- Email: [email protected]
Responses of monkey prefrontal neurons during a visual tracking task reinforced by substantia innominata self-stimulation
Brain Research, 276 (1983) 267-276
267
Elsevier
Responses of Monkey Prefrontal Neurons During a Visual Tracking Task Reinforced by Substantia Innominata Self-Stimulation SHINTARO FUNAHASHI*
Department of Neurophysiology, PrimateResearchInstitute, Kyoto University,Inuyama City, Aichi, 484 (Japan) (Accepted February 15th, 1983)
Key words: prefrontal neuron-- substantia innominata-- intracranial self-stimulation - visual tracking task - - movement-related neuron - - monkey
The role of the substantia innominata (SI) in the generation of a behavior mediated by the prefrontal (PF) cortex was examined in two Japanese monkeys. PF neuronal activities related to a visual tracking task by wrist movement were recorded and neural responses to SI stimulation were analyzed. Sixty-six neurons showed task-related activity and were classified into 3 types. Type 1 (n = 31) showed transient activation during the movement. Type 2 (n = 26) showed gradually increasing activity before the reward presentation. Type 3 (n = 9) showed tonic activation from the GO signal to the reward presentation. Antidromic and orthodromic responses to SI stimulation were observed in every type. In Type 1, the percentage of antidromically activated neurons (26%) was similar to that of orthodromically activated ones (16%), but in Type 2, 83% of responding neurons showed antidromic responses, and 56% of Type 3 showed orthodromic responses. These results show that the different types of PF neurons have different anatomical relations to SI. Although orthodromically activated neurons were fewer than antidromically activated neurons, many orthodromically activated neurons showed movement-related activity. This suggests that ICSS at SI facilitates the generation of the behavior through the afferent pathway to PF: INTRODUCTION Recent chronic single unit studies in monkeys have demonstrated a relation between neurons in the prefrontal cortex and the performance of a type of goaldirected behaviorg-12,14. In a delayed responsel0,11 or a delayed alternation task12A6, there was neural activation prior to the execution of the movement and these activities presumably reflected information about the ensuing movement. In a previous studyl0, the dorsolateral prefrontal neurons showed bidirectional or unidirectional activity before and during the task's movement, when the monkeys performed a visual tracking task with wrist extension and flexion. These movement-related neurons were thought to initiate the task's m o v e m e n t ~0. On the other hand, it has been reported that, in the monkey, the substantia innominata (SI) projects directly to the prefrontal cortex 6,7 and that SI is a positive area for intracranial self-stimulation (ICSS)13,18. As for the ICSS effect, the reward strength and the
response rate increase in proportion as the current intensity of the stimulation increases~, 5. Therefore, it is expected that ICSS at SI may activate the movementrelated neurons in the prefrontal cortex through the afferent pathway from SI and may facilitate the generation of ICSS behavior. In the present study, the role of SI in the generation of the behavior mediated by the prefrontal cortex is examined. The monkeys performed a twochoice visual tracking task reinforced by ICSS at SI. The prefrontal neuronal activities related to the task performance were recorded, and the types of neurons activated by SI stimulation were examined. A preliminary report has been presented2. MATERIALS AND METHODS Two young adult Japanese monkeys (Macaca fuscata), weighing at 4.0 kg (M19) and 6.2 kg (M20), were used. They sat in primate chairs, and performed a two-choice visual tracking task requiring wrist flex-
* Present address: Department of Physiology, Nara Medical College, Kashihara City, Nara, 634, Japan. 0006-8993/83/$03.00 © 1983 Elsevier Science Publishers B.V.
268 ion and extension. Instruments and m a n i p u l a n d a were the same as those used in a previous study ~0. The set-up, in short, is as follows. The m o n k e y rotates a vertical handle with his left hand. His right hand is tied gently to the p r i m a t e chair. To restrain arm movements, the left arm is inserted into two pieces of plexiglass tube interconnected at the elbow joint, leaving only the wrist joint movable. The handle is connected to a p o t e n t i o m e t e r (Copai, J45S, 50
A O
O
T3
T7
0 0 0 0 0 0 0 P2
B
P3
1
P4
P5
2
P6
P7
P8
3
I° o off Io° . "olI° o :i
/f
k~2). A b o u t 25 cm away from the m o n k e y ' s face. there is a panel with 9 light-emitting diodes ( L E D s ) (Fig. 1A). The upper two L E D s indicate the target position of the handle's m o v e m e n t and the lower 7 L E D s indicate the position of the handle. The middle part of the movable zone of the handle is divided into 8 ° zones. Each L E D indicates one zone, the n u m b e r of which corresponds to the n u m b e r of each L E D . The u p p e r L E D s are n u m b e r e d so as to correspond to the n u m b e r of the lower L E D s . The starting position was P5, the center of the movable zone, and the target position was T3 or 7"7. The sequence of a flexion trial, controlled by a minicomputer PDP-11/10, was as follows (Fig. 1B). The m o n k e y rotated the handle. As soon as the handle entered the start position (P5), a green lamp was lit at P5 (Fig. 1B, 1). It was lit as long as the m o n k e y held the handle at P5. If the handle was held at P5 for I s, the P5 color changed from green to red. This indicated a G O signal. Simultaneously with the G O signal, the T7 l a m p (red) was lit (Fig. 1B, 2). The monkey r o t a t e d the handle to the target position (P7) and held it there for 0.5 s (Fig. 1B, 3). Then, the reward (ICSS or a drop of water) was given. In an extension trial, the T3 lamp (red) was lit simultaneously with the G O signal presentation. Then, the m o n k e y rotated the handle to P3 and held it there for (1.5 s. The target position was selected r a n d o m l y according to a G e l l e r m a n series. In the early stages of the training, a drop of water (0.2 ml/trial) was used as reward.
lntracranial self-stimulation 1
2 - ~ - ~
iCSS
$
target zone
•
Fig. 1. Schematic illustrations of experimental set-ups and temporal sequence of a visual tracking task. A: a panel with 9 LEDs arranged in two rows and a handle coupled to the axis of a potentiometer. Upper LEDs (T3 and T7) are the target lamps. Lower LEDs (P2-P8) are the position lamps of the handle. P5 is also used as the start position lamp. The movable zone of the handle was divided into 80 zones. The numbering of the LEDs corresponds to that of the zones. B: schematic illustrations of the temporal sequence of the visual tracking task. Only 5 LEDs corresponded to T3, T7, P3, P5 and P7 are illustrated, Displacement (Disp.) of the handle was measured by the potentiometer. Two zones between broken lines are the start zone and the target zone. The filled square on the ICSS trace indicates the timing of ICSS. Fined circles in the panel indicate that the red lamp is lit and the hatched circle indicates that the green lamp is lit. Further details about the temporal sequence of the task are in the text.
A f t e r the m o n k e y learned to p e r f o r m securely this task reinforced by a drop of water, 3 or 4 stimulation electrodes were implanted in the right substantia innominata (SI) ( A 22-18, L 7 . 0 , V +2.8) under pentobarbital sodium anesthesia. During the surgery, the tip positions of the electrodes were estimated by observing lateral and frontal X-ray p h o t o g r a p h s of the skull, and these locations were later confirmed histologically. Concentric stainless steel electrodes (i.d. 0.2 mm, o.d. 0.6 mm) were used. P a r a m e t e r s of ICSS were 24 pulses of 100 Hz, 0.2 ms duration, squarewave, negative-going pulse generated by an electronic stimulator (ME-6012, M E Commercial). To measure the current intensity, the voltage between both ends of a 1 kg2 resistor, in series with the stimulation electrode, was m o n i t o r e d by an oscilloscope (5103N,
269 Tektronix). An indifferent electrode was a screw implanted in the occipital bone. Response rate (the rate of the reward presentation) was recorded by a cumulative recorder (Model C-3, Ralph Gerbrands). A response rate of 30 in 5 min with no intermission longer than 30 s was taken as the criterion for the positive ICSS effect 18. The threshold current intensity of ICSS was determined by gradually increasing the current intensity from 0.5 to 2.0 mA and then gradually decreasing from 2.0 to 0.5 mA. The threshold values of ICSS in the substantia innominata varied from 0.7 to 1.1 mA and did not alter during these experiments. When current intensity was more than 1.8 mA, neck movement or jaw movement was observed, indicating current spread to nearby structures such as the globus pallidus. Therfore, in these experiments, the current intensity of ICSS or SI stimulation was set at less than 1.5 mA.
were converted into 1-ms pulses by a window discriminator and stored in the computer's memory. Collected data were displayed on an oscilloscope screen as averaged histogram and dot displays, aligned at the moment of the GO signal presentation, the initiation of the movement, and the reward presentation. The responses to SI stimulation were displayed on a screen of the storage oscilloscope (5103N, Tektronix) and were photographed.
Histology At the end of the experiments, monkeys were also used in 2-deoxyglucose incorporation studies 15. They were sacrificed by an overdose of pentobarbital sodium. The brain was removed and serially sectioned (50 gm). Every third section was stained by the Nissl
A
Recording and data analysis After determining the threshold intensity for ICSS, the monkey was again anesthetized with pentobarbital sodium. A stainless steel cylinder (19 mm in diameter) and an L-shaped metal plate were implanted in the skull with dental resin. After recovery from the surgery, single unit recording was started. During recording sessions, the monkey was seated in the primate chair with his head tightly fixed to the chair's frame. A hydraulic microdrive (MO-8, Narishige) was attached to the cylinder. Glass-coated platinum-iridium microelectrodes (3-5 Mf~) were used. Single unit activities were recorded while the monkey was performing the task reinforced by ICSS. After these recordings, the responses to single or double pulse stimulation (10-ms or 3-ms intervals) at SI were recorded. Antidromic and orthodromic responses were observed. In the present study, if the responses were evoked repetitively with a fixed latency at the threshold, and at suprathreshold current intensity by 3-ms interval double shocks, the response was classified as the antidromic response 4. In 27 units, the threshold current intensity of the response was determined. Unit discharges, lamp signals, reward signals, and hand displacement were recorded on an FM tape recorder (R-260, TEAC). During recording sessions, unit discharges and signals were monitored by an oscilloscope. During off-line analysis, unit discharges
C 7
Put
Fig. 2. A: recorded area in monkey M19. B: recorded area in monkey M20. C: tip locations of ICSS electrodes implanted in monkey M19 (filled circles) and monkey M20 (open circles). Abbreviations: Amy, amygdala; CA, anterior commissure; Cd, caudate; GP, globus pallidus; HL, lateral hypothalamus; Put, putamen; S, Septum; SI, substantia innominata; TO, optic tract.
270 method. The tip positions of the stimulation electrodes and the recording sites in the prefrontal cortex were determined. The recorded areas of two monkeys are shown by circles in Fig. 2A and 2B. They were mainly the ventrolateral region of the prefrontal cortex (Walker's area 46). The recording region of one monkey (M19) included an anterior part of area 8 (Fig. 2A) and that of another monkey (M20) was mainly an area inferior to the principal sulcus (Fig. 2B). Tip positions of the stimulation electrodes are shown in Fig. 2C. In M19, two electrodes were in the substantia innominata, one was in the globus pallidus, and the remaining one was at the border region between the substantia innominata and the globus pallidus. Stimulation through each of 4 electrodes showed a positive effect. However, threshold current intensity of the electrode whose tip was in the globus pallidus was higher than that of the electrodes whose tips were in the substantia innominata. In M20, two electrode tips were in the substantia innominata and the other tip was in the lateral hypothalamus. Stimulation of the lateral hypothalamus did not show any positive effects of ICSS. In the present experiments, ICSS was delivered through the one electrode that was found to be most effective as a reward site and whose tip was in the substantia innominata. RESULTS
Prefrontal neuron activity The activities of 116 prefrontal (PF) neurons were recorded while monkeys were performing a visual tracking task. Sixty-six neurons changed their activities during the task. If a change of activity was recognized visually in the averaged histograms in relation to the events of the task, that change was taken as task-related. These task-related neurons were classified into 3 types (Types 1, 2 and 3) depending on the phase of the task during which the change of activity occurred. Among 66 task-related neurons, those whose activity increased around the initiation of the movement were classified as Type 1 (n = 31). These activities gradually increased around the GO signal presentation, peaked during the movement execution, and gradually decreased after the end of the movement. The increases in activity were bidirec-
tional (increase of the discharge rate for both movement directions) 10. Neurons with differential activity to the different movement directions (unidirectional type of activity) were not recorded in the present study. Fig. 3 is an example of Type 1 neuronal activity. Fig. 3 shows dot displays, averaged displacements, and histograms aligned at the moment of the GO signal presentation (A), at the initiation of the movement (B), and at reward presentation (C). Short continuous lines in the dot displays indicate the artefacts of 1CSS. The neuronal activity increased after the GO signal for both flexion and extension movements and the increase in activity started before the movement initiation. This activity, as seen in Fig. 3B, peaked during the movement execution and then decreased after the handle entered the target position. This neuron was not activated before or after the reward presentation. A weak difference in the activity magnitude depending on the difference of the movement directions was observed in Fig. 3A and 3B: more discharges were presented during extension (E) than during flexion (F). The difference of the activity between flexion and extension movements was observed in 42% of the neurons of this type. The neurons whose activity increased 300-500 ms prior to the reward presentation were classified as Type 2 (n = 26). Type 2 neuron activity was not related to movement execution, because these neurons were not activated during the movement and because the increase of the discharge rate began just after the end of the movement. In the example of Type 2 neuronal activity in Fig. 4, the gradual increase in activity started about 500 ms prior to reward presentation (Fig. 4C). This corresponded to the moment when the handle entered the target position. This increased activity was maintained until the reward presentation and then ceased. The magnitude of increased activity did not depend on the movement direction. In 18 neurons, the increasing activity prior to the reward presentation was compared between the natural reward (a drop of water) and ICSS. Fig. 5 shows the activity during task performance reinforced by ICSS through two stimulation electrodes (A and B) and by the natural reward (water, C). These dots and histograms include neuronal activity during both extension and flexion movements. Increasing activity prior to the reward presentation was observed during the
271
F
E 401~'ls .....
°,!
•
.
.
.
.
.
.
Q
. . . . . .
m°.~o.o .o .. . . . . . . . ,°.°~ .... •
A
B
....
°. •'.
"
m
• .....
.../
~
....
.
:
lllii~lli.lili ~Iilldiili~,.,
~ilL,,J,d..,,~,l,,....dd~. .....
C
' ~
.
.
.
.
.
.
.
.
.ill
• .......
i
I
120°
IlOOHz Fig. 3. An example of Type 1 activity. A: dot displays, averaged displacements, and histograms aligned at the moment of the GO signal presentation. B: those aligned at the moment of the imtiation of the movement. C: those aligned at the moment of the reward presentation• Figures on the left indicate the activities during the flexion movements (F) and those on the right indicate the activities during the extension movements (E). Short continuous lines in the dot displays indicate the artefacts of ICSS and these artefacts are included in the histograms.
task performance reinforced by ICSS (Fig• 5A and 5B), but increasing activity was not observed when a drop of water was used as reward (Fig. 5C). In 61% of examined Type 2 neurons, more increased activity was observed during the task reinforced by ICSS than by water. In the remaining neurons, the degree of in-
creasing activity was not different between ICSS and water. The neuronal activity which continued tonically from the GO signal presentation until the reward delivery was classified as Type 3 (n = 9). The activity of Type 3 neuron started to increase at about 60 ms after
272
F
E 400ms
~-..--,....~. ~,..
. _ ~ - "
,lil..l,l,a ,..,,, l.l , ,,,,II ], I..,
:i Iiiii!
B
7
,I
f
.... ~.....-:.-.:...->.. . . . .
~
'.
. .........
C J,,ilDd,,, I ml
,.,
,, ,.,
k.,d,~,l, m~d,,
z
Fig. 4. Type 2 activity• The arrangement of figures is the same as Fig. 3. Short continuous lines in the dot displays indicate the artefacts of ICSS.
the GO signal (Fig. 6A), and this activation was maintained tonically until the reward presentation. This increasing activity ceased after the reward presentation (Fig. 6C). Increasing activity ceased as soon as the natural reward was presented (see upper 10 trials in Fig, 6C). These changes were also observed when ICSS was used as reward. In this neuron, if the reward was not given at the moment of the reward presentation, increasing neuronal activity continued until the reward delivery.
Responses of prefrontal neurons to the substantia innominata stimulation The responses of the prefrontal neurons to SI stimulation were examined. A m o n g 66 task-related neurons, 37 neurons responded to single or double shock stimulation at SI. Twenty-four neurons showed antidromic responses (Fig. 7A) and 13 showed orthodromic responses (Fig. 7B). The antidromic responses occured with the same fixed latency at threshold stimulation (0.7 mA) by 3 ms interval
273 TABLE I
•400ms
Relationship between the responses to SI stimulation and the neuron types during the task performance
A
B ,n,m I,,, ~,,,al,
:i~.L..-....,~..i..i...j...i:....... 11:5 '..
12
' I
~ll~ll~u~ ...........
u. . . . .
IlOOHz
Fig. 5. Type 2 activities aligned at the reward presentation. The rewards were ICSS through one electrode in SI (A), ICSS through another electrode (B), and a drop of water (C). Each figure represents 10 trials of both directions of the movement.
double shocks as at 1.5 mA stimulation• The latency was 2.4 ms. In the antidromically responding neurons, the mean latency was 2.31 + 1.65 ms (mean + standard deviations). The orthodromic response in Fig. 7B shows early excitation (25-50 ms) followed by inhibition (50-100 ms). Except for two neurons, the exact latency was hard to be determined, because the latencies were usually long (20-80 ms) and variable and because the responses were inconstant. The latencies of the two exceptional neurons were 7.5 and 8.8 ms. Neuron types and responsiveness to S I stimulation
Table I shows the relationships between the responses to SI stimulation and the neuron types during
Type
Antidromic
Orthodromic
No response Total
1 2 3 Total
8 15 1 24
5 3 5 13
18 8 3 29
31 26 9 66
the task. Forty-two percent of Type 1 neurons responded to SI stimulation and the percentage of antidromically activated neurons (26%) was similar to that of orthodromically activated ones (16%). On the other hand, 69% of Type 2 neurons responded to SI stimulation and 83% of these neurons responded antidromically. These results show that there is a significant difference (P < 0.05, chi-square test) between Type 1 and Type 2 as for the percentage of responding neurons to SI stimulation and the ratio of antidromically activated neurons to orthodromically activated neurons. In Type 3 neurons, 6 responded to SI stimulation and, of these, 5 showed the orthodromic response. These results were significantly different from those of Type 2 (P < 0.02, chi-square test). On the other hand, of 66 task-related prefrontal neurons, 44% did not respond to SI stimulation. About 60% of these neurons belonged to Type 1. Thus, both antidromic and orthodromic responses were observed in every type of the prefrontal neurons. But percentages of responding Type 2 and Type 3 neurons were more than those of responding Type 1 neurons. DISCUSSION The present study showed that prefrontal neurons related to the visual tracking task were activated antidromically or orthodromically by rewarding stimulation at SI. There were 3 types of prefrontal neurons: Type 1 discharged during movement, Type 2 after movement onset and before reward presentation, and Type 3 from the GO signal to reward presentation. There were significant differences among these types as for the percentage of responding neurons and the ratio of antidromically activated neurons to orthodromically activated neurons. As to the relationships between ICSS positive sites
274
E
F
"'
.'
•
• .
."
'
.'..'.
T..'.".::.';'.:"
. . . . . . . . . . . . . . . . . . . . . .
A
~
.:.:.~..~.~,..~
7;.'..'.~""
....~
.~=.L~,
400ms
";'?." ";
...............
..................
,
. . . . . . . . . . .
• ' ~" "' :"
....................
B
.
................. .'. " ,:..'. . . . . . . . . . . . . . . . .
°. . . .
"-...
"".:::
,| .........
;'.':, .-.':'.'..':.'.':..L'..
. ...................
, -.- :.-;:::-: .;- ...-.;.. ;.2.- .:---...-:..~.--..-:-.-...--...-;---..-...--..--..--..~.
• :i "'. : ' .i.: ":-.:i'l'.':.'i.~"
~ ................
i": ..... : ..... ::::::::::-ri':i : ~" :.:: : ' " . .
C
-
.
.
.
.
.
.
.
.
.
.
.
.
.
~ - - . . . .
•.--.__.---:.-:-:-...-?...-.--..-.-...-. . . . . . . . . . . . . . . . . . . . . .
b .
.
.
|-o
.
.
• ,
.
...
.
.
.
.:
.
.':".: . .
f - - ~ - -'-'--''~
:.
.
....
1200
Fig. 6. Type 3 activity. In all dot displays, the upper 10 indicate the neuronal activities during the task reinforced by water and the lower 3 indicate the activities during the task reinforced by ICSS. In C, artefacts of ICSS are displayed in dots and histograms contain these artefacts•
275
A
B Fig. 7. Prefrontal neuronal responsesto SI stimulation. A: antidromic response. These show 10 superimposed traces at 10-ms interval double shocks (1.5 mA). Latency was 2.4 ms. B: orthodromic responses. These show 15 superimposedtraces at 10-ms interval double shocks (1.5 mA). These responses in the figure were triggered by first pulse of double pulse stimulation. These responses include early excitation (25-50 ms) followedby inhibition (50-100 ms).Callibrations: A, 1 ms, 200/~V; B, 20 ms, 200~V. and movement-related neurons, the rewarding brain stimulation may activate these movement-related neurons and facilitate the generation of the behavior, because the reward strength or response rate increased proportionally to the increase in current intensity of ICSS1, s. In the present study, Type 1 neurons were thought to belong to the movement-related neurons, because they were activated only around the time of movement execution and were not active after the handle entering the target position and because these activity patterns were similar to those of the bidirectional type in the choice task of the previous study 10. About 40% of these Type 1 neurons responded to SI stimulation. Though the percentage of Type 1 neurons which responded to SI stimulation was lower than those of Type 2 and Type 3 neurons, some Type 1 neurons (16%) did respond orthodromically to SI stimulation. About 40% of the orthodromicaUy activated task-related neurons belonged to Type 1. This result indicates the possibility that ICSS facilitates the generation of the behavior
through movement-related neurons which respond orthodromically. However, in this study, other types of prefrontal neurons responded orthodromically to SI stimulation. Thus another possibility may be that ICSS activates various types of neurons in the prefrontal cortex and such general activation of the prefrontal cortex by ICSS may facilitate the generation of the behavior. Many of the Type 2 neurons, whose activity increased 300-500 ms prior to reward presentation and ceased at reward presentation, were activated by SI stimulation. These Type 2 activities were similar to Sakai's gradually increasing activity during the waiting period before the GO signal 20, Niki and Watanabe's type IIIa and IIIb activities ~7, and Komatsu's Type B activity8. These increasing activities presumably reflected 'expectancy'S,20 or 'anticipatory change '17. In Type 2 neurons, the gradually increasing activity ceased at the reward presentation. This increasing activity appeared with ICSS. As shown in Fig. 5, the degree of activation was different in accordance with the nature of the reward (electrical stimulation or a drop of water). Therefore, these gradually increasing activities prior to reward presentation may reflect the expectancy of the reward or anticipatory change to the reward presentation. Most Type 2 neurons responded antidromically to SI stimulation. Accordingly, these Type 2 neurons may send anticipatory information about the reward to the positive site of ICSS from the prefrontal cortex. Routtenberg and Santos-Anderson 19 pointed out that the efferent fibers from the prefrontal cortex to other ICSS positive sites were important in the generation of ICSS behavior. Therefore, Type 2 neurons with the antidromic response and anticipatory activity to the reward presentation may play an important role in the generation of ICSS behavior. Type 3 neuronal activity was similar to Niki and Watanabe's type IIIc activity17, Fuster's C or D neurons 3, and Komatsu's Type C activityS. However, Type 3 activity differed from these studies in that Niki and Watanabe's and Fuster's activities were recorded during the delay phase without motor performance, and in that the beginning and termination of the increased activity of Type 3 neurons were relatively time locked to the GO signal presentation and the reward delivery, respectively. These Type 3 neurons might be activated, at least partially, by visual
276 inputs, because it has been r e p o r t e d that many neurons in the prefrontal cortex receive visual inputs8,11A4. H o w e v e r , these T y p e 3 neurons did not respond to the light signal at the time when the handle entered the starting zone. T y p e 3 activity may relate to both m o t o r p e r f o r m a n c e and r e w a r d presentation, because this type of neuron was activated by the specific stimulus related to the m o v e m e n t execution (the G O signal) and activated tonically during the task m o v e m e n t and because the termination of increasing activity was time locked to the reward presentation. Most of the Type 3 neurons r e s p o n d e d orthodromically to SI stimulation. This result also indicates the possibility that ICSS facilitates the generation of the behavior through o r t h o d r o m i c a l l y activated movement-related neurons. Thus, prefrontal neurons r e s p o n d e d antidromically or o r t h o d r o m i c a l l y to the r e w a r d stimulation. The
neurons which showed different types of activity during the visual tracking task showed different types of responses. These results show that functionally different neurons have different anatomical relations to SI. Though fewer m o v e m e n t - r e l a t e d neurons responded to SI stimulation than o t h e r types of neurons did, some m o v e m e n t - r e l a t e d neurons r e s p o n d e d orthodromically. This indicates the possibility that ICSS facilitates the generation of the behavior through the afferent pathway to the prefrontal cortex. ACKNOWLEDGEMENT I would like to express my thanks to Dr. K. K u b o t a for his guidance and e n c o u r a g e m e n t throughout the course of this study.
REFERENCES 1 Ettenberg, A., Conditioned taste preference and response rate as measures of brain-stimulafion reward: a comparison, Physiol. Behav., 24 (1980) 755-758. 2 Funahashi, S. and Kubota, K., Prefrontal neuronal activity during a visual tracking task and substantia innominata stimulation, Neurosci. Len., Suppl. 9 (1982) $62. 3 Fuster, J. M., Unit activity in prefrontal cortex during delayed-response performance: neuronal correlates of transient memory, J. Neurophysiol., 36 (1973) 61-78. 4 German, D. C. and Fetz, E. E., Responses of primate locus coeruleus and subcoeruleus neurons to stimulation at reinforcing brain sites and to natural reinforcers, Brain Research, 109 (1976) 497-514. 5 Hawkins, T. D. and Pliskoff, S. S., Brain-stimulafion intensity, rate of self-stimulation, and reinforcement strength: an analysis through chaining, J. exp. Anal. Behav., 7 (1964) 285-288. 6 Jones, E. G., Burton, H., Saper, C. B. and Swanson, L. W., Midbrain, diencephalic and cortical relationship of the basal nucleus of Meynert and associate structures in primates, J. comp. Neurol., 167 (1976) 385--420. 7 Kievit, J. and Kuypers, H. G. J. M., Basal forebrain and hypothalamic connections to frontal and parietal cortex in the rhesus monkey, Science, 187 (1974) 660--662. 8 Komatsu, H., Prefrontal unit activity during a color discrimination task with GO and NO-GO responses in the monkey, Brain Research, 244 (1982) 269-277. 9 Kubota, K., Neuron activity in the dorsolateral prefrontal cortex of the monkey and initiation of behavior. In M. Ito, N. Tsukahara, K. Kubota, and K. Yagi (Eds.), Integrative Control Functions of the Brain, Vol. I, Kohdansha-Elsevier, Amsterdam, 1978, pp. 407-417. 10 Kubota, K. and Funahashi, S., Direction specific activities of dorsolateral prefrontal and motor cortex pyramidal tract neurons during visual tracking, J. Neurophysiol., 47 (1982) 362-376.
11 Kubota, K., Iwamoto, T., and Suzuki, H,, Visuokinetic activities of primate prefrontal neurons during delayed-response performance, J. Neurophysiol., 37 (1974) 11971212. 12 Kubota, K. and Niki, H., Prefrontal cortical unit activity and delayed alternation performance in monkey, J. Neurophysiol., 34 (1971) 337-347. 13 Kubota, K. and Noda, T., Visually activated neurons of monkeys and substantia innominata self-stimulation, Neurosci. Lett., Suppl. 4 (1980) $61. 14 Kubota, K., Tonoike, M., and Mikami, A., Neuronal activity in the monkey dorsolateral prefrontal cortex during a discrimination task with delay, Brain Research, 183 (1980) 29-42. 15 Matsunami, K., Funahashi, S. and Kubota, K., Radioactive 2-deoxyglucose uptake into brains of monkeys performing intracranial self-stimulation task, J. Physiol. Soc. Ipn., 44 (1982) 405. 16 Niki, H., Prefrontal unit activity during delayed alternation in the monkey. I. Relation to direction of response, Brain Research, 68 (1974) 185--196. 17 Niki, H. and Watanabe, M., Prefrontal and cingulate unit activity during timing behavior in the monkey, Brain Research, 171 (1979) 213-224. 18 Roils, E. T., Burton, M. J. and Mora, F., Neurophysiological analysis of brain-stimulation reward in the monkey, Brain Research, 194 (1980) 339-357. 19 Routtenberg, A. and Santos-Anderson, R., The role of prefrontal cortex in intracranial self-stimulation: a case history of anatomical localization of motivational substrates. In L. L. Iversen, S. D, Iversen, and S, D. Snyder (Eds.), Handbook of Psychopharmacology, Vol. 8, Plenum, New York, 1977, pp. 1-24. 20 Sakai, M., Prefrontal unit activity during visually guided lever pressing reaction in the monkey, Brain Research, 81 (1974) 297-309.
267
Elsevier
Responses of Monkey Prefrontal Neurons During a Visual Tracking Task Reinforced by Substantia Innominata Self-Stimulation SHINTARO FUNAHASHI*
Department of Neurophysiology, PrimateResearchInstitute, Kyoto University,Inuyama City, Aichi, 484 (Japan) (Accepted February 15th, 1983)
Key words: prefrontal neuron-- substantia innominata-- intracranial self-stimulation - visual tracking task - - movement-related neuron - - monkey
The role of the substantia innominata (SI) in the generation of a behavior mediated by the prefrontal (PF) cortex was examined in two Japanese monkeys. PF neuronal activities related to a visual tracking task by wrist movement were recorded and neural responses to SI stimulation were analyzed. Sixty-six neurons showed task-related activity and were classified into 3 types. Type 1 (n = 31) showed transient activation during the movement. Type 2 (n = 26) showed gradually increasing activity before the reward presentation. Type 3 (n = 9) showed tonic activation from the GO signal to the reward presentation. Antidromic and orthodromic responses to SI stimulation were observed in every type. In Type 1, the percentage of antidromically activated neurons (26%) was similar to that of orthodromically activated ones (16%), but in Type 2, 83% of responding neurons showed antidromic responses, and 56% of Type 3 showed orthodromic responses. These results show that the different types of PF neurons have different anatomical relations to SI. Although orthodromically activated neurons were fewer than antidromically activated neurons, many orthodromically activated neurons showed movement-related activity. This suggests that ICSS at SI facilitates the generation of the behavior through the afferent pathway to PF: INTRODUCTION Recent chronic single unit studies in monkeys have demonstrated a relation between neurons in the prefrontal cortex and the performance of a type of goaldirected behaviorg-12,14. In a delayed responsel0,11 or a delayed alternation task12A6, there was neural activation prior to the execution of the movement and these activities presumably reflected information about the ensuing movement. In a previous studyl0, the dorsolateral prefrontal neurons showed bidirectional or unidirectional activity before and during the task's movement, when the monkeys performed a visual tracking task with wrist extension and flexion. These movement-related neurons were thought to initiate the task's m o v e m e n t ~0. On the other hand, it has been reported that, in the monkey, the substantia innominata (SI) projects directly to the prefrontal cortex 6,7 and that SI is a positive area for intracranial self-stimulation (ICSS)13,18. As for the ICSS effect, the reward strength and the
response rate increase in proportion as the current intensity of the stimulation increases~, 5. Therefore, it is expected that ICSS at SI may activate the movementrelated neurons in the prefrontal cortex through the afferent pathway from SI and may facilitate the generation of ICSS behavior. In the present study, the role of SI in the generation of the behavior mediated by the prefrontal cortex is examined. The monkeys performed a twochoice visual tracking task reinforced by ICSS at SI. The prefrontal neuronal activities related to the task performance were recorded, and the types of neurons activated by SI stimulation were examined. A preliminary report has been presented2. MATERIALS AND METHODS Two young adult Japanese monkeys (Macaca fuscata), weighing at 4.0 kg (M19) and 6.2 kg (M20), were used. They sat in primate chairs, and performed a two-choice visual tracking task requiring wrist flex-
* Present address: Department of Physiology, Nara Medical College, Kashihara City, Nara, 634, Japan. 0006-8993/83/$03.00 © 1983 Elsevier Science Publishers B.V.
268 ion and extension. Instruments and m a n i p u l a n d a were the same as those used in a previous study ~0. The set-up, in short, is as follows. The m o n k e y rotates a vertical handle with his left hand. His right hand is tied gently to the p r i m a t e chair. To restrain arm movements, the left arm is inserted into two pieces of plexiglass tube interconnected at the elbow joint, leaving only the wrist joint movable. The handle is connected to a p o t e n t i o m e t e r (Copai, J45S, 50
A O
O
T3
T7
0 0 0 0 0 0 0 P2
B
P3
1
P4
P5
2
P6
P7
P8
3
I° o off Io° . "olI° o :i
/f
k~2). A b o u t 25 cm away from the m o n k e y ' s face. there is a panel with 9 light-emitting diodes ( L E D s ) (Fig. 1A). The upper two L E D s indicate the target position of the handle's m o v e m e n t and the lower 7 L E D s indicate the position of the handle. The middle part of the movable zone of the handle is divided into 8 ° zones. Each L E D indicates one zone, the n u m b e r of which corresponds to the n u m b e r of each L E D . The u p p e r L E D s are n u m b e r e d so as to correspond to the n u m b e r of the lower L E D s . The starting position was P5, the center of the movable zone, and the target position was T3 or 7"7. The sequence of a flexion trial, controlled by a minicomputer PDP-11/10, was as follows (Fig. 1B). The m o n k e y rotated the handle. As soon as the handle entered the start position (P5), a green lamp was lit at P5 (Fig. 1B, 1). It was lit as long as the m o n k e y held the handle at P5. If the handle was held at P5 for I s, the P5 color changed from green to red. This indicated a G O signal. Simultaneously with the G O signal, the T7 l a m p (red) was lit (Fig. 1B, 2). The monkey r o t a t e d the handle to the target position (P7) and held it there for 0.5 s (Fig. 1B, 3). Then, the reward (ICSS or a drop of water) was given. In an extension trial, the T3 lamp (red) was lit simultaneously with the G O signal presentation. Then, the m o n k e y rotated the handle to P3 and held it there for (1.5 s. The target position was selected r a n d o m l y according to a G e l l e r m a n series. In the early stages of the training, a drop of water (0.2 ml/trial) was used as reward.
lntracranial self-stimulation 1
2 - ~ - ~
iCSS
$
target zone
•
Fig. 1. Schematic illustrations of experimental set-ups and temporal sequence of a visual tracking task. A: a panel with 9 LEDs arranged in two rows and a handle coupled to the axis of a potentiometer. Upper LEDs (T3 and T7) are the target lamps. Lower LEDs (P2-P8) are the position lamps of the handle. P5 is also used as the start position lamp. The movable zone of the handle was divided into 80 zones. The numbering of the LEDs corresponds to that of the zones. B: schematic illustrations of the temporal sequence of the visual tracking task. Only 5 LEDs corresponded to T3, T7, P3, P5 and P7 are illustrated, Displacement (Disp.) of the handle was measured by the potentiometer. Two zones between broken lines are the start zone and the target zone. The filled square on the ICSS trace indicates the timing of ICSS. Fined circles in the panel indicate that the red lamp is lit and the hatched circle indicates that the green lamp is lit. Further details about the temporal sequence of the task are in the text.
A f t e r the m o n k e y learned to p e r f o r m securely this task reinforced by a drop of water, 3 or 4 stimulation electrodes were implanted in the right substantia innominata (SI) ( A 22-18, L 7 . 0 , V +2.8) under pentobarbital sodium anesthesia. During the surgery, the tip positions of the electrodes were estimated by observing lateral and frontal X-ray p h o t o g r a p h s of the skull, and these locations were later confirmed histologically. Concentric stainless steel electrodes (i.d. 0.2 mm, o.d. 0.6 mm) were used. P a r a m e t e r s of ICSS were 24 pulses of 100 Hz, 0.2 ms duration, squarewave, negative-going pulse generated by an electronic stimulator (ME-6012, M E Commercial). To measure the current intensity, the voltage between both ends of a 1 kg2 resistor, in series with the stimulation electrode, was m o n i t o r e d by an oscilloscope (5103N,
269 Tektronix). An indifferent electrode was a screw implanted in the occipital bone. Response rate (the rate of the reward presentation) was recorded by a cumulative recorder (Model C-3, Ralph Gerbrands). A response rate of 30 in 5 min with no intermission longer than 30 s was taken as the criterion for the positive ICSS effect 18. The threshold current intensity of ICSS was determined by gradually increasing the current intensity from 0.5 to 2.0 mA and then gradually decreasing from 2.0 to 0.5 mA. The threshold values of ICSS in the substantia innominata varied from 0.7 to 1.1 mA and did not alter during these experiments. When current intensity was more than 1.8 mA, neck movement or jaw movement was observed, indicating current spread to nearby structures such as the globus pallidus. Therfore, in these experiments, the current intensity of ICSS or SI stimulation was set at less than 1.5 mA.
were converted into 1-ms pulses by a window discriminator and stored in the computer's memory. Collected data were displayed on an oscilloscope screen as averaged histogram and dot displays, aligned at the moment of the GO signal presentation, the initiation of the movement, and the reward presentation. The responses to SI stimulation were displayed on a screen of the storage oscilloscope (5103N, Tektronix) and were photographed.
Histology At the end of the experiments, monkeys were also used in 2-deoxyglucose incorporation studies 15. They were sacrificed by an overdose of pentobarbital sodium. The brain was removed and serially sectioned (50 gm). Every third section was stained by the Nissl
A
Recording and data analysis After determining the threshold intensity for ICSS, the monkey was again anesthetized with pentobarbital sodium. A stainless steel cylinder (19 mm in diameter) and an L-shaped metal plate were implanted in the skull with dental resin. After recovery from the surgery, single unit recording was started. During recording sessions, the monkey was seated in the primate chair with his head tightly fixed to the chair's frame. A hydraulic microdrive (MO-8, Narishige) was attached to the cylinder. Glass-coated platinum-iridium microelectrodes (3-5 Mf~) were used. Single unit activities were recorded while the monkey was performing the task reinforced by ICSS. After these recordings, the responses to single or double pulse stimulation (10-ms or 3-ms intervals) at SI were recorded. Antidromic and orthodromic responses were observed. In the present study, if the responses were evoked repetitively with a fixed latency at the threshold, and at suprathreshold current intensity by 3-ms interval double shocks, the response was classified as the antidromic response 4. In 27 units, the threshold current intensity of the response was determined. Unit discharges, lamp signals, reward signals, and hand displacement were recorded on an FM tape recorder (R-260, TEAC). During recording sessions, unit discharges and signals were monitored by an oscilloscope. During off-line analysis, unit discharges
C 7
Put
Fig. 2. A: recorded area in monkey M19. B: recorded area in monkey M20. C: tip locations of ICSS electrodes implanted in monkey M19 (filled circles) and monkey M20 (open circles). Abbreviations: Amy, amygdala; CA, anterior commissure; Cd, caudate; GP, globus pallidus; HL, lateral hypothalamus; Put, putamen; S, Septum; SI, substantia innominata; TO, optic tract.
270 method. The tip positions of the stimulation electrodes and the recording sites in the prefrontal cortex were determined. The recorded areas of two monkeys are shown by circles in Fig. 2A and 2B. They were mainly the ventrolateral region of the prefrontal cortex (Walker's area 46). The recording region of one monkey (M19) included an anterior part of area 8 (Fig. 2A) and that of another monkey (M20) was mainly an area inferior to the principal sulcus (Fig. 2B). Tip positions of the stimulation electrodes are shown in Fig. 2C. In M19, two electrodes were in the substantia innominata, one was in the globus pallidus, and the remaining one was at the border region between the substantia innominata and the globus pallidus. Stimulation through each of 4 electrodes showed a positive effect. However, threshold current intensity of the electrode whose tip was in the globus pallidus was higher than that of the electrodes whose tips were in the substantia innominata. In M20, two electrode tips were in the substantia innominata and the other tip was in the lateral hypothalamus. Stimulation of the lateral hypothalamus did not show any positive effects of ICSS. In the present experiments, ICSS was delivered through the one electrode that was found to be most effective as a reward site and whose tip was in the substantia innominata. RESULTS
Prefrontal neuron activity The activities of 116 prefrontal (PF) neurons were recorded while monkeys were performing a visual tracking task. Sixty-six neurons changed their activities during the task. If a change of activity was recognized visually in the averaged histograms in relation to the events of the task, that change was taken as task-related. These task-related neurons were classified into 3 types (Types 1, 2 and 3) depending on the phase of the task during which the change of activity occurred. Among 66 task-related neurons, those whose activity increased around the initiation of the movement were classified as Type 1 (n = 31). These activities gradually increased around the GO signal presentation, peaked during the movement execution, and gradually decreased after the end of the movement. The increases in activity were bidirec-
tional (increase of the discharge rate for both movement directions) 10. Neurons with differential activity to the different movement directions (unidirectional type of activity) were not recorded in the present study. Fig. 3 is an example of Type 1 neuronal activity. Fig. 3 shows dot displays, averaged displacements, and histograms aligned at the moment of the GO signal presentation (A), at the initiation of the movement (B), and at reward presentation (C). Short continuous lines in the dot displays indicate the artefacts of 1CSS. The neuronal activity increased after the GO signal for both flexion and extension movements and the increase in activity started before the movement initiation. This activity, as seen in Fig. 3B, peaked during the movement execution and then decreased after the handle entered the target position. This neuron was not activated before or after the reward presentation. A weak difference in the activity magnitude depending on the difference of the movement directions was observed in Fig. 3A and 3B: more discharges were presented during extension (E) than during flexion (F). The difference of the activity between flexion and extension movements was observed in 42% of the neurons of this type. The neurons whose activity increased 300-500 ms prior to the reward presentation were classified as Type 2 (n = 26). Type 2 neuron activity was not related to movement execution, because these neurons were not activated during the movement and because the increase of the discharge rate began just after the end of the movement. In the example of Type 2 neuronal activity in Fig. 4, the gradual increase in activity started about 500 ms prior to reward presentation (Fig. 4C). This corresponded to the moment when the handle entered the target position. This increased activity was maintained until the reward presentation and then ceased. The magnitude of increased activity did not depend on the movement direction. In 18 neurons, the increasing activity prior to the reward presentation was compared between the natural reward (a drop of water) and ICSS. Fig. 5 shows the activity during task performance reinforced by ICSS through two stimulation electrodes (A and B) and by the natural reward (water, C). These dots and histograms include neuronal activity during both extension and flexion movements. Increasing activity prior to the reward presentation was observed during the
271
F
E 401~'ls .....
°,!
•
.
.
.
.
.
.
Q
. . . . . .
m°.~o.o .o .. . . . . . . . ,°.°~ .... •
A
B
....
°. •'.
"
m
• .....
.../
~
....
.
:
lllii~lli.lili ~Iilldiili~,.,
~ilL,,J,d..,,~,l,,....dd~. .....
C
' ~
.
.
.
.
.
.
.
.
.ill
• .......
i
I
120°
IlOOHz Fig. 3. An example of Type 1 activity. A: dot displays, averaged displacements, and histograms aligned at the moment of the GO signal presentation. B: those aligned at the moment of the imtiation of the movement. C: those aligned at the moment of the reward presentation• Figures on the left indicate the activities during the flexion movements (F) and those on the right indicate the activities during the extension movements (E). Short continuous lines in the dot displays indicate the artefacts of ICSS and these artefacts are included in the histograms.
task performance reinforced by ICSS (Fig• 5A and 5B), but increasing activity was not observed when a drop of water was used as reward (Fig. 5C). In 61% of examined Type 2 neurons, more increased activity was observed during the task reinforced by ICSS than by water. In the remaining neurons, the degree of in-
creasing activity was not different between ICSS and water. The neuronal activity which continued tonically from the GO signal presentation until the reward delivery was classified as Type 3 (n = 9). The activity of Type 3 neuron started to increase at about 60 ms after
272
F
E 400ms
~-..--,....~. ~,..
. _ ~ - "
,lil..l,l,a ,..,,, l.l , ,,,,II ], I..,
:i Iiiii!
B
7
,I
f
.... ~.....-:.-.:...->.. . . . .
~
'.
. .........
C J,,ilDd,,, I ml
,.,
,, ,.,
k.,d,~,l, m~d,,
z
Fig. 4. Type 2 activity• The arrangement of figures is the same as Fig. 3. Short continuous lines in the dot displays indicate the artefacts of ICSS.
the GO signal (Fig. 6A), and this activation was maintained tonically until the reward presentation. This increasing activity ceased after the reward presentation (Fig. 6C). Increasing activity ceased as soon as the natural reward was presented (see upper 10 trials in Fig, 6C). These changes were also observed when ICSS was used as reward. In this neuron, if the reward was not given at the moment of the reward presentation, increasing neuronal activity continued until the reward delivery.
Responses of prefrontal neurons to the substantia innominata stimulation The responses of the prefrontal neurons to SI stimulation were examined. A m o n g 66 task-related neurons, 37 neurons responded to single or double shock stimulation at SI. Twenty-four neurons showed antidromic responses (Fig. 7A) and 13 showed orthodromic responses (Fig. 7B). The antidromic responses occured with the same fixed latency at threshold stimulation (0.7 mA) by 3 ms interval
273 TABLE I
•400ms
Relationship between the responses to SI stimulation and the neuron types during the task performance
A
B ,n,m I,,, ~,,,al,
:i~.L..-....,~..i..i...j...i:....... 11:5 '..
12
' I
~ll~ll~u~ ...........
u. . . . .
IlOOHz
Fig. 5. Type 2 activities aligned at the reward presentation. The rewards were ICSS through one electrode in SI (A), ICSS through another electrode (B), and a drop of water (C). Each figure represents 10 trials of both directions of the movement.
double shocks as at 1.5 mA stimulation• The latency was 2.4 ms. In the antidromically responding neurons, the mean latency was 2.31 + 1.65 ms (mean + standard deviations). The orthodromic response in Fig. 7B shows early excitation (25-50 ms) followed by inhibition (50-100 ms). Except for two neurons, the exact latency was hard to be determined, because the latencies were usually long (20-80 ms) and variable and because the responses were inconstant. The latencies of the two exceptional neurons were 7.5 and 8.8 ms. Neuron types and responsiveness to S I stimulation
Table I shows the relationships between the responses to SI stimulation and the neuron types during
Type
Antidromic
Orthodromic
No response Total
1 2 3 Total
8 15 1 24
5 3 5 13
18 8 3 29
31 26 9 66
the task. Forty-two percent of Type 1 neurons responded to SI stimulation and the percentage of antidromically activated neurons (26%) was similar to that of orthodromically activated ones (16%). On the other hand, 69% of Type 2 neurons responded to SI stimulation and 83% of these neurons responded antidromically. These results show that there is a significant difference (P < 0.05, chi-square test) between Type 1 and Type 2 as for the percentage of responding neurons to SI stimulation and the ratio of antidromically activated neurons to orthodromically activated neurons. In Type 3 neurons, 6 responded to SI stimulation and, of these, 5 showed the orthodromic response. These results were significantly different from those of Type 2 (P < 0.02, chi-square test). On the other hand, of 66 task-related prefrontal neurons, 44% did not respond to SI stimulation. About 60% of these neurons belonged to Type 1. Thus, both antidromic and orthodromic responses were observed in every type of the prefrontal neurons. But percentages of responding Type 2 and Type 3 neurons were more than those of responding Type 1 neurons. DISCUSSION The present study showed that prefrontal neurons related to the visual tracking task were activated antidromically or orthodromically by rewarding stimulation at SI. There were 3 types of prefrontal neurons: Type 1 discharged during movement, Type 2 after movement onset and before reward presentation, and Type 3 from the GO signal to reward presentation. There were significant differences among these types as for the percentage of responding neurons and the ratio of antidromically activated neurons to orthodromically activated neurons. As to the relationships between ICSS positive sites
274
E
F
"'
.'
•
• .
."
'
.'..'.
T..'.".::.';'.:"
. . . . . . . . . . . . . . . . . . . . . .
A
~
.:.:.~..~.~,..~
7;.'..'.~""
....~
.~=.L~,
400ms
";'?." ";
...............
..................
,
. . . . . . . . . . .
• ' ~" "' :"
....................
B
.
................. .'. " ,:..'. . . . . . . . . . . . . . . . .
°. . . .
"-...
"".:::
,| .........
;'.':, .-.':'.'..':.'.':..L'..
. ...................
, -.- :.-;:::-: .;- ...-.;.. ;.2.- .:---...-:..~.--..-:-.-...--...-;---..-...--..--..--..~.
• :i "'. : ' .i.: ":-.:i'l'.':.'i.~"
~ ................
i": ..... : ..... ::::::::::-ri':i : ~" :.:: : ' " . .
C
-
.
.
.
.
.
.
.
.
.
.
.
.
.
~ - - . . . .
•.--.__.---:.-:-:-...-?...-.--..-.-...-. . . . . . . . . . . . . . . . . . . . . .
b .
.
.
|-o
.
.
• ,
.
...
.
.
.
.:
.
.':".: . .
f - - ~ - -'-'--''~
:.
.
....
1200
Fig. 6. Type 3 activity. In all dot displays, the upper 10 indicate the neuronal activities during the task reinforced by water and the lower 3 indicate the activities during the task reinforced by ICSS. In C, artefacts of ICSS are displayed in dots and histograms contain these artefacts•
275
A
B Fig. 7. Prefrontal neuronal responsesto SI stimulation. A: antidromic response. These show 10 superimposed traces at 10-ms interval double shocks (1.5 mA). Latency was 2.4 ms. B: orthodromic responses. These show 15 superimposedtraces at 10-ms interval double shocks (1.5 mA). These responses in the figure were triggered by first pulse of double pulse stimulation. These responses include early excitation (25-50 ms) followedby inhibition (50-100 ms).Callibrations: A, 1 ms, 200/~V; B, 20 ms, 200~V. and movement-related neurons, the rewarding brain stimulation may activate these movement-related neurons and facilitate the generation of the behavior, because the reward strength or response rate increased proportionally to the increase in current intensity of ICSS1, s. In the present study, Type 1 neurons were thought to belong to the movement-related neurons, because they were activated only around the time of movement execution and were not active after the handle entering the target position and because these activity patterns were similar to those of the bidirectional type in the choice task of the previous study 10. About 40% of these Type 1 neurons responded to SI stimulation. Though the percentage of Type 1 neurons which responded to SI stimulation was lower than those of Type 2 and Type 3 neurons, some Type 1 neurons (16%) did respond orthodromically to SI stimulation. About 40% of the orthodromicaUy activated task-related neurons belonged to Type 1. This result indicates the possibility that ICSS facilitates the generation of the behavior
through movement-related neurons which respond orthodromically. However, in this study, other types of prefrontal neurons responded orthodromically to SI stimulation. Thus another possibility may be that ICSS activates various types of neurons in the prefrontal cortex and such general activation of the prefrontal cortex by ICSS may facilitate the generation of the behavior. Many of the Type 2 neurons, whose activity increased 300-500 ms prior to reward presentation and ceased at reward presentation, were activated by SI stimulation. These Type 2 activities were similar to Sakai's gradually increasing activity during the waiting period before the GO signal 20, Niki and Watanabe's type IIIa and IIIb activities ~7, and Komatsu's Type B activity8. These increasing activities presumably reflected 'expectancy'S,20 or 'anticipatory change '17. In Type 2 neurons, the gradually increasing activity ceased at the reward presentation. This increasing activity appeared with ICSS. As shown in Fig. 5, the degree of activation was different in accordance with the nature of the reward (electrical stimulation or a drop of water). Therefore, these gradually increasing activities prior to reward presentation may reflect the expectancy of the reward or anticipatory change to the reward presentation. Most Type 2 neurons responded antidromically to SI stimulation. Accordingly, these Type 2 neurons may send anticipatory information about the reward to the positive site of ICSS from the prefrontal cortex. Routtenberg and Santos-Anderson 19 pointed out that the efferent fibers from the prefrontal cortex to other ICSS positive sites were important in the generation of ICSS behavior. Therefore, Type 2 neurons with the antidromic response and anticipatory activity to the reward presentation may play an important role in the generation of ICSS behavior. Type 3 neuronal activity was similar to Niki and Watanabe's type IIIc activity17, Fuster's C or D neurons 3, and Komatsu's Type C activityS. However, Type 3 activity differed from these studies in that Niki and Watanabe's and Fuster's activities were recorded during the delay phase without motor performance, and in that the beginning and termination of the increased activity of Type 3 neurons were relatively time locked to the GO signal presentation and the reward delivery, respectively. These Type 3 neurons might be activated, at least partially, by visual
276 inputs, because it has been r e p o r t e d that many neurons in the prefrontal cortex receive visual inputs8,11A4. H o w e v e r , these T y p e 3 neurons did not respond to the light signal at the time when the handle entered the starting zone. T y p e 3 activity may relate to both m o t o r p e r f o r m a n c e and r e w a r d presentation, because this type of neuron was activated by the specific stimulus related to the m o v e m e n t execution (the G O signal) and activated tonically during the task m o v e m e n t and because the termination of increasing activity was time locked to the reward presentation. Most of the Type 3 neurons r e s p o n d e d orthodromically to SI stimulation. This result also indicates the possibility that ICSS facilitates the generation of the behavior through o r t h o d r o m i c a l l y activated movement-related neurons. Thus, prefrontal neurons r e s p o n d e d antidromically or o r t h o d r o m i c a l l y to the r e w a r d stimulation. The
neurons which showed different types of activity during the visual tracking task showed different types of responses. These results show that functionally different neurons have different anatomical relations to SI. Though fewer m o v e m e n t - r e l a t e d neurons responded to SI stimulation than o t h e r types of neurons did, some m o v e m e n t - r e l a t e d neurons r e s p o n d e d orthodromically. This indicates the possibility that ICSS facilitates the generation of the behavior through the afferent pathway to the prefrontal cortex. ACKNOWLEDGEMENT I would like to express my thanks to Dr. K. K u b o t a for his guidance and e n c o u r a g e m e n t throughout the course of this study.
REFERENCES 1 Ettenberg, A., Conditioned taste preference and response rate as measures of brain-stimulafion reward: a comparison, Physiol. Behav., 24 (1980) 755-758. 2 Funahashi, S. and Kubota, K., Prefrontal neuronal activity during a visual tracking task and substantia innominata stimulation, Neurosci. Len., Suppl. 9 (1982) $62. 3 Fuster, J. M., Unit activity in prefrontal cortex during delayed-response performance: neuronal correlates of transient memory, J. Neurophysiol., 36 (1973) 61-78. 4 German, D. C. and Fetz, E. E., Responses of primate locus coeruleus and subcoeruleus neurons to stimulation at reinforcing brain sites and to natural reinforcers, Brain Research, 109 (1976) 497-514. 5 Hawkins, T. D. and Pliskoff, S. S., Brain-stimulafion intensity, rate of self-stimulation, and reinforcement strength: an analysis through chaining, J. exp. Anal. Behav., 7 (1964) 285-288. 6 Jones, E. G., Burton, H., Saper, C. B. and Swanson, L. W., Midbrain, diencephalic and cortical relationship of the basal nucleus of Meynert and associate structures in primates, J. comp. Neurol., 167 (1976) 385--420. 7 Kievit, J. and Kuypers, H. G. J. M., Basal forebrain and hypothalamic connections to frontal and parietal cortex in the rhesus monkey, Science, 187 (1974) 660--662. 8 Komatsu, H., Prefrontal unit activity during a color discrimination task with GO and NO-GO responses in the monkey, Brain Research, 244 (1982) 269-277. 9 Kubota, K., Neuron activity in the dorsolateral prefrontal cortex of the monkey and initiation of behavior. In M. Ito, N. Tsukahara, K. Kubota, and K. Yagi (Eds.), Integrative Control Functions of the Brain, Vol. I, Kohdansha-Elsevier, Amsterdam, 1978, pp. 407-417. 10 Kubota, K. and Funahashi, S., Direction specific activities of dorsolateral prefrontal and motor cortex pyramidal tract neurons during visual tracking, J. Neurophysiol., 47 (1982) 362-376.
11 Kubota, K., Iwamoto, T., and Suzuki, H,, Visuokinetic activities of primate prefrontal neurons during delayed-response performance, J. Neurophysiol., 37 (1974) 11971212. 12 Kubota, K. and Niki, H., Prefrontal cortical unit activity and delayed alternation performance in monkey, J. Neurophysiol., 34 (1971) 337-347. 13 Kubota, K. and Noda, T., Visually activated neurons of monkeys and substantia innominata self-stimulation, Neurosci. Lett., Suppl. 4 (1980) $61. 14 Kubota, K., Tonoike, M., and Mikami, A., Neuronal activity in the monkey dorsolateral prefrontal cortex during a discrimination task with delay, Brain Research, 183 (1980) 29-42. 15 Matsunami, K., Funahashi, S. and Kubota, K., Radioactive 2-deoxyglucose uptake into brains of monkeys performing intracranial self-stimulation task, J. Physiol. Soc. Ipn., 44 (1982) 405. 16 Niki, H., Prefrontal unit activity during delayed alternation in the monkey. I. Relation to direction of response, Brain Research, 68 (1974) 185--196. 17 Niki, H. and Watanabe, M., Prefrontal and cingulate unit activity during timing behavior in the monkey, Brain Research, 171 (1979) 213-224. 18 Roils, E. T., Burton, M. J. and Mora, F., Neurophysiological analysis of brain-stimulation reward in the monkey, Brain Research, 194 (1980) 339-357. 19 Routtenberg, A. and Santos-Anderson, R., The role of prefrontal cortex in intracranial self-stimulation: a case history of anatomical localization of motivational substrates. In L. L. Iversen, S. D, Iversen, and S, D. Snyder (Eds.), Handbook of Psychopharmacology, Vol. 8, Plenum, New York, 1977, pp. 1-24. 20 Sakai, M., Prefrontal unit activity during visually guided lever pressing reaction in the monkey, Brain Research, 81 (1974) 297-309.