Abstract
We tested the hypothesis that the pallidum contributes to the control of both posture and movement. We recorded neuronal activity from the pallidum in a task in which male cats reached forward from a standing posture to depress a lever. In agreement with previous studies, we found that a majority of pallidal cells (91/116, 78%), including neurons in both the entopeduncular nucleus and the globus pallidus, showed significant modulation of their activity during reaching with the contralateral limb. Mostly different populations of cells were active during the transport (flexion) and lever press (extension) phase of the task. Most cells showed dynamic patterns of activity related to the movement. However, a modest proportion of modulated cells (18/91, 20%) showed properties consistent with a contribution to the control of anticipatory postural responses, whereas a further 10% showed activity consistent with a contribution to postural support during the movement. Although some cells that showed modified activity only during reaches with the contralateral forelimb, many cells (65/91, 71%) were also activated during reaches with the ipsilateral forelimb. This was particularly true for cells related to the lever press, many of which discharged similarly during reaches of either limb. This suggests a context-dependent control of movement and posture in which the same muscles are used for different functions during contralateral and ipsilateral reach. Comparison with the results from recordings made previously from the motor cortex and the pontomedullary reticular formation in the same task show more similarities with the former than the latter.
SIGNIFICANCE STATEMENT Pathologic changes in basal ganglia function frequently lead to problems with postural stability and gait initiation. Here, we show that some neurons in one of the output regions of the basal ganglia, the pallidum, show discharge activity compatible with a contribution to postural control. At the same time, we note that such cells are a minority in this region with most cells being related to movement rather than posture. We also show that many neurons are active during movements of both the contralateral and ipsilateral limbs, sometimes with identical discharge patterns. We suggest that this indicates a context-dependent regulation of movement and posture in the pallidum.
- anticipatory postural adjustment
- basal ganglia
- globus pallidus
- posture
- reaching
- single-unit recording
Introduction
Damage to the basal ganglia, whether the result of experimental lesion or of pathologic changes, produces deficits in the production of discrete voluntary movements (Mink and Thach, 1991a; Wenger et al., 1999; Desmurget and Turner, 2008) and the control of posture (Amalric et al., 1994; Burbaud et al., 1998; Grimbergen et al., 2004; Boonstra et al., 2008; Jacobs et al., 2009; Rogers et al., 2011). In agreement with the suggestion that these deficits are indicative of a contribution of the basal ganglia to the normal control of behavior, single-cell recordings in different nuclei of the basal ganglia, including the pallidum, show that many cells show changes in discharge activity that are tightly related to different events during the preparation (Nambu et al., 1990; Schultz and Romo, 1992; Jaeger et al., 1993) and execution (DeLong, 1971; Georgopoulos et al., 1983; Hamada et al., 1990; Mink and Thach, 1991b,c; Cheruel et al., 1994; Turner and Anderson, 1997) of voluntary movements. However, these studies leave some important questions.
First, on the basis of studies in Parkinson's patients (Grimbergen et al., 2004; Boonstra et al., 2008; Jacobs et al., 2009; Rogers et al., 2011), it has been suggested that the basal ganglia make a contribution not only to discrete movements but also to the anticipatory postural adjustments (APAs) that precede movement, as well as those that accompany it (Viallet et al., 1987; Massion, 1992; Dimitrova et al., 2004; Jacobs et al., 2009). However, most previous neuronal recording studies have been performed on seated animals making isolated arm movements (but see, Cheruel et al., 1994). Such movements require little in the way of postural activity. As such, there is little information at the single-cell level concerning the contribution of the basal ganglia to the control of posture, and, indeed, whether the need for postural support to maintain stability is reflected in basal ganglia cell discharge during movement. If the pallidum does contribute to postural control, then one might expect to observe neurons that discharge with respect to the APAs preceding a voluntary movement as well as to the postural support needed during the movement. In the case of APA-related activity, this might occur in concert with the command for movement as it does in both the pontomedullary reticular formation (PMRF) and the motor cortex (Schepens and Drew, 2006; Schepens et al., 2008; Yakovenko and Drew, 2009), or the signals for the APA and the movement might be kept separate, as suggested by Massion's (1992) theory of parallel streams, and then combined in other structures, such as the PMRF.
Second, with a few exceptions, there has been little comparison of the activity of basal ganglia neurons during movements of the contralateral and ipsilateral limb, the emphasis being generally placed on the contralateral limb. In those studies that have recorded activity to movement of the ipsilateral limb (DeLong, 1971; Iansek and Porter, 1980; Wannier et al., 2002), only relatively few (<30%) of cells were active for both contralateral and ipsilateral movements. However, in our recent locomotor study (Mullié et al., 2020), we found a much larger proportion of cells that discharged in relationship to the gait modifications of the ipsilateral limb. Possibly this contribution to the control of each forelimb may be specific to locomotion in which the movement of the two forelimbs (and, of course, the hindlimbs) is obligatorily coupled. However, it may also reflect a more general contribution to coordinated patterns of movement, in which case one might expect a greater proportion of cells discharging to ipsilateral movements when there is a need for coordination of movement and posture.
To address these questions, we recorded the activity of cells in the entopeduncular nucleus (EP) and the globus pallidus (respectively the feline equivalent of the internal, GPi, and external, GPe, segments of the primate globus pallidus) in a task requiring the production of APAs before a movement and postural support during that movement (Schepens and Drew, 2003; Schepens et al., 2008).
Materials and Methods
Tasks and training
Experiments were performed on two male cats (BG5 and BG6, weight 5.6 and 5.0 kg, respectively) that were initially trained to walk on a treadmill at a speed of 0.45 ms−1 and to step over two obstacles attached to that treadmill (Mullié et al., 2020). They were subsequently trained to perform a task in which they reached forward and depressed a lever with either the left or right forelimb (Fig. 1A; Schepens et al., 2008; Yakovenko and Drew, 2009). Cats were trained to stand quietly on four force platforms in response to an initial tone of 0.5 s. A second tone, given 1.5 s later, indicated which limb the cat should use to perform the task (left limb, tone 400 Hz; right limb, tone 4 kHz). After a random period of 0.5 −1.5 s the tone stopped, and a shutter opened to provide access to a lever positioned in line with either the left or right forelimb. If the cat pressed on the lever with the appropriate limb, a tray containing a food reward was advanced. Trials were immediately aborted if the cats used the wrong limb. Training continued until the cats performed the task with a minimum of 85% success rate. The entire training period covering both the locomotion task and the reaching task lasted between six and nine months. Both cats were used in previous experiments (Mullié et al., 2020).
Surgery
Once the cats were trained, they were prepared for surgery in aseptic conditions and under general anesthesia (Mullié et al., 2020). Cats were premedicated, intubated, and anesthetized with 2–3% isoflurane with oxygen. Respiration, heart rate, and temperature were monitored and controlled. Analgesia was provided by buprenorphine (5 µg/kg). Stainless steel, Teflon-insulated wires were inserted into the belly of selected muscles of the four limbs as well as into nuchal muscles to record electromyograms (EMGs). A stainless steel recording chamber was placed over the parietal cortex on the right side, providing access to the EP and the GPe. A connector was placed on the cranium to accept a small preamplifier during experimental sessions. All wires and connectors were held in place on the cranium with dental acrylic. Antibiotics and analgesia were proved postsurgery following the advice of the institutional veterinary service. All experiments followed the guidelines of the Canadian Council of Animal Care and were approved by the Université de Montréal Animal Ethics Committee.
Protocol
During each experimental session, an electrode was driven through the parietal cortex to the location of the EP or GPe. The nuclei were identified based on their depth and their relationship to the surrounding pattern of neuronal activity (Mullié et al., 2020). Once a single unit was isolated in the respective nucleus, the cell activity was recorded while the cat walked on the treadmill and stepped over each of the two obstacles attached to that treadmill (Mullié et al., 2020). Following these recordings, the cat was transferred to the reaching apparatus, which was positioned adjacent to the treadmill, and data were recorded either from the same cell, or from newly isolated cells, during the reaching task. We equally recorded EMG activity from muscles of all four limbs. In this article we use data only from the elbow flexor, brachialis (Br); the shoulder protractor and elbow flexor, cleidobrachialis (ClB); the elbow extensor; the lateral head of triceps (TriL); and the shoulder retractor and elbow extensor, the long head of triceps (Tri) (Krouchev et al., 2006 for location of these muscles). We collected data during 10 reaches with each forelimb, five reaches recorded with the left forelimb, 10 reaches with the right forelimb, and then another five with the left forelimb. Occasionally, further reaches were recorded for each limb. We then searched for other cells.
Data from the EMGs were digitized to disk at 1 kHz while cell activity was sampled at 100 kHz to allow the full form of the action potential to be recorded for off-line discrimination. All experiments were filmed (60 frames/s) and recorded to a CD disk or to the computer. A digital time code was recorded with both the analog data and the video to allow synchronization of the two records.
Data analysis
Data during the reaching task were analyzed using methods similar to those that we used in previous work using a similar or identical task (Schepens and Drew, 2003, 2004; Schepens et al., 2008; Yakovenko and Drew, 2009). In brief, single units were discriminated on the basis of principal component analysis by using an off-line spike sorter (Plexon). For most recordings, only a single cell was discriminated; in a few cases it was possible to discriminate two cells. We then examined each trial, for each cell, and eliminated those trials in which the cat moved a paw following the instruction cue but before the Go-signal. For the remaining trials, we identified the onset of the activity in the ClB of BG5 or the Br of BG6 of the reaching limb, which we defined as the onset of the reach; generally this occurred just before the time at which the paw left the support surface (Schepens and Drew, 2003) and the vertical ground-reaction force (Fv) fell to zero (Fig. 1B). We also identified the onset of the lever press, as detected by a mechanical switch on the lever arm. Other events (onset and offset of different muscles) were measured depending on the pattern of activity of the cell. Averaged changes in the activity of the cell discharge and the muscle activity were calculated for each cell, triggered on the Go-signal and/or the onset of the reach. Data were filtered at 25 Hz (dual-pass, second-order digital Butterworth filter), and significant changes in activity were defined as responses that deviated >2 SDs from the background resting activity during quiet standing. Background discharge was calculated from a 500 ms period occurring before the onset of the instruction cue. For all cells for which we were able to measure changes in activity from single trials, we calculated linear regressions between the onset and offset of the period of activity and the onset and offset of activity in the ClB/Br and/or Tri/TriL.
As noted in the Introduction, one of the important aims of this study was to determine whether cells in the EP and GPe showed responses related both to the APAs preceding movement and to the movement itself, as we previously found for cells in both the PMRF and in the motor cortex. To distinguish and classify these two types of cells, we initially identified all those significant changes in activity that occurred following the Go-signal and before or during the transport phase of the reach. This subset of cells includes all cells that are related to either the Go-signal (probable APA-related cells) and/or the transport phase of the reach.
We then measured the onset of cell activity with respect to the Go-signal in individual trials (for those cells for which it was possible) and plotted linear regressions of the onset of cell activity as a function of the onset of activity in the ClB/Br and as a function of the lead time of the cell (defined as the difference between the onset of the ClB/Br and the onset of the cell activity; Vicario et al., 1983; Schepens and Drew, 2004, 2006; Schepens et al., 2008). As illustrated in Figure 2, this analysis provides very different correlations for cells related to the transport phase of the reach and those related to the Go-signal. Cells related to the transport phase, such as the one illustrated in Figure 2, A and C, showed a sharp change in discharge activity that occurred at a variable time after the Go-signal (Fig. 2A) but kept a constant relationship with respect to the onset of activity in the ClB/Br (Fig. 2B). A linear regression of the onset of cell latency as a function of ClB onset, each measured with respect to the Go-signal (Fig. 2Ciii), showed a significant relationship (p < 0.05, R2 = 0.74). In contrast, the linear regression of cell onset activity as a function of lead time (Fig. 2Civ) was not significant.
Go-related cells showed the opposite pattern of activity (Fig. 2D–F). The initial change in cell discharge (Fig. 2D,E, arrows) in this example occurred at a relatively constant latency following the Go-signal (Fig. 2D). When discharge was synchronized to the onset of the period of activity in the contralateral ClB (coClB; Fig. 2E), this increase was less obvious and preceded ClB onset by several hundreds of milliseconds. Plotting cell onset against of ClB onset (Fig. 2Fiii) showed a nonsignificant relationship, whereas the plot of cell lead time versus ClB onset (Fig. 2Fiv) showed a highly significant relationship (R2 = 0.94). Note that this cell showed a larger, second period of activity that was movement related (linear regressions not illustrated).
All cells in which the onset of discharge showed a significant relationship (p < 0.05) with the onset of activity in the ClB/Br were defined as movement-related cells, associated with the transport phase of the movement (see below). On the other hand, a significant relationship (p < 0.05) between lead time and the onset of the ClB/Br was taken as being indicative of a constant time of discharge after the Go-signal. However, using only this method of defining a Go-related cell is compromised by the repetition of the same variable (onset of cell activity) on both axes. As a complementary method, we therefore calculated the variance of the onset of change in cell discharge for all those cells that showed significant relationships to the lead time. Cells that showed a significantly reduced variance with respect to the variance of the onset of the ClB/Br were classified as being related to the Go stimulus (Yakovenko and Drew, 2009; see below, Results).
For cells for which we could not measure the latency of the cell discharge from individual trials, Go-related cells were defined as those in which the latency of the significant change in discharge was <50% of the time between the Go-signal and the onset of the reach (see Yakovenko and Drew, 2009). In a later post hoc analysis we defined a boundary value based on the latencies of cells identified as Go-related and movement-related on the basis of the linear regression analysis.
Movement-related responses were classified into two major groups, those related to the transport phase and those related to the lever press period.
The transport phase was defined as the period from the onset of the reach (onset of ClB/Br) to the end of the period of activity in the ClB/Br (Fig. 1B, blue shading). Cells were classified as transport-related if their discharge was significantly modulated during this period. Cells were divided into those discharging dynamically or tonically based on the time at which the cell discharge frequency decreased (for a maximum) or increased (for a minimum) by ≥50% of the peak or trough value. For the onset of a peak or trough, the change in activity had to occur either within the transport period or within 50% of the average time between the Go-signal and the onset of the ClB/Br (Yakovenko and Drew, 2009). For the offset, this change had to occur within the first 50% of the lever press period. Cells were classified as having a more tonic discharge if the latency of the change in cell discharge either began earlier, and/or ended later, than the limits used to define the dynamic cells. For this analysis and for the analysis of cell discharge during the lever press (see below) activity was always triggered on the onset of the reach, as in Figure 2, B and E. Note that we classified the cell in Figure 2D–F as having a dynamic discharge during the transport phase following these definitions.
The lever press phase was defined as the period from the onset of the initial period activity in the Tri/TriL following reach onset until the end of the period of activity in the same muscle (Fig. 1B, red shading). Lever press cells were defined using the simple definition that the onset of discharge followed the onset of the reach, and the peak discharge fell within the lever press period but after the end of the transport phase. No criterion was set for the end of the discharge as the duration of the lever press was very variable with the cat occasionally keeping the lever depressed throughout the recording period. However, as some cells in this classification continued to discharge at higher levels than control for a prolonged period after the peak, we used a working definition that differentiated more phasically discharging cells from more tonic ones. Dynamic cells were defined as those in which the discharge declined to <50% of peak discharge within 2.5 s of the reach onset. A similar criterion was used for cells with decreased activity, for example, cell discharge activity had to increase by ≥50% of the minimal discharge frequency within 2.5 s.
In addition to these three classes of cells (Go related, transport phase related, and lever press related) we also identified two other classes, cue-related cells in which response activity was significantly modified following the cue signal but preceding the Go-signal and late-discharging cells in which the onset of activity occurred >2.5 s after the onset of the reach.
To examine the distribution of the population activity of the database we constructed heat maps using the Z score of cell discharge from the period of 1 s before the onset of the reach to 2.5 s after that point.
To compare the similarity of the pattern of cell activity during contralateral and ipsilateral reaches we calculated normalized cross-correlations using the cross-correlation functions in Systat software (version 13) and in the International Mathematics Scientific Library. Cross-correlations were made for a period of activity beginning 500 ms before the onset of the reach and for 2000 ms after. This included the major period of time encompassing cell activity in all cells except those that discharged after the lever press (see below, Results). Cross-correlations were made on data filtered at 25 Hz.
Histology
At the end of the series of experiments, the cats were premedicated with ketamine and then anesthetized with sodium pentobarbitol (Somnotol, 30 mg/kg). They were then perfused per cardium with a formaldehyde solution. The brain was sectioned at 40 µm, and sections were stained with cresyl violet. The location of recorded cells was interpolated on the basis of the depth of the recordings and the identification of the location of small lesions made during the recording session as well as larger lesions made just before the perfusion. As detailed in a previous article (Mullié et al., 2020), the location of the recorded cells was transposed to standard sections of the diencephalon taken from the atlas of Berman and Jones (1982) and then collapsed onto the horizontal plane.
Results
Behavioral activity
The changes in EMG activity and ground reaction forces observed during the reach were similar to those we have detailed previously in this identical task (Schepens et al., 2008; Yakovenko and Drew, 2009) as well as to those described by us and others in similar tasks involving postural destabilization produced by movement of a single limb (Bouisset and Zattara, 1981; Cordo and Nashner, 1982; Dufossé et al., 1982; Alstermark and Wessberg, 1985; Burleigh et al., 1994; Schepens and Drew, 2003, 2004); they therefore are not described in detail here. In brief, the onset of the reach was defined by the sharp onset of activity in the prime flexor muscles of the limb (ClB/Br) that shortly preceded the raising of the paw from the support surface, indicated by the fall of the vertical ground reaction force in the left(l) limb (lFv) to zero (Fig. 1B). The onset of the reach was preceded by an APA that has been shown to displace the center of vertical force and the center of mass of the cat in preparation for the upcoming movement (Ioffe et al., 1982; Schepens and Drew, 2003). In this typical example, the APA was best characterized by an increase in activity of lTri before the onset of reach (black arrowhead in Fig. 1B, see Fig. 5). This change was accompanied by a small, sharp, increase in the ground reaction force under the lifting leg (lFv) that was more or less pronounced in different recording sessions. At the same time, there was a small decrease in the activity of the elbow extensor muscle in the other forelimb (rTri) accompanied by a decrease in the ground reaction force in the supporting limb (Fig. 1B, rFV, white arrowhead). These changes in activity generally began at a relatively short latency after the Go-signal (mean of 83–122 ms in Schepens and Drew, 2003).
At reach onset, there was a cessation of activity in the extensor muscles of the reaching limb (Fig. 1B, lTri) as the paw was lifted from the ground. This was accompanied by an increase in the activity in the rTri and in the ground reaction force under the supporting limb (rFv). Subsequently, there was an increase in activity in the triceps of the reaching limb as the cat pressed on the lever; this partially unloaded the supporting limb (decrease in level of rFv). There was then a second period of activity in the lBr as the cat lifted the limb from the lever and replaced it on the support surface at its own rhythm.
As indicated in Table 1 the reaction times for the onset of the movement showed some variation between cats and between limbs, with movements made by the limb ipsilateral to the recording site (the right limb) of cat BG5 being the fastest and movements with the contralateral limb of the same cat being the slowest. Movements on the two sides of cat BG6 were more comparable. these values are similar to those obtained in other studies from this laboratory using the identical task (Schepens et al., 2008).
Activity during contralateral reach
We recorded the activity of 116 pallidal cells (81 from the EP and 35 from the GPe) for which we had at least four reaches with the contralateral left limb (Table 2). Of these, 91/116 (68/81 EP and 23/35 GPe cells) showed a significant modification of their discharge during the reach of the limb contralateral to the recording site at one or more periods of the motor activity (between the onset of the cue and the end of the recording period, ∼4–5 s after the Go-signal).
Many of the cells exhibited more than one period of significantly modified activity, so among the 91 cells showing task-related activity during the contralateral reach, there were 137 periods of significantly modified activity, distributed across the analyzed period. The most common significant changes were those that occurred in movement-related cells, either during the transport phase of the reach or during the lever press (Table 2). Cells recorded from the EP and the GPe showed a similar distribution of their activity patterns (Table 2), and in the following we have, therefore, included the data from these populations together.
Movement-related cells
The majority of cells that we recorded in the pallidum showed periods of activity that were related to the two major phases of the reach, namely the transport phase and the subsequent lever press (Table 2). These two populations are therefore addressed first.
Cells related to the transport phase of the reach
A total of 40/91 task-related cells showed a period of significantly modified activity in which the peak discharge occurred during the transport phase of the motor task (Table 2). Of these, 24 cells showed either a dynamic increase (12/24) or a dynamic decrease (12/24) in their discharge activity (as defined above in Materials and Methods).
An example of a cell showing a simple, unimodal increase in activity during the transport phase is illustrated in Figure 3A (also Fig. 2A–C). The cell shows a significant increase in activity just after the onset of the period of activity in the ClB (vertical black line) and the end of the discharge activity coincided with the end of the period of activity in the ClB (red staggered line). Linear regression analysis showed that the time of onset of the change in activity was well correlated with the onset of the coClB activity (Fig. 3Bi) as well as to the time at which the paw left the support surface (Fig. 3Bii, paw lift). The end of the period of cell discharge was likewise well correlated with the end of the period of activity in the ClB (Fig. 3Biii) and with the time that the paw contacted and depressed the lever (Fig. 3Biv, A, green staggered line).
More complex bimodal changes in activity were observed in some other cells (Fig. 3C,E). In the cell illustrated in Figure 3C, there was an increase in cell activity that was well correlated to the loss of contact of the support surface by the contralateral paw (Fig. 3Dii). As for the cell in Figure 3A, cell discharge stopped as the cat pressed the lever (Fig. 3Div) and then fell just below baseline (nonsignificant decrease) from that point until the end of the period of activity in the Br. There was then a subsequent significant increase in activity as the cat pressed on the lever, and activity in the extensor muscles increased (Fig. 1B).
The cell in Figure 3E showed a variation on this pattern in that the initial increase in activity occurred >200 ms before the onset of the activity in the Br. This increase was then followed by a discrete, and significant, decrease in activity that began at the onset of the transport phase of the reach. The beginning of this decrease in activity was well correlated with the onset of the period of activity in the Br (Fig. 3Fi) and the lifting of the paw from the support surface (Fig. 3Fii) while the end of this trough was well related to the end of the period of the coBr and the depression of the lever (Fig. 3Fiii,iv).
Of the 12 cells showing increased activity during the transport phase, 11/12 showed a significant linear relationship between the onset of the cell discharge and the onset of the period of activity in the ClB/Br, whereas another 11/12 (one cell different) showed a significant relationship between the end of the period of modified cell activity and the end of the period of ClB/Br activity (Fig. 4A). Linear regressions were compiled for 9/12 cells showing decreases in activity (and for which measures could be made from individual trials). Of these, 8/9 showed a significant relationship to ClB/Br onset, and all 9 showed a relationship to the EMG offset (Fig. 4B). Similar relationships to those with the onset of the period of activity in the ClB/Br were found with respect to paw lift, whereas similar relationships to those with respect to the offset of the period of activity in the ClB/Br were found with respect to the lever press (data not shown; Fig. 3).
Note that in most cases the onset of the significant change in activity of these cells occurred at or after the onset of activity in the ClB/Br, as indicated by the positive intercepts of many of the regressions (Fig. 4A–C). Indeed, the average onset latency of the 24 cells in this category was 61 ms after the ClB/Br onset. The earliest discharge was 146 ms before ClB/Br onset.
The distinction between dynamic and tonically discharging cells is supported by the scatter plots of Figure 4D. The dynamic transport-related cells (plotted in blue) form a tight homogeneous population, with respect to latency of onset and duration. The tonic cells (cyan symbols) are more heterogeneous and spatially distinct, with little or no overlap with the dynamic cells.
Cells related to the lever press
A substantial number of cells (53/91, 58%) showed task-related activity during the contralateral lever press; 38/53 (72%) showed an increase, and 15/53 (28%) showed a decrease. Of these, 34/38 showed a dynamic increase in activity, and 9/15 showed a dynamic decrease.
The majority of the cells with a dynamic increase in activity (28/34, 82%) exhibited a single peak of activity as shown for two cells recorded from the EP in Figure 5, A and C. Linear regression analysis (Fig. 5B,D) showed that the onset and offset of the period of increased cell discharge correlated well with the onset and offset of the contralateral lateral head of triceps (coTriL). In both cells, the increase in cell discharge occurred following the onset of activity in the coTriL (Fig. 5B, intercept of linear regression of 138 ms, D, 342 ms). The other 6/34 (18%) cells had bimodal activity profiles similar to those illustrated in Figure 3, C and E. In these cells, an initial increase in activity during (Fig. 3C) or before (Fig. 3E) the transport phase was followed by a decrease in activity (Fig. 3C, not significant) and then a second increase during the lever press. The other 4/38 cells with increased activity showed a tonic pattern of activity (see Fig. 7E)
An example of a cell with a dynamic decrease of activity during the lever press is illustrated in Figure 5E. This cell showed a reciprocal pattern of activity in which there was a dynamic increase of activity during the transport phase followed by a decrease in activity during the lever press. The onset of the decrease in cell activity showed a linear relationship with the onset and offset of the TriL (Fig. 5F). As for the cells with increased activity, some of the cells with decreased activity (6/15) showed a tonic decrease following the onset of the lever press period.
Only 4/34 cells with a dynamic decrease in activity during the lever press phase showed a dynamic increase in activity during the transport phase, whereas two other cells (2/34) showed the opposite pattern of activity. Overall, therefore, 28/34 (83%) of cells with dynamic activity during the lever press showed no dynamic activity during the transport phase.
We compiled linear regressions for 28 cells showing a dynamic increase during the lever press, 22 with a simple unimodal increase and 6 with a bimodal pattern of activity. Overall, 12/22 (55%) of the cells with a unimodal increase and 6/6 cells with a bimodal profile showed a significantly linear relationship between the onset of cell activity and the onset of Tri/TriL EMG activity (Fig. 6A,B). A similar linear relationship was found for 5/9 of the cells showing a significant dynamic decrease (Fig. 6C). In most of these cases, the intercepts were positive (Figs. 6D, 5A,B), indicating that increased cell discharge followed the onset of the muscle activity responsible for the lever press. A similar number (10/22, 45%) of unimodal cells was significantly related to the end of the Tri/TriL activity (Fig. 6A). A significant relationship between the end of the cell activity and the end of the Tri/TriL activity was equally found in 6/9 cells with decreases in activity and in 2/6 bimodal cells (data not shown).
Figure 6E plots the time of the beginning and end of the period of the discharge for those cells that were dynamically modified during the lever press (red, green, and black symbols, key at bottom) and compares them to those cells that were dynamically active during the transport phase (blue). All of the lever press cells are clearly separated in time from the transport cells, and there is an extensive overlap between the three types of lever press cells illustrated. Like the cells active during the transport phase, the cells showing a single unimodal period of activity (red and black symbols) during lever press show relatively homogeneous onsets and offsets of activity. The separation between transport-related and lever-press-related cells is also illustrated in the histograms plotting the time of averaged peak activity (Fig. 6F) and the time of the averaged onset of activity (Fig. 6G). The minimal overlap in activity (Fig. 6F, one case, G, two cases) is explained by differences in the average reach times during the recordings from different cells.
Cells related to the Go-signal
A total of 18/91 (20%) cells began to discharge at short latency following the Go-signal (<50% of the period between the Go-signal and the onset of the reach) and were identified as potentially Go-related cells. Of these, 14 cells (11 with increased activity and 3 with decreased activity) showed a significant linear relationship with lead time and coClB/Br onset, but not between cell onset and the onset of EMG activity (Fig. 2D–F). The linear regressions for these cells are illustrated in Figure 7A and are contrasted with the activity from the 19 dynamic, transport-related cells (Fig. 4A,B), which showed the opposite pattern—significant relationships between the onset of cell activity and ClB/Br onset but not between lead time and EMG onset (Fig. 7B).
To better contrast the properties of the putative Go-related cells with the transport-related cells, Figure 7, C and D (blue traces), plots the distribution of the mean latencies of the onset of the cell discharge (Fig. 7Ci,Di) and the variances of these mean values (Fig. 7Cii,Dii) for these two populations of cells. Figures 7, C and D (red traces), similarly show the distributions of the mean latency and the variance for the onset of the period of activity in the ClB/Br. For the putative Go-related cells, the latency for the onset of the discharge and the variance of that discharge was significantly different from the latency and variance of the onset of the ClB/Br (p < 0.001; Fig. 7Ci,ii, Table 2). In contrast, both the latency of the onset of the cell discharge of the transport-related cells, and the variance of that discharge, was not significantly different from that of the onset of the ClB/Br (Fig. 7Di, p = 0.68, Dii, p = 1.0). Further, in comparing the two populations (not directly illustrated), there was no significant difference between the mean latency of the onset of the ClB/Br measured from the population of transport-related and Go-related cells (p = 0.44) or for the variance (p = 0.09; Table 3). However, there were significant differences (p < 0.001) for the two populations of cells for both the mean of the latency of the cell discharge (Fig. 7Ci,Di, compare blue traces) and the variance of that discharge (Fig. 7,Cii,Dii compare blues traces, Table 3).
In sum, the population of Go-related cells is defined by the lack of variance in their discharge pattern, indicating that they fire at a relatively constant latency following the Go stimulus, independent of the reaction time.
In considering whether the 4/18 cells with short latency changes of discharge, but for which we could not measure onsets from individual trials, should also be considered as Go related, we measured the average onset latency, with respect to the Go-signal, from the perievent histograms (PEHs) compiled for the 24 dynamically active transport-related cells. Within this population, the shortest latency onset was 243 ms. For the 14 Go-related cells the longest onset latency (again measured from the average PEHs) was 161 ms. For the four cells for which we could not measure individual trials, the longest averaged latency was 110 ms, suggesting that these four cells are also Go related. We also re-examined all cells with onset latencies between the Go-signal and the onset of reach to determine whether other cells should be included as Go-related on the basis of this post hoc analysis, but no further cells fulfilling this criterion were found.
The 18 Go-related cells showed a heterogenous pattern of activity. In 7/18 cells, the Go-related discharge continued during the period of the transport phase; these cells were included in the tonically discharging transport-related cells. A further 3/18 cells showed both a Go-related discharge and a later, distinct, dynamic discharge during the transport phase (Fig. 2D–F). A third of the cells (6/18) showed significant Go-related activity and then a secondary period of activity during the lever press (Figs. 3E, 7E), whereas 2/18 cells showed sustained activity that continued from the Go-signal to the end of the lever press (Fig. 7F). Only 2/18 cells showed an isolated change in activity that lasted until the onset of the reach.
The scatter plot of Figure 7G compares the timing of the population of Go-related cells with that of the two populations of transport-related cells (dynamic and tonic). As dictated by their definition, the population of Go-related cells started to discharge earlier than the dynamically active transport-related cells, and a proportion of cells in which discharge onset was related to the Go-signal continued to discharge throughout the transport period and were, therefore, also identified as tonic transport-related cells (see above).
Cue-related cells
A small population of cells (8/91, 9%) discharged following the onset of the instruction cue (Fig. 1A) and continued to discharge either until the onset of the Go-signal (four cells) or until the end of the transport phase (four cells; Fig. 8A,B). Note that in the illustrated example (Fig. 8A), the cell increased its discharge activity consistently at the time of the instruction cue (indicated by the cyan squares), although there was no noticeable increase in Fv between this time and the Go-signal. This suggests that this cell was not discharging in association with changes in the general posture of the cat. Indeed, no clear changes in Fv were observed in any of the cells showing changes in activity during the cue period.
Other cells
A further population of cells (19/91) showed significant changes in discharge activity late in the behavior with the peak of activity occurring subsequent to the end of the lever press period. These cells were not analyzed further.
Population activity
A summary of the changes in discharge activity for all of those cells showing significantly modulated activity during the contralateral reach (68 cells in the EP and the 23 cells in the GPe) is illustrated in Figure 9. The figure shows phase plots that indicate the periods during which significant changes of activity were observed (Fig. 9A,B), as well as heat maps that illustrate the relative discharge frequency of the cells. Most of the dynamic changes in cell activity during the task, for both EP and GPe cells, occurred during the transport (red rectangle) and the subsequent lever press. The majority of cells in both the EP and the GPe showed increased activity over baseline during the task, although cells showing decreases are also well represented, especially for the EP cells during the transport phase of the reach (see above). We saw no major differences in our population of GPe cells that might indicate recordings from different functional populations (e.g., arkypallidal vs prototypic populations; Mallet et al., 2012); however, we note that the number of task-related cells that we recorded was relatively small.
Activity during ipsilateral reach
Most of the cells (108/116) recorded during the contralateral reach were also recorded during at least four reaches with the ipsilateral limb. A total of 75/108 (69%) cells showed at least one period of significant activity during the ipsilateral reach (60 EP cells and 15 GPe cells), and 65/75 (87%) of these cells were significantly modified during both contralateral and ipsilateral reach (Table 2).
In general, the distribution of periods of task-related activity during the ipsilateral reach was similar to that during the contralateral reach. This is demonstrated for the 60 task-related EP cells by the plots of Figure 10, A and B, with cells showing increases of activity preceding the onset of the reach, during the transport phase (red rectangle), and during the lever press. A similar pattern of activity is seen for those cells showing decreases in activity. A direct comparison of the total population discharge of the cells during the contralateral and ipsilateral reaches (Fig. 10C) shows that equal percentages of cells were significantly modified at similar periods of the reach during the performance of both the contralateral and ipsilateral reach. Data are not shown for the GPe because of the small size of the database (N = 15). However, the basic characteristics of the cells were similar to those illustrated in Figure 10B.
Cells active during the transport phase
Although the population activity of the pallidum was similar during contralateral and ipsilateral reaching, the same was not always true for individual cells. For example, only 19/40 cells active during the transport phase in the contralateral reach were also significantly modified during the transport phase in the ipsilateral reach. Among those cells that showed dynamic increases in discharge activity during the contralateral reach, only two also showed a dynamic increase during the transport phase of the ipsilateral reach, and five showed no task-related activity during this same period. (Table 4). For example, the cell illustrated in Figure 11A showed a strong increase in activity during the contralateral reach (red trace) but no change in activity during the ipsilateral reach (green trace). Cross-correlation of the signals during contralateral and ipsilateral reach showed a correlation coefficient close to zero at zero lag (Fig. 11Fi). The example illustrated in Figure 11B showed significant changes in activity during the ipsilateral reach, but the large increase in activity observed during the transport phase of the contralateral reach was replaced during the ipsilateral reach by a decrease in activity at the same phase and a small increase before onset of activity in the ClB/Br. In contrast, the activity observed during the lever press in this cell was maintained during the ipsilateral reach. The complex change in the activity pattern is reflected in the equally complex pattern of the cross-correlation (Fig. 11Fii).
Cells that showed dynamic decreases in activity during the transport phase of the contralateral reach equally showed variable responses during the ipsilateral reach. Only 4/12 cells showed a similar pattern during reaches with either limb (Table 4), one example of which is illustrated in Figure 11C, Fiii. Two other cells with decreased activity during the transport phase during the contralateral reach instead modified their activity during the ipsilateral lever press, and another three were not task related during the ipsilateral reach (Table 4).
The most homogeneous group comprised those cells showing a tonic discharge during the contralateral reach. In this group, 7/16 cells discharged in a similar manner during reaches of either limb as illustrated in Figure 11D,Fiv. Among the other tonic cells, 4/16 were not task related during the ipsilateral reach.
A comparison of the change in population activity can be obtained from the summary of the cross-correlation coefficients illustrated in Figure 11, G and H. Most of the correlation coefficients at zero lag were <0.5, and a proportion (13/40, 32%) were <0.0, indicating opposite patterns of activity in the contralateral and ipsilateral reach. This was especially evident for those cells showing increased dynamic activity in the contralateral reach (Fig. 11G, red bars). Among the cells that showed relatively high correlation coefficients (>0.5) and so were modified at the same phase of the movement (Fig. 11H, cells top right), the majority were cells with dynamic decreases in activity during the transport phase and those classified as having a tonic discharge during transport.
Although there were changes in the classification of many of the cells when comparing the significant changes in activity in contralateral and ipsilateral reach, overall this population of cells showed only minor changes in their time of activation in contralateral and ipsilateral reach. This is illustrated in the scatterplots of Figure 11I, where the blue and cyan symbols represent, respectively, dynamically and tonically discharging cells related to the transport phase during contralateral reach (as in Fig. 4D). The red and green symbols similarly represent, respectively, the time of activation of these same cells (for those that were modified) during the ipsilateral reach. Red and green symbols represent, respectively, cells that discharged dynamically during contralateral reach and those that discharged tonically.
Examining the cells active during transport during the ipsilateral reach in their own right (i.e., regardless of their activity during contralateral reach) showed that a total of 31/76 cells discharged during the transport phase during ipsilateral reach. This included the 19 cells that also discharged during transport in contralateral reach (Table 4). The other 12 cells included 6 cells that discharged during the lever press during the contralateral reach and 2 that were not modified during the contralateral reach (Table 5).
Cells active during the lever press
In contrast to the cells active during the transport phase, many of the cells (22/34, 65%) that had significant, dynamic increases in discharge activity related to the lever press during the contralateral reach showed similar patterns and magnitudes of activity when recorded during ipsilateral reach (Table 4). This can be observed in the two examples of Figure 12, A and B, which show overlapping waveforms during the contralateral and ipsilateral reach, as reflected in the high correlation coefficients in the cross correlograms at zero lag (Fig. 12D,E). In 7/34 other cells there was no significant change in activity during the ipsilateral reach (Fig. 12C,F).
Of those few cells with decreased activity during contralateral lever press, 4/9 also showed decreased activity during the ipsilateral lever press. Most cells with sustained increases or decreases in activity following the lever press during contralateral reach equally maintained their activity during ipsilateral reach (Table 4).
Altogether, 43 cells showed a significant discharge during the lever press during the ipsilateral reach (Table 5). Of these, 24/43 showed a dynamic increase in activity with 20/24 of these cells discharging in a similar manner during the ipsilateral reach (Table 5). Indeed, the majority of the cells active during the ipsilateral reach discharged in the same manner during the contralateral reach (Tables 4, 5, values in boldface).
Examining the entire population, the majority of cells had positive correlation coefficients at zero lag (Fig. 12G,H) indicating that they discharged with the same sign of activity (i.e., both increased their activity or both decreased). Moreover, in many cases, both for cells showing a unimodal decrease (blue symbols) or increase (red symbols) during the contralateral reach, the cells lay on the line of equivalence (Fig. 12H). This means that the maximum correlation coefficient occurred at zero lag, indicating that the activity pattern of the cells was similar in the two conditions. Only two cells showed a negative correlation at zero lag, indicative of a reciprocal change in activity during the ipsilateral reach.
The scatter plots of Figure 12I show that the timing of the responses during contralateral and ipsilateral reach remained relatively constant for most cells. This is not unexpected as many of the cells active during the lever press maintained the same pattern of activity in ipsilateral reach (Table 4). In some cases, however, there were differences in the time of the response leading to a greater dispersion of the responses during ipsilateral reach (Figure 12I, cyan and green symbols) than during the contralateral reach (blue and red symbols). This is shown more clearly by the relative size of the ellipses describing the dispersion of the responses.
Go-related cells
Among the 18 cells that were classified as discharging in response to the Go-signal during the contralateral reach, 9/18 also showed Go-related activity during the ipsilateral reach, although the overall pattern was not necessarily identical. A further 3/18 cells showed no significant activity during the ipsilateral reach, one cell was not recorded in ipsilateral reach. The other 5/18 cells showed diverse patterns of activity. A further three cells showed Go-related activity only during the ipsilateral reach.
Other cell types
During the contralateral reach, eight cells showed changes in discharge activity that followed the instruction cue and preceded the Go stimulus. Half of these cells (4/8) showed similar changes during the ipsilateral reach, whereas the other half showed no significant change in activity during the ipsilateral reach. A further seven cells showed changes during the instruction period only during the ipsilateral reach.
A summary of the timing of the different cell populations is provided in Figure 13, A–C. Figure 13A summarizes the activity of the major populations of cells in which the discharge activity was significantly modulated during reach of the contralateral forelimb. As detailed above, cue-related cells discharged initially with a range of onset latencies, followed by the Go-related population in which the onset of activity occurred in a narrow range but for which the end of the period of activity was more widely distributed. Cells with dynamic responses related to the transport and lever press periods discharged in relatively discrete clusters in the space. A similar distribution was observed during ipsilateral reach (Fig. 13B). Finally, Figure 13C classifies the population of cells according to the pattern of activity during contralateral reach and then plots the significant changes of activity of these cells during ipsilateral reach. This serves to emphasize that those cells that showed significant changes of activity both during contralateral and ipsilateral reach tended to discharge in a generally similar pattern of activity (at least with respect to timing).
Localization
The cells recorded in this study form a subset of those used in a previous report (Mullié et al., 2020) and covered almost the entire extent of the EP and the GPe (Fig. 14). Cells showing dynamic increases in activity during the transport phase of the reach were more localized in the caudal and lateral regions of the nucleus (Fig. 14A, red symbols). This region overlaps with the region from which we found cells having receptive fields including the contralateral forelimb, as well as cells that showed strong modulation of their activity with the peak of activity during the swing phase of the step cycle (Mullié et al., 2020, their Suppl. Fig. 4). Fewer cells showed dynamic decreases (black symbols) and there was less of an indication of any localization.
Cells with dynamic increases of activity during the lever press also showed a tendency to be more concentrated in the caudal and lateral regions of the EP, although other cells were found in anteromedial regions (Fig. 14C). The number of cells recorded in the GP was too small to determine any signs of localization, but data are shown in Figure 14, B and D.
Comparison with PMRF and motor cortex
To facilitate comparison with the results from other studies from this laboratory using the same task (Schepens et al., 2008; Yakovenko et al., 2011), we reanalyzed data obtained from the motor cortex and the PMRF using the same criteria as in the present study. Motor cortex data were taken from recordings contralateral to the reach, whereas recordings from the PMRF were ipsilateral. The results of this analysis (Table 6) show both similarities and differences in the patterns of activity in these structures with respect to those presented in this article. Similarities include the presence of neurons that discharged both during the transport and lever press periods of the activity, including cells that discharged both dynamically and tonically. In addition to the higher percentage of cells in both the motor cortex and the PMRF that showed significant changes in discharge activity during the transport phase than did cells in the pallidum, it should be noted that the nature of the discharge in these two structures was also quite different. A majority of cells in the motor cortex showed dynamic discharges during the transport phase of the reach, as illustrated in Figure 15A (Yakovenko et al., 2011). Cells with dynamic discharges were also observed during the lever press period, and 30 cells (27%) showed dynamic changes of activity during both the transport and lever press periods. Of these, 25/30 showed a reciprocal pattern of activity, increasing in one phase (transport or lever press) and decreasing in the other. This may be compared with the results from the pallidum where only 10 cells (11%) showed dynamic changes in activity in both the transport and lever press phase with six cells showing reciprocal changes in activity. In contrast to the cells in both the pallidum and the motor cortex, the majority of cells in the PMRF showed a more tonic pattern of discharge that began at the onset of the Go-signal and continued throughout the transport and lever phase (Fig. 15C). Such cells are defined using different definitions and terminologies in this article and in previous papers from this laboratory. Using the definition used here and in Yakovenko et al. (2011) such cells are classified as showing significant but tonic changes in activity during both the transport and the lever press periods. However, using the definitions in Schepens et al. (2008), they would be described as showing a tonic pattern of activity throughout the movement (Table 4, entire period).
In addition, we reported previously (Schepens et al., 2008) that many cells in the PMRF showed a bimodal pattern of activity, as illustrated in Figure 15D. In that study, cells were defined as bimodal on the basis that cell discharge during two peaks exceeded the intervening minima by ≥150%. Using this definition, bimodal cells were also found in the pallidum (Fig. 3), as well as in the motor cortex (Fig. 15B), although in these structures the percentage of bimodal cells was noticeably lower (Table 4). Two patterns of bimodal activity were observed in all three structures. In the first of these patterns, bimodal cells showed peaks during the transport phase and then again as the limb was lifted from the lever, as in Figure 15, B and D. In the second case, cells exhibited one peak before the onset of the Go-signal, a decrease in activity during the transport phase, and another peak during the lever press (Figs. 3E, 7E). Bimodal cells active during transport and during removal of the paw from the lever were most frequent in the motor cortex and the PMRF, whereas the two patterns of bimodal cell were more equally represented in the pallidum.
It should also be noted that there were cells in the PMRF that showed dynamic changes in activity that were restricted to either the transport or the lever press phases (Table 4; Schepens et al., 2008; Schepens and Drew, 2004). Cells showing dynamic discharges during the transport phase frequently discharged from the onset of the Go-signal to the end of the transport phase in a manner similar to the initial peak of activity illustrated in the bimodal cell of Figure 15D.
Discussion
The results in this article confirm and expand on previous studies showing a contribution of cells in the pallidum to the control of both flexor and extensor muscle activity, and this in both the limbs contralateral and ipsilateral to the recording site. The results also show a contribution of pallidal cells to the control of both posture and movement, although cells related to the dynamic aspects of the movement predominate.
Discharge patterns during contralateral reach
Although our task contained a major postural component (see below), the general characteristics of the cells recorded during the reach of the contralateral forelimb were compatible with those described previously in the pallidum in cats (Cheruel et al., 1994) and primates (DeLong, 1971; Iansek and Porter, 1980; Georgopoulos et al., 1983; Anderson and Horak, 1985; Hamada et al., 1990; Nambu et al., 1990; Mink and Thach, 1991a,c; Turner and Anderson, 1997; Schwab et al., 2020). Moreover, as in most studies (Mink and Thach, 1991a,c; Cheruel et al., 1994; Arimura et al., 2013; Cui et al., 2013; Mullié et al., 2020), the general properties of cells in the EP (GPi) and the GPe were similar, so the populations are described together in the following discussion.
In agreement with most studies, we observed cells that showed both dynamic increases and decreases in discharge activity when flexor muscles were activated during the transport phase of the reach. However, we also observed a larger proportion of cells that showed a dynamic discharge during the lever press when extensor muscles were activated. In most cases, cells were related to either one or the other, that is, the transport or the lever press phase. For example only 17% of cells showing dynamic changes in activity during the lever press also discharged dynamically during the transport phase. These properties are similar to those recorded from the motor cortex in a similar task (Yakovenko et al., 2011).
In most cells, the onset of the significant change in cell discharge was coincident with, or followed, the onset of activity of the prime muscles involved in the transport and lever press phases. This is in agreement with previous studies (Anderson and Horak, 1985; Mink and Thach, 1991c; Turner and Anderson, 1997; Schwab et al., 2020) and is consistent with the view that the pallidum is not involved in movement initiation but might instead serve to modify parameters of the movement such as scaling or vigor (Anderson and Horak, 1985; Desmurget and Turner, 2008; Turner and Desmurget, 2010). However, it cannot be discounted that the discharge might be related to other muscles that only become active later than those used to define movement onset in the current study (Yakovenko et al., 2011).
A contribution to the control of posture
On the basis of observation in patients with Parkinson's disease, it has been suggested that the basal ganglia might contribute to postural control, including the production of APAs (see above, Introduction). On this basis, we expected to find cells with similar profiles to those that we have previously recorded in the PMRF (Schepens and Drew, 2004, 2006; Schepens et al., 2008). In particular, cells involved in contributing to the APAs would be expected to discharge at short latency after the Go-stimulus, whereas cells involved in producing postural responses during the movement would be expected to show tonic, or bimodal responses, throughout the movement (Fig. 15C,D). In agreement with this hypothesis, we found both cell types in our sample, including some cells that discharged during the APA (Go related) and then again during the lever press. Such cells could contribute to postural control both before and during the movement.
However, although present, both tonic discharging cells (10% of the population) and Go-related cells (20%) were relatively infrequent. This is a much lower proportion of cells than those observed in the PMRF (Table 4). Possibly, we underestimated the number of Go-related cells because responses related to the APA were more difficult to identify in the pallidum because of the irregular background discharge of the cells. However, this would not affect the identification of cells discharging tonically throughout the movement sequence to produce postural support. The relative paucity of either cell type suggests that although the pallidum contributes to postural control, it may not be a major function. However, this does not rule out a more important basal ganglia contribution to the control of posture as other structures, including the substantia nigra or pars reticulata, which has been shown to exert strong effects on posture during locomotion (Takakusaki et al., 2003, 2004) may contribute.
It should also be noted that when Go-related responses were observed, they were frequently in combination with discharge activity related to the movement (Figs. 2C–E, 7E); only 2/18 Go-related responses during contralateral reach were observed in isolation from other significant discharges later in the movement. This suggests that movement and posture (at least with respect to signals related to the APA) are not represented as separate streams in the pallidum.
Last, these results suggest that the pallidum is not a major source of the posture-related discharges observed in the PMRF, although it is possible that the latter might be formed as the result of the integration of signals from pallidal cells discharging during the transport phase and the dynamic phase.
Cells are related to both ipsilateral and contralateral movements
As in our locomotor study (Mullié et al., 2020), we found a much larger proportion of cells active during ipsilateral movements in this task (76/116, 66%; Table 2) than in previous studies (29% in Iansek and Porter, 1980; 12–18% in DeLong, 1971). This difference with respect to the previous reaching studies probably relates to the nature of our task, which, as during locomotion, requires coordinated activity on both sides of the body.
The majority of cells that showed similar activity during both contralateral and ipsilateral movements were those showing significant changes of activity during the lever press. This might reflect a contribution of the pallidum to the production of extensor muscle activity in the contralateral and ipsilateral limb during lever press of the respective limb, although the contralateral projections from the GPi to the thalamus are relatively weak (Hazrati and Parent, 1991). Alternatively, it might indicate a contribution to the activation of the contralateral extensor muscles during both the contralateral and ipsilateral reach, reflecting the bilateral nature of the corticostriatal inputs (Künzle, 1975; Ragsdale and Graybiel, 1981). During the contralateral reach, the contralateral extensor muscles act to extend the limb and depress the lever. During the ipsilateral reach, these same contralateral muscles are activated in a postural context as the ipsilateral limb presses on the lever. This would suggest that some pallidal cells regulate activity in a given muscle group in a context-independent manner. However, it should be noted that the waveform of the neural response follows more the dynamic nature of the lever press (Fig. 12A–C) than the more tonic nature of the postural responses reflected in the ground reaction forces during support (Fig. 1B). This is in contrast to the more tonic nature of the responses observed in reticulospinal neurons in the PMRF (Fig. 15C). The pallidal neurons might, therefore, act preferentially on the posture support required specifically during the lever press than on the more tonic postural responses required during the entire period that the contralateral limb is removed from the support surface.
It is also noteworthy that limb-independent cells were less frequent among those cells activated dynamically during the transport phase. Only 2/12 cells with a dynamic increase and 4/12 with a dynamic decrease during the contralateral reach showed the same pattern of response during the ipsilateral reach. This suggests a more specific, limb-dependent control over the activity of the flexor muscles than of extensor muscles.
Summary
This study allowed us both to examine the relative contribution of the pallidum to the control of movement and posture and to compare these results with the activity of two structures that we have previously examined under the same task conditions, namely, the motor cortex and the PMRF. In general, it is clear that the general characteristics of the cell activity in the pallidum are more similar to those recorded in the motor cortex than in the PMRF. In both the motor cortex and the pallidum, cells discharged primarily with a dynamic pattern of activity, both with respect to the transport phase and the lever press. We suggest, therefore, that most cells in the pallidum are likely to be influencing motor activity via the thalamocortical pathway. Nonetheless, a postural component, related to both the APAs and the overall postural control was observed in a population of pallidal cells, and this may modulate postural control via the pallidal projections to the pedunculopontine nucleus (Filion and Harnois, 1978; Garcia-Rill, 1986; Parent et al., 2001; Caggiano et al., 2018).
Footnotes
- Received March 2, 2022.
- Revision received May 15, 2022.
- Accepted June 7, 2022.
This work was supported by an operating grant (PJT-156281) from the Canadian Institutes for Health Research and a studentship from the Fonds de Recherche Santé, Quebec, to Y.M. We thank T. Ariel, M. Bourdeau, N. De Sylva, P. Drapeau, F. Lebel, and J. Soucy for technical assistance in the performance and analysis of these experiments; Dr. Stephane Menard for veterinary assistance; Dr. Elaine Chapman for comments on this manuscript; and Nabiha Yahiaoui, who participated in some of the experiments.
The authors declare no competing financial interests.
- Correspondence should be addressed to Trevor Drew at Trevor.Drew{at}umontreal.ca
- Copyright © 2022 the authors