflexes by as-suming a dual mechanism (as yet un-identified) in cord internuncials.

Perhaps the most important pointabout the theory is that it may not benecessary to assume a model-compari-son process to account for dishabitua-tion by reduced stimulus intensity oraltered stimulus characteristic. Cod-ing of intensity in neural networks iscomplex, and there is not always asimple relation between stimulus in-tensity and neural activity. Thus inthe inferior colliculus and auditory cor-tex, for any given frequency, someneurons fire to weak tones but not tostrong tones (Hind, Rose, Davies,Woolsey, Benjamin, Welker, & Thomp-son, 1960; Rose, Greenwood, Goldberg,& Hind, 1963). Hence a weak tonewill activate cells and synapses notactivated by the stronger tone. Ifhabituation is simply some sort of de-crease in synaptic efficacy, then a largerresponse would be expected to a weaktone after habituation to a strong tonebecause different neurons are activated.

B It should be noted that this is a some-what special use of the term dishabituationin that the extra stimulus is subsequentlyused as the test stimulus.


Consequently no sort of comparisonagainst a model is necessary; the pro-cess can equally be construed asstraightforward decreases in the effi-cacy of certain pathways, together withdiffering degrees of pathway overlapresulting from differing stimulus con-ditions. Dishabituation by change toa weaker or otherwise altered stimulusthen becomes an instance of incompletestimulus generalisation of habituation.A stimulus having somewhat differentcentral connections than the habituat-ing stimulus yields a larger net re-sponse because generalization of habit-uation to the new stimulus is onlypartial.

Afferent Neuronal Inhibition

This theory has been elaborated byHernandez-Peon (1955) and has re-ceived considerable attention (e.g., Liv-ingston, 1959; Worden & Livingston,1961). In essence it is assumed thatcentrifugal fibers from the reticularformation and related structures cours-ing to various peripheral sensory re-ceptors or nuclei act to inhibit respon-siveness at the periphery, thus reduc-ing activity in sensory systems. Thisinhibition is said to develop duringrepetitive stimulation and results inevoked response habituation.

There are sound neurophysiologicalreasons for believing that some types ofefferent control systems do exist, al-though they do not always involve thereticular formation. Two such sys-tems have been described for the audi-tory system, one acting on the cochlea(Galambos, 1965; Rasmussen, 1946)and one on the cochlear nucleus (Des-medt, 1960; Desmedt & Michelse,1958). However neither is coexten-sive with the reticular formation. Forthe optic system, Granit (1955) hasdemonstrated both excitatory and in-hibitory effects on retinal ganglion cellsby stimulation of the midbrain tegmen-tum. In the somatic system, Hag-barth and Kerr (1954) showed that

reticular stimulation could yield bothfacilitatory and inhibitory effects on thesecond-order ascending sensory neu-rons in the spinal cord. Kerr andHagbarth (1955) found diminution ofolfactory-bulb activity when stimulat-ing a variety of nonreticular structuressuch as prepyriform cortex, amygda-loid nucleus, and olfactory tubercle.

In sum, efferent control systems ap-pear to exist for all sensory modalitiesstudied. Control is sometimes exertedon early central relays rather than onperipheral receptors and does not al-ways involve the reticular formation.These systems could well provide thebasis for habituation of sensory evokedprimary responses. Unfortunately, asyet there seems to be no entirely satis-factory or consistent evidence for theoccurrence of such habituation (exceptfor rhythmic activity of the olfactorybulb, Moulton, 1963). The furtherpossibility that EEC arousal habitua-tion is so mediated is negated bystudies showing comparable habitua-tion phenomena for EEC arousal in-duced by central electrical reticularstimulation (Glickman & Feldman,1961) where peripheral gating couldnot have any role, and by Sharplessand Jasper's (1956) finding that pri-mary evoked responses did not habitu-ate during habituation of EEC arousalto auditory stimuli.

Alterations in Sy nap tic Transmission

In the course of this article it hasbeen suggested at several points thatcommon neuronal mechanisms may ac-count for many varieties of responsehabituation. Sherrington (1898,1906), Prosser and Hunter (1936),Sharpless and Jasper (1956), andKennedy and Mellon (1964), amongothers, have all raised the possibilitythat some type of decrease in synaptictransmission may form the basis ofhabituation. An hypothetical mechan-ism will be developed here for thespinal flexion reflex, and its possible


applicability to other systems consid-ered.

It was proved above that habituationof the flexion reflex in the neurally iso-lated spinal cord must occur in thetransmission through interneurons be-tween incoming afferent fibers and out-put motor neurons. At least threegeneral classes of neural events maybe postulated to yield decreased neu-ron response tendency: synaptic inhibi-tion, decreased synaptic efficacy, ormembrane desensitization.

Synaptic Inhibition. Within thecategory of synaptic inhibition, twogeneral types have been described,postsynaptic inhibition (PostSI), re-flected in hyperpolarizing inhibitorypostsynaptic membrane potentials, andpresynaptic inhibition (PreSI), whichacts by depolarizing presynaptic ter-minals to yield a smaller excitatorypostsynaptic potential (Eccles, 1964).These inhibitory mechanisms have beendeveloped for several types of neuronsin the spinal cord and are now beingestablished for certain neurons in otherportions of the CNS (Eccles, 1964).

Both types of inhibition could bepostulated to act on interneurons medi-ating the flexion reflex. However, themeasured time courses of both thesetypes of inhibition (Eccles, 1964) aremuch too short to summate with I/second stimulation and to account foreffects lasting several minutes. Itcould be postulated that complex net-works involving tonic or cascading andcumulative PreSI or PostSI are acti-vated by the repeated-stimulus pattern.Time relations are still a problem sinceGerard and Forbes (1928) were ableto detect some homonymous depressionup to .8 second after only a singleshock (to mixed nerves in their experi-ment), and we have found a similardepression with a 1-second interval.The time course of known types ofPre- and PostSI is less than about 500milliseconds. Evidence from pharma-cological studies is also against an in-

hibitory hypothesis. Strychnine hasbeen shown to attenuate greatly alltypes of spinal-cord PostSI on whichit has been tested (Eccles, 1964). Pic-rotoxin has been shown to partiallyblock PreSI (Eccles, 1964). Wehave found that habituation of theflexion reflex is not significantly al-tered following administration of thesetwo agents in doses sufficient to obtainthe reductions in inhibition just noted.Finally, it may be mentioned that elec-trical indexes of phasic PreSI (suchas the cord-dorsum P wave) andphasic PostSI (hyperpolarizing phasesof intracellularly and electrotonicallyrecorded motor-neuron responses)themselves show decrement upon re-peated activation of cutaneous afferents(Spencer et al., 1964). The A^ com-ponent of the cord-dorsom response,which itself is highly susceptible topresynaptic inhibition, often shows nodecrement during the development ofhabituation, thus providing further evi-dence against PreSI. To summarize,all existing evidence points away frominhibition as the agent underlying thistype of habituation.6

Decreased Synaptic Efficacy. In con-sidering alternative explanations, it ishelpful to recall that there are at leasttwo other well-known consequences ofreflex action which are quite differentfrom inhibition. The first is a speciesof synaptic depression variously termed"low frequency depression," "defacili-tation," "homosynaptic depression," or"subsynaptic depression" which hasbeen studied in the monosynaptic reflexpathway (Curtis & Eccles, 1960; Jef-ferson & Schlapp, 1953; Lloyd & Wil-son, 1957) and also in some of thesimplest disynaptic systems (Wilson,

6 However it is to be noted that our ex-periments were carried out with single-shockstimuli at 1-3-second intervals and that in-hibition, particularly the presynaptic variety,may play an important role when higher fre-quencies are used, as they were in Sherring-ton's (1899, 1906) studies on fatigue or whenintermittent stimulus trains are employed.


1958). Several features of this depres-sion are noteworthy within the contextof our discussion: (a) It is manifestedas a decrease of PSP amplitude due tointermittent low-frequency synaptic ac-tivation and is either pre- or sub-synaptic rather than postsynaptic inorigin; (b) its time course is too longto be accounted for by presynaptic in-hibition (Eccles, Kostyuk, & Schmidt,1962); and (c) it is detectable at thosestimulus intervals required to producethe polysynaptic decrement we havedescribed using single-shock stimuli.The degree of low-frequency depres-sion is greater with increased stimulusfrequency, as is the case with habitu-ation. Although Jefferson and Schlapp(1953), recording ventral-root volleys,did report diminished depression withstrong stimuli, which would furtherparallel habituation, Curtis and Eccles(1960), recording motor-neuron exci-tatory postsynaptic potentials, did notfind this effect even with stimuli supra-maximal for Group la fibers. The de-pression produced in monosynapticpathways by low-frequency depressionis briefer in duration (less than 20seconds) than that which we have de-scribed. However, it must be remem-bered that the time course of the de-pression might be modified in ourexperiments due to the fact that inter-neurons characteristically fire in brieftrains of discharges, even to single pre-synaptic volleys, and that this wouldconstitute a distinctly different testpattern which might have a moreprolonged recovery period.

The second relevant consequence ofreflex action, "after discharge"(Forbes, 1922; Ranson & Hinsey,1930), is opposite in effect and char-acteristic of polysynaptic systems.This is manifested as a postsynapticdepolarization and prolonged dischargeof inter and motor neurons (Fuortes,1954; Hunt & Kuno, 1959), is gen-erally more prolonged with strongerstimuli, and may build up with repeated

stimuli (Frank & Fuortes, 1956).Preliminary evidence we have obtainedfrom extracellular interneuron record-ings indicates that such afterdischargesare present and build up with stimulusrepetition during the processes de-scribed.

We would therefore suggest, as aworking hypothesis, that the ampli-tude of repeated polysynaptic test re-sponses depends upon the balance be-tween low-frequency depression andafterdischarge existing at the time eachstimulus is presented. Viewed in thisway, the frequency sensitivity of thedecrement we have described would berelated to the well-known frequencysensitivity of low-frequency depression.Stimulus generalization would dependupon depression of interneuron synap-ses shared by the two input channels.The greater response stability seen atthe strong stimulus strengths might re-flect a greater preponderance of after-discharge onto interneurons relayingthe reflex. This would increase in du-ration with strong stimuli and wouldexist during the presentation of succes-sive stimuli, thereby facilitating thepostsynaptic interneuron responses tothem, in a sense compensating for thelow-frequency depression. Moreover,if the phenomenon of sensitization (i.e.,response restoration following a strongextrastimulus) be considered a specialcase of very prolonged afterdischargedue to a strong stimulus, then many ofthe puzzling features of this phenome-non could be more readily understood.Wilson's (1955) postulated mechanismof polysynaptic posttetanic potentiation(PTP) is an alternative possibility forsensitization, as is the recently de-scribed phenomenon of "presynapticfacilitation" (Kandel & Tauc, 1964;Mendell & Wall, 1965), since there arecertain obvious similarities to the proc-ess of response restoration we havedescribed.

Membrane Desensitisation. Wenoted earlier that some forms of re-


sponse habituation, even of spinal re-flexes, may last for days or weeks. Inthe experiment by Kozak et al. (1962),for example, a crossed scratch reflexof the chronic spinal kitten, elicited byrubbing an area of skin for 10 minutesdaily, exhibited marked habituation, re-quiring a period of 8 days to recover.The synaptic processes discussed above,namely inhibition and decreased effi-cacy due to the balance of low-fre-quency depression and afterdischarge,would both seem to have too short atime course, per se, to account forsuch long-lasting habituation. Sharp-less (1964) has suggested that a typeof membrane desensitization process,involving decreased responsiveness ofthe postsynaptic membranes of nervecells due to repeated applications of thetransmitter substances, may be involvedin long-term habituation. In hyper-sensitized denervated smooth musclesand glands, for example, administrationof appropriate excitatory agents suchas epinephrine several times daily pro-duces a desensitization effect that maylast for several days (Emmelin &Muren, 1952; Simeone, 1938). Thistype of hypothesis may be particularlyappropriate to the chronic spinal prep-aration in that spinal reflexes becomemarkedly exaggerated or hypersensi-tized, as do responses of dennervatedsmooth muscles and glands. Two dif-ferent classes or components of re-sponses have been distinguished on thebasis of habituation and spontaneous-recovery time courses. Sokolov (1955)noted that the "orienting reflex" ex-hibited both rapidly habituating andslowly habituating components. Sharp-less and Jasper (1956) differentiatedbetween a "phasic" EEC arousal thatoccurs immediately after a stimulus, isof short duration, and is relatively re-sistant to habituation and a "tonic"component that habituates much morerapidly. Lesions of the brachia of theinferior colliculus appeared to abolishthe phasic component (to auditory

stimuli) without eliminating the toniccomponent. Thompson and Welker(1963) found that ablation of auditorycortex prevented the development oflong-term intersession habituation oforienting reflexes to sounds, but did notalter short-term intrasession habitu-ation.

It may well be necessary to postulateat least two different types of neuronalprocesses to account for short-term andlong-term habituation. Our suggestedmechanisms of low-frequency depres-sion and afterdischarge are perhaps ap-propriate to short-term effects andSharpless' "membrane desensitization"to long-term effects. Alternatively, allthese processes may interact to agreater or lesser extent in all instancesof habituation. The fact that, exceptfor time course, both short-term andlong-term habituation exhibit the sameparametric features may imply someidentity of underlying processes.

CONCLUSION

In summary, we would suggest that,at least for the spinal flexion reflex aswe have studied it, the process of ha-bituation may represent the cumulativeeffect of a polysynaptic low-frequencydepression and that dishabituation is atype of superimposed sensitization, pos-sibly resulting from facilitatory after-discharge.

In amplifying this scheme, one mightpostulate, on the basis of other evidence(Frank & Fuortes, 1956), that ratherlong iterative trains of interneurondischarges are produced by each stimu-lus, and that the long duration of thechanges noted depends on this featureof the test response. If one provision-ally adopts this point of view, severalaspects of the literature on habituationreviewed here may be seen in somewhatdifferent perspective. The absence orpoor development of evoked responsehabituation in primary sensory sys-tems would be consistent with the factthat the interneurons exhibiting very


prolonged discharge patterns to singlestimuli are not typical of the long re-lays from receptors to cortex in thesesystems. However such iterative firingis more characteristic of neurons in thesystem mediating EEG arousal; thereticular formation (Moruzzi, 1954),which does exhibit habituation (Bell,Sierra, Buendia, & Segundo, 1964;Sharpless & Jasper, 1956). In likemanner, such interneurons are un-doubtedly present in all functional sys-tems of the intact organism partici-pating in behavioral response habitu-ation.

Other things equal, complex re-sponses tend to habituate more rapidlythan simpler responses. If degree ofresponse complexity is reflected in in-creased numbers of interneurons inter-posed between stimulus and response,and in an increase in the duration oftheir repetitive discharges, these fac-tors themselves might account for morerapid habituation. The effects of CNSlesions on behavioral response habitu-ation may also be reconsidered in thisway : Effective lesions remove many in-terneurons normally involved in com-plex responses and hence could lead toslower response habituation.

Many of these extrapolations fromdata obtained in the study of responsedecrement in the spinal cord to habitu-ation as it appears in the more intactorganism are obviously speculative.They do have the virtue, however, ofindicating the directions which cellu-lar theories of habituation could take.Quantitative theoretical models wouldclearly be of great help in making suchcellular explanations useful alternativesto the more molar theories commonlyapplied to behavior of the intact or-ganism.

REFERENCESADRIAN, E. D. The electrical activity of the

mammalian olfactory bulb. Electroen-cepkalography and Clinical Neurophysi-ology, 1950, 2, 377-388.

AFELT, Z. Variability of reflexes in chronicspinal frogs. In E. Gutmann & P. Hnik

(Eds.), Proceedings of the conference oncentral and peripheral mechanisms ofmotor functions. Prague, Czechoslovakia:Czechoslovak Academy of Sciences, 1963.Pp. 37-41.

AFFANNI, J., MANCIA, M., & MARCHIAFAVA,P. L. Role of the pupil in changes inevoked responses along the visual path-ways. Archives Italiennes de biologic,1962, 100, 287-296.

ALTMAN, I. A. Electrophysiological exami-nation of different parts of the auditorysystem in the cat during sustained rhythmi-cal stimulation. Sechenov PhysiologicalJournal of the USSR, I960, 46, 617-629.

APELBAUM, J., SILVA, E. E., FRICK, O., &SEGUNDO, J. P. Specificity and biasingof arousal reaction habituation. Electro-encephalography and Clinical Neurophysi-ology, 1960,12, 829-839.

BARD, P., & MACHT, M. B. The behavior ofchronically decerebrate cats. In G. E. W.Wolstenholme & C. M. O'Connor (Eds.),Neurological basis of behavior. London:Churchill, 19S8. Pp. 55-71.

BARTLEY, S. H., & CHUTE, E. Fatigue andimpairment in man. New York : McGraw-Hill, 1947.

BELL, C., SIERRA, G., BUENDIA, N., & SE-GUNDO, J. P. Sensory properties of neu-rons in the mesencephalic reticular forma-tion. Journal of Neurophysiology, 1964,27, 961-987.

BEHLYNE, D. E. Novelty and curiosity asdeterminants of exploratory behavior.British Journal of Psychology, 19SO, 41,68-80.

BERLYNE, D. E. Conflict, arousal, and curi-osity, New York: McGraw-Hill, 1960.

BINDRA, D. Stimulus change, reactions tonovelty, and response decrement. Psycho-logical Review, 1959, 66, 96-103.

BROGDEN, W. J. Animal studies of learning.In S. S. Stevens (Ed.), Handbook of ex-perimental psychology. New York: Wiley,1951. Pp. 568-612.

BROWN, J. S. The motivation of behavior.New York: McGraw-Hill, 1961.

BUCHWALD, J. S., HALAS, E. S., & SCHRAMM,S. Progressive changes in unit responsesto a repeated stimulus. Federation Pro-ceedings, 1964, 23, 919.

BUCHWALD, J. S., HALAS, E. S., & SCHRAMM,S. Progressive changes in efferent unitresponses to repeated cutaneous stimula-tion in spinal cats. Journal of Neuro-physiology, 1965, 28, 200-215.

BUTLER, R. A., & HARLOW, H. F. Per-sistence of visual exploration in monkeys.Journal of Comparative and Physiologi-cal Psychology, 1954, 47, 258-263.


CASPERS, H., LERCHE, E., & GREUTER, H.Adaption-serscheinungen der akustich aus-gelosten Wechreaktion bei Reizung mit de-finierten Tonimpulsen. Pfiiiger's Archives,1958, 267,128-141.

CAVAGGIONI, A., GIANNELLI, G., & SANTI-BANEZ, H. G. Effects of repetitive photicstimulation on responses evoked in thelateral geniculate body and the visual cor-tex. Archives Italiennes de biologie, 1959,97, 226-275.

COOMBS, C. H. Adaptation of the galvanicresponse to auditory stimuli. Journal ofExperimental Psychology, 1938, 22, 244-268.

CRAMPTON, G. H., & SCHWAM, W. J. Ef-fects of arousal reaction on nystagmushabituation in the cat. American Journalof Physiology, 1961, 200, 29-33.

CURTIS, D. R., & ECCLES, J. C. Synaptic ac-tion during and after repetitive stimula-tion. Journal of Physiology, 1960, ISO,374-398.

DAVIS, R. C. Modification of the galvanicreflex by daily repetition of a stimulus.Journal of Experimental Psychology, 1934,17, 504-535.

DAVIS, R. C., BUCHWALD, A. M., & FRANK-MANN, R. W. Autonomic and muscularresponses, and their relation to simplestimuli. Psychological Monographs, 1955,69(20, Whole No. 405).

DEMBER, W. N. Response by the rat to en-vironmental change. Journal of Compara-tive and Physiological Psychology, 1956,49, 93-95.

DEMBER, W. N. Alternation behavior. InD. W. Fiske & S. R. Maddi (Eds.),Functions of varied experience. Home-wood, 111.: Dorsey, 1961. Pp. 227-252.

DEMBER, W. N., & EARL, R. W. Analysisof exploratory, manipulatory, and curiositybehavior. Psychological Review, 1957, 64,91-96.

DESMEDT, J. E. Neurophysiological mech-anisms controlling acoustic input. In G. L.Rasmussen & W. Windle (Eds.), Neuralmechanisms of the auditory and vestibularsystems. Springfield, 111.: Charles CThomas, 1960. Pp. 152-164.

DESMEDT, J. E,, & MICHELSE, K. Suppres-sion of acoustic input by thalamic stimula-tion. Proceedings of the Society for Ex-perimental Biology and Medicine, 1958, 99,772-775.

DEUTSCH, J. A. Higher nervous function:The physiological bases of memory. An-nal Review of Physiology, 1962, 24, 259-286.

DYKMAN, R. A., REESE, W. G., GALBRECHT,C. R., & THOMASSON, P. J. Psycho-physiological reactions to novel stimuli:

Measurement, adaptation, and relationshipof psychological variables in the normalhuman. Annals of the New York Acad-emy of Sciences, 1959, 79, 43-107.

ECCLES, J. C. The physiology of synapses.New York: Academic Press, 1964.

ECCLES, J. C., KOSTYUK, P. G., & SCHMIDT,R. F. Presynaptic inhibition of the centralactions of flexor reflex afferents. Journalof Physiology 1962, 161, 258-281.

EMMELIN, N., & MUREN, A. The sensitiv-ity of submaxillary glands to chemicalagents studied in cats under various condi-tions over long periods. Acta PhysiologicaScandinavica, 1952, 26, 221-231.

ENGEN, T., & LIPSITT, L. P. Decrement andrecovery of responses to olfactory stimuliin the human neonate. Journal of Com-parative and Physiological Psychology,1965, 59, 312-316.

FERNANDEZ-GUARDIOLA, A., RoLDAN, E.,FANJUL, L., & CASTELLS, C. Role of thepupillary mechanism in the process ofhabituation of the visual pathways. Elec-troencephalography and Clinical Neuro-physiology, 1961, 13, 564-576.

FORBES, A. The interpretation of spinal re-flexes in terms of present knowledge ofnerve conduction. Physiological Review,1922, 2, 361-414.

FRANK, K., & FUORTES, M. G. F. Unitaryactivity of spinal interneurons of cats.Journal of Physiology, 1956, 131, 424-436.

FRENCH, G. M. Locomotor effects of re-gional ablations of frontal cortex in rhesusmonkeys. Journal of Comparative andPhysiological Psychology, 1959, 52, 18-24.

FRENCH, G. M., & HASLOW, H. F. Loco-motor reaction decrement in normal andbrain-damaged monkeys. Journal of Com-parative and Physiological Psychology,1955, 48, 496-501.

FUORTES, M. G. F. Activity of single motor-neurons during some reflex reactions ofmammals. Comptes-Rendus 5th Interna-tional Neurological Congress (Lisbon),1954, 4, 36-51.

GALAMBOS, R. Suppression of auditorynerve activity by stimulation of efferentfibers to cochlea. Journal of Neurophysi-ology, 1956, 19, 424-437.

GALAMBOS, R., SHEATZ, G., & VERNIER, V. G.Electrophysiological correlates of a condi-tioned response in cats. Science, 1956, 123,331-332.

GARCIA-AUSTT, E. Influence of the states ofawareness upon sensory evoked potentials.Electroencephalography and Clinical Neu-rophysiology Suppl. 24, 1963, 76-89.

GERARD, R. W., & FORBES, A. "Fatigue"of the flexion reflex. American Journalof Physiology, 1928, 86, 186-205.


GERSHUKI, G. V., KOZHEVNIKOV, V. A.,MARUSEVA, A. M., AVAKYAN, R. V.,RADIONOVA, E. Z., ALTMAN, J. A., &SOROKO, V. I. Modifications in electricalresponses of the auditory system in differ-ent states of the higher nervous system.Electroencephalography and Clinical Neu-rophysiology Suppl. 13, I960, 115-124.

GLANZER, M. Stimulus satiation: An ex-planation of spontaneous alternation andrelated phenomena. Psychological Review,1953, 60, 257-268.

GLASER, E. M., & GRIFFIN, J. P. Influenceof the cerebral cortex on habituation.Journal of Physiology, 1962, 160, 429-445.

GLICKMAN, S. E. Deficits in avoidancelearning produced by stimulation of theascending reticular formation. CanadianJournal of Psychology, 1958, 12, 97-102,

GLICKMAN, S. E., & FELDMAN, S. M. Ha-bituation of the arousal response to directstimulation of the brainstem. Electro-encephalography and Clinical Neurophysi-ology, 1961, 13, 703-709.

GRANIT, R. Centrifugal and antidromic ef-fects on ganglion cells of retina. Journalof Neurophysiology, 1955, 18, 388-411.

GREEN, J. D., & ARDUINI, A. A. Hippo-campal electrical activity in arousal. Jour-nal of Neurophysiology, 1954, 17, 533-557.

GRIFFIN, J. P. The role of the frontal areasof the cortex upon habituation in man.Clinical Science, 1963, 24, 127-134.

GRIFFITH, C. R. The effect upon the whiterat of continued bodily rotation. AmericanNaturalist, 1920, 54, 524-534.

GUZMAN-FLORES, C., ALCARAZ, M., & HAR-MONY, T. Role of the intrinsic ear musclesin the process of acoustic habituation.Boletin del Institute de estudios medicos ybioldgicos, Universidad nacional de Mexico,1960, 18, 135-149.

HAGAMEN, W. G., LANCE, E. M., & UNGE-WITTER, L. H. Increased responsiveness tostimuli following lesions of the forebrain.Anatomical Record, 1959, 133, 387-388.

HAGBARTH, K. E., & FINER, B. L. Theplasticity of human withdrawal reflexes tonoxious skin stimuli in lower limbs. InG. Moruzzi, A. Fessard, & H. H. Jasper(Eds.), Progress in brain research. Vol.1. Brain mechanisms. New York: Else-vier,1963. Pp. 65-81.

HAGBARTH, K. E., & KERR, D. I. B. Centralinfluences on spinal afferent conduction.Journal of Neurophysiology, 1954, 17, 295-307.

HAGBARTH, K. E., .& KUGELBERG, E. Plas-ticity of the human abdominal skin reflex.Brain, 1958, 81, 305-318.

HARRIS, J. D. Habituatory response decre-

ment in the intact organism. Psychologi-cal Bulletin, 1943, 40, 385-422.

HERNANDEZ-PE6N, R. Central mechanismscontrolling conduction along central sen-sory pathways. Ada neurologica latino-americana, 1955, 1, 256-264.

HERNANDEZ-PEON, R. Neurophysiologicalcorrelates of habituation and other mani-festations of plastic inhibition. Electro-encephalography and Clinical Neurophysi-ology Suppl. 13, 1960, 101-114.

HERNANDEZ-PE6N, R., GuZMAjST-FLORES, C.,ALCARAZ, M., & FERNANDEZ-GUARDIOLA,A. Habituation in the visual pathway.Acta neurologica latinoamericana, 1958,4, 121-129.

HERNANDEZ-PE6N, R., JoUVET, M., &SCHERRER, H. Auditory potentials atcochlear nucleus during acoustic habitua-tion. Acta neurologica latinoamericana,1957, 3, 144-156.

HERNANDEZ-PE6N, R., LAVIN, A., ALCOCER-CUAR6N, C., & MARCELIN, J. P. Electricalactivity of the olfactory bulb during wake-fulness and sleep. Electroencephalographyand Clinical Neurophysiology, 1960, 12,41-58.

HERNANDEZ-PEON, R., & SCHERRER, H.Habituation to acoustic stimuli in cochlearnucleus. Federation Proceedings, 1955, 14,71.

HERNANDEZ-PEON, R., SCHERRER, H., &VELAsco, M. Central influences on affer-ent conduction in the somatic and visualpathways. Acta neurologica latinoameri-cana, 1956, 2, 8-22.

HILGARD, E. R., & MARQUIS, D. G. Condi-tioning and learning. New York: Apple-ton-Century-Crofts, 1940.

HIND, J. E., ROSE, J. E., DAVIES, P. W.,WOOLSEY, C. N., BENJAMIN, R. M.,WELKER, W. L, & THOMPSON, R. F. Unitactivity in the auditory cortex. In G. L.Rasmussen & W. Windle (Eds.), Neuralmechanisms of the auditory and vestibularsystems. Springfield, 111.: Charles CThomas, 1960. Pp. 201-210.

HINDE, R. A. Changes in responsiveness toa constant stimulus. British Journal ofAnimal Behavior, 1954, 2, 41-55.

HUGELIN, A., DUMONT, S., & PAILLAS, N.Tympanic muscles and control of auditoryinput during arousal. Science, 1960, 131,1371-1372.

HULL, C. L. Principles of behavior. NewYork: Appleton-Century-Crofts, 1943.

HUMPHREY, G. Extinction and negativeadaptation. Psychological Review, 1930,37, 361-363.

HUMPHREY, G. The nature of learning.New York: Harcourt, Brace, 1933.


HUNT, C. C, & KUNO, M. Properties ofspinal interneurons. Journal of Physi-ology, 1959, 147, 346-363.

HUTTENLOCHER, P. R. Effects of state ofarousal on click responses in the mesen-cephalic reticular formation. Electroen-cephalography and Clinical Neurophysi-ology, 1960, 12, 819-827.

JAMES, W. Principles of psychology. NewYork: Holt, 1890.

JEFFERSON, A., & SCHLAPP, W. Some effectsof repetitive stimulation of afferents onreflex conduction. In J. L. Malcolm &J. A. B. Gray (Eds.), The spinal cord.London: Churchill, 1953. Pp. 99-119.

JOUVET, M., & HERNANDEZ-PEIJN, R. Mec-anismes neurophysiologiques concernant1'habituation, 1'attention, et le conditionne-ment. Electroencephalography and Clini-cal Neurophysiology Suppl. 6, 1957, 39-49.

KANDEL, E. R., & TAUC, L. Mechanisms ofprolonged heterosynaptic facilitation. Na-ture, 1964, 202, 145.

KENNEDY, D., & MELLON, D., JR. Synapticactivation and receptive fields in crayfishinterneurons. Comparative Biochemistryand Physiology, 1964, 13, 275-300.

KERR, D. I. B., & HAGBARTH, K. E. An in-vestigation of olfactory centrifugal fibersystem. Journal of Neurophysiology, 1955,18, 362-374.

KIMBLE, G. A. Hilgard and Marquis' condi-tioning and learning. New York: Apple-ton-Century-Crofts, 1961.

KNOTT, J. R., & HENRY, C. E. The condi-tioning of the blocking- of the alpha rhythmof the human electroencephalogram. Jour-nal of Experimental Psychology, 1941, 28,134-144.

KONORSKI, J. Conditioned reflexes and neu-ron organisation. New York: CambridgeUniver. Press., 1948.

KOZAK, W., MACFARLANE, W. W., & WEST-

ERMAN, R. Longlasting reversible changesin the reflex responses of chronic spinalcats to touch, heat and cold. Nature, 1962,193, 171-173.

KRECHEVSKY, I. Brain mechanisms and vari-ability: I. Variability within a means-end-readiness. Journal of ComparativePsychology, 1937, 23, 121-138. (a)

KRECHEVSKY, I. Brain mechanisms and vari-ability : II. Variability where no learningis involved. Journal of Comparative Psy-chology, 1937, 23, 139-163. (b)

KRECHEVSKY, I. Brain mechanisms and vari-ability: III. Limitations of the effect ofcortical injury upon variability. Journalof Comparative Psychology, 1937, 23, 351-364. (c)

LEBEDINSKAIA, S. I., & ROSENTHAL, J. S.Reactions of a log after removal of the

cerebral hemispheres. Brain, 1935, 58,412-419.

LEHNER, G. F. J. A study of the extinctionof unconditioned reflexes. Journal of Ex-perimental Psychology, 1941, 29, 435-456.

LETHLEAN, A. K. Habituation in the ratspinal cord—a triggered process. Federa-tion Proceedings, 1965, 24, 517.

LIVINGSTON, R. B. Central control of re-ceptors and sensory transmission. In J.Field (Ed.), Handbook of physiology:Section 1. Neurophysiology. Vol. 1.Washington, D. C.: American Physiologi-cal Society, 1959. Pp. 741-760.

LLOYD, D. P. C, & WILSON, V. J. Reflexdepression in rhythmically active mono-synaptic reflex pathways. Journal of Gen-eral Physiology, 1957, 40, 409-426.

LOMBARD, W. P. The variations of the nor-mal knee jerk and their relation to theactivity of the central nervous system.American Journal of Physiology, 1887, 1,5-71.

MAGOUN, H, W. Central neural inhibition.In M. R. Jones (Ed.), Nebraska sympo-sium on motivation: 1963. Lincoln: Uni-ver. Nebraska Press, 1963. Pp. 161-196.

MANCIA, M., MEULDERS, M., & SANTI-BANEZ-H., G. Changes of photically evokedpotentials in the visual pathway of thecerveau isole cat. Archives Italiennes debiologie, 1959, 97, 378-398.

MARSH, J. T., MCCARTHY, D. A., SIIEATZ,G., & GALAMBOS, R. Amplitude changesin evoked auditory potentials during habit-uation and conditioning. Electroenceph-alography and Clinical Neurophysiology,1961, 13, 224-234.

MARSH, J. T., & WOHDEN, F. G. Auditorypotentials during acoustic habituation:Cochlear nucleus, cerebellum and auditorycortex. Electroencephalography and Clin-ical Neurophysiology, 1964, 17, 685-692.

MARSH, J. T., WOSDEN, F. G., & HICKS, L.Some effects of room acoustics on evokedauditory potentials. Science, 1962, 137,281-282.

MARTIN, I. Adaptation. Psychological Bul-letin, 1964, 61, 35-44.

MELZACK, R. On the survival of Mallardducks after "habituation" to the hawk-shaped figure. Behavior, 1961, 17, 1-16.

MENDELL, L. M., & WALL, P. D. Pre-synaptic hyperpolarization: A role for fineafferent fibers. Journal of Physiology,1965, 172, 274-294.

MERTON, P. A. Voluntary strength andfatigue. Journal of Physiology, 1954, 123,553-564.

MONTGOMERY, K. C. Exploratory behavioras a function of "similarity" of stimulussituations. Journal of Comparative and


Physiological Psychology, 1953, 46, 129-133.

MORGAN, C. T., & WOOD, W. M. Corticallocalization of symbolic processes in therat: II. Effect of cortical lesions upondelayed alternation in the rat. Journal ofNeurophysiology, 1943, 6, 173-180.

MORRELL, F. Electrophysiological contribu-tions to the neural basis of learning. Phys-iological Review, 1961, 41, 443-494.

MORUZZI, G. The ascending reticular acti-vating system and wakefulness. In J. F.Delafresnaye (Ed.), Brain mechanismsand consciousness. Oxford: Blackwell,19S4. Pp. 1-20.

MOULTON, D. G. Electrical activity in theolfactory system of rabbits with indwellingelectrodes. In Y. Zotterman (Ed.), Olfac-tion and taste. New York: MacMillan,1963. Pp. 71-84.

MOUSHEGAN, G., RUPERT, A., MARSH, J., &GALAMBOS, R. Evoked cortical potentialsin absence of middle ear muscles. Science,1961,133, 582.

MOWRER, O. H. The modification of vestibu-lar nystagmus by means of repeated elici-tation. Comparative Psychology Mono-graphs, 1934,9(45).

NAQUET, R., REGIS, H., FISCHER-WILLIAMS,M., & FERNANDEZ-GUARDIOLA, A. Vari-ations in the responses evoked by lightalong the specific pathways. Brain, 1960,83, 52-56.

NESMEIANOVA, T. N. The inhibition of themotor reflex in spinal dogs under condi-tions of chronic experimentation. SechenovJournal of Physiology of the USSR, 1957,42, 281-288.

PALESTINI, M., DAVIDOVICH, A., & HERNAN-DEZ-PE6N, R. Functional significance ofcentrifugal influences upon the retina.Acta neurologica latinoamericana, 1959, 5,113-131.

PAVLOV, I. P. Conditioned reflexes. (Trans,by G. V. Anrep) London: Oxford, 1927.

PORTER, J. M. Adaptation of the galvanicskin response. Journal of ExperimentalPsychology, 1938, 23, 553-557.

PRIBRAM, K. H. Reinforcement revisited: Astructural view. In M. R. Jones (Ed.),Nebraska symposium on motivation: 1963.Lincoln: Univer. Nebraska Press, 1963.Pp. 113-160.

PROSSER, C. L., & HUNTER, W. S. The ex-tinction of startle responses and spinal re-flexes in the white rat. American Journalof Physiology, 1936, 117, 609-618.

RANSON, S. W., & HINSEY, J. C. Reflexesin the hind limbs of cats after transsectionof the spinal cord at various levels. Amer-

ican Journal of Physiology, 1930, 94, 471-495.

RASMUSSEN, G. L. The olivary peduncleand other fiber projections of the superiorolivary complex. Journal of ComparativeNeurology, 1946, 84, 141-219.

RHINEBERGER, M., & JASPER, H. H. Electri-cal activity of the cerebral cortex in theunanesthetized cat. American Journal ofPhysiology, 1937, 119, 186-196.

ROBERTS, W. W., DEMBER, W. N., & BROD-WICK, M. Alternation and exploration inrats with hippocampal lesions. Journal ofComparative and Physiological Psychol-ogy, 1962, 55, 695-700.

ROBINSON, E. S. Work of the integrated or-ganism. In C. Murchison (Ed.), A hand-book of general experimental psychology.Worcester, Mass.: Clark Univer. Press,1934. Pp. 571-650.

ROBINSON, J., & GANTT, W. H. The orient-ing reflex (questioning reaction) : Cardiac,respiratory, salivary, and motor com-ponents. Bulletin of the Johns HopkinsHospital, 1947, 80, 231-253.

RODGERS, W. L., MELZACK, R., & SEGAL,J. R. "Tail flip response" in goldfish.Journal of Comparative and PhysiologicalPsychology, 1963, 56, 917-923.

ROIG, J. A., & SOMMER-SMITH, J. A.Atenuacion especifica del sobresalto pro-vacado par excitacion nerviosa central.Anales de la Facultad de medicina, Uni-versidad de Montevideo, 1959, 44, 410-413.

ROSE, J. E., GREENWOOD, D. D., GOLDBERG, J.M., & HIND, J. E. Some discharge charac-teristics of single neurons in the inferiorcolliculus of the cat: I. Tonotopical or-ganization, relation of spike-counts to toneintensity, and firing patterns of single ele-ments. Journal of Neurophysiology, 1963,26, 294-320.

SANTIBANEZ, G., TROUCHE, E., & ALBE-FESSARD, D. fitude de 1'evolution au coursdu temps de 1'amplitude des activites corti-cales et thalamiques evoquees par desstimuli somatique repetes. Journal dePhysiologie, 1963, 55, 335.

SCHOLANDER, T. Habituation of autonomicresponse elements under two conditions ofalterness. Acta physiologica scandinavica,1960, 50, 259-268.

SCHWARTZBAUM, J. S., WlLSON, W. A., JR.,

& MORRISSETTE, J. R. The effects ofamygdalectomy on locomotor activity inmonkeys. Journal of Comparative and'Physiological Psychology, 1961, 54, 334-336.

SEGUNDO, J. P., ROIG, J. A., & SOMMER-SMITH, J. A. Conditioning of reticular-


formation stimulation effects. Electro-encephalography and Clinical Neurophysi-ology, 1959, 11, 471-484.

SHARPLESS, S. K. Reorganization of func-tion in the nervous system—use and disuse.Annual Review of Physiology, 1964, 26,357-388.

SHARPLESS, S., & JASPER, H. Habituationof the arousal reaction. Brain, 1956, 79,655-682.

SHAW, J. A., & THOMPSON, R. F. Depend-ence of evoked cortical association re-sponses on behavioral variables. Psycho-nomic Science, 1964, 1, 153-154.

SHERRINGTON, C. S. Experiments in exami-nation of the peripheral distribution of thefibers of the posterior roots of some spinalnerves. Royal Society of London, 1898,190, 45-186.

SHERRINGTON, C. S. The integrative actionof the nervous system. New Haven: YaleUniver. Press, 1906.

SHURRAGER, P. S., & CULLER, E. Condi-tioned extinction of a reflex in the spinaldog. Journal of Experimental Psychology,1941, 28, 287-303.

SHURRAGER, P. S., & SHURRAGER, H. C.Converting a spinal CR into a reflex.Journal of Experimental Psychology, 1941,29, 217-224.

SIMEONE, F. A. Responsiveness of de-nervated nictitating membrane. AmericanJournal of Physiology, 1938, 122, 650-658.

SOKOLOV, E. N. The higher nervous activityand the problem of perception. Proceed-ings of the XIX International Congress ofPsychology, Montreal, 1955.

SOKOLOV, E. N. Neuronal models and theorienting influence. In M. A. B. Brazier(Ed.), The central nervous system andbehavior: III. New York: Macy Foun-dation, 1960.

SOLOMON, R. L. The influence of work onbehavior. Psychological Bulletin, 1948,45, 1-40.

SPENCE, K. W. Behavior theory and learn-ing. Englewood Cliffs, N. J.: Prentice-Hall, 1960.

SPENCER, W. A., THOMPSON, R. F., & NEIL-SON, D. R. Analysis of polysynaptic re-flex response decrement in the acute spinalcat. Physiologist, 1964, 7, 262.

SYMMES, D. Anxiety reduction and noveltyas goals of visual exploration by monkeys.Journal of Genetic Psychology, 1959, 94,181-198.

SYMMES, D. Effect of cortical ablations onvisual exploration by monkeys. Journalof Comparative and Physiological Psychol-ogy, 1963, 56, 757-763.

THOMPSON, R. F., & SHAW, J. A. Be-havioral correlates of evoked activity re-corded from association areas of the cere-bral cortex. Journal of Comparative andPhysiological Psychology, 1965, 60, 329-339.

THOMPSON, R. F., & WELKER, W. I. Roleof auditory cortex in reflex head orienta-tion by cats to auditory stimuli. Journalof Comparative and Physiological Psychol-ogy, 1963, 56, 996-1002.

THORPE, W. H. Learning and instinct inanimals. London : Methuen, 1956.

VORONIN, L. G., & SOKOLOV, E. N. Corticalmechanisms of the orienting reflex and itsrelation to the conditioned reflex. Electro-encephalography and Clinical Neurophysi-ology Suppl. 13, I960, 335-346.

WALKER, E. L. Action decrement and itsrelation to learning. Psychological Re-view. 1958, 65, 129-142.

WEBSTER, W. R., DUNLOP, C. W., SIMONS,L. A., & AITKIN, L. M. Auditory habitu-ation: A test of a centrifugal and aperipheral theory. Science, 1965, 148, 654-656.

WELKER, W. I. Variability of play andexploratory behavior in chimpanzees.Journal of Comparative and PhysiologicalPsychology, 1956, 49, 181-185.

WELKER, W. I. An analysis of exploratoryand play behavior in animals. In D. W.Fiske & S. R. Maddi (Eds.), Functionsof varied experience. Homewood, 111.:Dorsey, 1961. Pp. 175-226.

WENDT, G. R. Vestibular functions. InS. S. Stevens (Ed.), Handbook of ex-perimental psychology. New York: Wiley,1951. Pp. 1191-1223.

WILSON, V. J. Post-tetanic potentiation ofpolysynaptic reflexes of the spinal cord.Journal of Genetic Physiology, 1955, 39,197-206.

WILSON, V. J. Early post-tetanic potentia-tion and low frequency depression of somegroup I reflex actions. Journal of GeneticPhysiology, 1958, 41, 1005-1018.

WORDEN, F. G., & LIVINGSTON, R. B. Brain-stem reticular formation. In D. E. Sheer(Ed.), Electrical stimulation of the brain.Austin: Univer. Texas Press, 1961. Pp.263-276.

WORDEN, F. G., & MARSH, J. T. Amplitudechanges of auditory potentials evoked atcochlear nucleus during acoustic habitua-tion. Electroencephalography and ClinicalNeurophysiology, 1963, 15, 866-881.

(Early publication received April 5, 1965)

psychological review 73, no. 1,...

Documents