| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 Department of Physiology, Tohoku University School of Medicine, Sendai 980-8575, Japan; 2 Core Research for Evolutional Science and Technology Program, Kawaguchi 332-0012, Japan
Submitted 14 April 2003; accepted in final form 5 August 2003
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONDISCLOSURESACKNOWLEDGMENTSREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONDISCLOSURESACKNOWLEDGMENTSREFERENCES |
|---|
; Rosenbaum 1991; Zingale and Kowler 1987). When organisms act to achieve an objective, they usually need to execute more than one action. Several individual actions need to be organized for purposeful behavior, and the incorrect ordering of actions can lead to disastrous consequences. Therefore the organization of a series of actions into an appropriate temporal sequence is particularly important in the voluntary control of behavior (Tanji 2001). In the present study, we addressed the cortical control of saccade sequences as an experimental model with which to elucidate the role of the cerebral cortex in organizing behavioral sequences.
Among the multiple motor areas within the cerebral cortex, two principal oculomotor fields have been established (Schall 1997), namely the frontal eye field (FEF) (Bizzi 1967; Bruce and Goldberg 1985; Stanton et al. 1989), located laterally, and the supplementary eye field (SEF) (Schlag and Schlag-Rey 1985, 1987), located medially. Although these two areas share several anatomical connections (Huerta and Kaas 1990; Huerta et al. 1987; Parthasarathy et al. 1992; Shook et al. 1990, 1991; Stanton et al. 1988a,b) and exhibit similar physiological properties (Bruce and Goldberg 1985; Russo and Bruce 2000; Schall 1991a,b; Schlag and Schlag-Rey 1987), some functional aspects are known to differ between the two areas. Based largely on studies examining effects of electrical stimulation, it has been proposed that the FEF codes saccades in an oculocentric coordinate system, whereas the SEF codes saccades in a craniocentric coordinate system (Bon and Lucchetti 1992; Mann et al. 1988; Mitz and Godschalk 1989; Schall 1991b; Tehovnik and Lee 1993; but see Russo and Bruce 1993, 1996). On the other hand, it is also hypothesized that SEF neurons code saccades with reference to an object-centered coordinate (Olson and Gettner 1995, 1999), although FEF neurons have not yet been tested under the same experimental condition. Other intriguing hypotheses propose that the SEF is more involved in the learning of stimulus-saccade associations (Chen and Wise 1995b), eye-hand coordination (Mushiake et al. 1996), anti-saccade production (Everling and Munoz 2000; Schlag-Rey et al. 1997), decision-making saccade processes (Coe et al. 2002), and supervisory control of sensorimotor processes (Hanes et al. 1998; Schall et al. 2002; Stuphorn et al. 2000).
Although we have now some knowledge about aspects of functional differences between the two areas, little is known about how differently these areas take part in organizing the temporal sequence of saccades (Isoda and Tanji 2002; Lu et al. 2002). Having considered previous reports indicating the involvement of medial cortical motor areas in sequential control of limb movements in both humans (Dick et al. 1986; Gaymard et al. 1990; Hikosaka et al. 1996; Kawashima et al. 1998; Muri et al. 1995; Petit et al. 1996; Roland et al. 1980; Shibasaki et al. 1993) and nonhuman primates (Clower and Alexander 1998; Mushiake et al. 1990, 1991; Nakamura et al. 1998, 1999; Schiller and Chou 1998; Shima and Tanji 1998, 2000; Tanji and Shima 1994), we hypothesized that neuronal activity in the SEF, more than the FEF, represents signals that are relevant to the temporal structuring of multiple saccades. To test this hypothesis, we analyzed and compared neuronal activity in the SEF and FEF while monkeys performed multiple saccades in sequences that were instructed and memorized. We found a variety of neuronal activity that appeared preferentially in each of the two areas. Based on our findings, we propose a hypothesis for the use of the SEF and FEF in the orderly performance of multiple saccades.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONDISCLOSURESACKNOWLEDGMENTSREFERENCES |
|---|
We used two male Japanese monkeys (Macaca fuscata, 5–6 kg). The monkeys were kept in individual primate cages in an air-conditioned room with food available ad libitum. They were given restricted amount of fluid during periods of training and recording. Their body weight and appetite were checked daily, and supplementary water and fruit were provided as appropriate. During experimental sessions, each monkey was seated in a primate chair and placed in a sound-attenuated room and faced a display monitor on which three potential saccade targets (red, 0.5° in diameter) were illuminated above (U), to the left (L), and to the right (R) of center (12° each). The experiments were performed while the monkey's head and arms were restrained, and its eye movements were monitored using an infrared monitor system (R-21C-A; RMS, Hirosaki, Japan), sampling at 250 Hz. For this purpose, we implanted nonmetallic head-fixation bolts and two head holders on the skull under deep pentobarbital sodium anesthesia. All surgical and experimental protocols were in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals and the guidelines for Institutional Animal Care and Use published by our institute.
Behavioral procedures
We used a behavioral task that has been described previously (Isoda and Tanji 2002). Using operant conditioning with positive reinforcement, monkeys were trained to perform three saccades to the peripheral lights in six different orders. The correct sequence was indicated during visually guided trials (Fig. 1). Before each saccade, the monkey fixated on a central, white fixation point (FP, 0.5° diam) for 1,450–1,600 ms. When the FP was extinguished (the go signal), one of three red peripheral lights turned white, thereby indicating a saccade target. If the animal acquired the target within 500 ms and maintained fixation on it for 
350 ms, the target color was changed to red, a high-pitch tone sounded, and the central FP was illuminated again. The animal was then required to capture the central FP within 1,000 ms (re-fixation). For capturing the FP, monkeys were allowed to make either single or multiple saccades with variable trajectories. In the same way, a period of central fixation was followed by the second and third saccade triggers. A series of correct saccades was rewarded with a drop of juice after the third target-fixation period. After four trials had been accomplished under visual guidance, the monkey was required to repeat the same sequence of three saccades to the peripheral lights (center-out saccades) based on memory. At this phase of the task (memory-guided trials), the three correct directions of saccades had to be remembered because the peripheral targets remained red while the central FP was extinguished (Fig. 1). After the monkey had completed four trials from memory, flashing lights (duration: 700 ms) signaled both the end of the current sequence and the beginning of the next sequence, followed by an intertrial interval (ITI) of 2,500 ms. Thus a particular sequence was performed in blocks, each of which comprised four trials under visual guidance followed by four trials using memory (Fig. 1). Six different sequences were presented in a pseudorandom order.
|
Training procedures
The monkeys were first trained to generate a single saccade to one of the three peripheral lights within 300 ms after the central FP was turned off. Saccades to each direction were repeated in a block of eight trials: four trials under visual guidance and then four trials from memory. When the performance of each center-out saccade was stable after 2 wk of training, the monkey learned to make two center-out saccades in a correct order with an intervening re-fixation eye movements to capture the FP (within 1,000 ms), followed by a period of central fixation. The order of two center-out saccades employed was U-L, U-R, R-L, and L-R; each order was repeated in blocks. In this task phase, the time restriction to acquire peripheral targets was relaxed to 500 ms, but the reaction time of saccades remained mostly within 250 ms. When performing the center-out saccades, the monkeys were required to acquire a correct target with a single saccade. Turn-around saccades were not allowed even if the correct target was acquired within 500 ms. This strict restriction was not applied to eye movements to recapture the central fixation target. The monkeys had only to come back to the FP within 1,000 ms, and multiple saccades with a variety of trajectories were allowed. The monkeys acquired stable performance of this two center-out saccade task after about 1 mo of training. Thereafter, the final stage of task training that required three center-out saccades was introduced. It took 5–6 mo for both monkeys to become proficient in performing three saccades in six different orders that were selected based on memory.
Neuronal recording
Neuronal activity was recorded with glass-insulated Elgiloy microelectrodes (1.0–2.5 M
at 333 Hz). An electrode was inserted through the dura using a hydraulic microdrive (MO-951S; Narishige, Tokyo, Japan). Neuronal activity was amplified and band-pass filtered (1–6 kHz) and single-unit potentials were discriminated using a template-matching algorithm (MCP-Plus and MSD; Alpha Omega Engineering, Nazareth, Israel). Behavioral events and raster displays of action potentials were displayed on-line and data were stored in a microcomputer for analysis off-line.
Recording sites
We recorded neuronal activity in the SEF and FEF. The localization of cortical sites was based on physiological criteria established previously (Bruce and Goldberg 1985; Bruce et al. 1985; Mushiake et al. 1996; Russo and Bruce 1996; Schlag and Schlag-Rey 1987). The SEF was defined as a cortical area in the dorsomedial convexity immediately lateral to the presupplementary motor area (pre-SMA) (Matsuzaka et al. 1992; Picard and Strick 1996), in which saccadic eye movements were elicited at a low threshold (<50 µA) of intracortical microstimulation (ICMS, a train of 12–50 cathodal pulses of 0.2-ms duration at 333 Hz). The FEF was defined as an area in the anterior bank of the arcuate sulcus in which low-threshold ICMS evoked saccades. The recording sites were confirmed by magnetic resonance imaging (OPART 3D-System; Toshiba, Tokyo, Japan).
Data analysis
Our initial aim was to determine the relationship of presaccadic activity to the direction, rank, and sequence of three center-out saccades during trials guided by memory. Eye movements required to bring the gaze back to the central FP was not included in this analysis because the re-fixation eye movements were not memory based and frequently contaminated with saccades having variable trajectories. For this purpose of analysis, we selected neurons that had task-related presaccadic activity. If the number of discharges during a presaccadic period (200–0 ms before saccade onset) was significantly different from that during a control period (the last 500 ms of ITI as a default time window) for at least one sequence (Mann-Whitney U test, 
= 0.05), the neuron was judged to be active during the presaccadic period. The selection of the time window for presaccadic activity was based on our finding that most part of phasic activity preceding saccades fell on that interval. Furthermore, we obtained similar results by selecting the interval of 100–0 ms preceding saccade initiation for the quantitative analysis. The time window of 200–0 ms was comparable to previous studies (Chen and Wise 1995a; Mushiake et al. 1996; Schlag-Rey et al. 1997). The selection of the last ITI for the control period, instead of a fixation epoch after the start of each trial, was based on the following reasons. Because each sequence of saccades was performed in blocks, the monkey had prior knowledge about which sequence was currently required before the initiation of the first saccade. This "cognitive set" occasionally led to changes in neuronal activity during the epoch of central fixation but not during ITI. Therefore we considered that ITI was more suitable as a control period. A separate analysis revealed that neuronal activity during the 500-ms ITI was not significantly influenced by the metrics of the first saccade or the sequence of saccades, lending support to the choice of the ITI period for the control.
Because we used only six different sequences out of all of the possible 27, we could not apply three-way ANOVA. Instead, we first performed two-way ANOVA for relatedness to the direction and rank ( = 0.01). The rank factor included three levels (1st, 2nd, and 3rd) and the direction factor included two levels (rightward and leftward), because upward saccades appeared only as the first saccade (see Fig. 1). In the next step of the analysis, we examined the effect of saccade sequences on neuronal activity. For this analysis, we selected five subgroups of data, each of which comprised saccades of equal direction and rank (see saccade category in Fig. 4). We then applied one-way ANOVA to each saccade category ( = 0.01). Significant ANOVA outcomes were further analyzed with Fisher's least significant difference (LSD) post hoc test for multiple comparisons ( = 0.05).
|
|
In these equations, Fpd and Fnpd refer to the firing rate preceding saccades in the preferred and nonpreferred directions. Fpr and Fnpr refer to the firing rate for saccades in the preferred and nonpreferred rank. Fps and Fnps refer to the firing rate for saccades in sequences in which neurons were most and least active. A sequence index was calculated for each of the five saccade categories described in the preceding text, each of which was treated as a separate case.
A separate analysis was carried out to detect selectivity in neuronal activity for trial numbers within each trial block (n = 8). For this analysis, we used all neurons sampled in the SEF (n = 443) and FEF (n = 259) and examined their activity during four epochs, specifically the 2,000 ms preceding the trial start (onset of the fixation period) and the three presaccadic periods. We performed one-way ANOVA on neuronal activity in each trial epoch to examine the effect of trial number ( = 0.01). Where appropriate, Fisher's LSD test was carried out to determine whether selectivity was determined by individual trial numbers ( = 0.05).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONDISCLOSURESACKNOWLEDGMENTSREFERENCES |
|---|
We studied 443 neurons in the SEF and 259 in the FEF that were isolated for long enough to perform a statistical analysis. Among these cells, 323 SEF and 167 FEF neurons exhibited presaccadic activity. We found a striking difference in the characteristics of presaccadic activity during the performance of memory-guided sequential tasks (Fig. 1). For half of the FEF neurons, presaccadic activity was direction selective, irrespective of whether a saccade was first, second, or third in a sequence of three saccades. By contrast, for the majority of SEF neurons, presaccadic activity was selective for, or influenced by, the position of a saccade within a sequence. Based on two-way ANOVA (see METHODS), presaccadic activity was classified into three major types according to whether activity was selective for saccade direction, the numerical position of a saccade within each sequence (rank), or a combination of direction and rank. The first type of activity was observed in 81 (49%) FEF and 60 (19%) SEF neurons. A typical example is shown in Fig. 2A, in which the presaccadic activity of an FEF neuron appears to be selective for the leftward saccade, irrespective of rank. The second type of activity was observed in 58 (18%) SEF and 8 (5%) FEF neurons in which presaccadic activity depended only on the saccade rank, regardless of direction. In an example of an SEF neuron exhibiting this type of activity (Fig. 2B), the activity was most prominent when the monkey made the first saccade, irrespective of the saccade direction. When the monkey made a second or third saccade, the neuron was much less active. The third type of activity was observed in 123 (38%) SEF and 25 (15%) FEF neurons in which presaccadic activity depended on both the direction and rank or the interaction between these two factors. Neuronal activity of this type is shown in Fig. 2C, where a neuron exhibited presaccadic activity that was preferential for both the direction (leftward saccades) and rank (2nd saccades), being most active before the second, leftward saccades. When we compared the distribution of each type of activity between the two cortical fields, we found that direction-selective neurons were significantly more prevalent in the FEF (P < 0.001, 
2 test), whereas rank-selective neurons and combination-selective neurons were observed more frequently in the SEF (P < 0.001, 2 test).
To quantify the selectivity of individual cells, we computed an index for direction selectivity and rank selectivity for presaccadic activity (see METHODS). A scatterplot of the direction selectivity index versus the rank selectivity index (Fig. 3A) revealed that the selectivity of SEF neurons was greater for rank than for direction (P < 0.001, Mann-Whitney U test), although both indices exhibited considerable variability. By contrast, for FEF neurons, selectivity of FEF neurons was greater for direction than rank (P < 0.001, Mann-Whitney U test). To compare the two indices in each of the two cortical areas more directly, we constructed cumulative frequency histograms (Fig. 3B). The rank index was significantly larger for SEF neurons (P < 0.001, Kolmogorov-Smirnov test; Fig. 3B, left). The difference in the distribution between the SEF and FEF for the direction index was smaller than for the rank index, but was significantly larger for FEF neurons (P < 0.05, Kolmogorov-Smirnov test; Fig. 3B, right).
|
Next, we examined the selectivity of presaccadic activity for the saccade sequence. For this analysis, neuronal activity preceding saccades of equal direction and rank were examined. This analysis was applicable to five subgroups of saccades. We used one-way ANOVA to examine each of the five subgroups to elucidate the effect of saccade sequence on neuronal activity (see METHODS). An example of sequence selectivity that preceded the first saccade is shown in Fig. 4A, in which activity was selective for an up-right-left (U-R-L) sequence. Although the monkey made upward saccades as the first saccade in four different sequences (U-L-L, U-L-R, U-R-R, and U-R-L), presaccadic activity was most prominent for the U-R-L sequence. Differences in activity could not be attributed to either the direction or rank of a saccade but rather reflected the specific order of saccades that were currently required to be performed. For neurons active prior to the saccade, we found that the activity of 69 (21%) SEF and 15 (9%) FEF neurons differed significantly, depending on the sequence of saccades. The proportion of sequence-selective neurons was significantly greater in the SEF than in the FEF (P < 0.001, 2 test). Two examples of sequence selectivity for the second saccade are shown in Fig. 4, B and C, for saccades directed to left and right, respectively. In Fig. 4. D and E, examples of sequence selectivity for the third saccade are shown. The presaccadic activity of these sequence-selective neurons changed significantly, depending on the sequences of saccades, even though the subject made saccades to the same direction for saccades in the same position within a sequence. We computed a sequence index for each of the five saccade subgroups (each of which is referred to as a case), using data from all neurons that were active prior to the saccade. This produced a total of 1,615 and 835 cases in the SEF and FEF, respectively. As is apparent in the cumulative frequency histogram shown in Fig. 5A, the distribution of the sequence index for SEF neurons lies to the right of that of the FEF neurons, indicating that SEF neurons had significantly larger indices (P < 0.001, Kolmogorov-Smirnov test).
|
We found that the sequence selectivity appeared at different phases of the saccade sequence in the SEF and FEF. In the SEF, sequence selectivity preceding the first saccades was more prevalent than sequence selectivity preceding the third saccades (Fig. 5B). By contrast, in the FEF, sequence selectivity was more prevalent for the third saccade rather than the first saccade. The distribution of the frequency of sequence selectivity preceding the first and third saccades differed significantly between the two areas (P < 0.001, 2 test).
To determine whether the selectivity of presaccadic activity for rank and sequence could be ascribed to differences in kinematic saccade parameters (reaction times, movement times, and peak saccade velocity), we carried out an analysis of covariance (ANCOVA) using each of these parameters as a covariate. None of the saccade parameters covaried with either rank or sequence in 97% of cases.
Neuronal activity at the transition of sequences
In the present study, each sequence of saccades was performed in blocks of eight trials: four trials under visual guidance followed by four trials from memory (Fig. 1). After the eighth trial, the monkey was required to prepare for switching from one sequence to another. During this period, we found that a discrete set of neurons exhibited a profound increase in activity as in an example shown in Fig. 6A. When the monkey performed a U-L-L sequence under visual guidance, neurons within this set exhibited a gradual increase in activity toward the initiation of the first trial (arrows in Fig. 6A, top). In later trials, however, the increase in activity was less marked and had disappeared by the time the sequential performance was based on memory (Fig. 6A, bottom). This trend was observed every time the saccade sequence was updated. For example, neuronal activity also increased before the onset of the first trial in the visually guided U-L-R sequence (Fig. 6A, 2nd panel) but became less noticeable soon thereafter. To demonstrate selectivity for the first trial more clearly, we rearranged the same data set according to the trial number within a block, irrespective of the saccade sequence. As is apparent in Fig. 6B, the increase in activity occurred primarily in the first trial of each sequence. The modulation at sequence transitions started 1.6 s before the start of a trial (900 ms after the end of the light flash that signaled a transition) and peaked near the initiation of the trial. This time course suggests that such selective activity was prospective for the new sequence rather than a direct response to the visual flash. We observed this type of neuronal activity in 112 SEF and 9 FEF neurons (25.3 and 3.5% of all neurons recorded in isolation, respectively). The proportion of neurons exhibiting first-trial selectivity was significantly higher for the SEF as compared with the FEF (P < 0.001, 2 test). Because the monkey was not required to control its gaze during the ITI before the trial onset (before the initiation of gaze fixation), we analyzed the frequency, direction and amplitude of saccades during this period (i.e., in the 2,000-ms period prior to the start of the trial). These parameters did not differ according to the number of trials within a visual block.
|
In approximately half of the neurons that exhibited first-trial selectivity, an increase in activity was observed before the initiation of the trial (Table 1) as for the cell shown in Fig. 6. In other cells, selectivity was observed during the presaccadic period. For such cells, the novel sequence-selective activity appeared to either precede a saccade at one specific position within the sequence (68%) or in more than one position (32%). Figure 7 illustrates an example in which first-trial selectivity was observed largely in the first of three saccades. ANCOVA revealed that such first-trial selectivity during presaccadic periods could not be ascribed to differences in kinematic saccade parameters within a block.
|
|
Cortical localization of neuronal activity belonging to different types
In the present study, we recorded neuronal activity in three hemispheres of SEF and two hemispheres of FEF, using two monkeys. Figure 8A shows surface reconstructions of the penetration sites of monkey 1 from which saccadic eye movements could be evoked by low-threshold ICMS (<50 µA). Saccades were elicited primarily in a small region of the dorsomedial convexity of the frontal lobe (SEF) and in the anterior bank of the arcuate sulcus (FEF). To examine whether activity properties of any types were localized to a particular cortical portion within the SEF and FEF, we analyzed the distribution of each type of neuronal activity within each area (Fig. 8B). Although rank-selective, sequence-selective, and transition-selective activities were encountered significantly more rostrally than other types of activity in the right SEF of monkey 1 (P < 0.01, Kruskal-Wallis test; Fig. 8B), we could not find significant differences in distribution in the left SEF or in the FEF. For monkey 2, the difference in distribution was not observed in either area.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONDISCLOSURESACKNOWLEDGMENTSREFERENCES |
|---|
Our results revealed that the FEF and SEF exhibit a variety of neuronal activity that appeared to be useful for the correct performance of memorized saccade sequences. In addition, the distribution of each type of activity differed among neurons in the two fields. Sequence-, rank-, and combination-selective (direction and rank selective) activities were more prevalent in the SEF than in the FEF. As a population, selectivity for the rank and sequence was expressed to a greater degree in neurons in the SEF than in the FEF. By contrast, direction selectivity was more prevalent in the FEF and direction selectivity was the dominant signal encoded by FEF neurons. Direction-selective activity would appear to be useful for encoding a forthcoming saccade at the behavioral stage of output signaling, but in itself contains no information about the time structure of multiple saccades. By contrast, other types of activity appear to be relevant to the temporal sequencing of saccades. Rank-selective activity allows the discrimination of the numerical position of saccades within a sequence, specifying which of the first, second, or third saccades a subject should perform. This type of activity was independent of saccade direction and, therefore more abstract in nature as a control signal. Combination-selective activity, which was selective for both the saccade direction and rank, appears to be useful for specifying direction-selective saccades at a specific rank within a sequence. Because saccades of equal direction in the task used in the present study appeared twice in each of the four sequences (specifically, U-L-L, L-L-R, U-R-R, and R-R-L), activity that distinguished the same direction of saccades seems to be a component of a task-regulating signal. Furthermore, sequence-selective activity, which was more prevalent in the SEF, would appear to be useful for structuring the temporal order of the three saccades as a sequence.
Apart from the difference in the distribution of neuronal activity and the magnitude of selectivity, we found a difference in the sequence selectivity between the two areas. In the SEF, sequence selectivity tended to appear earlier in a sequence, whereas in the FEF it appeared later. As is illustrated in Fig. 5B, for SEF neurons, sequence selectivity preceding the first saccades was more prevalent than selectivity preceding the third saccades. By contrast, for FEF neurons, selectivity was more prevalent in the period preceding the third saccades rather than the first. How might this finding be interpreted? The selective activity preceding the first saccades was selective for the sequence of upcoming saccades, whereas the activity preceding the third saccades was specific for the sequence of saccades that had already been performed. It is possible that SEF neurons are more involved in prospective coding of the saccade sequence, whereas FEF neurons are influenced more by what the animal has already performed. However, such a prospective/retrospective coding hypothesis requires further investigation, because we used a limited number of sequences in this study (6 of 27 possible sequences).
Functional significance of the transition-selective activity in the SEF
During visually guided task performance, we observed another type of neuronal activity that characterized SEF neurons. Approximately one-quarter of neurons sampled in the SEF exhibited activity selectively at the transition of sequences (i.e., selectively for the first trial in any sequence). Can this activity be ascribed to factors such as level of arousal, general attention or the task difficulty? The following findings argue against this possibility. First, first-trial selectivity was much less frequent in the FEF. Second, the reaction times at individual saccades were unrelated to the magnitude of transition-selective activity. Finally, neither the reaction times nor the error rates for the first visual trials were different from those of other trials (P > 0.05). Consequently, other behavioral factors should be considered in trying to explain transition-selective activity. In the behavioral task used in this study, the monkeys learned six sequences extensively and were completely familiar with the task. Therefore it seems reasonable to assume that sequential saccade performances were based on information that had been acquired and stored, which we shall tentatively refer to as a "motor plan." During trials guided by memory, the task performance seemed to be based on motor plans for individual sequences. After the completion of the final memory-guided trial, the monkey had to discard the current motor plan and then call for the next motor plan. It was at this particular point in the task performance that selective neuronal activity was observed only in the SEF. It is plausible, therefore that activity at the transition of sequences reflects the process of updating motor plans to organize the sequence of a subsequent saccade. This activity is likely to be the oculomotor equivalent of the activity observed in the presupplementary motor area, which is selective for the renewal of limb motor sequences (Shima et al. 1996).
Relevance of the transition-selective activity to previous studies
Some aspects of cellular activity in the medial cortical areas reported previously resemble the characteristic activities observed in the present study: learning-selective activity in the SEF (Chen and Wise 1995a), preference for a new behavioral set in the pre-SMA (Nakamura et al. 1998), and search-related activity in the anterior cingulate cortex (Procyk et al. 2000). Learning-selective activity was observed while monkeys learned the correct association of novel visual stimuli and the direction of saccades (Chen and Wise 1995a). Activity associated with preference for a new behavioral set was observed while subjects acquired novel visuo-skeletomotor associations to construct a spatial pattern (Nakamura et al. 1998). Search-related activity was observed during the period in which monkeys were required to find the correct spatiotemporal pattern for combined eye-arm movements (Procyk et al. 2000). It is important to note that the activity in each of these reports was observed during the acquisition of novel associations between visual stimuli and motor responses through trial-and-error exploration. Although entirely new association was not required in Procyk's task, the trial-and-error process was a necessary prerequisite for search-preferring cells to become active. In our study, by contrast, neither novel associations nor trial-and-error processes were involved in the behavioral task. For both animals in the present study, all saccade sequences were sufficiently familiar to have become a behavioral routine through extensive training. In addition, the correct sequence was initially prompted by means of visual cues and was not acquired by searching. Taken together, we interpret these results to mean that the sequence-transition selective activity observed in the present study differs from the activities reported previously despite their apparent resemblance.
Methodological consideration and limitation of interpretation
The essence of the oculomotor task in this study was to select the direction of center-out saccades and perform three saccades in a correct sequence based on memorized information. In the process of behavioral training, care was taken to establish that the metrics of individual saccades and their latencies did not differ depending on the temporal sequence of saccades. In contrast to strictly controlled saccades to the three peripheral targets, no strict requirements were imposed for eye movements that brought the gaze back to the central FP. Monkeys had only to capture the fixation target within 1,000 ms after the illumination of the FP. Therefore saccades at this stage were not necessarily directed straight to the FP. Instead, re-fixation movements often included two saccades with a corrective-saccade component or turn-around saccades accompanied by eye blinks. Thus quantitative analysis was made for the three center-out saccades but not for saccades at the time of re-fixation. We found that neuronal activity accompanying the re-fixation process was not apparent, as exemplified in Fig. 2, or highly variable. The above behavioral factors (nondemanding task requirements and variable metrics) seem to account for the paucity of neuronal activity during the period of re-fixation eye movements.
On the other hand, we should discuss two issues with regard to limitation of interpretation of the present results due to the design of the behavioral task employed in this study. The first concerns the nonrandom nature of the six sequences of saccades. Among six sequences employed in this study, four sequences started with the UP target. Thus prior to the first target appearance in a new sequence, there was a 67% probability of UP target appearance. If the SEF has preferential access to probability signals, then the bias to the upward target could produce the increase in activity at the transition of sequences preceding the start of the first trial. In fact, Coe et al. (2002) showed that, using a free-choice saccade task, SEF activity preferentially reflected anticipatory bias of subsequent saccade-target. This interpretation, however, could not account for all the mechanisms underlying the transition-selective activity during ITI because the modulation of the activity at the transition was observed regardless of gaze direction during ITI, whereas anticipatory bias reported by Coe et al. was receptivefield bound. Accessibility of SEF neurons to probability signals per se, like midbrain dopamine neurons (Fiorillo et al. 2003), remains to be studied.
Second, we relied on three possible saccade directions (up, left, and right) to determine directional preference of neuronal activity. Considering the small number of directions to analyze the directional preference, it seems likely that the number of directionally tuned neurons was underrepresented, especially in the FEF. It is possible that the difference in the distribution between the two areas for the direction index (Fig. 3B, right) might actually be greater than that shown.
Hypothetical operation of neural elements in the SEF and FEF for orderly arrangement of three memorized saccades
We now propose a hypothesis to explain how the variety of neuronal activity observed in the SEF and FEF might be coordinated to provide the signals that are required to perform three saccades in an orderly manner. Let us consider the sequence U-L-L as an example (Fig. 9A). First, the subject needs to be informed that the sequence that is currently required is U-L-L; sequence-selective activity could serve to retrieve and carry such information. If the sequence information is combined with information about rank to indicate that the first saccade is required, then the information about saccade direction would be provided, in this case, an upward saccade. During this decoding process, the combination of information about direction and rank could signal that an upward saccade should be executed first (combination-selective activity). After the first saccade has been correctly performed, a signal specifying a leftward saccade could be provided by combining the sequence and rank (specifying the second saccade) signals. In this decoding process, the signal specifying a leftward saccade in the second rank of the sequence could be generated. Thereafter, a signal for the third (leftward) saccade could be generated in an analogous manner to complete the U-L-L sequence. Once one block of the U-L-L sequence has been accomplished, transition-selective activity could serve to discard the current sequence and prepare the next motor plan, thereby enabling a smooth and nonpersistent transition to the next sequence. We are aware that the processes illustrated in Fig. 9A might be too simple to account for the actual functioning of neuronal elements. Further study is required to fully elucidate the details involved in each of the processes that involve the two cortical eye fields.
|
In conclusion, if we compare the distribution of the different types of neuronal activity observed in the present study (Fig. 9B), it can be argued that the SEF is involved more in planning, decoding, and updating saccade sequences, whereas the FEF plays a major part in determining the direction of forthcoming saccades. The possible participation of other cortical and subcortical neural structures (Gaymard et al. 1998; Hikosaka et al. 1999; Munoz 2002; Tanji 2001) in the temporal organization of multiple saccades remains to be determined.
|
DISCLOSURES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONDISCLOSURESACKNOWLEDGMENTSREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONDISCLOSURESACKNOWLEDGMENTSREFERENCES |
|---|
|
FOOTNOTES |
|---|
Address for reprint requests and other correspondence: J. Tanji, Dept. of Physiology, Tohoku University School of Medicine, 2-1 Seiryo-cho, Aoba-ku, Sendai 980-8575 Japan (E-mail: tanjij{at}mail.cc.tohoku.ac.jp).
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONDISCLOSURESACKNOWLEDGMENTSREFERENCES |
|---|
Bon L and Lucchetti C. The dorsomedial frontal cortex of the macaca monkey: fixation and saccade-related activity. Exp Brain Res 89: 571–580, 1992.[ISI][Medline]
Bruce CJ and Goldberg ME. Primate frontal eye fields. I. Single neurons discharging before saccades. J Neurophysiol 53: 603–635, 1985.
Bruce CJ, Goldberg ME, Bushnell C, and Stanton GB. Primate frontal eye fields. II. Physiolological and anatomical correlates of electrically evoked eye movements. J Neurophysiol 54: 714–734, 1985.
Chen LL and Wise SP. Neuronal activity in the supplementary eye field during acquisition of conditional oculomotor associations. J Neurophysiol 73: 1101–1121, 1995a.
Chen LL and Wise SP. Supplementary eye field contrasted with the frontal eye field during acquisition of conditional oculomotor associations. J Neurophysiol 73: 1122–1134, 1995b.
Clower WT and Alexander GE. Movement sequence-related activity reflecting numerical order of components in supplementary and presupplementary motor areas. J Neurophysiol 80: 1562–1566, 1998.
Coe B, Tomihara K, Matsuzawa M, and Hikosaka O. Visual and anticipatory bias in three cortical eye fields of the monkey during an adaptive decision-making task. J Neurosci 22: 5081–5090, 2002.
Dick JPR, Benecke R, Rothwell JC, Day BL, and Marsden CD. Simple and complex movements in a patient with infarction of the right supplementary motor area. Mov Disord 1: 255–266, 1986.[Medline]
Everling S and Munoz DP. Neuronal correlates for preparatory set associated with pro-saccades and anti-saccades in the primate frontal eye field. J Neurosci 20: 387–400, 2000.
Fiorillo CD, Tobler PN, and Schultz W. Discrete coding of reward probability and uncertainty by dopamine neurons. Science 299: 1898–1902, 2003.
Gaymard B, Pierrot-Deseilligny C, and Rivaud S. Impairment of sequences of memory-guided saccades after supplementary motor area lesions. Ann Neurol 28: 622–626, 1990.[ISI][Medline]
Gaymard B, Ploner CJ, Rivaud S, Vermersch AI, and Pierrot-Deseilligny C. Cortical control of saccades. Exp Brain Res 123: 159–163, 1998.[ISI][Medline]
Hanes DP, Patterson WF, and Schall JD. The role of frontal eye field in countermanding saccades: visual, movement, and fixation activity. J Neurophysiol 79: 817–834, 1998.
Hikosaka O, Nakahara H, Rand MK, Sakai K, Lu X, Nakamura K, Miyachi S, and Doya K. Parallel neural networks for learning sequential procedures. Trends Neurosci 22: 464–471, 1999.[ISI][Medline]
Hikosaka O, Sakai K, Miyauchi S, Takino R, Sakai Y, and Pütz B. Activation of human presupplementary motor area in learning of sequential procedures: a functional MRI study. J Neurophysiol 76: 617–621, 1996.
Huerta MF and Kaas JH. Supplementary eye field as defined by intracortical microstimulation: connections in macaques. J Comp Neurol 293: 299–330, 1990.[ISI]
Huerta MF, Krubitzer LA, and Kaas JH. Frontal eye fields as defined by intracortical microstimulation in squirrel monkeys, owl monkeys and macaque monkeys. II. Cortical connections. J Comp Neurol 265: 332–361, 1987.[ISI][Medline]
Isoda M and Tanji J. Cellular activity in the supplementary eye field during sequential performance of multiple saccades. J Neurophysiol 88: 3541–3545, 2002.
Kawashima R, Tanji J, Okada K, Sugiura M, Sato K, Kinomura S, Inoue K, Ogawa A, and Fukuda H. Oculomotor sequence learning: a positron emission tomography study. Exp Brain Res 122: 1–8, 1998.[ISI][Medline]
Lashley KS. The problem of serial order in behavior. In: Cerebral Mechanisms in Behavior, edited by Jeffress LA. New York: Wiley, 1951, p. 112–136.
Lu X, Matsuzawa M, and Hikosaka O. A neural correlate of oculomotor sequences in supplementary eye field. Neuron 34: 317–325, 2002.[ISI][Medline]
Mann SE, Thau R, and Schiller PH. Conditional task-related responses in monkey dorsomedial frontal cortex. Exp Brain Res 69: 460–468, 1988.[ISI][Medline]
Matsuzaka Y, Aizawa H, and Tanji J. A motor area rostral to the supplementary motor area (presupplementary motor area) in the monkey: neuronal activity during a learned motor task. J Neurophysiol 68: 653–662, 1992.
Mitz AR and Godschalk M. Eye-movement representation in the frontal lobe of rhesus monkeys. Neurosci Lett 106: 157–162, 1989.[ISI][Medline]
Munoz DP. Commentary: saccadic eye movements: overview of neural circuitry. Prog Brain Res 140: 89–96, 2002.[ISI][Medline]
Müri RM, Rivaud S, Vermersch AI, Léger JM, and Pierrot-Deseilligny C. Effects of transcranial magnetic stimulation over the region of the supplementary motor area during sequences of memory-guided saccades. Exp Brain Res 104: 163–166, 1995.[ISI][Medline]
Mushiake H, Fujii N, and Tanji J. Visually guided saccade versus eye-hand reach: contrasting neuronal activity in the cortical supplementary and frontal eye fields. J Neurophysiol 75: 2187–2191, 1996.
Mushiake H, Inase M, and Tanji J. Selective coding of motor sequence in the supplementary motor area of the monkey cerebral cortex. Exp Brain Res 82: 208–210, 1990.[ISI][Medline]
Mushiake H, Inase M, and Tanji J. Neuronal activity in the primate premotor, supplementary, and precentral motor cortex during visually guided and internally determined sequential movements. J Neurophysiol 66: 705–718, 1991.
Nakamura K, Sakai K, and Hikosaka O. Neuronal activity in medial frontal cortex during learning of sequential procedures. J Neurophysiol 80: 2671–2687, 1998.
Nakamura K, Sakai K, and Hikosaka O. Effects of local inactivation of monkey medial frontal cortex in learning of sequential procedures. J Neurophysiol 82: 1063–1068, 1999.
Olson CR and Gettner SN. Object-centered direction selectivity in the macaque supplementary eye field. Science 269: 985–988, 1995.
Olson CR and Gettner SN. Macaque SEF neurons encode object-centered directions of eye movements regardless of the visual attributes of instructional cues. J Neurophysiol 81: 2340–2346, 1999.
Parthasarathy HB, Schall JD, and Graybiel AM. Distributed but convergent ordering of corticostriatal projections: analysis of the frontal eye field and the supplementary eye field in the macaque monkey. J Neurosci 12: 4468–4488, 1992.[Abstract]
Petit L, Orssaud C, Tzourio N, Crivello F, Berthoz A, and Mazoyer B. Functional anatomy of a prelearned sequence of horizontal saccades in humans. J Neurosci 16: 3714–3726, 1996.
Picard N and Strick PL. Motor areas of the medial wall: a review of their location and functional activation. Cereb Cortex 6: 342–353, 1996.
Procyk E, Tanaka YL, and Joseph JP. Anterior cingulate activity during routine and non-routine sequential behaviors in macaques. Nat Neurosci 3: 502–508, 2000.[ISI][Medline]
Roland PE, Larsen B, Lassen NA, and Skinhoj E. Supplementary motor area and other cortical areas in organization of voluntary movements in man. J Neurophysiol 43: 118–136, 1980.
Rosenbaum DA. Human Motor Control. San Diego, CA: Academic, 1991.
Russo GS and Bruce CJ. Effect of eye position within the orbit on electrically elicited saccadic eye movements: a comparison of the macaque monkey's frontal and supplementary eye fields. J Neurophysiol 69: 800–818, 1993.
Russo GS and Bruce CJ. Neurons in the supplementary eye field of rhesus monkeys code visual targets and saccadic eye movements in an oculocentric coordinate system. J Neurophysiol 76: 825–848, 1996.
Russo GS and Bruce CJ. Supplementary eye field: representation of saccades and relationship between neural response fields and elicited eye movements. J Neurophysiol 84: 2605–2621, 2000.
Schall JD. Neuronal activity related to visually guided saccadic eye movements in the supplementary motor area of rhesus monkeys. J Neurophysiol 66: 530–558, 1991a.
Schall JD. Neuronal activity related to visually guided saccades in the frontal eye fields of rhesus monkeys: comparison with supplementary eye fields. J Neurophysiol 66: 559–579, 1991b.
Schall JD. Visuomotor areas of the frontal lobe. In: Extrastriate Cortex of Primates, Cerebral Cortex, edited by Rockland KS, Kaas JH, and Peters A. New York: Plenum, 1997, vol. 12, p. 527–638.
Schall JD, Stuphorn V, and Brown JW. Monitoring and control of action by the frontal lobes. Neuron 36: 309–322, 2002.[ISI][Medline]
Schiller PH and Chou I. The effects of frontal eye field and dorsomedial frontal cortex lesions on visually guided eye movements. Nat Neurosci 1: 248–253, 1998.[ISI][Medline]
Schlag J and Schlag-Rey M. Unit activity related to spontaneous saccades in frontal dorsomedial cortex of monkey. Exp Brain Res 58: 208–211, 1985.[ISI][Medline]
Schlag J and Schlag-Rey M. Evidence for a supplementary eye field. J Neurophysiol 57: 179–200, 1987.
Schlag-Rey M, Amador N, Sanchez H, and Schlag J. Antisaccade performance predicted by neuronal activity in the supplementary eye field. Nature 390: 398–401, 1997.[Medline]
Shibasaki H, Sadato N, Lyshkow H, Yonekura Y, Honda M, Nagamine T, Suwazono S, Magata Y, Ikeda A, Miyazaki M, Fukuyama H, Asato R, and Konishi J. Both primary motor cortex and supplementary motor area play an important role in complex finger movement. Brain 116: 1387–1398, 1993.
Shima K, Mushiake H, Saito N, and Tanji J. Role for cells in the presupplementary motor area in updating motor plans. Proc Natl Acad Sci USA 93: 8694–8698, 1996.
). Other conventions are as in 
) and FEF (
). For the SEF, n = 84. For the FEF, n = 22. Before calculating the relative frequency, the number of selective cases that appeared in each rank was normalized to the number of sequences examined, i.e., by dividing by 4 for the 1st saccade and by 6 for the 2nd and 3rd saccades.
) in individual sequences under visual guidance, regardless of saccade order. In this case, modulation of activity was observed before the trial was initiated (i.e., before the initiation of gaze fixation). The increase in activity had disappeared by the time the saccade sequences became guided by memory (bottom). The raster diagrams are aligned with the trial start. B: activity of the same neuron as in A rearranged according to the trial number within a block. Trials for the different sequences have been combined. This format illustrates more clearly that the cellular discharge for this cell was most prominent during the 1st trials in the visually guided performance. Sac1, 1st saccade; sac2, 2nd saccade; sac3, 3rd saccade.
, points of electrode entry from which saccadic eye movements were evoked by low-threshold (<50 µA) intracortical microstimulation. PS, principal sulcus; ARC, arcuate sulcus; CS, central sulcus. B: recording sites of neurons exhibiting five different properties in the FEF and SEF. Each horizontal row of rectangles corresponds to the 3 rectangles shown in A. The number of task-related neurons with different properties was plotted separately at each penetration sites. Each circle's size is proportional to the number of neurons, selective for saccade direction (D), direction and rank (DR), rank (R), sequence (S), and sequence transition (T).