cortical networks for visual reaching: paul b. johnson,1 ...€¦ · a result of this study, area...

Cortical Networks for Visual Reaching:Physiological and AnatomicalOrganization of Frontal and Parietal LobeArm Regions

Paul B. Johnson,1 Stefano Ferraina,2 Luigi Bianchi,2 andRoberto Caminiti2

'Department of Anatomy and Cell Biology, University ofNorth Carolina, Chapel Hill, North Carolina 27599, and^stituto di Fisiologia umana, Universita degli Studi di Roma"La Sapienza," 00185 Rome, Italy

The functional and structural properties of the dorsolateral frontal lobeand posterior parietal proximal arm representations were studied inmacaque monkeys. Physiological mapping of primary motor (Ml), dor-sal premotor (PMd), and posterior parietal (area 5) cortices was per-formed in behaving monkeys trained in an instructed-delay reachingtask. The parietofrontal corticocortical connectivities of these sameareas were subsequently examined anatomically by means of retro-grade tracing techniques.

Signal-, set-, movement-, and position-related directional neuronalactivities were distributed nonuniformly within the task-related areasin both frontal and parietal cortices. Within the frontal lobe, movingcaudally from PMd to the Ml, the activity that signals for the visuo-spatial events leading to target localization decreased, while the ac-tivity more directly linked to movement generation increased.

Physiological recordings in the superior parietal lobule revealed agradient-like distribution of functional properties similar to that ob-served in the frontal lobe. Signal- and set-related activities were en-countered more frequently in the intermediate and ventral part of themedial bank of the intraparietal sulcus (IPS), in area MIP. Movement-and position-related activities were distributed more uniformly withinthe superior parietal lobule (SPL), in both dorsal area 5 and in MIP.

Frontal and parietal regions sharing similar functional propertieswere preferentially connected through their association pathways. Asa result of this study, area MIP, and possibly areas MDP and 7m aswell, emerge as the parietal nodes by which visual information maybe relayed to the frontal lobe arm region. These parietal and frontalareas, along with their association connections, represent a potentialcortical network for visual reaching. The architecture of this networkis ideal for coding reaching as the result of a combination betweenvisual and somatic information.

The act of reaching to visual targets requires the combinationof information regarding the locations of external objectswith information concerning the configuration of our ownbody segments with respect to those objects. The presentstudies were undertaken to investigate, within the corticalareas involved in reaching, the relationships between the rep-resentations of sensorimotor information and the anatomicalsubstrates by which this information is combined and trans-formed in selected cortical areas that are directly involved inthe control of movement: the primary motor cortex (MI), thedorsal premotor cortex (PMd), and parts of area 5 in the su-perior parietal lobule (SPL). The relationship between thephysiological properties and the extrinsic cortical connectiv-ity of these cortical regions was studied by combining phys-iological and anatomical approaches, to obtain data on boththe representation and the flow of information to, andthrough, the cortical frontal and parietal areas involved in thecontrol of reaching.

The proximal arm representations of MI and PMd havebeeq well identified by neurophysiological studies (Geo'rgo-poulbs et al., 1982; Weinrich and Wise, 1982; Kurata and Tanji,1986; Schwartz et al., 1988; Caminiti et al., 1991) and manyneurons in both regions are active with arm-reaching move-ments. A difference between these areas has been observed

during the performance of instructed-delay tasks. Neurons ex-ist in the proximal arm regions that are preferentially activeduring the delay period between the "cue" signal and the "go"signal in such tasks (for reviews, see Wise, 1984,1985). Whilethese types of signal- and set-related activities are present inMI (Tanji and Evarts, 1976; Lecas et al., 1986; Georgopouloset al., 1989; Alexander and Crutcher, 1990), they have beenfound to be more prevalent in PMd (Weinrich and Wise, 1982;Weinrich et al., 1984; Godschalk et al., 1985; Riehle and Re-quin, 1989; Riehle, 1991). Although these activities cannot beregarded as purely sensory related, their existence, along withother data, have led to the hypothesis that PMd is involved inthe sensory guidance of movement (see Wise, 1984, 1985).Studies employing experimentally induced lesions or inacti-vations of PMd have shown that, while this cortical motorarea is not essential for the generation of movement, impair-ment of PMd seems to affect the processing of the sensoryinformation used to elicit movement (Sasaki and Gemba,1986; Rea et al., 1987; Passingham, 1988; Kurata and Hoffman,1994).

Early degeneration and autoradiographic tracing studies re-vealed anatomical connections between PMd and MI (Pandyaand Kuypers, 1969; Pandya and Vignolo, 1971; Kiinzle,1978a,b). These studies showed reciprocal connections be-tween the two areas, although the projection from PMd to MIhas gained more attention in subsequent retrograde tracingexperiments (Matsumura and Kubota, 1979; Muakkassa andStrick, 1979; Godschalk et al., 1984; Ghosh et al., 1987). Theanatomical experiments of recent years have revealed thatboth MI and PMd form part of an extensively interconnectednetwork of frontal motor areas that also include the arcuatepremotor area, the supplementary motor area (SMA), and thecingulate motor areas (Muakkassa and Strick, 1979; Matelli etal., 1984; Primrose and Strick, 1985; Barbas and Pandya, 1987;Dum and Strick, 1991; Tokuno and Tanji, 1993). Furthermore,there is evidence of substantial projections from PMd to thespinal cord segments controlling the proximal arm (He et al.,1993), indicating that MI is not the only output region of thefrontal lobe areas involved in reaching. As a result, the intrin-sic functional organization of the frontal lobe areas control-ling reaching is currently in question.

Knowledge of the extrinsic anatomical connectivities ofthese areas could provide insight into their functional roles.Previous anatomical studies of PMd and MI have been accom-plished either without physiological characterization of the in-vestigated region or, at most, with cortical stimulation mappingin anesthetized animals. The first objective of the present studywas to characterize physiologically the MI and PMd reaching-related regions using a directional instructed-delay task andthen to correlate the distributions of functional properties withthe extrinsic corticocortical connectivity of the region.

The motor and premotor areas receive anatomical projec-tions from the SPL (Chavis and Pandya, 1976; Jones et al.,1978; Strick and Kim, 1978; Petrides and Pandya, 1984; Ghoshet al., 1987; Cavada and Goldman-Rakic, 1989b; Kurata, 1991)

Cerebral Cortex Mar/Apr 1996;6:1O2-119; 1047-3211/96/S4.00

by guest on September 23, 2016

http://cercor.oxfordjournals.org/D

ownloaded from

http://cercor.oxfordjournals.org/

as well as motor areas of the medial frontal lobe (Pandya andKuypers, 1969; Jones et al., 1978; Kiinzle, 1978a; Matsumuraand Kubota, 1979; Muakkassa and Strick, 1979; Leichnetz,1986; Ghosh et al., 1987; Dum and Strick, 1991; Morecraft andvan Hoesen, 1992). The SPL has traditionally been viewed asbeing primarily somatosensory and somatomotor in function(Sakata et al., 1973; Mountcastle et al., 1975; Kalaska et al.,1983, 1990; Kalaska, 1988), and the projection from this areato the frontal lobe is a likely pathway by which informationconcerning arm configuration reaches the motor areas.

An issue that remains open, however, concerns the path-way by which visual information about target location is re-layed to the frontal lobe motor centers controlling reaching.Recent anatomical studies have shown that the SPL is corti-cocortically connected with the newly recognized visual area(area PO) in the parieto-occipital sulcus (POs). The regions ofthe SPL that have been shown to receive from area PO arethe medial intraparietal area (MIP) in the medial bank of theintraparietal sulcus (IPS; Blatt et al., 1990) and the medial wallof the SPL (area 7m; Cavada and Goldman Rakic, 1989a). Inaddition to receiving from area PO, areas MIP and 7m, as wellas an area at the caudal pole of the SPL (the medial dorsalparietal area, MDP), also project back to area PO (Colby et al.,1988; Cavada and Goldman-Rakic, 1989a).

The physiological properties of areas MIP, MDP, and 7mhave not been well characterized. There have been no re-cording studies of area 7m, whereas one study has reportedvisual and oculomotor activity in portions of area MDP (Gal-letti et al., 1991). Area MIP, which loosely corresponds to areaPEa of Pandya and Seltzer (1982), has not been considered asdistinct from the remainder of Brodmann's area 5 by mostauthors. Therefore, many physiological studies of area 5 mayhave included neurons that could now be classified as beinglocated in MIP. Some recent physiological studies of the SPLhave noted differences in properties between superficial(area PE, Pandya and Seltzer, 1982) and deep (MIP) regions ofthe SPL (Crammond and Kalaska, 1989; Burbaud et al., 1991;Colby and Duhamel, 1991). The data from these experimentssuggest that activity in PE is more related to somatosensoryfunction, while that in MIP may be more related to motor andvisual functions.

As a result, another goal of the present study was to relatethe distributions of visuomotor activity in the areas of thefrontal lobe involved in reaching to the distribution of func-tional properties in the parietal regions where the parieto-frontal projections originate.

A preliminary report has appeared (Johnson et al., 1993).

Materials and Methods

Behavioral Apparatus and TaskThe behavioral apparatus consisted of nine metal rods oriented to-ward the animal. The end of each rod was fitted with a translucentpush button that could be illuminated by a red/green LED mountedbehind the button. Eight of the push buttons were positioned at thevertices of an imaginary cube of 10 cm on a side. The ninth pushbutton was positioned at the center of the cube (8.7 cm from thevertices). The center button of the cube was located in the midsa-gittal plane at shoulder height and 25 cm from the animal. Duringrecording sessions, the animal sat in a primate chair with the headfixed and made free arm-reaching movements. A personal computerwas used to monitor the status of the push buttons and to turn onand off the LEDs.

The monkeys were trained to perform an instructeoSdelay reachingtask. At the beginning of each trial the center button was illuminatedred, and the animal was required to press and hold it depressed fora variable period of time (1000-1500 msec) until the LED was extin-guished and one of the peripheral targets was illuminated green (in-struction stimulus, IS). The animal was required to maintain pressingthe center button for a variable instructed-delay time (IDT; 600-1800

msec) until the green light turned to red. The change in color of theTE1~> served as a "go" stimulus (GS) for the animal to reach towardthat push button, and to press it for a specified target-holding time(THT; 150 msec) in order to receive a liquid reward. Predefined re-action (120 and 250 msec, lower and upper limits) and movementtimes (1000 msec upper limit) were used to control the animal'sperformance. Different target buttons were then presented in a ran-domized block design. Randomization of trials, sequencing of targets,and monitoring of the animals' task performance were all performedby computer.

Animals

Experiments in the Frontal LobeTwo juvenile female Macaca nemestrina monkeys (body weights 4.2kg and 5.7 kg) were trained in die above task and used for the neu-rophysiological and retrograde tracing experiments to map activityand inputs of task-related regions of frontal cortex. The first animal(animal MO) performed the task with the right arm, the second (an-imal TY) with the left arm. For consistency in presentation, all datafrom animal TY have been mirror transformed.

Experiments in the Parietal LobeTwo additional juvenile female Macaca nemestrina monkeys (bodyweights 3-2 and 3.5 kg) were used for parietal recording, to providefunctional mapping of die SPL in the same task. The animals weretrained to use either arm to perform the task. Recordings were always

' made in the hemisphere contralateral to the performing arm.

Neuropbysiological RecordingsAfter each animal was trained in the task, a recording chamber anda head holder were surgically implanted under aseptic conditions.Surgical and extracellular recording procedures were similar to thosereported in a previous study (Caminiti et al., 1990a) to which diereader is referred for details.

For each isolated neuron, a qualitative examination of die animalwas performed in an effort to determine whedier die cell's activitywas related to arm movements at die shoulder and/or elbow jointsand to determine the presence and quality of passive inputs to thecell from skin, deep tissues, muscles, and joints. Only those cells diatwere related to proximal arm movements were selected for furtherstudy since die task involved primarily the proximal musculature (seeCaminiti et al., 1990b).

Relative loci of die recording sites were reconstructed based upondie recording chamber coordinates of the microelectrode penetra-tions. Recording chamber coordinates were related to die histologicaltissue and to brain surface features by means of a marking dye (Alcianblue) injected into die cortex at known chamber coordinates (seeAnatomical Metiiods below).

Quantitative Physiological AnalysesIn each trial, four epochs of interest were identified: (1) die first 300msec of the instructed-delay time (IDT;DT1); (2) the remainder ofthe IDT (DT2); (3) die 400 msec epoch extending from 200 msecbefore movement onset until 200 msec after movement onset (re-action/movement-time; RTMT); (4) and the target-holding time; THT.For each cell, die mean neuronal spike frequency during each epochof each trial was calculated by dividing the number of spikes in anepoch by its duration. As a statistical measure of whedier a cell wasmodulated during a particular epoch, a one-factor ANOVA (Zar, 1984)was performed oh these frequencies. The ANOVA tested die nullhypothesis diat there was no effect of movement direction on meanfiring frequency during die given epoch (see Schwartz et al., 1988).Probability levels were approximated from the F statistic (Press etal., 1992).

A quantitative measure was used in this study to examine die dis-tribution of die amount of neuronal activity across the tangentialdimension of die cortex. For each neuron n and each epoch i, theamount of directional activity, Dm, was computed as

max, - min, (F^),

where Fnf/ is die mean firing frequency of cell n during epoch i formovements in direction/ Max, and min, are die functions that takedie maximum and minimum over all directions/ respectively. Meanlevels of this measure for each penetration were computed by aver-

Cerebral Cortex Mar/Apr 1996, V 6 N 2 103



ownloaded from


aging the individual measures of all neurons recorded in that pene-tration.

hi the frontal lobe, the rostrocaudal positions of neurons weredetermined by relating the recording chamber coordinates of themicroelectrode penetrations to the histological tissue and to brainsurface features by means of a marking dye (Alcian blue) injectedinto the cortex at known chamber coordinates. In order to pool dataacross animals, data were combined using the cytoarchitectonicallydefined area 4/6 border as an alignment guide.

In the parietal lobe, most penetrations were performed within thecortex forming the rostral (dorsomedial) bank of the intraparietalsulcus (IPS). These penetrations entered the cortex on the crown ofthe sulcus and traversed the cortex tangentially for varying distances(e.g., see Fig. 9). Therefore, the position of each neuron was deter-mined relative to the location where the first sign of neural activitywas recorded (top of the neural activity). The latter was determined,for each penetration, both on entering and exiting the cortex.

For the analysis of neuronal latencies with respect to behavioralevents, times of onset of changes in firing frequency were deter-mined using the cumulative sum (cusum) of the perievent time his-togram (PETH). Activities of neurons were aligned to the time ofbehavioral events and 20 msec bin PETHs were calculated separatelyfor each direction class of each neuron. A significance threshold forthe cusum of the PETH was set at three times the standard deviationof the PETH during a control period (Ellaway et al., 1983; Davey etal., 1986). For PETHs aligned to instruction stimulus (IS) or go stim-ulus (GS) delivery, the control period was denned as the 1 sec pre-ceding stimulus delivery. For PETHs aligned to movement onset, thecontrol period was defined as the last 1 sec preceding the presen-tation of the GS. The cusum was calculated by subtracting the meancontrol period bin height from each bin of the PETH and then form-ing the cumulative sum of bins. The moment in time when the cusumcrossed the significance threshold, departure (p < 0.01) from theprestimulus firing frequency is judged to have occurred. If the sub-sequent two bins showed changes from the control rate in the samedirection (increase or decrease), this departure was taken as the on-set of neuronal activity. This latter condition was added as a measureof consistency of the observed change in activity. For each neuronstudied, the earliest onset in activity across movement directions wasused as the latency for that neuron. Comparison of the distributionsof onset times of firing of frontal and parietal neurons were per-formed using the Kolmogorov-Smirnov test.

Anatomical Tracer Injectionshi the two frontal lobe animals, after approximately 30 d of recording,a second surgery was performed under aseptic conditions. The ani-mal was anesthetized (35-40 mg sodium pentobarbital i.p. per kg ofbody weight), and the dura within the recording chamber was in-cised and reflected. A 30 gauge hypodermic needle mounted directlyto the microdrive was used to inject the cortex with small quantities(0.5-1.0 u.1) of an inert but visible dye (Alcian blue) at known cham-ber coordinates outside the cortical area of interest. The visible dyemarks were used to guide the injection of retrograde tracers. All trac-ers were pressure injected -with a microsyringe (Hamilton) througha glass micropipette tip (approximately 100 jim diameter). After theinjections, the dura was sutured closed and sealed with fibrinogenglue (Tissucol, Immuno). The animal was returned to its cage for theremainder of the survival period.

The following tracers were used in these animals- red (rhodamine>labeled fluorescent latex microspheres (RLM; Lumafluor; undiluted);green (fluorescein>labeled fluorescent latex microspheres (GLM; Lu-mafluor; undiluted); cholera toxin B-subunit conjugated to colloidalgold (CTB-Au; Gilt; one vial reconstituted in 25 ^1 distilled water); andwheat germ agglutinin-apo horseradish peroxidase conjugated to col-loidal gold (WGA-apoHRP-Au; undiluted, prepared according to Bas-baum and Menetrey, 1987). The locations and quantities of injectedmaterials and the survival times are given in Table 1. During thesurvival period the animals were maintained under wide-spectrumantibiotic coverage (Spectrum, Sigma-Tau).

Histological ProceduresAt the end of the survival period, the monkeys were deeply anesthe-tized (40-50 mg sodium pentobarbital i.p./kg body weight) and per-fused through the heart first with 0.9% saline followed by 2 liters of4% paraformaldehyde in 0.1 M phosphate buffer (pH 7.2). The brains

Table 1Retrograde tracer injections

Animal

MO

TY

Physiology

IDT taskIDT taskIDT task

IDT taskIDT taskIDT task

Hemi-sphere

leftleftleft

rightrightright

Area

MlMI/PMdPMd

MlMI/PMdPMd

Tracer

RLMHRP-AuGLM

RLMCTB-AuGLM

Number xamount

4 x 0.5 MJ4xO.3|d4x0.5 pi

4 x 0.5 td3 x 0.5 (J4 x 0.5 MJ

Sur-vival(days)

444

181818

were removed and postfixed in the same fixative. After infiltrationwith 30% sucrose, the brains were cut on a freezing microtome. Sec-tions were cut at 40 u.m in the coronal plane. Colloidal gold conju-gated tracers were visualized in a series of sections by silver inten-sification GntenSE, Amersham). Intensified and unintensified sectionswere mounted, dried, cleared briefly in xylene, and coverslipped.From each animal, a series of sections for cytoarchitectonic analysiswas counterstained with thionin (0.025%).

Analysis of Anatomical DataFor each animal, a series of coronal sections at 800 urn intervals fromeach animal was plotted using a computer-based plotting system. Xand Y coordinates of labeled neurons, tracer injection sites, dye in-jection sites, and other landmarks were measured by transducersmounted on the stage of a epifluorescence/dark-field microscope.The outputs of the transducers were fed to an interactive plottingprogram. Gold-conjugated, silver-intensified tracers were viewed un-der dark-field illumination, while the microspheres were viewed us-ing appropriate epifluorescence filter sets.. Flattened cortical maps showing the tangential distribution of la-beled neurons in the SPL were generated from plots of individualsections. The computer-based flattening procedure that was used hasbeen described in detail previously (Johnson et al., 1989). Briefly, thecomputer first generated a series of radial line segments extendingfrom the layer Wwhite matter boundary to the cortical surface. Sec-ond, a "trend line" was generated by connecting the midpoints of theradial line segments. Third, the tangential locations of the items ofinterest were projected onto the trend line by using the radial linesegments as "guides." Fourth, within each section, the cell distribu-tions along the mediolateral trend line were binned into 0.1 mm bins.Finally, two-dimensional gray scale maps of cell densities were cre-ated by aligning successive sections to fixed anatomical landmarksand interpolating the cell densities at 0.1 mm intervals in the rostro-caudal dimension.

Cytoarchitectonic ClassificationTraditionally, the distribution of giant pyramidal cells has been usedalmost exclusively for the determination of the border between areas4 and 6 (Sessle and Wiesendanger, 1982; Weinrich and Wise, 1982).To allow comparison with other studies, the area 4/6 border wasestablished at the rostral extent of the distribution of giant pyramidalcells (soma widths > 29 u.m). hi this study, we refer to MI, PMd, andthe MI/PMd border region, loosely defined as the cortex surrounding4/6 border. The size distributions of counterstained pyramidal cellbodies were compiled from thionin-stained coronal sections. The celldiameter was measured along the minor axis (90 degrees relative tothe major axis) of all neurons in layer V using a computer-based videomicroscopy system. The diameters of all neurons with minor axesgreater than 15 jun were plotted as a function of their anterior-pos-terior position.

Results

Frontal Lobe Reacb-Related Activity

Neuronal Data BaseThe frontal lobe physiology portions of this article are basedon quantitative analyses of the activities of 208 neurons re-corded in 93 successful microelectrode penetrations in two

104 Intrinsic Organization of Frontal and Parietal Cortices •Johnson et al.



ownloaded from


Animal MO Animal TY

Hgure 1. Locations of recording penetrations. Center panels shows enlargements of the recorded region. The location of the enlarged areas are indicated on the drawings below.Entry points [filled circled of microelectrode penetrations in which neuronal activity was recorded were reconstructed using locations of marking dye injections [open squares!) foralignment of recording chamber. Open circles indicate locations of retrograde tracer injections. The approximate border between areas 4 and 6 is indicated by the broken line.These borders were derived from the data in the graphs above the enlarged areas that plot the minor axis dimension of layer V pyramidal cells throughout the recorded region.Each vertical cluster of data points is the size distribution of pyramidal cells in a 5 mm segment through the recorded region of one 40 p.n\ coronal section. The clusters are alignedto the enlarged center panels to indicate the rostrocaudal position of the section from which they were measured. AS, arcuate sulcus; CS, central sulcus; SPcS, superior precentralsulcus. Scale bars are 3 mm.

hemispheres of two monkeys (92 neurons in 44 penetrationsin monkey MO, 116 neurons in 49 penetrations in monkeyTY). Most penetrations were made in the flat, exposed surfaceof cortex lying rostral to the central sulcus (CS), medial to thespur of the arcuate sulcus (spAS), and lateral to the superiorprecentral sulcus (SPcS). The locations of microelectrode en-try points are shown in Figure 1 for the two animals. In bothanimals, the recording penetrations spanned a caudal to ros-tral extent of cortex in which there was a marked decreasein the size of the largest pyramidal cells (Fig. 1).

Types of Neuronal ActivityNeurons in the recorded regions displayed activity related tovarious aspects of the instructed-delay task. In accord withnumerous previous studies, certain characteristic types of ac-tivity were observed during four temporally distinct epochsof the task: phasic signal-related activity immediately follow-ing the presentation of the IS; tonic set-related activity during

the interval between IS and GS; movement-related activitythat usually began before the initiation of movement; and po-sition-related activity that occurred while the monkey held its.hand at the different target positions. The defined time ep-ochs (DTI, DT2, RTMT, THT) used in the study were selectedto distinguish these four types of activity. Figure 2A-D showsprototypical examples of these activity types. Signal- and set-related activities were broadly tuned to the direction of theupcoming movement as has been previously reported formovement-related activity (Schwartz et al., 1988; Caminiti etal., 1991). The majority of recorded neurons exhibited morethan one of these types of activity. Some neurons were foundto display all four of the above-defined activity types.

Effects of Movement DirectionNeurons were found to vary their activity during these fourepochs in relation to the direction of movement. Such direc-tional modulation was not the only type of task-related activ-




ownloaded from


aB

Ia

a2.

is M

18

• i I I i •

11 II

I 1 ,1111

!' ' ' \ '\ if• i m 1 1 ' i i i J i n i i

i . | n

, i »iIII I III I III•! •III I I I I

•! •I I , • Ian I I ,1

4 - I I n I • • T «ll t I I • I I1 7 ' i , ' ~ ; i. i

I I • • » • • i i In

• • , • , • • ' > : " " •

i " 5 i i • • I H ' I I ,• I i I I 1 5 Ii

I I l! B i l l I

•' i • - • '• •:• • "I ii. ' i ^ . ' TT1.1

, : ' , ! " • " • • 1 4 • ' ' • / " ' i1 ' ' ' i i 1 1 i i .

'it

i i h • • • a • i , J • ! • ] i I I ._

•' ' IS • • i i ' , ' I V ' 1 1 3

• II i r . • • , , 1 1 , , i • 11

ftI I I,

, i i 1 1 ' 12 ' ' • ",

" I • • ' , _ , ! 11 , II' . . . . —.: i-.i ••'rt ' ' 11

• •• t • - . - . • • i * , ."i ' * "

I , II I •

•M•»........Jil /I

i I I

-1000 0 1000 2000ms -3000 -2000 -1000 Oms

Bis

ini

i

i iI

II •

i

hl

II

1000

II

iII i

I

i ii

i ia in

' . .• n i

rai i i

i I

i ii11 i

•i

i ii n

111il I

i i• i I

i iI I it III at i

an nun • • • • 1

i mi itn HI HI i f i

'Ki'iiTi'.r"!i

. ' i . i , 1 I , I

M ' V I m i ' I ' I

.1 ' ' i i i

:• '• s1 \in minii 1 1 1

aammimmC^T'

• ' :'l Imiian • • •••

mill i 1 ITi mi i i al i

III I I

. . , , . . ' . fI ' I

0 1000

it.

\ ,

11

inL

i

• p" l i

t:

i i

r\1

' I . ' " '

'!• I 1 , .fi ' 'i i i

i' ' Ji i i

I'"'I 'i '

| l ; V l . , ,

I.1!1 "

1 il

2000ms

18

17

16

15

14

13

12

11

M

l i i i i innn HI • i i i rI , ill in I mi BBI • 11I , I I i li inn i |nI il IHI II In

I , i i , I I i i I,I i i i i I I

i I I i ia in it I i

I in i I * • • I I I I i I,I I I iT II i i I

I . I i H n i In I I il •

I!

• inn iI ni i I i •

* \ > •• \ ft"• it i 1 1 r i i i

Ii i I uiminaiI I I i l l i n m i i i n m a i i mi nil

1 , • • • Tiiiaaiafli n BI • i r • 11I i T u u a i m i n ! • • • • •

III I l l l B l l H i h BI

i l II, i il m a i m • i MI ii i t i l l i I IIi . i II f l i p n • • • • n 11 I I

till I l " B i t B i l l IIHJU I I II I I I I Him r > J

I I I I 1 i • m i i af I I (

. . . ' II III. . . ' I • •I I I I H I i I

-3000 -2000 -1000 Oms

IS

*

1S

I I iii

13

12

11

1 . . ' : . ' .

i

i i

i i

-1000

MT

lj'Siil'1

I ) * , • - ! ( . I.

I'.1! r ( , '

i II i1 ' I'ISII'IIII

v 7. I II I I I,

I I

II I

. ' : • • • • ! ' 1 ^

1000 2000ms -3000 -2000 -1000 Oms

IS

, ' , I'ff" , , I V n i ' l * 'i IMII i m i ik i I litntBM

i t i i l t i l n m i l

1 '

I M i,I I II II I I I II I

II I I I III I IllllllllIII III II I

I Iia

Liw

• i

i n I I i II in ntai it I I 11; n t • iota ilJMMW

in I nil I I 11 11 i i i i i in IB in 11 n In t, taiI I • i i

11 11 11 11 I

-1000 1000 2000ms

18

17

15

14

13

12

11 I I I I I U t i l I1 Ti11 ;1":,1 , ' ! , ,

1 i ' 'm 1 , , , , 1 , ! , ! , , ! 1

" " " " I ' I , ' , '"•„"•,,i

, ''V.V.1. »'-"• '! l'.'E'1C li inn n I I ilnn i n i t i t l nkIw | || i i a i IM h i in f li tCi

I il i i ti I I t n I I III I MBI

II I I

"l.i , il I I . , • , . «I , 11 11 i I ,1 ,1 i II1 I I I , II t i l l ,1 IB

II til I ha ia

I In i i i n In , i 1, I ii, I I < i nun I I i mi i ,1 11

" ' " • " ' i , ' " i " " i1, ,1 l,I I «i i H I M I lirm

I i n I I i I I in nkii I I I I MI n i • nita 11 iinHH

i ''."'l/'r" ij";"""'l'iif:!l11 l r , , , , , , , . n | 4 , , J n , i , , , , | . . n h , l u j . i ! J J |

I I, ii i i ii i i i il i , i imi i n Iii i n

-3000 -2000 -1000 Oms

Figure 2. Rasters showing prototypical neuronal activity. A, Signal-related neuron. Each horizontal row of tick marks corresponds to one behavioral trial. Same data is shown twice. At left each trial is aligned to thepresentation of tha IS. At right, the same data is aligned to tha onset of movement Rasters are grouped by movement direction as indicated by the number at center. Short, thick tick marks indicate the time of occurrenceof single-action potentials. Longer, thin tick marks correspond to behavioral events during the evolution of the trial; from left beginning of neuronal recording, presentation of the IS, presentation of the GS, onset ofmovement, target acquisition, end of trial/reward. The time scale is in milliseconds. Note the burst of activity following the presentation of the IS in some movement directions. B, Set-related neuron. Conventions as in ACell discharged with a tonic frequency during the IDT in some movement directions. For other directions, the cell was silent. C, Movement-related neuron. Conventions as in A Neuron fired only around the time of movementgeneration. D, Position-related neuron. Conventions as in A except that at right rasters are aligned to the beginning of the target hold time (7). Cell fired differentially while the animal's hand was positioned at the varioustarget positions.



ownloaded from


-1

-2-8

B1 r

E

O

-2- 8 - 4 0 4 8

distance rostral to 4/6 border (mm)

Figure 3. Plots of directional activity (measure 0) ratios as functions of rostrocauda! po-sition. A, Plot of the logarithm of the ratio of signal- to movement-related activity as afunction of cell position relative to the area 4/6 border. Positive values indicate a highsignal-related activity relative to movement-related activity. Negative values indicate rel-atively more movement-related activity. Solid line is a linear regression with slope signif-icantly different from zero (p < 0.0001, regression Ftest r = 0322). B, Plot of the logarithmof the ratio of set- to movement-related activity. Conventions as in A Regression slope issignificantly different from zero (p < 0.0001, regression f t es t r - 0.295).

ity observed in the regions studied, as some neurons exhib-ited nondirectional increases or decreases in mean firing ratefollowing stimulus presentation or in association with thegeneration of movement. The directional design of the task,however, provided a simple method to quantify those changesin activity associated with changes in movement direction. AnANOVA was used to determine significance of the effect ofdirection on neuronal activity during each behavioral epoch.Neurons with statistically significant directional modulationwere considered task related. Because of the existence of non-directional neuronal modulation, the set of neurons that fulfillthis criterion is only a subset of all task-related neurons.

Most neurons were found to be modulated in more thanone epoch. While 189 of 208 (90.9%) neurons were direc-tional in at least one epoch, only 45 of these (45/189, 238%)neurons were directional in only one single epoch. Neuronsexhibiting significant directional modulation in more thanone epoch tended to do so in temporally adjacent epochs. Ofthe 144 neurons with directional modulation in more thanone epoch, only 18 (12.5%) exhibited temporally nonadjacentdirectional activities.

Distributions of Functional Properties across theFrontal Lobe

Distributions of Directional ModulationThe region of significant directional modulation (ANOVA,/> <O.O1, see Materials and Methods) extends from the CS rostrally

S.

80

60

40

20

n

- signal

-

.r—,r~i.-8 -4

B80

60

40

20

set

-

80

60

40

20

-4 J

movement

80

60

40

20

position

- 8 - 4 0 4 8

distance rostral to 4/6 border (mm)

Figure 4. Percentages of neurons with significant directional modulation (p < 0.01,ANOVA) as a function of rostrocaudal location. Percentages were calculated for 2 mmwide rostrocaudal bins relative to the area 4/6 border. A, Signal-related directionalmodulation. B, Set-related directional modulation. C, Movement-related directional mod-ulation. 0, Position-related directional modulation.

for approximately 15 mm. In the mediolateral dimension, thisregion extends at least from the SPcS, medially, to the level ofthe spAS, laterally. Given that directionally modulated neuronsare only a subset of all task-related neurons, the region iden-tified by this method is a conservative estimate of the entireextent of the task-related region of Ml and PMd. A few direc-tional neurons were encountered in penetrations medial tothe SPcS and lateral to the spAS. The extent of reach-relatedactivity in these latter two areas is not addressed by the pres-ent study.

The distributions of significant directional modulation dur-ing each of the four epochs covered wide areas within thestudied region. A common feature of these distributions wasthe degree to which they overlapped in the cortex. Often,usually within the central areas of the reach representation,significant directional modulation for all four epochs was en-countered within the same penetration (27/93 = 29.0% of




ownloaded from


Figure 5. Micrographs showing cross-sections of injection sites. All images are of coronal sections through the MI/PMd arm representation. Broken lines indicate the layer Vl/whitematter border. Scale bar in panel D is 1.0 mm. A, Dark-field/fluorescence image of a section through PMd showing four injections [arrowheads] of green fluorescent latex microspheres.B, Dark-field/fluorescence image showing three injections [arrowheads] of red fluorescent latex microspheres in Ml. C, Dark-field image of a silver-intensified section through theMI/PMd border region showing three injections [arrowhead^ of cholera toxin B subunit conjugated to colloidal gold. D, Bright-field image of a silver-intensified, thionin-stainedsection showing a single MI/PMd border region injection of WGA-apoHRP conjugated to colloidal gold.

108 Intrinsic Organization of Frontal and Parietal Cortices •Johnson et al.



ownloaded from


A. Animal MO B. Animal TY

area 5

A C E A C

Figure 6. Series of coronal sections through the parietal lobes of two monkeys. A, Data from animal MO. Locations of frontal lobe tracer injections are indicated by colors in the inset (red, FILM; yellow, CTB-Au; green, GLM).Locations of neurons labeled by each of these tracers are indicated by the corresponding colors in the sections. Levels of sections are indicated on the brain figurine. Abbreviations: AS, arcuate sulcus; CS, central sulcus; SPcS,superior precentral sulcus; IPS, intraparietal sulcus; MIP, medial intraparietal area; MDP, medial dorsal parietal area; m, medial; r, rostral. B, Data from animal TY. Yellow indicates WGA-apoHRP-Au injections and labeled neurons.All other conventions as in A



ownloaded from


penetrations). In spite of this overlap, these distributions tend-ed to exhibit some consistent differences from one another,hi both animals, the caudalmost penetrations were devoid ofsignal (DTI) directional modulation. Set (DT2) modulationwas more widely spread throughout the studied region. Move-ment-related (RTMT) and positional (THT) directional mod-ulation was concentrated in the more caudal regions.

Degree of Neuronal ModulationThe rostrocaudal distributions of the amounts of directionalneuronal activity (measure £>) during each of the four dennedtask epochs was quantified. As most neurons had some degreeof directional activity during both the IDT and the movementphase of the task, it was possible to compute the ratios ofthese activities. In this way, the relative activation of neuronscould be analyzed. The plots of these ratios (Fig. 3) revealedsignificant trends (p < 0.0001, regression F tests) for the rel-ative amount of directional activity during the earlier phasesof the task to increase as one moves rostrally through thecortex of the exposed surface of the frontal lobe.

Distributions of Directional NeuronsThe frequency of occurrence of neurons with significant di-rectional modulation during a given epoch varied with ros-trocaudal level (Fig. 4). Neurons with significant movement-related modulation are common in the caudal portions of therecorded region, between the central sulcus (CS) and the4/6 border. Rostral to the 4/6 border the incidence of move-ment-related modulation gradually decreased. The distributionof significant set-related modulation showed a converse ar-rangement: set modulation was common rostral to the 4/6border but gradually became less frequent approaching theCS. Significant signal-related modulation is, jn general, lesscommon than the other activity types but showed a distri-bution similar in shape to that of set-related modulation. Sig-nificant position-related modulation was spread throughoutthe rostrocaudal extent of cortex studied, and showed no con-sistent rostrocaudal trend. The incidences of all types of neu-ronal modulation were relatively high in the region aroundthe border between areas 4 and 6.

Retrograde Tracer InjectionsA total of 23 tracer injections were made at six sites in thefrontal lobes of the two monkeys used for MI and PMd re-cordings (see Table 1). The tracers used in these experimentswere chosen for their limited diffusion at the injection site.Upon histological inspection (Fig. 5), the fluorescent micro-sphere injection sites exhibited a dense core of injected ma-terial surrounded by a zone of extracellular label. Surroundingthis zone was an area of dense cellular labeling. The minoraxis diameters of the cores ranged from 250 to 800 u,m. Theinjections sites of CTB-Au and WGA-apoHRP-Au possessed acentral dense core and a "halo." Surrounding this halo was alarger area of dense labeling of cell bodies. The diameters ofthe core regions for CTB-Au and WGA-apoHRP-Au rangedfrom 300 to 650 n-m. The core regions of all of the injectionswere found to be limited to the gray matter of the cortex. Inaddition, multiple injections were made at six sites in twoother monkeys. Results from these two animals were consis-tent with the results from the animals studied in the instituteddelay task.

In both animals MO and TY there was one type of retro-grade tracer injected in each of Ml, PMd, and die MI/PMdborder region (Table 1, Fig. 1). In animal MO the injectionsites were located slightly more laterally and rostrally than inanimal TY. In each animal the injection sites were locatedwithin the physiologically characterized regions.

Tangential Distributions of Labeled NeuronsFollowing all injections, labeled cortical neurons were foundin die areas immediately surrounding the injection sites, invarious frontal motor areas, and in the superior parietal lobule.Many labeled cells were found in the posterior wall of thegenu of the arcuate sulcus (AS) and the medial wall of itsspur. There was a slight tendency for the cells projecting tothe more caudal portions of the proximal arm representationto be located more laterally in the AS. Neurons projecting tothe rostral injection sites tended to be located in the lateralwall of the spur of the AS. The cortex in the depth of thesuperior precentral sulcus was also labeled following mostinjections. Label in this region was notable for the tendencyof the distributions of cells labeled with the different tracersto overlap.

Labeled Neurons in the Parietal LobeWithin the parietal lobe, labeled neurons were located almostexclusively in the SPL. A few were found in area 7b followingthe MI and MI/PMd injections. Within the SPL, labeled cellswere found in the exposed cortex on the surface of the lob-ule and extending caudoventrally into the medial wall of theIPS. There were additional sites of labeling at the caudalmosttip of the SPL and within the caudal portions of the SPL cor-tex located on the medial wall of the hemisphere.

Neurons Projecting to MIFollowing the injections in the caudal regions of the frontallobe reaching-related area, labeled cells were found dorsallyin the exposed cortex of the SPL and in the medial crown ofthe IPS. In animal TY the label extended caudolaterally from

. the level of the postcentral dimple to the crown of the IPSand ventrally into the dorsalmost 5 mm of the medial wall ofthe IPS (Figs. 6B, red; ID). Although the border between areas2 and 5 is difficult to determine with certainty, it is probablethat labeled neurons were located within both of these areas.Labeling in animal MO was limited to a small focus in themedial crown of the IPS (Figs. 6A, red; 74).

Neurons Projecting to the MI/PMd Border RegionInjections in intermediate locations at the MI/PMd border la-beled SPL neurons at intermediate locations between the cellslabeled by MI and PMd injections. In animal TY, neurons werelocated caudolaterally on the exposed surface and extendedinto the dorsal part of the medial wall of the IPS (Figs. 6B,yellow; IE). In animal MO, labeled cells occupied much of themedial wall of the IPS as well as a restricted region of theventral wall of the posterior cingulate sulcus (CiS; Figs. 6A,yellow; IB). In this latter animal, some label also extendedonto die exposed surface of the SPL.

Neurons Projecting to PMdInjections in die PMd proximal arm area labeled neurons inseparate regions of the SPL. In animal TY, labeled neurons werefound primarily in die medial wall of die IPS, in area MIP (Figs.6B, green; IF). Some additional labeled cells were observed indie most caudal and dorsal extent of the SPL, in area MDR Inanimal MO, where die injection site was more rostrally located,die labeling was observed in area MDP and on die medial as-pect of die SPL (Figs. 6A, green; 7O- The medial SPL label indiis animal extended from die ventral wall of die most caudalportion of die CiS onto die mesial surface of die hemisphere(area 7m; Cavada and Goldman-Rakic, 1989a). Rostral injectionsin die frontal lobe arm region of the two additional animalslabeled neurons in die medial wall of die IPS. In one of dieseadditional cases, injections were made in a further rostral re-gion, which corresponds approximately to die cortex lying im-mediately anterior to die proximal arm representation. These

110 Intrinsic Organization of Frontal and Parietal Cortices-Johnson el al.



ownloaded from


B 5mm

Animal MO

medial

lateral

rostral caudal

cIPS

caudal rostral

-4 0 4

injection site location relative to 4/6 border (mm)

\

Animal TY

medial

lateral

rostral caudal

U •= 5mm r

Figure 7. Rattened reconstructions of a portion of the superior parietal lobule showing densities of labeled neurons on a gray scale. A-C, Data from animal MO. A, Map of densityof cells labeled by PMd injection of GLM. 8, Map of density of cells labeled by MI/PMd injection of CTB-Au. C, Map of density of cells labeled by RLM injection in Ml. D-F, Datafrom animal TY. D, Map of density of cells labeled by PMd injection of GLM. f, Map of density of cells labeled by MI/PMd injection of WGA-apoHRP-Au. f, Map of density of cellslabeled by RLM injection in Ml. Rgurines at left identify landmarks indicated in the maps [UPS, fundus of intraparietal sulcus; cIPS, crown of intraparietal sulcus; clF, crown ofinterhemispheric fissure). Scale at center shows the rostrocaudal locations of the injection sites relative to the area 4/6 border.




ownloaded from


Figure 8. Brain figurine displaying locations of successful microelectrode penetrationsin the superior parietal lobule of four hemispheres of two animals. Insets are enlarge-ments of the region outlined in the brain figurine. Dots are entry points of microelectrodepenetrations. IPS and CS indicate intraparietal and central sulci, respectively. PCD in-dicates postcentral dimple.

injections labeled cells deep within the most posterior loca-tions within the SPL, extending into area PO.

Physiological Trends in the Parietal LobeThe analysis of the functional organization of the SPL is basedon quantitative analyses of the activities of 307 neurons re-corded during 74 microelectrode penetrations (Figs. 8, 9) intwo animals. The penetrations traversed the cortex tangen-tially from die dorsal exposed part of area 5, on the crownof the medial bank of the EPS, into area M1P, deeper in themiddle portion of die same bank. The types of neuronal ac-tivities encountered during these long, dorsoventrally orient-ed penetrations are illustrated in Figure 9. Cells displayingdirectional activity during the movement epochs of the task(RTMT) were observed at all tangential locations studied (Fig.9A-E). Neurons directionalty active during the IDT, whilepresent also in the more dorsal regions explored (Fig. 9B),were encountered more consistently as one moved deeper(more ventrally) toward the cortex located within the bank(Fig. 9C-E). This was particularly true for directional activityoccurring early during IDT, after the presentation of die IS(Fig. 9D,E).

Degree of Neural ActivityThe dorsoventral distribution of die amount of directionalneuronal activity (spikes/sec) during each of the four epochs

of the task was quantified for the studied population of neu-rons by using the same measure of activity used in the frontalcortex. There is a relative lack of signal- and set-related direc-tional activity in the more dorsal regions of the medial wallof the IPS. The distributions of movement- and position-relat-ed directional activity show a converse trend: there is a rela-tive lack of these activity types in the deeper, more ventralregions.

As in the frontal cortex, the relative activation of neuronswas analyzed by plotting the ratios of those types of direc-tional activity coexisting within the same neuron as a func-tion of the tangential location of each neuron studied (Fig.10). The results show significant (p < 0.0005, regression Ftest) trends for the directional activity during the earlier ep-ochs of the task to increase relatively as one moves ventrallythrough the cortex toward area MIP.

Distribution of Directional NeuronsThe frequency of occurrence of significant directional mod-ulation (ANOVA,/> < 0.01) during each of the behavioral ep-ochs was not uniform across the dorsoventral extent of thestudied region (Fig. 11). Cells with significant signal and set-related modulations were more frequent at the more ventrallocations. Significant movement- and position-related modu-lation were distributed more evenly throughout the dorso-ventral extent of the medial wall of the IPS.

Comparison of Neuronal Activity Onset LatenciesIn the frontal lobe, the mean onset latencies of neuronal ac-tivities with respect to the presentation of the IS, GS, andmovement onset were 166 msec, 166 msec, and —69 msec,respectively. In the parietal cortex, respective mean onset la-tencies were 179 msec, 191 msec, and —37 msec. The distri-butions of onset times were compared across frontal and pa-rietal lobes (Fig. 12). Onset time distributions with respect tothe IS were practically identical (p = 0.8266, Kolmogorov-Smirnov test; Fig. 12/4). Onset times with respect to the GSand with respect to movement onset displayed significant dif-ferences between frontal and parietal lobes (p = 0.001,/> <0.0001, Kolmogorov-Smirnov test). For these epochs, whiledie times of firing onset of the earliest neurons were similarin the frontal and parietal lobes, a population of neurons withlate onset times was present in the parietal lobe (Fig. 12Z?,Q.Neurons in the frontal lobe were more tightly locked tem-porally to the onset of movement (Fig. 12Q. Note that boththe PETH aligned to GS and the PETH aligned to movementonset are derived from the same neuronal activity during thereaction time. Thus, these two measures provide two comple-mentary views of die same physiological event.

DiscussionThe aim of die present study was to combine the techniquesof behavioral neurophysiology widi those of neuroanatomyto characterize the functional organization of the areas in-volved in reaching of both frontal and parietal cortices. Thereare two principal findings of this study. First, while character-istic activity types were observed in both frontal and parietalcortices, these different types were generally recorded inoverlapping cortical regions and often from the same individ-ual neuron. Second, the regions of SPL projecting into diefrontal lobe functional gradient contain functional trends sim-ilar to those observed in the frontal cortex, suggesting diatgradients of inputs from the SPL underlies die functional gra-dients observed in the frontal lobe. Interestingly, similar trendshave been observed in other studies (Alexander and Crutcher,1990; Matsuzaka et al., 1992) of SMA.

These results will be discussed in relation to the functionaldifferentiation of motor cortical areas, to die information flow

112 Intrinsic Organization of Frontal and Parietal Cortices • Johnson et al.



ownloaded from


700

1350-

2800

3000

3850

O

IPS

18 ,

17 ,

16

15

14 ,

13 ,

12

11 ••»; ••

.;' .,7'| #'&'.'•'} f . ' | ' ' • : ." ;• '

• l " ' l ^

• • ) • ] • & £

••':

.' r'i

'.?:-'.•

' • ,•.*»'

" A ,

-1000 0 1000 2000

18

17

16

15

14

13

12

11

'.n..',.",v>,r,!isjai' I • . - T •. 1s3:iii!!!i.V.'-Si'ilj"A1i'«, ..• \

B-1000 0 1000 2000

18 .

17 .

16 .

15 .

14 .

13 .

12 .

11 .

1^••^T:';vSi!f:|v:::;:i•^••yj.••];••:i

,'. •, ' • j V ^ . r . 1 . . ^ ^ - , , ^ ; •• • •>: ; ; • •.',,

,'<; '.v» •„•.••; . " j I ' • / ( •;;'••;•;;•' '"'i,

;-; .v^jiiSiiiijijj; ;|,;, VJ{IVI j

^ S "i"."i••:•.' '.''''('"'I' V'1'1' ' * ! ' ' {• • • , " . i . . • ' . ' . . I ;!•,.• MB • ••• I

,,n. •") I1 .".'a,1.11 'J

-1000 0 1000 2000

vl. •lA'ril '/: :Xi1,: • k '•' i" j'ii'/ .! f}

•".v.'-V.'m.fite'iai'T.1.?;!!."., <

v"';T»i;t,"ri,f/a-1000 0 1000 2000

D

18 ,

17 ,

16 .

15 .

14

13 .

12 .

11 .

• •i'F;t-

• I • • • : •

," I : • • : • . . • •

•

, . J . . . . .••:=

• / ' ' t i dS^? :^ 1 -• ' / ' ' •

0 1000 2000 3000

msec

Figure 9. Example of a typical microelectrode penetration within the rostral bank of the IPS showing the different activity types (A-£) encountered by recording cell activity at

different cortical tangential locations within the sulcus. Rve replications of neuronal activity were collected for each movement direction (71-78). Rasters in A-D are aligned to

movement onset, in f to presentation of the IS. Sections at the lower left display the microelectrode tracks of the same penetration as seen in the histological material.




ownloaded from


1 r

-22000 4000 6000 8000 10000

B

-20 2000 4000 6000 8000 10000

dorsoventral position relative to top of activity (urn)

Figure 10. Plots of directional activity (measure 0) ratios as functions of dorsoventralposition. A, Plot of the logarithm of the ratio of signal- to movement-related activity asa function of cell position relative to the top of neuronal activity. Positive values indicatea high signal-related activity relative to movement-related activity. Negative values in-dicate relatively more movement-related activity. Solid line is a linear regression withslope significantly different from zero (p < 0.0005, regression Ftest r = 0.197). B, Plotof the logarithm of the ratio of set- to movement-related activity. Conventions as in ARegression slope is significantly different from zero (p < 0.0002, regression F test; r =0.212).

between them, and to the possible parietofrontal mechanismsinvolved in the composition of motor commands for reaching.In addition, combined with other results available in the lit-erature, they assign to the superior parietal lobule a novel roleas intermediate link in the corticocortical connections for vi-sual reaching.

Differentiation of Motor Cortical AreasThe results of the present study are in agreement with pre-vious studies in indicating different amounts of signal-, set-,and movement-related activities in MI and PMd. Our data,however, do not support the existence of a sharp functionalboundary between MI and PMd under the behavioral condi-tions imposed. While some studies have reported set-relatedactivity primarily confined to PMd (Weinrich and Wise, 1982;Weinrich et al., 1984), we have found wide distributions ofset-related activity that extend caudally into MI. The existenceof set-related activity in MI has been reported previously (Tan-ji and Evarts, 1976; Lecas et al., 1986; Georgopoulos et al.,1989; Riehle and Requin, 1989; Alexander and Crutcher, 1990).In addition, movement-related activity was found to extendrostrally into PMd, overlapping the distributions of set- andsignal-related activity.

There has been a long-running debate over the existenceand definitions of the "premotor" cortex (for reviews, seeWise, 1984,1985). Considerable recent evidence has been ac-

8 80 -

° 6 0O)

2 4020g.

signal

2000 4000 6000 8000 10000

set

8

1

80

60

40

20

0

-

-

•

8 80

1 60O)

| 40

| 20

* o

8 80

1 60O)

| 40

| 20

2000 4000 6000 8000 10000

movement

2000 4000 6000 8000 10000

position

0 2000 4000 6000 8000 10000

dorsoventral position relative to top of activity (urn)

Figure 11. Percentages of parietal neurons with significant directional modulation (p <0.01, AN OVA) as a function of dorsoventral location. Percentages were calculated for2 mm wide dorsoventral bins along each penetration relative to the top of neural activity.A, Signal-related directional modulation. 8, Set-related directional modulation. C, Move-ment-related directional modulation. D, Position-related directional modulation.

cumulated to support the existence of a premotor cortex and,in fact, the existence of multiple premotor cortices (see Heet al., 1993). Although our data do not indicate a clear-cutfunctional border between MI and PMd, they confirm the ex-istence of anatomical and physiological differences betweenthe rostralmost and caudalmost portions of the frontal lobeproximal arm representation investigated.

Physiological Trends in the Frontal LobeTo measure the degree of relationship between neuronal ac-tivity and task performance we quantified the incidence ofsignificant directional modulation for each of the task epochs.Our results show the same general trends observed in previ-ous studies (Weinrich and Wise, 1982; Weinrich et al., 1984).The data of the present study, however, indicate more gradualchanges in the incidences of the signal- and set-related mod-ulation than was reported in those studies. Our data agreeswith studies that have shown a consistent amount of delaytime modulation in MI (Tanji and Evarts, 1976; Lecas et al.,1986; Georgopoulos et al., 1989).

114 Intrinsic Organization of Frontal and Parietal Cortices*Johnson et al.



ownloaded from


" 1.0

0.8<a•5

o•5 0.6i••e8. 0.4

sa.

0.20.0

B

frontalparietal

1.0

0.8

•5 0-6c

"I 0.42Q.

0.2

0.0>

100 200 300 400 500

neuronal onset relative to IS (ms)

frontal

parietal

100 200 300 400

neuronal onset relative to GS (ms)

500

-400 -300 -200 -100 0 100 200

neuronal onset relative to movement (ms)

300 400

Hgure 12. Comparison of distributions of frontal and parietal neuronal onset times as determined by the cusum technique. A, Cumulative distributions of onset times with respectto the presentation of the instruction stimulus. Filled squares indicate frontal cell distribution, open squares indicate parietal cell distribution. Distributions were not significantlydifferent (p = 0.8266, Kolmogorov-Smirnov test). B, Cumulative distributions of onset times with respect to go stimulus presentation. Conventions as in A Distributions were different(p - 0.001, Kolmogorov-Smirnov test). C, Cumulative distributions of onset times with respect to movement onset Negative times indicate neuronal onsets occurring before movementonset Conventions as in A Distributions were different (p < 0.0001, Kolmogorov-Smirnov test).

The physiological data of the present study confirm theexistence of a gradual functional trend within the proximalarm representation of the dorsolateral frontal lobe. Thesetrends were not obvious in terms of the absolute firing fre-quency of neurons. However, they were evident in relativemeasures of directional activity and in terms of the incidenceof significant modulation corresponding to behavioral param-eters of the task (directional ANOVA).

Anatomofunctional Trends in the SPLThe gradient of connectivity from sources in the SPL suggeststhat these anatomical pathways may be anatomical substratesof the observed functional trends in MI and PMd. If this is thecase, parallel functional trends could be expected in theseother cortical areas. So far, physiological studies had providedonly limited data on the distribution of functional propertieswithin the regions of the SPL labeled in this study. The resultsof the present study show that functional trends in the dis-tribution of activity types exist also in the parietal lobe. Inthe cortex of the medial bank of the intraparietal sulcus,while neurons displaying significant movement- and position-related activity are distributed throughout the tangential ex-tent of area 5 and MIP, neurons rich in signal and set-relatedactivity predominates in MIP. We should stress that our studiesshows that MIP projects preferentially to the more rostral por-

tions of the MI and PMd arm region, those frontal regions thatare rich in signal- and set-related neurons. Thus, -within thecortical distributed system controlling reaching, regions of dif-ferent cortical areas displaying the same functional propertiestend to be preferentially connected through association cor-ticocortical fibers.

We know of no physiological studies of MDP or 7m at-tempting to characterize these areas with respect to visuomo-tor control. Galletti et al. (1991, 1993) recorded visual andoculomotor activity in a region that they considered partiallyoverlapping with area MDP. Within area MIP, a gradient ofphysiological properties has been reported. Colby and Du-hamel (1991) found that, as they moved their recording elec-trode down through the medial wall of the posterior portionof the IPS, they successively passed through regions with dif-fering activity types. Dorsally, neurons respond to passive so-matosensory stimulation. Deep within the sulcus, neuronshave purely visual driving. At intermediate depths they foundcells that were related to active reaching movements and "bi-modal" cells that were driven by combinations of passive so-matosensory stimuli, visual stimuli, and active reaching.

The SPL as a Source of Visually Derived InformationAnatomical tracing studies have indicated visual area PO (Cov-ey et al., 1982; Gattass et al., 1985; Colby et al., 1988) to be a




ownloaded from


Figure 13. Diagram of connections of thecortical visual areas modified from Colbyand Ouhamel (1991). Bold lines havebeen added to indicate the connectionsrevealed in the present study. Note thatwhile areas MIP and 7m receive projec-tions from visual area PO, their involve-ment in the traditional visual informationprocessing streams is minimal. AreasMDP and 7m are indicated here as aunique area, 7m.

potential source of the visual input to the SPL. Area PO liesdeep in the medial aspect of the hemispheres in the anteriorwall of POs and has been shown to receive direct projectionsfrom a number of visual areas including VI, V2, V3, V4, andthe medial temporal visual area (area MT; Colby et al., 1988).Injections of retrograde tracers in the physiologically identi-fied visual area PO labeled neurons in MIP and MDP (Colbyet al., 1988),indicating a connection between these areas.Thereciprocal pathway was shown by experiments in which aretrograde tracer was injected in MIP (Blatt et al., 1990).These authors found labeled neurons in areas MDP and PO.Injections in area 7m, just rostral to MDP on the medial sur-face of the hemisphere, labeled a reciprocal pathway to andfrom area PO (Cavada and Goldman-Rakic, 1989a). These stud-ies indicate that area PO is capable of supplying MIP, 7m, andprobably MDP with visual information.

In area PO, the organization of the visual field representa-tion (Gattass et al., 1985) and the functional properties ofneurons (Galletti et al., 1991, 1993; Fattori et al., 1992) seemadequate to build an internal representation of target locationin space. Area PO is the only known visual area in the ma-caque that does not possess the foveal magnification charac-teristic of VI and other visual areas (Covey et al., 1982; Colbyet al., 1988; Galletti et al., 1991). This relative emphasis on thevisual periphery suggests a role of this area in visuospatialprocessing (Colby et al., 1988). Other properties of PO neu-ronal activity are consistent with this interpretation (Gallettiet al., 1991, 1993; Fattori et al., 1992). In addition, PO is con-nected with posterior parietal area LIP, which is thought tobe involved in the analysis of visual space (Andersen et al.,1990; Blatt et al., 1990; Barash et al., 1991).

By virtue of their afferent connections from visual area POand of their projections to the frontal lobe arm-relatedregions, area MIP, MDP, and 7m emerge as potential interme-diate link in the corticocortical pathways for visual reaching(Fig- 13).

Parietofrontal Network for Reaching to Visual TargetsFor the generation of motor commands appropriate for car-rying the hand to a visually located target, information con-cerning both the target location and the current configurationof the arm must be combined. Psychophysical studies havesuggested that these types of information are linked within a

reference frame that is centered on the shoulder joint(Soechting and Flanders, 1989a,b). Data from neurophysiolog-ical experiments have shown that such a combination of in-puts can be detected in the activities of populations of neu-rons within MI (Caminiti et al., 1990a), PMd (Caminiti et al.,1991), and posterior parietal cortex (Ferraina and Bianchi,1994).

The results of the present study suggest that these differentsources of information can be combined within the parietaland frontal network. The frontal lobe motor centers areshown by the data of the present study to receive projectionsfrom parietal areas currently thought to be involved in theprocessing of either somatosensory or visual information, ora combination of both. The existence of SPL neurons withsuch response properties (Seal and Commenges, 1985; Colbyand Duhamel, 1991) indicates that visual and somatosensoryinputs are already combined at the parietal level. Projectionsto MIP from the somatosensory areas in the exposed dorsalportion of area 5 and in area 2 (Pandya and Seltzer, 1982; Ponsand Kaas, 1986) and from the more visually related areas MDPand PO (Pandya and Seltzer, 1982; Cavada and Goldman-Rakic,1989a; Blatt et al., 1990) provide plausible anatomical sub-strates for this combination of information within the SPL.However, the data of the present study do not support thehypothesis that the parietal lobe is the sole region involvedin the combination of these types of information. The distri-bution of functional properties in the frontal regions and thepattern of corticocortical connectivity suggest that informa-tion from throughout the visual to somatosensory continuumis transmitted from the parietal to frontal lobes for the com-position of motor commands.

The temporal properties of cell activity relative to the pre-sentation of the instruction-stimulus are very similar in bothfrontal and parietal populations, displaying virtually identicalrecruitment curves. This suggests that the recruitment is gen-erated by mechanisms operating in parallel in both corticalregions, mechanisms based on local excitatory connectionsamong assemblies of neurons within each population and bycorticocortical connections between the two populations.These last connections, if reciprocal, play a role similar to thatof local connections within die parietal and frontal popula-tions, which can therefore be considered as a unitary distrib-

116 Intrinsic Organization of Frontal and Parietal Cortices -Johnson et al.



ownloaded from


uted assembly. The onset time of cell activity with respect tothe go signal has similar shape in both parietal and frontalcortex, but the rise time is significantly faster in the frontalthan in the parietal cortex, suggesting that the recruitmenttime of the parietal cells population depends on an efferentcopy of the motor command composed in the frontal cortex.

It is interesting to note that IS and GS are both visual stim-uli, with the crucial difference, however, that, within the be-havioral context of the task, the first one signals for targetlocation, the second one for movement generation. It is there-fore not surprising that the recruitment time of neuronal pop-ulations is similar in both cortices with respect to the eventsleading to target location, while it is faster in the frontal cor-tex -when the events concerning movement generation areconsidered. This suggest that the use of visually derived in-formation concerning target location for the generation of theappropriate motor command, in other words, the visual tomotor transformation, occurs in a parallel fashion in bothfrontal and parietal cortices, while the decision-making pro-cess for movement occurs first in the frontal cortex and theninfluences the parietal lobe, probably through corollary sig-nals. Finally, it is worth stressing that in the process leadingfrom target location to movement generation, very little timein spent in the relay of the visual information to the precen-tral motor fields, while most of the time is devoted to theintracortical processing within the frontal and parietal net-work.

Portions of this study have focused on the parietal to fron-tal projection. While the detailed organization of the frontalto superior parietal projection is not known, such a projec-tion is known to exist (Pandya and Kuypers, 1969; Jones etal., 1978;Caminiti et al., 1985; Johnson et al., 1989;Blatt et al.,1990). Communication along this pathway has been proposedas a means to relay and transform movement-related infor-mation (Caminiti et al., 1985; Johnson et al., 1989). The func-tional significance of these fronto-parietal projections, on the-oretical grounds, could be expected to be important for shap-ing the tuning of parietal neurons (Burnod et al., 1992a,b).The temporal lead of frontal over parietal cell activity withrespect to those events concerning movement generation isin accord with this view. It has been hypothesized (Mel, 1991)that the projections from posterior parietal cortex back toarea PO should result in the activity of PO neurons beingmodified by arm orientation. The significance of these "mo-tor" to "sensory" projections should not be underestimated.

The entire MI and PMd proximal arm representation hasrecently been shown to project to the spinal cord (He etal., 1993)- As such, the motor centers of the spinal cord andconvergence and divergence within the corticospinal projec-tion could represent additional layers of the network withinwhich visual, somatosensory, and motor information are com-bined.

NotesWe thank Dr. Aldo Rustioni for his support of and contributions tothis project and Dr. Yves Burnod for his ideas and comments. Thisresearch was supported by the Commission of the European Com-munity and the Ministry of University and Scientific and Technolog-ical Research of Italy.

Address correspondence to Paul B.Johnson, Division of Neurosur-gery, Box 3807, Duke University Medical Center, Durham, NC 27710.

Appendix

Abbreviations7m Medial area 7APA Arcuate premotor areaAS Arcuate sulcusCTB-Au Cholera toxin subunit B labeled with colloidal gold

CiS Cingulate sulcusCS Central sulcusDTI First 300 msec of instructed-delay timeDT2 Instructed-delay time, excluding first 300 msecGLM Green (fluorescein isothiocyanate>labeled micro-

spheresGS Go stimulusIDT Instructed-delay timeIPS Intraparietal sulcusIS Instruction stimulusMDP Medial dorsal parietal areaMI Primary motor cortexMIP Medial intraparietal areaPETH Perievent time histogramPMd Dorsal premotor cortexPO Parieto-occipital visual areaPOs Parieto-occipital sulcusRIM Red (rhodamine>labeled microspheresRT Reaction timeRTMT 400 msec time interval centered on movement on-

set (reaction time/movement time)SMA Supplementary motor areaspAS Spur of the arcuate sulcusSPL Superior parietal lobuleSPcS Superior precentral sulcusTHT Target holding timeWGA- Inactivated horseradish peroxidase conjugated toapoHRP- wheat germ agglutinin and labeled with colloidalAu gold

ReferencesAlexander GE, Crutcher MD (1990) Preparation for movement: neu-

ral representations of intended direction in three motor areas ofthe monkey. J Neurophysiol 64:133-150.

Andersen RA, Asanuma C, Essick G, Siegel RM (1990) Corticocorticalconnections of anatomically and physiologically denned subdivi-sions within the inferior parietal lobule. J Comp Neurol 296:65-113.

Barash S, BraceweU RM, Fogassi L, Gnadt JW, Andersen RA (1991)Saccade-related activity in the lateral intraparietal area. 2. Spatialproperties.J Neurophysiol 66:1109-1124.

Barbas H, Pandya DN (1987) Architecture and frontal cortical con-nections of the premotor cortex (area 6) in the rhesus monkey. JComp Neurol 256:211-228.

Basbaum AJ, Menetrey D (1987) Wheat germ agglutinin-apoHRPgold: a new retrograde tracer for light- and electron-microscopicsingle and double-label studies. J Comp Neurol 261:306-318.

Blatt GJ, Andersen RA, Stoner GR (1990) visual receptive field or-ganization and cortico-cortical connections of the lateral intrapar-ietal area (area LIP) in the macaque.J Comp Neurol 299:421-445.

Burbaud P, Doegle C, Gross C, Bioulac B (1991) A quantitative studyof neuronal discharge in areas 5, 2, and 4 of the monkey duringfast arm movements.J Neurophysiol 66:429-443.

Burnod Y, Grandguillaume P, Otto I, Johnson PB, Caminiti R (1992a)Reaching toward visual targets, n. Computational studies. In: Con-trol of arm movement in space (Caminiti R, Johnson PB, BurnodY, eds), pp 159-174. Berlin: Springer.

Burnod Y, Granguillaume P, Otto I, Ferraina S, Johnson PB, Caminiti R(1992b) Visuomotor transformations underlying arm movementstoward visual targets: a neural network model of cerebral corticaloperations.J Neurosci 12:1435-1453.

Caminiti R, Zeger S, Johnson PB, Urbano A, Georgopoulos AP (1985)Corticocortical efferent systems in the monkey: a quantitative spa-tial analysis of the tangential distribution of cells of origin. J CompNeurol 241:405-419.

Caminiti R, Johnson PB, Urbano A (1990a) Making arm movementswithin different parts of space: dynamic aspects in the primatemotor cortex.J Neurosci 10:2039-2058.

Caminiti R, Johnson PB, Burnod Y, Galli C, Ferraina S (1990b) Shiftof preferred directions of premotor cortical cells with arm move-ments performed across the workspace. Exp Brain Res 83:228-232.




ownloaded from


Caminiti R, Johnson PB, Galli C, Ferraina S, Burnod Y (1991) Makingarm movements within different parts of space: the premotor andmotor cortical representation of a coordinate system for reachingto visual targets. J Neurosci 11:1182-1197.

Cavada C, Goldman-Rakic PS (1989a) Posterior parietal cortex inrhesus monkey: I. Parcellation of areas based on distinctive limbicand sensory corticocortical connections.J Comp Neurol 287:393-421.

Cavada C, Goldman-Rakic PS (1989b) Posterior parietal cortex inrhesus: n. Evidence for segregated corticocortical networks link-ing sensory and limbic areas with the frontal lobe. J Comp Neurol287:422-445-

Chavis DA, Pandya DN (1976) Further observations on corticofrontalconnections in the rhesus monkey. Brain Res 117: 369-386.

Colby CL, Duhamel J-R (1991) Heterogeneity of extrastriate visualareas and multiple parietal areas in the macaque monkey. Neu-ropsychologia 29:517-537.

Colby CL, Gattass R, Olson CR, Gross CG (1988) Topographical or-ganization of cortical afferents to extrastriate visual area PO in themacaque: a dual tracer study. J Comp Neurol 269:392-413.

Covey E, Gattass R, Gross CG (1982) A new visual area in the parieto-occipital sulcus of the macaque. Soc Neurosci Abstr 8:681.

Crammond DJ, Kalaska JF (1989) Neuronal activity in primate pari-etal cortex area 5 varies with intended movement direction dur-ing an instructed-delay period. Exp Brain Res 76:458-462.

Davey NJ, FJlaway PH, Stein RB (1986) Statistical limits for detectingchange in the cumulative sum derivative of the peristimulus timehistogram.J Neurosci Methods 17:153-166.

Dum RP, Strick PL (1991) The origin of corticospinal projectionsfrom the premotor area in the frontal lobe. J Neurosci 11:667-689.

Ellaway PH, Gardner-Medwin AR, Pascoe JE (1983) Decision limitsfor significance of changes in the cumulative sum (cusum) of theperistimulus time histogram.J Physiol (Lond) 341:4-5.

Fattori P, Galletti C, Battaglini PP (1992) Parietal neurons encodingvisual space in a head-frame of reference. Boll Soc Ital Biol Sper68:663-670.

Ferraina S, Bianchi L (1994) Posterior parietal cortex: functionalproperties of neurons in area 5 during an instructed-delay reach-ing task within different parts of space. Exp Brain Res 99:175-178.

Galletti C, Battaglini PP, Fattori P (1991) Functional properties ofneurons in the anterior bank of the parieto-occipital sulcus of themacaque monkey. EurJ Neurosci 3:452-461.

Galletti C, Battaglini PP, Fattori P (1993) Parietal neurons encodingspatial locations in craniotopic coordinates. Exp Brain Res 96:221-229

Gattass R, Sousa APB, Covey E (1985) Cortical visual areas of themacaque: possible substrates for pattern recognition mechanisms.In: Pattern recognition mechanisms (Chagas C, Gattass R, Gross C,eds), pp 1-20. Rome: Pontificiae Acadamiae Scientarium ScriptaVaria.

Georgopoulos AP, Kalaska JF, Caminiti R, Massey JT (1982) On therelations between the direction of two-dimensional arm move-ments and cell discharge in primate motor cortex. J Neurosci 2:1527-1537.

Georgopoulos AP, Crutcher MD, Schwartz AB (1989) Cognitive spa-tial-motor processes. 3- Motor cortical prediction of movementdirection during an instructed delay period. Exp Brain Res 75:183-194.

Ghosh S, Brinkman C, Porter R (1987) A quantitative study of thedistribution of neurons projecting to the precentral motor cortexin the monkey (M. fascicularis).] Comp Neurol 259:424-444.

Godschalk M, Lemon RN, Kuypers HGJM, Ronday HK (1984) Corticalafferents and efferents of monkey postarcuate area: an anatomicaland electrophysiological study. Exp Brain Res 56:410-424.

Godschalk M, Lemon RN, Kuypers HGJM, van der Steen J (1985) Theinvolvement of monkey premotor cortex neurones in preparationof visually cued arm movements. Behav Brain Res 18:143-157.

He SQ, Dum RP, Strick PL (1993) Topographic organization of cor-ticospinal projections from the frontal lobe: motor areas on thelateral surface of the hemisphere. J Neurosci 13:952-980.

Johnson PB, Angelucci A, Ziparo R, Minciacchi D, Bentivoglio M, Cam-initi R (1989) Segregation and overlap of callosal and association

neurons in frontal and parietal cortices of primates. A spectraland coherency analysis.J Neurosci 9:2313-2326.

Johnson PB, Ferraina S, Caminiti R (1993) Cortical networks for vi-sual reaching. Exp Brain Res 97:361-365.

Jones EG, Coulter JD, Hendry SHC (1978) Intracortical connectivityof architectonic fields in the somatic sensory, motor and parietalcortex of monkeys. J Comp Neurol 181:291-348.

Kalaska JF (1988) The representation of arm movements in post-central and parietal cortex. Can J Physiol Pharmacol 66:455-463.

Kalaska JF, Caminiti R, Georgopoulos AP (1983) Cortical mechanismsrelated to the direction of two-dimensional arm movements: re-lations in parietal area 5 and comparison with motor cortex. ExpBrain Res 51:247-260.

Kalaska JF, Cohen DAD, Prud'homme M, Hyde ML (1990) Parietalarea 5 neuronal activity encodes movement kinematics, not move-ment dynamics. Exp Brain Res 80:351-364.

Kunzle H (1978a) Cortico-cortical efferents of primary motor andsomatosensory regions of cerebral cortex in Macaca fascicularis.Neuroscience 3:25-39.

Kunzle H (1978b) An autoradiographic analysis of the efferent con-nections from premotor and adjacent prefrontal regions (areas 6and 9) in Macaca fascicularis. Brain Behav Evol 15:185-234.

Kurata K (1991) Corticocortical inputs to the dorsal and ventralaspects of the premotor cortex of macaque monkeys. NeurosciRes 12:263-280.

Kurata K, Tanji J (1986) Premotor cortex neurons in macaques: ac-tivity before distal and proximal forelimb movements. J Neurosci6:403-411.

Kurata K, Hoffman DS (1994) Differential effects of muscimol mi-croinjection into dorsal and ventral aspects of the premotor cor-tex of monkeys.J Neurophysiol 71:1151-1164.

Lecas JC, Requin J, Anger C, Vitton N (1986) Changes in neuronalactivity of the monkey precentral cortex during preparation formovement.J Neurophysiol 56:1680-1702.

Leichnetz GR (1986) Afferent and efferent connections of the dor-solateral precentral gyrus (area 4, hand/arm region) in the ma-caque monkey, with comparisons to area 8. J Comp Neurol 254:460-492.

Matelli M, Carmarda R, Glickstein M, Rizzolatti G (1984) Intercon-nections within the postarcuate cortex (area 6) of the macaquemonkey. Brain Res 310:388-392.

Matsumura M, Kubota K (1979) Cortical projection to hand-arm mo-tor area from post-arcuate area in macaque monkeys: a histologi-cal study of retrograde transport of horseradish peroxidase. Neu-rosci Lett 11:241-246.

Matsuzaka Y, Aizawa H, Tanji J (1992) A motor area rostral to thesupplementary motor area (presupplementary motor area) in themonkey: neuronal activity during a learned motor task. J Neuro-physiol 68:653-662.

Mel BW (1991) A connectionist model may shed light on neuralmechanisms for visually guided reaching. J Cognit Neurosci 3:273-292.

Morecraft RJ.Van Hoesen GW (1992) Cingulate input to the primaryand supplementary motor cortices in the rhesus monkey: evi-dence for somatotopy in areas 24c and 23c. J Comp Neurol 322:471-489.

Mountcastle VB, Lynch JC, Georgopoulos AP, Salcata H, Acuna C(1975) Posterior parietal association cortex of the monkey: com-mand functions for operations within extrapersonal space. J Neu-rophysiol 38:871-908.

Muakkassa KF, Strick PL (1979) Frontal lobe inputs to primate motorcortex: evidence for four somatotopically organized 'premotor' ar-eas. Brain Res 177:176-182.

Pandya DN, Kuypers HGJM (1969) Cortico-cortical connections inthe rhesus monkey. Brain Res 13:13-36.

Pandya DN, Seltzer B (1982) Intrinsic connections and architectonicsof posterior parietal cortex in the rhesus monkey. J Comp Neurol204:196-210.

Pandya DN, Vignolo LA (1971) Intra- and interhemispheric projec-tions of the precentral, premotor and arcuate areas in the rhesusmonkey. Brain Res 26:217-233.

Passingham RE (1988) Premotor cortex and preparation for move-ment. Exp Brain Res 70:590-596.

Petrides M, Pandya DN (1984) Projections to the frontal cortex from

118 Intrinsic Organization of Frontal and Parietal Cortices 'Johnson et al.



ownloaded from


the posterior parietal region in the rhesus monkey. J Comp Neurol228:105-116.

Ports TP, Kaas JH (1986) Corticocortical connections of area 2 ofsomatosensory cortex in macaque monkeys: a correlative anatom-ical and physiological study. J Comp Neurol 248:313-335.

Press WH, Teukolsky SA, Vetterling WT, Flannery BP (1992) Numer-ical recipes in Fortran. Cambridge, MA: Cambridge UP.

Primrose DC, Strick PL (1985) The organization of interconnectionsbetween the premotor areas in the primate frontal lobe and thearm area of the primary motor cortex. Soc Neurosci Abstr 11:1274.

Rea G, Ebner TJ, Bloedel JR (1987) Evaluations of combined pre-motor and supplementary motor cortex lesions on a visually guid-ed arm movement. Brain Res 418:58-67.

Riehle A (1991) Visually induced signal-locked neuronal activitychanges in precentral motor areas of monkey: hierarchical pro-gression of signal processing. Brain Res 540:131-137.

Riehle A, Requin J (1989) Monkey primary motor and premotorcortex: single unit activity related to prior information about di-rection and extent of an intended movement. J Neurophysiol 61:534-549.

Sakata H, Takaoka Y, Kawarasaki A, Shibutani H (1973) Somatosen-sory properties of neurons in the superior parietal cortex (area5) of the rhesus monkey. Brain Res 64:85-102.

Sasaki K, Gemba H (1986) Effects of premotor cortex cooling uponvisually initiated hand movements in the monkey. Brain Res 374:278-286.

Schwartz AB, Kettner RE, Georgopoulos AP (1988) Primate motorcortex and free arm movements to visual targets in three-dimen-sional space. 1. Relations between single cell discharge and direc-tion of movement. J Neurosci 8:2913-2927.

Seal J, Commenges D (1985) A quantitative analysis of stimulus- andmovement-related responses in the posterior parietal cortex ofthe monkey. Exp Brain Res 58:144-153.

Sessle BJ, Wiesendanger M (1982) Structural and functional defini-tion of the motor cortex in the monkey (Macacafascicularts).}Physiol (Lond) 323:245-265.

Soechting JF, Flanders M (1989a) Sensorimotor representations forpointing to targets in three-dimensional space. J Neurophysiol 62:582-594.

Soechting JF, Flanders M (1989b) Errors in pointing are due to ap-proximations in sensorimotor transformations.J Neurophysiol 62:595-608.

Strick PL, Kim R (1978) Input to primate motor cortex from poste-rior parietal cortex (area 5). I. Demonstration by retrograde trans-port. Brain Res 157:325-330.

Tanji J, Evarts EV (1976) Anticipatory activity of motor cortex neu-rons in relation to direction of an intended movement. J Neuro-physiol 1062-1068.

Tokuno H, Tanji J (1993) Input organization of distal and proximalforelimb areas in the monkey primary motor cortex: a retrogradedouble labeling study. J Comp Neurol 333:199-209.

Weinrich M, Wise SP (1982) The premotor cortex of the monkey. JNeurosci 2:1329-1345.

Weinrich M, Wise SP, Mauritz KH (1984) A neurophysiological studyof the premotor cortex in the rhesus monkey. Brain 107:385-414.

Wise SP (1984) The nonprimary motor cortex and its role in thecerebral control of movement. In: Dynamic aspects of neocorticalfunction (Edelman GM, Gall WE, Cowan WM, eds), pp 525-556.New York: Wiley.

Wise SP (1985) The primate premotor cortex: past, present, and pre-paratory. Annu Rev Neurosci 8:1-19.

Zar JH (1984) Biostatistical analysis. Englewood Cliffs, NJ: Prentice-Hall.




ownloaded from


cortical networks for visual reaching: paul b. johnson,1 ...€¦ · a result of this study, area...

Documents