Exp Brain Res (1987) 67:460-468 Expedrn tal Brain Research �9 Springer-Vedag 1987

Prefrontal unit activity during a delayed oculomotor task in the monkey

J.P. Joseph and P. Barone

Laboratoire de Neuropsychologie Exp6rimentale, INSERM Unit6 94, 16, Avenue du Doyen L6pine, 1=-69500 Bron, France

Summary. 1. Unit activity was recorded in the prefrontal cortex of Rhesus monkeys during per- formance of a delay task with two motor compo- nents, an ocular saccade and an arm movement, following a complex pattern of auditory and visual stimuli. A special feature of the paradigm was that onsets of the different sensory stimuli, orienting saccade and arm movement were dissociated in time at predetermined time intervals. 216 task-related units were recorded. Discussion of the data focuses on two groups of cells the activity of which is modified by the saccade: the signal-related pre- saccadic tonic cells and the post-saccadic tonic cells. 2. Activity of signal-related pre-saccadic tonic cells was initiated with the onset of peripheral stimuli, visual and/or auditory, and terminated with an orien- ting saccade. Spatial selectivity was a feature of most units. They seemed to encode the region of space cued by the stimulus. 3. Many units were visually responsive. Activation of these depended both upon retinal locus of the stimulus and the requirement they imposed on subsequent behavior. Termination of their activity demanded foveation of some visual targets, not necessarily the one which had initiated their response. 4. The majority of these signal- related pre-saccadic tonic cells responded to onset of auditory stimuli. The termination of tonic activity in these cells with foveation of the region in space from which the signal originated suggests a spatial memory process that is no longer used when the visual fixation response can signal equivalent spatial information. 5. Post-saccadic tonic cells were activated as the animals fixated a target; spatial selectivity was a feature of most of these. We suggest that this group of cells are part of the system which encodes the location of a target only if it is the goal of a visually guided movement. 6. A substantial proportion (20-30%) of

Offprint requests to: J.P. Joseph (address see above)

signal-related tonic cells and post-saccadic tonic cells were activated differentially by use of one arm versus the other. These cells appeared to be more directly involved in motor preparation.

Key words: Delayed task - Prefrontal cortex - Unit - Monkey

Introduction

The earliest and best documented finding on the effects of lesions in dorsolateral prefrontal cortex is an impairment in delayed response and delayed alternation tasks (Jacobson 1935; Jacobson and Nis- sen 1937). The deficits in spatial tasks were inter- preted as disorders of spatial or kinesthetic short- term memory for, and/or deficiencies in the integra- tion of, temporally discontinuous spatial events. See Fuster (1980, 1981, 1985) for reviews.

The neural correlates of these behaviors have been examined through unit recordings in awake monkeys during performance of delayed spatial and non spatial tasks. Activity changes have been associ- ated with specific aspects of the tasks such as, attention (Fuster 1973; Kubota et al. 1980), mnemonic function (Kubota et al. 1974; Niki 1974) and movement initiation (Sakai 1974; Kubota and Funahashi 1982). A common finding in these studies was the presence of units whose activity was sus- tained during the interval either between the onset of two sensory events, or between a sensory and a motor event. It has been suggested that this sustained activity reflects the temporal contingencies, estab- lished during learning, between the sensory and motor events (Fuster 1985) which allow the integra- tion of temporally discontinuous events into goal- directed activities.

In spa t i a l t a sks o r i e n t i n g s a c c a d e s a n d s u b s e q u e n t

f ixa t ions a r e h igh ly c o r r e l a t e d w i t h t h e t i m i n g o f

b e h a v i o r . N o c o r r e l a t i o n has b e e n o b s e r v e d in t h e

p r e f r o n t a l c o r t e x b e t w e e n s p o n t a n e o u s e y e m o v e -

m e n t s a n d cel l d i s c h a r g e d u r i n g t h e i n t e r t r i a l o r t h e

de l ay p e r i o d o f a d e l a y t a s k ( F u s t e r 1973) o r d u r i n g a

v i s u o m o t o r t a sk w h i c h d o e s n o t i n c o r p o r a t e a d e l a y

(Saka i 1974). N e v e r t h e l e s s w e p r o p o s e t h a t t h e

c o n t r i b u t i o n o f p r e f r o n t a l c o r t e x to e n c o d i n g a n d

d e c o d i n g o f spa t i a l p a r a m e t e r s m a y b e c l a r i f i ed if,

i n s t ead o f s p o n t a n e o u s e y e m o v e m e n t s , t e m p o r a l l y

p r o g r a m m e d e y e m o v e m e n t s a r e s t ud i ed .

Methods

Two rhesus monkeys served as subjects in this experiment. They were seated in a primate chair with head restrained and faced a panel located at arm's reach. The animals were trained in two behavioral tasks. The basic paradigm was a fixation task. In each trial the touch of a bar, located medially in the lower half of the panel, illuminated a central light-key on the panel at a constant level for 1-3 s, with the level then reduced for about 1 s. If the monkeys released the bar and depressed the light-key during the dim period they received a drop of apple juice. If they either released the bar or depressed the key outside of the period they were neither punished nor rewarded. A second paradigm was added to induce the monkeys to make delayed saccades and arm movements towards either of two light-keys. These were located at 20 ~ eccentricity to the right or to the left of the central light-keys. In this paradigm all sensory events delivered to the monkeys, as well as their motor responses were dissociated in time. While the monkeys fixated the central light-key (part 1, Fig. 1), a short auditory stimulus (clicks at 20 pulses per second during one second) was delivered through one of two loudspeakers placed directly above the peripheral light-keys (part 2, Fig. 1). This stimulus cued the monkeys to the required direction of its next eye and arm movements. Then, while the monkeys fixated the central point, the peripheral light-keys were illuminated, one after the other (part 3 and 4, Fig. 1). Order of illumination was randomized with a fixed time interval of 1 s between onset of the two lights. They remained illuminated until the end of the trial. When illumination of the central light-key was terminated, the monkeys were allowed to break fixation and make an orienting saccade towards that peripheral light-key which had been cued by the auditory stimulus (part 5, Fig. 1). They had to release the bar and depress it as soon as the peripheral key dimmed (part 6, Fig. 1). They were then rewarded. The trial was terminated if the monkeys made eye movements outside a 10 ~ x 10 ~ window centered on the fixation point, before it was extinguished. An orienting saccade directed toward the wrong side, or release of the bar before the dimming also terminated the trial. Successive trials could be separated by an interval of several seconds by inactivation of the bar. During recording periods, a minimum of 5 trials for each direction of response were grouped together in blocks.

Eye movements were monitored with the aid of Ag/AgC1 EOG electrodes. Horizontal deviations of gaze, and the sequence of sensory and motor events, were displayed on a 608 Tektronix Monitor and on a paper chart. This made it possible 1. to adjust the gain of the horizontal EOG by comparing the amplitudes of the recorded saccades between known positions to their actual distance on the Screen of the monitor and 2. to correct amplifier drift within a session.

461

1 2 3

4 5 6

Fig. 1. Behavioral paradigm. Parts 1 to 6 show the sequence of signal presentation and motor activity. See text for description

The animals were trained using the correction procedure for about 300-400 reinforced trials every day until they reached a criterion of better than 95% correct responses. It took 4-5 months for the animals to master the task. During training and recording periods, the animals were deprived and received their fluid ration in the testing apparatus. Monkey-pellets were available ad libitum in their home cage.

On completion of the training, the animals were surgically prepared under sodium pentobarbital anesthesia. Stainless steel cylinders (18 mm in diameter) were implanted on the skull. The center of the cylinders were aimed at A = 30 mm and L = 10 mm in the stereotaxic coordinate system. During the recording ses- sions, the animal's heads were held rigidly. Activity of single cells was recorded in both hemispheres through movable tungsten microelectrodes plated with platinum black and driven by a Trent Wells motorized microdrive. Electrode penetrations were made on a millimeter grid of points usually separated by 1 mm but when a unit showed an activity related to the task, subsequent electrode penetrations were made in the immediate vicinity of the penetra- tion where the unit was found.

Unit action potentials were monitored on an oscilloscope, converted into square-wave pulses and displayed on the paper chart along with the task events. Changes in unit activity to a certain task event were assessed by visual inspection of the paper chart, Those units which showed clear changes in discharge rate in relation to one or more changes of the task events were selected for graphic representation (raster display).

At the end of the experiments lesions were made by passing anodal current (20 ~tA for 1 ran) through a microelectrode. After receiving an overdose of nembutal the monkeys were perfused with isotonic saline followed by 4% formaldehyde saline. Pins were inserted at known electrode coordinates. The brains were removed from the skulls, photographed and postfixed for 1 week. They were then sectioned on a freezing microtome at 40 ~m and stained with cresyl violet. Sections were observed to determine the areas of penetration.

Results

T w o h u n d r e d s ix ten (216) s ing le uni t s in t h e d o r s o l a t - e ra l p r e f r o n t a l c o r t e x r e l a t e d to t h e t a sk w e r e s t u d i e d

( T a b l e 1). T h e s e w e r e c lass i f i ed o n t h e basis o f t h e i r

ac t iv i ty in r e l a t i o n to o n s e t o f t h e d i f f e r e n t s e n s o r y

s ignals , s accad ic m o v e m e n t s a n d t a r g e t acqu i s i t i on .

N

Others

%

Presaccadic signal-related tonic �9 auditory 24 11 �9 visual 14 6 �9 auditory and visual 3 1

Postsaccadic tonic 77 35 Signal-related phasic

.. .::~: z*xx..

�9 auditory 7 3 �9 visual 21 10

70 33

216 99

462

Table 1. Number and classification of cells

Fig, 2. Outline drawing of the fronto-latral part of the brain of a macaque monkey. The shaded area encloses the portion of prefrontal cortex (area 9) which was explored in the present study

Base line was taken as the mean neural firing frequency between trials. Signal-related phasic units were those which responded with a transient dis- charge synchronized with, and following, either the auditory signal or one of the visual signals. Units were classified as signal-related presaccadic tonic cells if they showed a sustained change in activity between the onset of the peripheral signals (auditory and/or visual) and the orienting saccade. Movement related units were defined as those whose activity changed in synchrony with onset of eye or arm movement. Post-saccadic tonic cells were defined as units becoming active when the animals turned their gaze upon a peripheral target light which they attended in order to detect the dimming.

Signal-related presaccadic tonic cells

Forty one cells exhibited signal-related tonic activity. Spatial selectivity was encountered in most of these.

Twenty four neurons were specific to auditory signals. Sixteen were activated only by the contralat- eral auditory cue (Fig. 31) and four by the ipsilateral cue. Four cells were activated by both left and right

A o--[~ I

. . . | . �9 . �9 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . "2" - ' ' . . . , . .

t[ . . . . . . . . . . . . . . . . . , . , J . " .

I . . .

. . . . . . . . . . ~: . . . . 7 . . ;

[ ~ " O

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . �9 . �9 l . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . �9 . . . . . . . . t . . . . .

II A

. . . . . . . . . . . . . . . . . . . . . . . . . . . t " J ' "

. . . . . . . . . . . . . . . . . . . . . . . . . . y . . . l . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I t . . . I . . . .

B

[ ~ . O

i 2 . ...................... " i " . . . . . . . . . . . . . . . . . . . . . . . . . . . . v . . .

�9 " " " "

. . . . . . L.

t

0 : 7 . . . . . . . . . _ ~ . . . . . . . . . .

Fig. 3I, II. Response of two signal-related tonic cells, specific to auditory signals with (I) and without (II) spatial selectivity�9 Duration of the central fixation is indicated by the horizontal line at the top of the figure. The wavy line above indicates onset and duration of the auditory stimulus and the two vertical arrows pointing downwards onset of illumination of the right and the left light-keys. The large vertical arrow indicates onset of the dim period of the peripheral light-key cued by the auditory stimulus. Each line of the rasters represents unit activity during a trial and is centered on the onset of the fixation point. The dots represent the occurence of action potentials�9 One action potential out of two is represented. On each line, the arrowhead pointing downwards indicate onset of the orienting saccade, and the small vertical arrows, time of occurence of key pressing. On top of each raster, a little inset represents schematically the panel with the fixation point (central dot) and the two peripheral light-keys (large circle). The square indicates the location of the auditory cue and the arrow the direction of the orienting saccade. All subsequent figures use the same conventions. In IA, the auditory cue was delivered on the right and in IB, on the left side�9 In IIA, the tone was delivered on the right and, in I l l ] , on the left side. I n l i e , n o peripheral stimulus was delivered. The animal had to press the central target. The blocks of trials A-C are consecutive. Order of illumination of the peripheral lights is not indicated since it did not modify the firing of these two cells

auditory stimuli (Fig. 3II). We often observed activa- tion of cells with spatial selectivity (Figs. 3I, 4 and 5) as well as of cells without spatial selectivity (Fig. 3II), slightly anticipating onset of the auditory stimuli. In

,::::::~.::~

A ~ - - o

. . . . . . . . i . . . . . . . . . . . . . . . . . . . . . . a . . . . + . . . . . . . . . . . . . . ". ' :" " . ' " . "." . " : : ' : ' . ' . "2 : " : : ' 7 " : .77 : .7 . . - : : " " , ' " ' . " 2 , " "" . " * '~ : ' . . " : " . : -

B i ~ - . o

' : : 1 " :::':'J . ": .-.2--F"::."':-:':_s : ; i . i ' -" . 17.'-" . : - . ' . 2 " : - . ' 7 . ] . ..7 . . : . : : _ . . . " - " - . : . . . . 2"2"_.7"::2 . . . . . . . ~ _ "" '.2.. . . . . .

'" 7 " : " : " . . ' . : . 2::_.-7_-" 2 .2 . ' .22 -2 : : i -2 2..::2_7..'~:,.. " : 2: " ~ 2 : . ' . " . . . "

C o . - ~

Fig. 4A-C. Modification of response of a signal-related tonic cell by visual signals and by saccades. In A, firing rate activated by the tones located on the left was reset to intertrial level with onset of a left visual signal�9 In B, the left visual signal was not illuminated. On the first four lines of this raster, arrowheads pointing down- wards and upwards indicate orienting saccade towards respectively the left-handed and the right-handed light-key�9 During the first four trials, the animal oriented first toward the left-handed key, then the right-handed and then again towards the left-handed key. In the last three trials, he oriented directly and only towards the left-handed key. In C, the auditory stimulus was located on the right side�9 Conventions are the same as in Fig. 3. R and L on top of each raster indicate respectively the right and the left light-keys

o . ~ f &L &R

A . . . . . . s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . : 2:.,.. , ; : . . . . . . . . . ::"'d.. . . . . . . . . . , .

CR ,9-

. . . . . . . . . . . . . . . . . . . . . . . . . . z . . . . . . . . . . * . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Y . . r . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . �9 . . . . . . I . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .',t . . . . . ~ . . . . . . . . .

j t . . . . . . . . . . . . . . . . . . . . . . . . . . . . . :::-r . . . . . . . . . . . . . . . . . . . . . . . . . .

i t . ~n

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . v.,. . . . . . . r . . . . . . . . . .

l s

Fig. 5A, B. Response of a signal-related tonic cell responding to auditory and visual signals located on the right. In A, activation of the cell lasts up to foveation of the stimulus on the right. In B, activation lasts up to the end of the trial if the orienting saccade is directed to the le f t Conventions are the same as in F ig 3

Fig. 3II, for instance, after a series of trials in which the auditory stimuli were delivered in turn to the right and to the left side, the animal was required to press the central light-key. In this case onset of the central light was not followed by onset of a peripheral auditory stimulus However , during the first trial directed toward the central light-key, the cell

463

A

~ . o

tt iR

B

# IL - . . . I - - . . . . . * . . . . . . . . . .

�9 +

. . . . . . . . . . . . . . . v _ . , i . . . . . . . . . . . . . .

o . - [ ] o cR

. . . . . . . . . . . . . . . . . ~ , . . r . . . . . . .

�9 .. : ' . . . . : . . . . '".:"::..--.."7":27.:'_--"~'., "'" "." "': . ,' . . . . . . .

r . . . . . . . . .

I s

Fig. 6A, B. Response of a signal-related tonic cell, specific to visual signals. The cell is activated by visual signals located on the left side whatever the direction of the saccade to the left (A) or to the right (B). Conventions are the same as in Fig. 3

A . .

~,WW, A~A' , ,W

r~--. o t iR p

I-. ................. "

. . . . . . . . v , 1

. . . . . . . . . . . . . . . . . ~ ~ 1 7 6 I .

. . . . . . . . . . . . T . & .

�9 I

B..,.:...:::I:.. " ""':" ":" :': "

~L | R . . . . . . . . . . . . . . �9 . . . . & ,

. . . . . . . . . . ~ t . . . . . . . . v , . ~ I . . .

I . . . . . .

o . - - [ ]

~R ~L

i t 1 . . . . . . �9 ~ �9 �9 . ~ I . , �9

v I . . . . o _ .

1 s

~3--. o IL le f

A . ~ . . . . . . . . . . . . . . . . . ~ " a . . . . . . . & . . . . . + ,

. . . . . . . . . . . . . . . v . . &. o . . . . . . . . . . . . . . . . . . . . . . . . . a . . . . . . 1

~e ~ t

. . . . . . . . . . . . . . v . . J . . ,

. . . . . . . . �9 . . . . I

. . . . . . . . . . . �9 -~ �9 o .

. . . . . . . . . . . . . . . . . . . . . . . . W . m . . . ,&

O " - - I ~ t

IR IL

t . . . . . . . . . . . . . . ; ..... . . . . . . . . . . . . . . . . �9 I

�9 �9 . . . . . . I . . . . . . . . . . . �9 l l

. . . . . . . . . . . v ~ . .

I s

Fig. 7I, II. Modification of response of two signal-related tonic cell (I, II) specific to visual signals, by direction of saccades. In I, the cell is activated by visual signals located on the left if the forthcoming saccade is directed to the left (A). If it is directed to the right (B), there is no activation. In II, the cell is activated by visual signals located on the right if the saccade is directed to the left (A). If it is directed to the right (B), there is no activation. Conventions are the same as in Fig. 3

464

& � 9 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . r . . . . . �9

. . . . . . . . . . . . . . . . . . . . . . . . . . . ~

. . . . . . . . . . . . . . . . . . . . . . . . . . . . � 9 , , 4

II R R

I ~ 1 - - O

B l . . .'.. :. .I," �9 , � 9 . I . �9 �9 . ~ .

. . . . . �9 t .

. 1 1

o .-I~ ! . . . . . , . . . . .

RL " / . . . . . " 1 ~ .

. . . . . . . . i I . . . . . . . . . . , . �9 1 �9

- L R

o '--fN ~ J

, , �9 . . . . t

. . . . . . . . . "�9149 .-.-;~ . . " �9 �9 o 1 . .

[ ~ - . O

LL " ' . " . . . . . . . . . " � 9 - . . . . . �9 . . . . . t . . . . . .

. . . . . . �9 . . . . d . . . .

. , a ,

. . . . . �9 . . . . . . . h . .

1 II

Fig. 8I, II. In I, modification of response of a signal-related tonic cell specific to auditory signals by motor set. In IA, the animal prepared the key pressing with the left arm and in B, with the right arm. In II, modification of spatial selectivity of a post-saccadic tonic cell by motor set. Activity pattern of the cell when the animal uses first his right arm towards the right-handed (RR) and the left- handed (RL) light-key, and then his left arm towards the right- handed (LR) and left-handed (LL) light-key. Conventions are the same as in Fig. 3

responded as if a peripheral auditory stimulus had been delivered�9

In 5 cells, tonic activity elicited by the onset of the auditory signal on one side was reduced to resting level by onset of the visual stimulus on the same side (Fig. 4A), but most signal-related tonic cells returned to the resting level of discharge only when the animal foveated either of the two light-keys. Reset of the cells to their resting level of discharge did not occur once and for all after the orienting saccade if, under special experimental conditions, the animal made a saccade away from the light-key that he had foveated first�9 Figure 4B illustrates the effect

of a first saccade towards the side cued by the auditory stimulus when the corresponding light-key had not been illuminated before the dimming period�9 After a few hundred milliseconds the animal made a second saccade towards the illuminated key on the opposite side. Firing rate of the cell returned to its resting level after the first saccade but regained a sustained level of discharge with the second saccade, away from the side cued by the auditory stimulus�9 In the same way, in cells responding to auditory and visual signals, activation persisted to the end of the trials if the encoded stimulus was not foveated.

Three neurons were activated by auditory and visual signals located on the same side. An example is shown on Fig. 5 in which the neuron responded to stimuli located on the right side. When the auditory stimulus was delivered on the right, activity was sustained until onset of the rightward orienting saccade. When the auditory stimulus was delivered on the left - and the orienting saccade was also to the left - the right visual stimulus elicited sustained activity which persisted to the end of the trial�9

Fourteen cells were specific to visual signals�9 Thirteen were activated by contralateral and one by ipsilateral signals�9 Activation of these cells also depended on the direction of the saccade. Five cells were activated whatever the direction of the forth- coming saccade (Fig. 6). Six cells were activated only if the forthcoming saccade was directed toward the corresponding light-key (Fig. 71), three if it was directed toward the opposite light-key (Fig. 711).

To detect a possible involvement of these cells in motor preparation, twenty were studied while the animal performed the task first with one arm and then with the other. Six cells showed a clear differ- ence between the right and left arm (Fig. 81).

Post-saccadic tonic cells

In our experimental conditions, these cells composed the largest class of units recorded in dorsolateral prefrontal cortex (77/216 = 35%)�9 Their activity was modified with onset of the orienting saccade and remained thereafter nearly constant. Fifty one neurons were differentially activated or, sometimes, suppressed by fixation of the right and left targets. We call this property spatial selectivity. Two exam- ples of spatial selectivity between right and left fixations are shown in Fig. 9. The number of units driven by fixation of the target which was presented eon~',alateral to the recording site (24) was dose to the ~a~mber driven by the ipsilateral target (22). Five other cells showed sustained activity with beginning of fixation on the contralateral side and a marked

.. ~ :S.~. . . . . . . . . . . I . . . . . . . . . . A �9 .......{.'_" .. -. . . . . . . . . . . . . . . . . . . .

�9 " " " � 9 2 " 2 � 9 t . " . . ~ ' . . . . . . . . . " . ; 2 _ . - _ " _ - - " _ . , . . " . . � 9

�9 . �9 . . . . . . . . a . . . . . . . . . . . . .

�9 - . . . . . . ; . - . . . , �9

�9 - . . .~ . . . . . . ~ . . . . . . . . _ . . . . . . t . . .

iR I t

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . s " - d ' " - " : " 7 " / 2 ~--~.- .~. . . - : , ' . ' ' "" - - 7 .

E~'-" 0

"'" : ~ ' " . ' . ' 2 L " " - - ' : " S . " ~ B � 9 1 4 9 1 4 9 . . . . t . . . . �9

L . . . . . . . . . . . . . . . . . . + . . . . . . . . . . .

l s

Fig. 91, II. Response of two post-saccadic neurons with spatial selectivity. In I, the cell is activated by fixation of the right-handed light-key (A). In II, the cell is suppressed by fixation of the left- i handed (B) light-key. Conventions are the same as in Fig. 3

1 i t

2 : " ' L ' . . . : � 9 . . . . . . . . . . , . . t -

�9 . . . . . . . . ~ i i . " �9 �9 . . . . . . . . . . . . . 4 .

II ................................ ~ ? ........ ! . . . . . . . . L ....... .~ ......... : .,J .. . . :..._: .:.:..: ...... ":.:_..:...2;'--v '; �9 : '7 '~..z..2~..." ......

: ' " ". " " . " . 7 " ; "." " - ' ~ . . . " .7 " ".. . . L ; , . - ; . _ ~ - " : ' ~ ' . . 2 . : ; . ' : " - ., " . . ~" . .

F i g . 10I, II. Reset of firing rate to intertrial level in two post- saccadic tonic cells (I and II). On each line, the small vertical bar indicates time of occurence of juice delivery�9 In I, activity returns to intertrial level at the time of the key-pressing and in II, at the onset of the juice reward�9 Conventions are the same as in Fig. 3

suppression of discharge when fixation was directed ipsilaterally. In 26 additional cells we observed no spatial selectivity. Sustained increased activity occur- red with fixation both on right and on left target lights.

For most neurons, the activity level returned to that of the intertrial level with complet ion of arm movement toward the target light. However , an interesting result was obtained at the end of the experiment when a delay of 1 s was introduced between the t ime of the key-pressing and the reward. With this new paradigm, 11 of the 27 cells studied showed a reduction in activity to baseline directly after the key pressing. In the remaining 16 cells the activity reduced to this level only after reward (Fig. 10).

465

To detect a possible involvement of these neurons in motor set, they were studied first with the animal using his right and then his left arm in the arm-reach task. We obtained reliable results for 31 neurons. For the majori ty (23), sustained activity did not depend on the arm used. For 8 neurons, firing depended on the location of the target and on the arm used for the movement . Figure 811 shows the activation of such a cell by fixation of the left key when the animal prepares a movemen t with the left arm, and by fixation of the right key when he prepares the movemen t with the left arm. We see that the same cell remained silent with crossed association: the right arm towards the left key, the left arm towards the right key.

Only seven cells showed mixed activity patterns. These combined tonic activity to the tone on one side with fixation related activity to the light-key on the other.

Signal-related phas i c units. Seven neurons showed distinct phasic activity following presentat ion of the auditory cue, within or without a trial. Six were activated by the contralateral, and one by the ipsilat- eral, auditory signal.

Fourteen cells were phasically activated by pre- sentation of the central fixation light and seven by the onset of the peripheral visual signals during fixation of the central light. Among these seven cells, four were activated differentially if the arm movemen t was - or was not - directed towards the corre- sponding visual key.

Saccade or m o v e m e n t related units. Activity of thir- teen neurons were related to saccadic eye movement . Most (10) showed definitive short postsaccadic acti- vation. Three displayed some evidence of presac- cadic activation.

A r m movement and/or dimming of the target light was accompanied by increased activity in 8 cells.

Units related to other aspects o f behavior . Twenty eight (28) neurons were activated or suppressed throughout a trial, f rom onset of the central fixation point to the reward. Two of fifteen which were tested showed differential activity depending on the arm used in the hand-reach task.

Twenty one (21) units were activated only at the end of trials. The data show that about half of these neurons were responsive of the presentat ion of juice reward, and, perhaps, to licking and swallowing. The other half responded to termination of the trial regardless of the reward and subsequent motor activity.

466

Histology

The area from which we recorded was characterized by a well developped internal granular layer IV. Each type of unit was found throughout the area and no distinct localization of each type was observed (Fig. 2).

Discussion

The results show that an ocular saccade which occurs at a pre-determined time in a delayed task induces enduring modification of firing in many cells of the Dorsolateral Prefrontal cortex. In 41 cells tonically activated by onset of peripheral stimuli, the firing rate was reduced to the intertrial level by orienting saccades towards the corresponding targets. In con- trast, tonic activation was induced in 77 cells by the orienting saccade towards the peripheral targets. We encountered only three cells which were activated before and during execution of a goal-directed sac- cade. Our analysis will focus on the role of those cells tonically activated before or after an orienting sac- cade.

The firing characteristics of cells recorded in the Frontal Eye field in a comparable experimental situation are different (Bruce and Goldberg 1985). In this area post-saccadic tonic cells are poorly repre- sented and no population of cells has been reported which were tonically activated between onset of a peripheral target and the orienting saccade. In con- trast, many cells are activated with execution of goal- directed saccades (Bruce and Goldberg, 1985).

Role of signal-related presaccadic tonic cells

These sustained units were activated between onset of a peripheral signal (auditory and/or visual) and the orienting saccade.

Tonic sustained activity was elicited by the audit- ory stimuli in 27 neurons. However, when the stimuli were presented on the same side during successive trials, the tonic activity sometimes preceeded onset of the stimulus. This suggests that recall of the location of the stimulus in previous trials or other cognitive and behavioral correlates of the recall may activate the cells.

For most cells, the significance of the long-lasting activation between onset of the auditory stimulus and the orienting saccade, is not clear. However, the hypothesis that it reflects short-term spatial memory is attractive. Reset of firing rate to base-line, with foveation of the corresponding light-key, would indi-

cate that spatial memory is deleted when the corre- sponding locus is foveated. Alternatively this activa- tion might reflect selective attention to a non- foveated region in space (Bruce and Goldberg 1985) or the oculomotor set. Any one of these functions could account for the firing characteristics of cells unilaterally or bilaterally activated. In a few of these cells, reset of firing rate to baseline occurs with onset of the ipsilateral visual stimulus (Fig. 4A). This change in activity may be understood as the replace- ment of an oculomotor set, based on the auditory cue, by another set based on the visual cue or as a reduction in attention, or memory demand, when the target of the forthcoming saccade becomes illumi- nated.

In three cells activated by both the auditory and the visual stimuli located in the same hemifield, the firing seemed to be associated to a memory process (Fig. 5). If activation by the auditory stimulus reflected saccade preparation or attention towards that hemifield, activation by the visual stimulus should subserve the same roles. Since such activation occured when the animal prepared a saccade towards the contralateral hemifield, this does not appear likely. The hypothesis which better fits the data is that such cells encode one region in space, cued either by the remembered auditory stimulus or by the current visual signal. As for tonic cells specific to vision, foveation of the corresponding region deleted the neural trace of the stimuli.

In conclusion, for most neurons tonically acti- vated by auditory stimuli, differentiation of those related to the oculomotor set from those subserving other cognitive functions will require further study. The hypothesis that all these cells subserve memory would fit our data but the memory function could be demonstrated in only three cells. Further study is also required to clarify the contribution of those cells which incorporate information regarding the arm later used for the key pressing (Fig. 8I).

Tonic activity was elicited by onset of a visual stimulus in 17 cells. In 8 of these, the response to onset of visual stimuli was independent of the direc- tion of the forthcoming saccade. These cells did not appear to participate in the central code specifying the direction of the correct response. Their firing rate was reset to intertrial level as soon as the animal foveated the light-key which activated the cell. This property is reminiscent of the foveal sparing of light- sensitive neurons discovered in the monkey parietal cortex (Motter and Mountcastle 1981). Reset of firing rate to intertrial level also occurred when the animal foveated the other light-key. It thus appears that foveation of any of the two light-keys deleted the neural trace of all contextual visual information.

467

Activation, in 9 cells, depended both on locus of retinal stimulation and on direction of the forthcom- ing saccade. This observation supports earlier obser- vations (Watanabe 1981; Komatsu 1982) that firing of certain prefrontal units in response to onset of a stimulus depended on the significance of the stimulus for behaviour. That is, the response characteristic depends upon whether the stimulus initiated an eye movement towards, or an eye movement away from it. Although the orienting saccade was prepared at the onset of the auditory stimulus, i.e. before onset of the visual stimulus, these cells could participate in the oculomotor set.

Role of the postsaccadic tonic cells

Neurons have been found in the prefrontal cortex which fire in anticipation of - and during - goal- directed arm (Sakai 1974; Kubota and Niki 1971) or wrist (Kubota and Funahashi 1982) movements in response to sensory cues. These results have sug- gested that the prefrontal cortex was involved in the effective execution of voluntary movements.

We recorded only three cells which fired while the arm movement was executed. In contrast, we recorded many which were tonically activated after the orienting saccade towards the peripheral targets. These cells might have a role in preparation of the subsequent movements. The results showed that I. the post-saccadic activation was interrupted by execution of the arm movement in less than half of the cells. In the remaining half it was disrupted by the end of the trials and/or the drinking; 2. within these two groups, activation of the cells depended mostly on the location of the target and not on the arm used for the key-pressing; 3. activation of the cells was not associated with visual fixation since the animals sometimes broke fixation during individual trials, both before and after the key pressing, without affecting unit activity.

Target location determined, through the whole interval, activation of most cells which fired between the saccade and the end of the trial. This observation precludes any suggestion that activity is related to contraction of jaws and face muscles or motor set associated with drinking. Since after the key pressing, the animals kept on looking primarily at the previously illuminated key, we suggest that these cells participate in the system which maintains atten- tion in a restricted region in space.

In most cells which began firing with the saccade and terminated with key-pressing, spatial selectivity appears to be similar to that often described in comparable situations (Niki 1974; Kojima and Gold-

man-Rakic 1984). No clear modulation by choice of the arm used for the movement was observed; isometric muscle contractions were thus not the main source of activation. We conclude therefore that this postsaccadic firing reflects preparatory activity, either of sensory or of motor systems to forthcoming events (the dimming of the target light or the arm movement) from or towards a particular region of space. Our results do not distinguish between these sensory and motor alternatives.

However , an experiment has been reported by Susuki and Azuma (1977) whose results may help. They recorded visual fixation neurons in the dorso- lateral prefrontal cortex in an experimental paradigm which included bar pressing in response to the dimming of peripheral lights. In the absence of a visually guided reaching they observed either no directional selectivity among the fixation neurons in a region ventral to our recording site or no fixation neuron in the region from which we recorded. The data of Suzuki and Azuma suggest that activation of directionally selective tonic neurons of the dorsolat- eral prefrontal cortex intervenes only during prepara- tion of visually guided reaching. This could be done by specifying the final target position of the hands.

Most of our cells were activated in association with the two arms. They appear to encode the location of the goal regardless of the arm that will perform the movement. A small proportion of the cells was activated in association with one arm only, whether ipsi- or contralateral to the recording site. These neurons might have a more direct role in motor set. Their profile of activation is similar to that of set-related cells in the premotor cortex (area 6A) except that activity in this area is associated only with preparation of the movement of the contralateral arm (Weinrich and Wise 1982).

Acknowledgements. The authors are grateful for the technical assistance of P. Prince, F. Giroud, S. Bello and M. Rouvi6re. We thank A. Hein, Dept. of Psychology, M. I.T. Cambridge, Mass., for his critical reading of the manuscript.

References

Bruce CJ, Goldberg ME (1985) Primate frontal eye fields: single neurons discharging before saccades. J Neurophysiol 53: 603--635

Fuster JM (1980) The prefrontal cortex. Raven, New York, pp 1-209

Fuster JM (1981) Prefrontal cortex in motor control. In: Brooks VB (ed) Handbook of physiology, Section 1, Vol II. Am Physiol Soc Washington DC, pp 1149-1178

Fuster JM (1973) Unit activity in prefrontal cortex during delayed response performance: neural correlates of transient memory. J Neurophysiol 36:61-78

468

Fuster JM (1985) The prefrontal cortex, mediator of cross- temporal contingencies. Human Neurobiol 4:16%179

Jacobson CF (1935) Functions of the frontal association area in primates. Arch Neurol Psychiatr 33:558-569

Jacobson CF, Nissen HV (1937) Studies of cerebral function in primates. IV. The effects of frontal lobe lesions on the delayed alternation habit in monkeys. J Comp Physiol Psychol 23:101-112

Kojima S, Goldman-Rakic PS (1984) Functional analysis of spatially discriminative neurons in prefrontal cortex of rhesus monkey. Brain Res 291:22%240

Komatsu H (1982) Prefrontal unit activity during a color discrimi- nation task with go and no-go responses in the monkey. Brain Res 244:26%277

Kubota K, Funahashi S (1982) Direction-specific activities of dorsolateral prefrontal and. motor cortex pyramidal tract neurons during visual tracking. J Neurophysiol 47:362-376

Kubota K, Niki H (1971) Prefrontal cortical unit activity and delayed alternation performance in monkeys. J Neurophysiol 34:337-347

Kubota K, Iwamoto T, Suzuki H (1974) Visuokinetic activities of primate prefrontal neurons during delayed response perform- ance. J Neurophysiol 37:1197-1211

Kubota K, Tunoike M, Mikami A (1980) Neuronal activity in the monkey dorsolateral prefrontal cortex during a discrimination task with delay. Brain Res 183:2%42

Motter BC, Mounteastle VB (1981) The functional properties of the light-sensitive neurons of the posterior parietal cortex studied in waking monkeys: foveal sparing and opponent vector organization. J Neurosci 1:3-26

Niki H (1974) Differential activity of prefrontal units during right and left delayed response trials. Brain Res 70:346-349

Sakai M (1974) Prefrontal unit activity during visually guided lever pressing reaction in the monkey. Brain Res 81:297-309

Susuki H, Azuma M (1977) Prefrontal neuronal activity during gazing at a light spot in the monkey. Brain Res 126:497-508

Watanabe M (1981) Prefrontal unit activity during delayed condi- tional discriminations in the monkey. Brain Res 225:51-65

Weinrich M, Wise SP (1982) The premotor cortex of the monkey. J Neurosci 2:132%1345

Received September 15, 1986 / Accepted February 27, 1987