Brain Research Bulletin,
0361.9230191 $3.00 + .oO
Vol. 26, Pp. 583-592. 0 Pergamon Press plc, 1991. Prmted in the U.S.A
Neuronal Activity in Visual, Auditory and Polysensory Areas in the Monkey Temporal Cortex During Visual Fixation Task JOJI WATANABE
Department
visual
fixation
IWAI
of Behavioral Physiology, Tokyo Metropolitan Institute for Neurosciences 2-6 Musashidai, Fuchu-City, Tokyo 183, Japan
Received
WATANABE,
AND EIICHI
I8 June
1990
J. AND E. IWAI. Neuronal acfivit~ in visual. auditory and polyensory areas in the monkey temporalcwtex during tusk. BRAIN RES BULL 26(4) 583-592, 1991, --The activity of 252 neurons in the inferotemporal visual area TEO,
the superior temporal auditory area (AA), and the superior temporal polysensory area (STP) during the performance of a visual spot-fixation task and two variations, blink and tone tests, was examined in two behaving monkeys. A considerable number of not only TEO cells (45%) but also AA (29%) and STP (34%) cells were activated during the spot-fixation task. but unresponsive to the blanking of the spot during the fixation stage in the blink test. In addition. it was found that the activity of a third of the TEO. AA and STP cells which fired during the task-start stage in the spot-fixation task was modulated by cross-interaction between spot and tone simultaneously presented in the tone test: among these. the spot-induced activity of all TEO cells was enhanced by the tone. whereas the spot-induced activity of all AA and STP cells was suppressed by the tone. These findings are discussed in relation to the process of attending selectively to a fixation-spot. Inferotemporal area TEO Superior temporal auditory Attention Visual fixation task Monkey
area
Superior temporal polysensory
THE inferotemporal (IT) and superior temporal [which will be referred to as area AA (31)] cortices are known to be involved modality-specifically in higher order visual and auditory functions, respectively (12, 14, 16, 25, 44). The posterior IT area (PIT) (10, 18. l9), or the cytoarchitectonic areaTE (2, 21, 43). has been shown to be involved in visual perception (17, 20, 47). In addition, this TEO area is thought to be concerned with visual selective attention (8, 10, 26, 29), important in the perception of visual stimuli. The AA area is also thought to be involved in selective attention to an auditory stimulus in the environment (4-6, 24). The upper bank and bottom of the superior temporal sulcus (STS) makes up the polysensory area STP (3, 9, 11, 15), and it is thought that many STP cells are involved in orientation and attention to visual stimuli (9). In a visual fixation task. animals are required to fixate on a tiny light-spot until the spot is subsequently dimmed. Such a fixation task and its variations have been used in behaving monkeys to investigate the role of frontal, parietal and prelunate (area V4) cells in attention mechanisms (13, 27, 37, 41). As described above, areas TEO, AA and STP are all thought
area STP
to be involved in the mechanism of attention. Thus, not only area TEO but also areas AA and STP would play a significant role even in attending to a particular visual stimulus among the batches of information of various modalities in the environment. To our knowledge, however, there has been only one report (23) on the activity of TEO cells during the performance of a visual fixation task, and none for AA and STP cells. The first aim of the present study was to find out whether AA and STP cells are activated by a visual fixation task, and then to compare their activity with that of TEO cells. As a result, a considerable number of AA cells as well as STP cells were activated by a fixation-spot. Thus, secondly, we investigated the cross-interaction of visual and auditory stimuli in TEO, AA and STP cells by presenting spot and tone simultaneously. METHOD
Animals The animals used were two male adult rhesus monkeys (Mucaca mulurta). weighing 5.3 and 6. I kg. They normally lived
583
WATANABE
584
A spot-fixation task
Tone spot Lever
where the spot had been presented. It was judged erroneous when the eye moved more than I” during the blanking of the spot. Tone rest (Fig. ICi. For the tone test. the only difference from
fi i I-T
Reward
BBlillk test
cht?ne test
fl
Tone
fl
spot
I-l-~
Tone spot Lever Reward
AND IWAl
the fixation task was that the tone and fixation-spot were presented simultaneously. After learning each task to a criterion of 8.5% accuracy, the animals were retrained to learn these three tasks presented in an intermingled order. Surgical Preparation
n I I n
FIG. 1. Diagmnunatic line-illustrations of the fixation task (A) and its two variations, the blink (B) and tone (C) tests. Deflections in the lines indicate the timing of stimulus applications: upward and downward deflections, for presentation and extinction, respectively, of tone and spot, with a halfway deflection indicating dimming of the spot; upward and downward deflections, for pressing and release of the lever, respectively; and for the reward, upward deflection indicates presentation of juice.
in individual cages, but were seated in primate chairs during training and recording of neuronal activity. Tasks and Training Procedures
The animals were trained for three tasks, a spot-fixation task (46) and two modified tasks incorporating blink (36.42) and tone tests. The blink test was designed to investigate whether cells were activated by receiving information of a spot of retinal origin, i.e., whether cells were concerned with visual sensory information processing or with a~ention process to a ~xation-sat (42). The tone test was used to investigate whether the response to a spot was modulated by a tone. The sequence of events in each behavioral paradigm is shown in Fig. 1. Spot-&ation task (Fig. IA). This task was the standard one. Firstly, a brief tone of 700 Hz was presented for 225 ms, upon which the animals were required to press a lever. Then, a fixation-spot was presented 300 to 1000 ms after the lever press. The spot was a yellow light emitting diode of 10 cd/m’ covered by a black plate with a hole of 0.05” in diameter, and background illumination was 1.5 cd/m’. The animals were required to hold the lever and to fixate the spot, which was presented for 1.5-4.5 s, until the spot was dimmed from 10 to 6 cd/m2. The dimmed spot, upon which the animals had to release the lever quickly, was presented for 380 ms. When the animals failed to do this within 380 ms, the test was reset without punis~ent. Success was rewarded with a drop of juice. The intertrial interval (ITI) was 3.55.5 s. The fixation on the spot was monitored with an electrooculogram (EGG), and it was judged erroneous if the eye moved more than lo during the fixation period. Blink test (Fig. IB). The blink and tone tests were variations of the standard fixation task. For the blink test, the only difference was that the fiition-spot was extinguished for 500 ms in the middle of the fixation period. However, the animals were required to hold the lever continuously and to fixate the position
After learning the tasks, the animals underwent surgery for chronic, single unit recording. The animals were fitly given intramuscular injection of ketamine hy~~~o~de (Ketalar, I5 mg/ kg) and atropine sulfate (0.1 mg), and then they were deeply anesthetized with intraperitoneal injection of sodium pentobarbital (35 mg/kg). During surgery, heart and breathing rates were constantly monitored, and to maintain deep anesthesia, Ketalar (10 mg/kg) was given when the heart and breathing rates increased. In addition, a local anesthetic (Xylocaine) was infiltrated into the sites of skin incision and pressure points. Bolts for fixing the head were buried under the skull, and recording chambers were attached to the lateral skull. Unit activity was recorded from areas TEO, STP and AA. To monitor horizontal and vertical eye movements, two pairs of silver-silver chloride EOG electrodes (7) were buried around the orbit. Data Co~~ec~ia~and Analysis
A few days after the animals recovered from the implant surgery, chronic single unit recording sessions for unilateral hemisphere were conducted for about 2 months. The daily recording sessions usually lasted for 4-6 hours. Glass-coated Elgiloy electrodes with a tip exposure of about 20 p,rn were advanced with a hydraulic micromauipulator (MO-90, Narishige) to record single neuronal activity. A unit activity was identified with monitoring superimposed waveforms triggered by the pulses from a window discriminator, and was fed to a computer (DPS-1, InterSystems), which also received the signals of EGGS, tone-on/off, spot-on/ off, and lever press/release. The data were also recorded on magnetic tape for further analyses. For a given cell, three tasks were presented in a randomized order: the number of all three test trials was usually 90-120 in about half an hour. Data were collected only as long as the unit’s isolation remained good, its maintained discharges (spontaneous discharges during the ITI) and taskevoked activity remained stable, and behavioral performance remained constant. Data from erroneous trials were excluded from the analysis. Peristimulus time histogram (PSTH) for each task was usually compiled for 30-40 trials. The mean rates of maintained discharges of TEO, STP and AA cells were 7.2, 6.8 and 7.5 spikes/s, respectively. However, we sometimes observed cells showing low and irregular rates of spontaneous discharges, and such cells were neither tested further with the tasks nor taken into account of examined cells. The activity and maintained discharges of a single cell were stable throughout the testing and recording period for that cell [see Fig. 3A; (32,33)]. It was found in this study that the activity of most cells occurred transiently during the performance of the tasks. Therefore, neuronal discharges were counted and evaluated using an activity index (AI) by comparison with the maintained dischatge during the IT1 period, Erom 800 ms after the disappearance of the fixation-spot until the tone presentation in next test trial. The AI value was calculated by the following formula, which is modified from that of Mountcastle et al. (28):
585
TEO. AA AND STP NEURONS IN FIXATION TASK
AI =
-
where M, and M, denote mean firing rates for the samples a and b to be compared, respectively, N, and Nbr the number of trials for a and b, and S, and S,, standard deviations of a and b. The activity during the task was judged as significant when the AI value was greater than 3. this value corresponding to 0.005 in two-tailed t-test.
animals were deeply anesthetized with sodium ~ntobarbitai (45 mgikg) and perfused with normal saline followed by 10% paraformaldehyde with 2% ferrocyanide. The brain was frozen and sectioned at 75 km. and the sections were stained with cresyl violet. Almost all recording penetrations were identified, and their locations were dete~ined by histological reconst~ction (Fig. 2). It should be noted here that the location of area STP in this study was defined after the delineation of the STP of Bruce et al. (91, which is physiologic~ly determined. Thus, it included the dorsal bank as well as the dorsal lip of the STS, which is a ventral part of the cytoarchitectonic area Ts in the auditory cortex (30). RESULTS
In this study. the brief tone as signal for the lever press was presented about 500-1200 ms prior to the onset of spot in the spot-fixation and blink tasks (Fig. IA and B). In addition, as will be described in the Results section, the activity induced by the tone was transient and did not last until the spot presentation. Thus, the same cells could be examined and classified in two ways according to the presence or absence of activity induced by the tone alone and by the spot alone. TWO&asses ctfneurons based art t5~e-ev~ked activity (see Table 2). Cells responsive and unresponsive to the tone, respectively, were termed f~~e-~~~~~~~~~~v~ and ~~r~e_~F~~e~~~~~j~~e cells, regardless of the classification of spot-responsive and -unresponsive cells. Three classes of neurons based o~~~at~~n spot-induced activity (see Table Ii. Neurons evoked by the spot were classified into
three groups according to whether they were activated during the spot-fixation task and/or blink test, regardless of the classification of tone-responsive and -unresponsive cells. Of the cells activated during the spot-fixation task. those unresponsive to the blink test were classified as spot-fixation t~sk”r~lut~d cells, while those responsive to the blink test were termed ~~i~k-res~unsi~~e cells. Cells which were not activated by the sit-~xation task were defined as rus~-~~~~e~~u~~~~~e cells, To distinguish task-related cells more strictly, however, blink-responsive cells were defined as cells with an AI value of greater than 2 by ~omp~son with the discharge during the spot-~xation period, this value corresponding to the level of 0.05 in r-test. Seven t)f>es ~~‘s~ot~~~at~#n t~sk~r~lated neurons g’see Table I atzd Fz’gs. 3 and 4). It was observed that the activity of most spot-fixation task-related cells occurred transiently during task performance. Thus. in this study, the spot-fixation task was arbitrarily divided into three stages. The first stage was the tusk-start period, and this stage was defined as the period of 500 ms after the fixation-spot was turned on. The second stage was the @xution period, the period from 500 ms afrer turning on the spot until spot~dimmin~. The third stage was the end period of the task, which lasted 700 ms after spot-dimming. The presence of cell activity during the start and end stages was judged when their AI values were greater than 3 by comparing with the discharge during the fixation stage and also with the m~ntained discharge. According to the stage or stages during which cells fired, the spot-fixation task-related cells were further classified into seven types: I ) start, 2) fixation, 3) end, 4) start+ fixation, 5) start+ end, 6) fixation + end and 7) start + fixation + end types.
To estimate the locations of the units recorded, 3.5 volt DC was applied for 2 min at a few depths along an electrode track at the end of recording, so that iron deposits were made in the tissue. After completion of recording from both hemispheres. the
Two hundred and fifty-two neurons from 29 electrode penetrations into four cerebral hemispheres of two monkeys were examined (Fig. 2). From the histology, it was established that 83 cells were recorded from area TEO in the middle temporal gyms. The recording sites of 98 cells were located in a middle third of the crown of the superior temporal gyrus (area AA), and those of 71 cells in the superficial half of the upper bank of the superior temporal sulcus (area STP).
Of 252 cells examined, nearly half (46%) were activated during the ~~o~an~e of the sit-fixation task. The incidence of task-related, blink-responsive and task-unresponsive cells in areas TEO, STP and AA is shown in Table 1, and the PSTHs of representative task-related cells of the seven types are shown in Figs. 3 and 4, respectively. In the standard spot-fixation task, the onset of the fixationspot and saccadic eye movement to the spot occurred during the start stage. while the sat-djmmiug, cessation of eye fixation, the delivery of juice as reward, lever release and spot extinction occurred during the end stage, with no events during the fixation stage (Fig. 1). However, as may be seen in EOG in Fig. 3A. eye fixation to the spot occurred at 3~~ ms after the spot onset. and the activity of most cells began prior to the eye fixation and was time-locked to the onset of the spot. Eye fixation continued until 400-550 ms after the sit-dimming. and the activity of most cells disappeared before the end of the eye fixation. Thus. the activity of these cells did not depend on eye movement. In addition, although several events occurred within a short period in the task-end stage. the PST& of most end type cells appeared to be time-locked to the spot-dimming. Spot-fixation task-related neurons. As may be seen in Fig. 3A and 3, sit-fixation task-related cells were unres~ns~ve to the blink test; their activity was not changed by the disappearance and reappearance of the spot in the blink test. The dis~ibutions of task-related (hatched bars) and task-unresponsive cells (open bars) in areas TEO, STP and AA as a function of their AI values are shown in Fig. 5. The mean values of the AIs of task-related TEO, STP and AA cells were 6.8,4.5 and 4.3, respectively, and the mean value of TEO cells was significantly greater than those of STP and AA cells (r-test, r=4.2, p
0361.9230191 $3.00 + .oO
Vol. 26, Pp. 583-592. 0 Pergamon Press plc, 1991. Prmted in the U.S.A
Neuronal Activity in Visual, Auditory and Polysensory Areas in the Monkey Temporal Cortex During Visual Fixation Task JOJI WATANABE
Department
visual
fixation
IWAI
of Behavioral Physiology, Tokyo Metropolitan Institute for Neurosciences 2-6 Musashidai, Fuchu-City, Tokyo 183, Japan
Received
WATANABE,
AND EIICHI
I8 June
1990
J. AND E. IWAI. Neuronal acfivit~ in visual. auditory and polyensory areas in the monkey temporalcwtex during tusk. BRAIN RES BULL 26(4) 583-592, 1991, --The activity of 252 neurons in the inferotemporal visual area TEO,
the superior temporal auditory area (AA), and the superior temporal polysensory area (STP) during the performance of a visual spot-fixation task and two variations, blink and tone tests, was examined in two behaving monkeys. A considerable number of not only TEO cells (45%) but also AA (29%) and STP (34%) cells were activated during the spot-fixation task. but unresponsive to the blanking of the spot during the fixation stage in the blink test. In addition. it was found that the activity of a third of the TEO. AA and STP cells which fired during the task-start stage in the spot-fixation task was modulated by cross-interaction between spot and tone simultaneously presented in the tone test: among these. the spot-induced activity of all TEO cells was enhanced by the tone. whereas the spot-induced activity of all AA and STP cells was suppressed by the tone. These findings are discussed in relation to the process of attending selectively to a fixation-spot. Inferotemporal area TEO Superior temporal auditory Attention Visual fixation task Monkey
area
Superior temporal polysensory
THE inferotemporal (IT) and superior temporal [which will be referred to as area AA (31)] cortices are known to be involved modality-specifically in higher order visual and auditory functions, respectively (12, 14, 16, 25, 44). The posterior IT area (PIT) (10, 18. l9), or the cytoarchitectonic areaTE (2, 21, 43). has been shown to be involved in visual perception (17, 20, 47). In addition, this TEO area is thought to be concerned with visual selective attention (8, 10, 26, 29), important in the perception of visual stimuli. The AA area is also thought to be involved in selective attention to an auditory stimulus in the environment (4-6, 24). The upper bank and bottom of the superior temporal sulcus (STS) makes up the polysensory area STP (3, 9, 11, 15), and it is thought that many STP cells are involved in orientation and attention to visual stimuli (9). In a visual fixation task. animals are required to fixate on a tiny light-spot until the spot is subsequently dimmed. Such a fixation task and its variations have been used in behaving monkeys to investigate the role of frontal, parietal and prelunate (area V4) cells in attention mechanisms (13, 27, 37, 41). As described above, areas TEO, AA and STP are all thought
area STP
to be involved in the mechanism of attention. Thus, not only area TEO but also areas AA and STP would play a significant role even in attending to a particular visual stimulus among the batches of information of various modalities in the environment. To our knowledge, however, there has been only one report (23) on the activity of TEO cells during the performance of a visual fixation task, and none for AA and STP cells. The first aim of the present study was to find out whether AA and STP cells are activated by a visual fixation task, and then to compare their activity with that of TEO cells. As a result, a considerable number of AA cells as well as STP cells were activated by a fixation-spot. Thus, secondly, we investigated the cross-interaction of visual and auditory stimuli in TEO, AA and STP cells by presenting spot and tone simultaneously. METHOD
Animals The animals used were two male adult rhesus monkeys (Mucaca mulurta). weighing 5.3 and 6. I kg. They normally lived
583
WATANABE
584
A spot-fixation task
Tone spot Lever
where the spot had been presented. It was judged erroneous when the eye moved more than I” during the blanking of the spot. Tone rest (Fig. ICi. For the tone test. the only difference from
fi i I-T
Reward
BBlillk test
cht?ne test
fl
Tone
fl
spot
I-l-~
Tone spot Lever Reward
AND IWAl
the fixation task was that the tone and fixation-spot were presented simultaneously. After learning each task to a criterion of 8.5% accuracy, the animals were retrained to learn these three tasks presented in an intermingled order. Surgical Preparation
n I I n
FIG. 1. Diagmnunatic line-illustrations of the fixation task (A) and its two variations, the blink (B) and tone (C) tests. Deflections in the lines indicate the timing of stimulus applications: upward and downward deflections, for presentation and extinction, respectively, of tone and spot, with a halfway deflection indicating dimming of the spot; upward and downward deflections, for pressing and release of the lever, respectively; and for the reward, upward deflection indicates presentation of juice.
in individual cages, but were seated in primate chairs during training and recording of neuronal activity. Tasks and Training Procedures
The animals were trained for three tasks, a spot-fixation task (46) and two modified tasks incorporating blink (36.42) and tone tests. The blink test was designed to investigate whether cells were activated by receiving information of a spot of retinal origin, i.e., whether cells were concerned with visual sensory information processing or with a~ention process to a ~xation-sat (42). The tone test was used to investigate whether the response to a spot was modulated by a tone. The sequence of events in each behavioral paradigm is shown in Fig. 1. Spot-&ation task (Fig. IA). This task was the standard one. Firstly, a brief tone of 700 Hz was presented for 225 ms, upon which the animals were required to press a lever. Then, a fixation-spot was presented 300 to 1000 ms after the lever press. The spot was a yellow light emitting diode of 10 cd/m’ covered by a black plate with a hole of 0.05” in diameter, and background illumination was 1.5 cd/m’. The animals were required to hold the lever and to fixate the spot, which was presented for 1.5-4.5 s, until the spot was dimmed from 10 to 6 cd/m2. The dimmed spot, upon which the animals had to release the lever quickly, was presented for 380 ms. When the animals failed to do this within 380 ms, the test was reset without punis~ent. Success was rewarded with a drop of juice. The intertrial interval (ITI) was 3.55.5 s. The fixation on the spot was monitored with an electrooculogram (EGG), and it was judged erroneous if the eye moved more than lo during the fixation period. Blink test (Fig. IB). The blink and tone tests were variations of the standard fixation task. For the blink test, the only difference was that the fiition-spot was extinguished for 500 ms in the middle of the fixation period. However, the animals were required to hold the lever continuously and to fixate the position
After learning the tasks, the animals underwent surgery for chronic, single unit recording. The animals were fitly given intramuscular injection of ketamine hy~~~o~de (Ketalar, I5 mg/ kg) and atropine sulfate (0.1 mg), and then they were deeply anesthetized with intraperitoneal injection of sodium pentobarbital (35 mg/kg). During surgery, heart and breathing rates were constantly monitored, and to maintain deep anesthesia, Ketalar (10 mg/kg) was given when the heart and breathing rates increased. In addition, a local anesthetic (Xylocaine) was infiltrated into the sites of skin incision and pressure points. Bolts for fixing the head were buried under the skull, and recording chambers were attached to the lateral skull. Unit activity was recorded from areas TEO, STP and AA. To monitor horizontal and vertical eye movements, two pairs of silver-silver chloride EOG electrodes (7) were buried around the orbit. Data Co~~ec~ia~and Analysis
A few days after the animals recovered from the implant surgery, chronic single unit recording sessions for unilateral hemisphere were conducted for about 2 months. The daily recording sessions usually lasted for 4-6 hours. Glass-coated Elgiloy electrodes with a tip exposure of about 20 p,rn were advanced with a hydraulic micromauipulator (MO-90, Narishige) to record single neuronal activity. A unit activity was identified with monitoring superimposed waveforms triggered by the pulses from a window discriminator, and was fed to a computer (DPS-1, InterSystems), which also received the signals of EGGS, tone-on/off, spot-on/ off, and lever press/release. The data were also recorded on magnetic tape for further analyses. For a given cell, three tasks were presented in a randomized order: the number of all three test trials was usually 90-120 in about half an hour. Data were collected only as long as the unit’s isolation remained good, its maintained discharges (spontaneous discharges during the ITI) and taskevoked activity remained stable, and behavioral performance remained constant. Data from erroneous trials were excluded from the analysis. Peristimulus time histogram (PSTH) for each task was usually compiled for 30-40 trials. The mean rates of maintained discharges of TEO, STP and AA cells were 7.2, 6.8 and 7.5 spikes/s, respectively. However, we sometimes observed cells showing low and irregular rates of spontaneous discharges, and such cells were neither tested further with the tasks nor taken into account of examined cells. The activity and maintained discharges of a single cell were stable throughout the testing and recording period for that cell [see Fig. 3A; (32,33)]. It was found in this study that the activity of most cells occurred transiently during the performance of the tasks. Therefore, neuronal discharges were counted and evaluated using an activity index (AI) by comparison with the maintained dischatge during the IT1 period, Erom 800 ms after the disappearance of the fixation-spot until the tone presentation in next test trial. The AI value was calculated by the following formula, which is modified from that of Mountcastle et al. (28):
585
TEO. AA AND STP NEURONS IN FIXATION TASK
AI =
-
where M, and M, denote mean firing rates for the samples a and b to be compared, respectively, N, and Nbr the number of trials for a and b, and S, and S,, standard deviations of a and b. The activity during the task was judged as significant when the AI value was greater than 3. this value corresponding to 0.005 in two-tailed t-test.
animals were deeply anesthetized with sodium ~ntobarbitai (45 mgikg) and perfused with normal saline followed by 10% paraformaldehyde with 2% ferrocyanide. The brain was frozen and sectioned at 75 km. and the sections were stained with cresyl violet. Almost all recording penetrations were identified, and their locations were dete~ined by histological reconst~ction (Fig. 2). It should be noted here that the location of area STP in this study was defined after the delineation of the STP of Bruce et al. (91, which is physiologic~ly determined. Thus, it included the dorsal bank as well as the dorsal lip of the STS, which is a ventral part of the cytoarchitectonic area Ts in the auditory cortex (30). RESULTS
In this study. the brief tone as signal for the lever press was presented about 500-1200 ms prior to the onset of spot in the spot-fixation and blink tasks (Fig. IA and B). In addition, as will be described in the Results section, the activity induced by the tone was transient and did not last until the spot presentation. Thus, the same cells could be examined and classified in two ways according to the presence or absence of activity induced by the tone alone and by the spot alone. TWO&asses ctfneurons based art t5~e-ev~ked activity (see Table 2). Cells responsive and unresponsive to the tone, respectively, were termed f~~e-~~~~~~~~~~v~ and ~~r~e_~F~~e~~~~~j~~e cells, regardless of the classification of spot-responsive and -unresponsive cells. Three classes of neurons based o~~~at~~n spot-induced activity (see Table Ii. Neurons evoked by the spot were classified into
three groups according to whether they were activated during the spot-fixation task and/or blink test, regardless of the classification of tone-responsive and -unresponsive cells. Of the cells activated during the spot-fixation task. those unresponsive to the blink test were classified as spot-fixation t~sk”r~lut~d cells, while those responsive to the blink test were termed ~~i~k-res~unsi~~e cells. Cells which were not activated by the sit-~xation task were defined as rus~-~~~~e~~u~~~~~e cells, To distinguish task-related cells more strictly, however, blink-responsive cells were defined as cells with an AI value of greater than 2 by ~omp~son with the discharge during the spot-~xation period, this value corresponding to the level of 0.05 in r-test. Seven t)f>es ~~‘s~ot~~~at~#n t~sk~r~lated neurons g’see Table I atzd Fz’gs. 3 and 4). It was observed that the activity of most spot-fixation task-related cells occurred transiently during task performance. Thus. in this study, the spot-fixation task was arbitrarily divided into three stages. The first stage was the tusk-start period, and this stage was defined as the period of 500 ms after the fixation-spot was turned on. The second stage was the @xution period, the period from 500 ms afrer turning on the spot until spot~dimmin~. The third stage was the end period of the task, which lasted 700 ms after spot-dimming. The presence of cell activity during the start and end stages was judged when their AI values were greater than 3 by comparing with the discharge during the fixation stage and also with the m~ntained discharge. According to the stage or stages during which cells fired, the spot-fixation task-related cells were further classified into seven types: I ) start, 2) fixation, 3) end, 4) start+ fixation, 5) start+ end, 6) fixation + end and 7) start + fixation + end types.
To estimate the locations of the units recorded, 3.5 volt DC was applied for 2 min at a few depths along an electrode track at the end of recording, so that iron deposits were made in the tissue. After completion of recording from both hemispheres. the
Two hundred and fifty-two neurons from 29 electrode penetrations into four cerebral hemispheres of two monkeys were examined (Fig. 2). From the histology, it was established that 83 cells were recorded from area TEO in the middle temporal gyms. The recording sites of 98 cells were located in a middle third of the crown of the superior temporal gyrus (area AA), and those of 71 cells in the superficial half of the upper bank of the superior temporal sulcus (area STP).
Of 252 cells examined, nearly half (46%) were activated during the ~~o~an~e of the sit-fixation task. The incidence of task-related, blink-responsive and task-unresponsive cells in areas TEO, STP and AA is shown in Table 1, and the PSTHs of representative task-related cells of the seven types are shown in Figs. 3 and 4, respectively. In the standard spot-fixation task, the onset of the fixationspot and saccadic eye movement to the spot occurred during the start stage. while the sat-djmmiug, cessation of eye fixation, the delivery of juice as reward, lever release and spot extinction occurred during the end stage, with no events during the fixation stage (Fig. 1). However, as may be seen in EOG in Fig. 3A. eye fixation to the spot occurred at 3~~ ms after the spot onset. and the activity of most cells began prior to the eye fixation and was time-locked to the onset of the spot. Eye fixation continued until 400-550 ms after the sit-dimming. and the activity of most cells disappeared before the end of the eye fixation. Thus. the activity of these cells did not depend on eye movement. In addition, although several events occurred within a short period in the task-end stage. the PST& of most end type cells appeared to be time-locked to the spot-dimming. Spot-fixation task-related neurons. As may be seen in Fig. 3A and 3, sit-fixation task-related cells were unres~ns~ve to the blink test; their activity was not changed by the disappearance and reappearance of the spot in the blink test. The dis~ibutions of task-related (hatched bars) and task-unresponsive cells (open bars) in areas TEO, STP and AA as a function of their AI values are shown in Fig. 5. The mean values of the AIs of task-related TEO, STP and AA cells were 6.8,4.5 and 4.3, respectively, and the mean value of TEO cells was significantly greater than those of STP and AA cells (r-test, r=4.2, p
