| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Department of Neurology, Emory University School of Medicine, Atlanta, Georgia 30322
Submitted 31 December 2002; accepted in final form 4 September 2003
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
, 1990; Hoover and Strick 1993; Middleton and Strick 1997, 2002). These include motor, oculomotor, limbic, and associative circuits, which are thought to remain segregated throughout their subcortical course (see, e.g., Middleton and Strick 2002 for a recent review). Knowledge about the distribution of these circuits, and about their interaction at the level of the basal ganglia output nuclei, the internal pallidal segment (GPi) and the substantia nigra pars reticulata (SNr), is central to understanding basal ganglia function.
There is strong anatomic evidence that motor, limbic, and associative circuits are represented in GPi and SNr, although the distribution of the circuits and the degree of convergence of information among them remains controversial. It appears that GPi can be grossly divided into a large motor and lesser associative and limbic regions (e.g., Parent and Hazrati 1995), but in the SNr, motor, limbic, and associative circuits appear to intermingle. The cortical sources of information that reach the different territories in the SNr may differ from those reaching the equivalent territories in GPi with a possible preponderance of premotor and associative circuits in the SNr and activity related to movement execution in GPi. Nigral projections are in large part directed at portions of the ventrolateral and ventral anterior thalamus that have strong interactions with frontal and prefrontal cortical territories that differ from those targeted by GPi output (e.g., Ilinsky et al. 1985, 1993). Recent studies using retrograde transneuronal transport of viral tracers have shown, however, that both, GPi and SNr also project (via the thalamus) to primary motor cortex (Hoover and Strick 1999).
The involvement of the basal ganglia output nuclei in motor functions is particularly interesting because these structures are thought to participate in the development of movement disorders, such as Parkinson's disease. GPi neurons clearly participate in motor functions, but the involvement of the SNr in motor control is less clear. Previous electrophysiologic studies investigating task responses of SNr neurons in primates (DeLong et al. 1983; Lestienne and Caillier 1986; Magarinos-Ascone et al. 1992; Mora et al. 1977; Nishino et al. 1985b, 1991; Schultz 1986) or responses to somatosensory examination (DeLong et al. 1983; Magarinos-Ascone et al. 1994; Schultz 1986) have suggested that a portion of the motor circuit passes through the SNr. In addition, we have recently shown that in parkinsonian animals, SNr neurons show discharge abnormalities akin to those appearing in GPi and STN (Wichmann et al. 1999) and that some of the parkinsonian motor abnormalities can be reversed by inactivation of the SNr (Wichmann et al. 2001). On the whole, however, the contribution of the SNr to motor functions remains enigmatic because many of the published studies are contradictory, perhaps due to technical differences (such as sampling biases) or to the possible existence of significant interactions between motor and nonmotor functions in the SNr, which may impact task performance.
With the experiments described here, we sought to re-examine the role of the primate SNr in motor functions by investigating the neuronal responses of SNr neurons in two different motor tasks.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
Two juvenile Rhesus monkeys (Macaca mulatta, 4–6 kg), were each trained to perform two different visuomotor tasks (see following text). All surgical and experimental protocols were approved by the Institutional Animal Care and Use Committee at Emory University and were in accordance with guidelines of the National Institutes of Health (Guide for the Care and Use of Laboratory Animals 1996).
Prior to behavioral training, the animals were acclimated to a primate chair and trained to tolerate a physical examination that consisted of manipulation of hip, knee, ankle, shoulder, elbow, wrist, and trunk as well as light touch of multiple body regions (including limbs, trunk, and face).
Behavioral paradigm
Each animal was seated in a primate chair, facing a computer monitor, with the left forearm placed in a low-friction manipulandum and the right arm lightly restrained in an armrest at the animal's side. A torque motor was attached to the manipulandum for application of torque pulses to the elbow. In addition, the manipulandum was coupled to a precision potentiometer for recording of the position of the monkey's forearm. For the behavioral training, the animals received rewards consisting of liquid food delivered through a spout which was positioned close to the animal's mouth.
The behavioral tasks were controlled by a computer using custom-designed software. The animal controlled the position of a small cursor on the computer screen through horizontal forearm movements and was required to align the cursor within circular targets (diameter 18 mm) by moving the manipulandum in either the flexion or extension directions. All timing parameters, as well as movement and torque directions, were randomized.
Two different tasks, a step-tracking task (STT) and a delayed-response task (DRT), were used (see Fig. 1). For both tasks, four trial classes (elbow flexion and extension movements, each with either flexion or extension torque pulses) were presented in a pseudorandom sequence in blocks of 20 trials/class. Both STT and DRT trials were presented with an inter-trial interval of 2–4 s (randomized).
|
In the DRT, the animal also had to capture the center target, followed by torque application and re-capture of the center target, as described in the preceding text. After a randomized hold period, a 250-ms instruction cue was displayed. The instruction cue indicated the location of the correct side target for the upcoming elbow movement. During the instruction cue and thereafter, the center target remained illuminated and the monkey had to maintain the alignment of the cursor on the center target. The `Go'-signal consisted in this case of extinction of the center target and simultaneous illumination of both side targets, 5–7 s (randomized) after the instruction cue. The maximum time allowed to capture the side targets was 450 ms, followed by another hold period of 1–1.5 s, after which the reward was given.
The two tasks differed substantially. In both tasks, the animal knew that a movement had to be made to obtain a reward, and the kinematic specifics of the required movements were the same. However, in the STT, the target location and the direction of movement was not known prior to target presentation, whereas the animal received advance information regarding the target location in the DRT. Therefore the animal could prepare for the specific upcoming movement in the DRT but not in the STT. Another difference is that at "target-on," the target location was unequivocally shown in the STT, while the animal had to use the information provided by the preceding cue to select the correct target (of the 2 shown) in the DRT. Above-chance performance in the DRT depended, therefore, on memory, whereas performance in the STT did not.
Surgical procedures
On completion of behavioral training, chronic metal recording chambers (16 mm ID) were stereotaxically positioned over a trephine hole under aseptic conditions and gas anesthesia with isoflurane (1–3%). The target area, the SNr, was approached in the parasagittal plane 25° posterior from the vertical. The chamber was affixed to the skull with dental acrylic. Bolts were embedded in the acrylic assembly to permit head fixation during the recording experiments. In monkey B, two commercial Ag/AgCl EOG electrodes (In Vitro Metric, Healdsburg, CA) were implanted during the chamber placement surgery. For this, a small (5 mm) incision was made lateral to the orbit, and a 2- to 3-mm-deep defect was drilled in the underlying bone. The electrooculogram (EOG) electrodes were fixed to these shallow bone defects to record horizontal eye movements during task performance.
Recording experiments
Monkeys sat in the primate chair with their heads restrained while performing the behavioral tasks. Glass-coated platinum-iridium microelectrodes (impedances 0.5–1 M
at 1 kHz) were used to record extracellular potentials. A hydraulic microdrive (MO-95B, Narishige, Tokyo), coupled to a potentiometer to yield a digital depth readout was used to advance the microelectrode into the brain through a guide tube. The signal was amplified (DAM-80 amplifier, WPI, Sarasota, FL; MD-2 amplifier, BAK, Germantown, MD), displayed on a digital oscilloscope (DL1540, Yokogawa, Tokyo), and audio-amplified. SNr cells were isolated and identified by their characteristic high-frequency discharge and by their spatial relationship to neighboring structures, such as thalamus, STN, and cerebral peduncle. Neuronal spikes were isolated using a spike-sorting device that selects neuronal spikes based on their waveform (MSD, Alpha-Omega Engineering, Nazareth, Israel). The device produces a TTL pulse whenever a spike is detected. The timing of these pulses, measured in the form of inter-spike intervals (ISIs), was stored to computer disk, along with the forearm position signal from the manipulandum and with digital codes that indicated the timing of task events.
Possible relationships between the neuronal activity of a given cell and passive joint manipulations were evaluated by listening to the audio representation of the recorded signal while a complete sensorimotor examination of the body was carried out (see preceding text).
Histology
At the end of the experiments, the animals were deeply anesthetized with pentobarbital (100 mg/kg) and perfused transcardially with cold oxygenated Ringer solution, followed by a mixture of 4% paraformaldehyde and 0.1% glutaraldehyde in phosphate buffer (0.1 M, pH 7.4). The brains were blocked, frozen, and sectioned in the parasagittal plane in 50-µm sections. Alternating sections were stained with a cresyl violet or with a tyrosine hydroxylase staining protocol. Recording sites were reconstructed based on the linear gliosis associated with each microelectrode track and electrophysiological information pertaining to the borders of neighboring nuclei.
Data analysis
As mentioned in the preceding text, neuronal data (in the form of ISIs) were collected on-line to computer disk along with the timing of all task events and with analog data representing manipulandum position and EOG (in monkey B only). Throughout the text, the term "event" will encompass both sensory cues provided to the animal and movements made by the animal in response to these cues.
The activity of each neuron was analyzed off-line using statistical methods implemented as an interactive program in the Matlab environment (Mathworks, Natick, MA). Although for some aspects of the analysis it would have been interesting to compare SNr responses between successful (rewarded) and unsuccessful (unrewarded) trials, the number of unsuccessful trials was too small for a meaningful analysis; thus only successful trials were considered further. The data set for a cell needed to contain at least five successful trials of a given task type to be entered into the data base.
Each trial was divided into epochs, centered around the "torque on," "cue on," target on, and reward events. In addition, the epoch 300–800 ms after the initial center target capture was defined as the control epoch. During this epoch (labeled CP in Fig. 1), the monkey was required to hold the manipulandum steady, waiting for the next event (the torque-on event). The control epoch represents a period of stable performance; the requirements in successful trials are consistent from trial to trial. In addition, this epoch is not likely to be influenced by cue or target presentation or by the application of reward, because those events occur later in the trial. The discharge rate from this epoch was used to express the average discharge rate of neurons and for statistical comparisons of the peri-event activity of all other epochs.
Spike trains from individual trials were aligned according to their temporal relationship to the task event (in the case of the control epoch, the cue on event, and the reward on event), or to the onset of movement (in the case of the torque on and the target on events), which was calculated as the first deflection of the manipulandum position trace that was >5 SDs away from the mean of the position trace during the 250-ms epoch immediately preceding the task event. Throughout the analysis, the neuronal data were binned in 5-ms bins, and the resulting discharge rate estimates were smoothed using a 10-point moving average. A given cell was classified as responding to an event if the discharge rate during three or more consecutive bins during the peri-event epoch deviated by >2 SDs from the mean discharge rate of the control epoch. The analysis program presented raster diagrams, binned neuronal data, and collapsed analog trials. All neuronal responses were also subjected to visual inspection. Only those responses that were also convincing by visual inspection were considered further.
Saccadic activity was occasionally identified on the basis of EOG data in monkey B. In all cases, these cells displayed abrupt and highly characteristic pauses of activity. All cells with such pauses were excluded from further analysis in both animals.
Neurons were classified according to their responses to either the task events themselves or to the movements occurring after the event. Thus responses related to target presentation in the STT and DRT were classified as anticipatory if they began >500 ms before the onset of movement, whereas responses starting within 200 ms of the movement were presumed to be directly related to the movement itself. Responses to cue presentation in the DRT were classified similarly. For instance, responses that started before or 
100 ms after the presentation of the cue were classified as anticipatory (this was done under the presumption that cortical sensory processing takes 
100 ms). Postcue responses which started between 100 and 2,000 ms after cue presentation were classified as persistent if their cue-related changes in neuronal discharge persisted for 1,000 ms, or as related to the cue itself, if the duration was shorter. In a few cells, it was difficult to distinguish between persistent postcue activity and anticipatory activity related to the target on event. These cells were excluded from the analysis.
Responses to reward application were classified in a similar manner. Such responses may reflect the reward itself, but may also reflect licking or other orofacial movements. Our experiments did not distinguish between these possibilities. Reward-related responses that started before or 100 ms after reward application were classified as anticipatory, whereas responses that started 100 ms after the reward were classified as postreward responses.
Because of the fact that the timing of the events was randomized, it would not have been meaningful to determine onset and offset times for anticipatory responses. Thus timing information is only given for neuronal responses that occurred after the events.
The same analysis program that was used for the analysis of neuronal responses was also used to determine reaction and movement times in the STT and DRT. These were calculated from the analog position trace, averaged across trials in which the same type of movement was performed. Movement onset was defined as in the preceding text. Movement offset was defined as the time when the manipulandum position trace entered its final position (within ± 5 SDs from the final position mean). The reaction time was operationally defined as the time between presentation of the target on event and the onset of movement, whereas movement time was defined as the amount of time between movement onset and movement offset.
The onset and offset, as well as mean discharge rates, during individual responses were entered into a database program (Microsoft Access). The significance of differences between means of firing rates, response latencies and durations, and movement and reaction times of different groups of data were assessed with an ANOVA, followed by t-test. P levels of <0.05 were assumed to be significant. Statistical tests were done using SPSS.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
The experiments were always carried out at the same time of the day to minimize circadian differences in task performance. The animals were over-trained and performed the task with a high rate of success (>90%). Reaction and movement times were generally similar between the monkeys, although for most measures, monkey N performed the task slightly slower than monkey B (see Table 1).
|
A total of 261 SNr neurons were recorded (141 in monkey B, 120 in monkey N). The average discharge rate of all SNr neurons was 66.4 ± 26.9/s (mean ± SD). An ANOVA did not show significant differences of either discharge rates or task responses between monkeys B and N, so that for the subsequent analyses the data from both monkeys were pooled. In some cases, the data sets were incomplete because the neurons could not be held long enough with good recording quality. Thus the n numbers are smaller in most portions of the RESULTS.
Response to somatosensory examination
Of the 219 cells tested (130 in monkey B, 89 in monkey N), 63 (29%) responded to somatosensory examination. These responses were typically subtle and tended to habituate quickly. The majority of responses were elicited by passive joint rotation or pressure involving proximal joints or trunk. Figure 2 shows a map of the distribution of neurons responding to manipulation of arm, leg, trunk, or face. Most of the responses were found in the dorsal half of the nucleus and did not show a distinct somatotopic distribution within the nucleus. Perhaps more importantly, many cells responded to manipulations of several adjacent joints, or even different body parts, and were thus rather nonspecific. As is true for most responses in the basal ganglia, the majority of responses to somatosensory examination in the SNr were seen with joint rotation or deep muscle palpation rather than skin touch.
|
Responses to torque application. Torque pulses, were applied after a randomized delay (between 1,000 and 1,700 ms) after center target onset in both tasks. These torque pulses resulted in a brief elbow deflection. The monkey had to recapture the center target within 1,000–1,500 ms. Data to assess responses to application of extension and flexion torques are available in 258 cells (see Table 2). Six cells (2.3%) responded to the torque-induced elbow deflection. Two cells (0.8%) responded to extension torques. Both responses consisted of increases in discharge (starting 5 and 105 ms after the beginning of the torque-induced movement and lasting 130 and 145 ms, respectively). The remaining four cells (1.6%) responded to flexion torques. Two of these responses consisted of increases in discharge (starting 5 and 35 ms after the beginning of the torque-induced movement and lasting 50 and 70 ms, respectively), and two consisted of decreases in discharge (both starting at 5 ms after elbow deflection onset and lasting 540 and 100 ms, respectively). All of these responses were subtle and were strictly directional, occurring only with torque movements in one of the two directions. None of the cells that responded to torque was found to also respond to elbow movements during manual somatosensory examination. The discrepancy between the examination and the results of the torque trials is most likely due to the fact that the responses to elbow movements were not consistent enough to be verified by examination.
|
|
Some cells showed additional responses starting around the time of movement that persisted without significant modulation until the time of reward application. These cells were not considered as movement-related but were felt to be related to the expectation of reward.
RESPONSES ASSOCIATED WITH CUE PRESENTATION IN THE DRT. Anticipatory activity preceding the onset of the cue was frequently seen. Although it was known to the animal at this time that a cue would be shown, it had no advance information regarding the location of the upcoming cue. Therefore flexion and extension trials were combined for the analysis of anticipatory responses. All cells that showed pausing activity, which is characteristic of a saccadic response, related to cue onset, were presumed to respond to eye movements, and were excluded from this analysis. Of the remaining 228 cells, 55 (24.1%) responded to the instruction cue. Of these, 38 (69.1%) showed anticipatory activity. The cells shown in Figs. 4, A and B, demonstrate typical precue changes of activity. In most cells, anticipatory activity was characterized by an increase or decrease in discharge, which was terminated abruptly at or after the time of cue onset.
|
Twenty-five cells (45.5% of all cells responding to the cue) responded to cue presentation, as is shown in the example cell in Fig. 7 (top). In these cells, 42 responses were evaluated (some cells responded to cues in either direction). Cue-related responses started 220.0 ± 169.9 ms after cue presentation and lasted 535.7 ± 327.3 ms.
|
RESPONSES ASSOCIATED WITH THE TARGET-ON EVENT IN THE DRT. As mentioned in the preceding text, sufficient data for the DRT were available for 228 cells. Forty-seven cells responded around the time of the target-on event. Of these, 19 cells (40.4% of responding cells) showed anticipatory activity as defined in METHODS. All in all, 30 anticipatory responses were seen (some cells responded with movements in both directions). An example of this can be seen in Fig. 5C. This cell showed anticipatory activity [an increase in flexion direction (left) and a low discharge rate in the extension direction (right)], which was terminated at the time of target presentation.
|
100 ms after the target-on event and ended at or after the onset of movement. We recorded 71 responses in these cells (some cells responded to movements in both directions). On average, these responses started 112.5 ± 130.0 ms prior to movement onset, and lasted for 442.1 ± 211.4 ms. Few cells showed monophasic responses, such as the one shown in Fig. 5A. This cell shows an increase in activity starting prior to movement onset, which persists during the first phase of movement. The increase is more substantial in flexion than in extension direction. In most cells, however, the responses were multiphasic and tended to be quite subtle, sometimes even showing significant weakening during the block of trials. An example of a complex multiphasic response is shown in Fig. 5B. This cell shows a substantial decrease in activity following the extension movement (right). However, in flexion direction, the movement is associated with an oscillatory response pattern, which appears to reset around the time of the movement. The movement itself represents a complex sequence of acceleration and deceleration phases. In some cases, the animal made additional small corrective movements at the end of the target capture phase which also may have contributed to the oscillatory nature of some of the responses. Part of the variability of the onset and offset times stems from the fact that many cells responded to a particular phase of the movement (which was not always the first portion of the movement). None of the cells responding to the target-on event also responded to the application of elbow torque pulses.
RESPONSES ASSOCIATED WITH REWARD APPLICATION. We initially analyzed reward-related responses separately for the two different tasks and for the two different movement directions. In all 252 cells tested in this manner, the reward-related responses were independent of the task or the preceding movement. For the subsequent analysis, all successful trials from each cell were pooled to assess the reward responses. Twenty-four of these cells (9.5%) showed a response to reward application (see Figs. 6 and Fig. 7, bottom row, and Table 2). In 13 cells (54.2%), the activity was classified as anticipatory. In each of these cases, the activity lasted into the postreward phase (as defined in the METHODS). The remaining 11 cells (45.8%) showed postreward responses, starting 175.0 ± 73.5 ms after the reward application and lasted 366.4 ± 196.5 ms. Most of the responses (22/24, 91.7%) were monophasic with 15 inhibitions and 7 excitatory responses. The remaining responses were biphasic (1 inhibition-inhibition sequence and 1 excitation-inhibition sequence).
|
|
All 17 reward-responding cells in this sample of 208 cells responded in both tasks. Nine of these cells (52.9%) responded also with movement in one or both tasks, whereas the remaining 8 cells (47.1%) responded to reward alone.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
A comparison of these results with previous experiments in which GPi responses were studied under similar conditions reveals substantial differences in the task-related activity in GPi and SNr. The results support the notion that the two basal ganglia output nuclei process distinct and complementary types of information relevant to behavior.
Responses related to limb movement
There is substantial evidence from anatomic studies supporting a role of the SNr in the control of movement. For instance, virtually all anatomic studies of the striatonigral pathway have shown that the striatal motor territory, the putamen, projects to the SNr (e.g., Francois et al. 1987; Hedreen and DeLong 1991; Lynd-Balta and Haber 1994; Parent and Hazrati 1994; Percheron et al. 1984; Selemon and Goldman-Rakic 1990; Szabo 1967). Anatomic details regarding the topography of this projection are contradictory, but it appears that the putamenal efferents reach the central and lateral portions of the SNr throughout the entire anterior-posterior extent of the nucleus [see, e.g., case 5L of the paper by Hedreen and DeLong (1991)]. Although most SNr output appears to be directed toward thalamic areas, which then project to prefrontal cortical areas (see following text), SNr output is also directed toward premotor areas (e.g., Ilinsky et al. 1985) and even primary motor cortex (Hoover and Strick 1999). Indirect evidence for the functional significance of the putamen-SNr projection comes from studies showing an increase of metabolic activity in the SNr in motor phenomena such as focal motor seizures (Hosokawa et al. 1983) or experimental hemiplegia in monkeys (Gilman et al. 1987).
One of the criteria that characterizes neurons in the motor territory of striatum, STN and GPi is that they respond vigorously and specifically to passive manipulation of body parts. By this criterion, few of the SNr neurons in our study would be classified as typical motor circuit neurons because their responses tended to be subtle, habituated quickly, and could be evoked by movements around multiple joints or body regions. Similar nonspecific and inconsistent responses of neurons in the primate SNr to somatosensory examination have also been reported by other authors (Magarinos-Ascone et al. 1994; Mora et al. 1977; Schultz 1986). Although our study is in agreement with most of the published results, some of the studies (e.g., DeLong et al. 1983) have emphasized the prominence of orofacial responses, which was not apparent in our sample of cells. This discrepancy is likely due to the fact that in the study by DeLong et al. (1983), neurons were classified as "orofacial" if they responded to active licking and mouth movements, whereas we focused on orofacial touch or jaw manipulation. The results presented by Magarinos-Ascone et al. (1994) also differ from our results and from those of the other published studies. For unclear reasons, these authors found strong responses of SNr neurons to touch in circumscribed skin regions.
A second characteristic of neurons in the motor territory of striatum, STN and GPi is that their activity is directly related to active movements. In our experiments as well as previously published studies (DeLong et al. 1983; Schultz 1986; Schwarz et al. 1984), some SNr neurons indeed responded preferentially during the execution of proximal forelimb movements. In this respect, our findings are closest to those of DeLong et al. (1983), whereas a somewhat larger proportion of movement-related neurons was identified in some of the other studies that either used a less constrained task than ours (Schultz 1986) or a task that required self-initiated sinusoidal ballistic arm movements (Magarinos-Ascone et al. 1992). The differences between our data and the results of the latter study suggest that the SNr may be more directly engaged in self-initiated rather than externally triggered movements.
The proportion of neurons that strongly responded to examination or to active movements was relatively small, whereas premotor or nonmotor aspects of behavior appeared to have a very significant impact on the activity of SNr cells. Anticipatory activity preceding the movement associated with the target-on event (interpreted as premotor activity) differed in many cases between the DRT and the STT. The majority of cells showed a movement-related response only in one of the tasks (STT or DRT) but not in both. In addition, if activity in both tasks was found, the timing of the activity change relative to movement onset was often different between STT and DRT. The different response patterns may reflect differences between the two task conditions. In the DRT, the instruction cue provided information relating to the direction of the upcoming movement, and the animal had to use this information to perform above chance. In the STT, no advance information was provided. Thus in the DRT, it was possible for the animal to prepare for the kinematic specifics required to conduct the upcoming movement, whereas this was not possible in the STT. Premotor activity in the STT may reflect a more general form of motor preparation or state of readiness.
Responses similar to those shown here for appendicular movements have also been documented in the oculomotor system. Discharge of eye-movement-related neurons in the SNr often precede saccades, and many saccade-related cells may also be concerned with aspects other than control of saccadic movements alone, such as spatial memory (Hikosaka and Wurtz 1983) or reward (Sato and Hikosaka 2002). However, some of the oculomotor functions of the SNr appear to be uniquely linked to superior colliculus activity by virtue of the phylogenetically old nigrocollicular pathway. Thus the involvement of the SNr in limb and eye movements may not be fully comparable. This is also apparent from the fact that SNr inactivation results in excessive eye (and neck) movements but does not induce involuntary limb movements (Burbaud et al. 1998; Lestienne and Thullier 1998; Wichmann et al. 2001).
Nonmotor functions of the SNr
Anatomic studies have demonstrated that the SNr receives not only motor-related inputs but also prominent projections from the associative and limbic striatal territories (e.g., Francois et al. 1987; Haber et al. 1994, 1995; Hedreen and DeLong 1991; Lynd-Balta and Haber 1994) and from the associative circuits in the parafascicular nucleus of the thalamus (Sadikot et al. 1992a,b; Sidibe et al. 2002). Output of the SNr is directed in large part toward areas of thalamus that project to nonmotor regions of cortex (Ilinsky et al. 1985, 1993). More recent anatomic studies using viral tracers have supported this concept with the demonstration that a portion of SNr output is targeted toward associative and limbic prefrontal areas of cortex (e.g., Middleton and Strick 2002) and perhaps also to the parietal area 7b (Clower et al. 2002). These projections were shown to arise from SNr neurons that are separate from those projecting to premotor areas (Middleton and Strick 2002), attesting to the segregation of motor and nonmotor circuits in the SNr. The general principle that motor and nonmotor circuits remain segregated throughout their subcortical path may, however, only partly apply to the SNr, because of its interactions with subcortical nuclei, such as the SNc or brain stem areas, which are less specifically organized.
A substantial portion of SNr cells in our study showed responses that were not directly related to sensory stimuli or movements and may thus have been reflected phenomena such as memory or attention (neither of which were directly assessed). In addition, many neurons showed responses to more than one event. These patterns of task responses of SNr neurons go beyond a pure motor function of the SNr and may reflect circuit activity in associative and limbic areas of cortex, striatum, and thalamus. For instance, the nonmotor striatal areas known to project to the SNr have been shown to contain neurons with complex responses in limb motor tasks that were very similar to those observed in our study. Thus many caudate nucleus neurons show preparatory or persistent postcue activity (Jaeger et al. 1993; Schultz and Romo 1992) or respond to shifts in attention (Boussaoud and Kermadi 1997). Anticipatory and reward responses are now also well documented for striatal neurons participating in oculomotor functions (Hikosaka et al. 1989; Lauwereyns et al. 2002; Takikawa et al. 2002). In addition, some neurons in the associative and limbic areas of the striatum have been shown to possess activity specific to a sequence of limb movements (Kermadi and Joseph 1995; Kermadi et al. 1993; Miyachi et al. 1997), to reward (Apicella et al. 1991; Hollerman et al. 1998), to the "newness" of a stimulus in learning paradigms (Miyachi et al. 2002) or specifics of food items presented to the animal (Nishino et al. 1981). Most of these higher-order aspects of behavior were not specifically assessed in the present study, but it is likely that task-related neurons in the SNr would show similar responses.
Nonmotor aspects of the task such as responses to the cue-on event were common in our sample of cells. Interestingly, many neurons showed responses to both the movement and the preceding cue (or reward). The mixed responses may signify the presence of neurons in the SNr that participate in associative circuits, or, alternatively, the presence of convergence of motor and nonmotor information onto single SNr neurons.
Comparison with GPi
It has been argued that both basal ganglia output nuclei function as a unit (e.g., DeLong and Georgopoulos 1981), which is divided by the internal capsule. The distribution of different portions of the basal ganglia circuitry between these nuclei has not been fully elucidated, however. It is therefore instructive to compare the responses recorded in the SNr to responses recorded in GPi in similar tasks. In some respects, the involvement of the two nuclei in motor functions is comparable. For instance, inactivation of SNr or GPi does not result in significant involuntary limb movements (Baron et al. 2002; Burbaud et al. 1998; Inase et al. 1996; Lestienne and Thullier 1998; Wenger et al. 1999; Wichmann et al. 2001). In addition, some of the movement-related responses in either nucleus appear to be strongly modulated by the context in which the movement occurs. Thus similar to the findings presented here for the SNr, task-specific movement-related responses have been described in GPi (Brotchie et al. 1991b; Mink and Thach 1991a). In addition, Gdowski et al. recently described movement-related responses of GPi neurons that were more vigorous in situations that were predictive of reward than in situations in which the movement occurred without reward (Gdowski et al. 2001).
In general, however, movement-related responses in GPi and SNr appear to be more dissimilar than they are alike. Thus few of our SNr neurons responded strongly to examination, but vigorous responses to passive movements of specific joints are common in the ventroposteriorlateral motor territory of GPi (e.g., Allum et al. 1983; Anderson and Horak 1985; Anderson and Turner 1991; DeLong 1969; Georgopoulos et al. 1983; Hamada et al. 1990; Mushiake and Strick 1995; Nishino et al. 1985a). The two nuclei also differ in their responses to active movement. Such responses were uncommon in the SNr (see preceding text), whereas many pallidal neurons discharge in relation to kinematic parameters, such as the direction, velocity, or amplitude of movement (Brotchie et al. 1991a; Georgopoulos et al. 1983; Mitchell et al. 1987; Turner and Anderson 1997). Most movement-related responses in GPi occur at or after EMG onset and are likely related to the execution of movement (Brotchie et al. 1991a; Georgopoulos et al. 1983; Mink and Thach 1991b; Turner and Anderson 1997; Wannier et al. 2002). In contrast, early preparatory activity was common in our sample of cells.
Conclusion
In agreement with previous anatomic studies, our results support the notion that the primate SNr subserves functions that are distinct from those represented in GPi. Whereas GPi neurons may be more concerned with motor activity per se, SNr circuitry may have a larger role in premotor, associative, and limbic functions. Further study is needed to better define its involvement in higher-order motor processing such as sequencing of movement, or its involvement in nonmotor functions, such as specific types of memory, attention, or valence.
|
ACKNOWLEDGMENTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
GRANTS
This work was supported by National Institute of Neurological Disorders and Stroke Grant NS-34872 to T. Wichmann.
|
FOOTNOTES |
|---|
Address for reprint requests and other correspondence: T. Wichmann, Emory University, Dept. Neurology, Suite 6000, Woodruff Memorial Research Bldg., 1639 Pierce Dr., Atlanta, GA 30322 (E-mail: twichma{at}emory.edu).
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
Alexander GE, DeLong MR, and Strick PL. Parallel organization of functionally segregated circuits linking basal ganglia and cortex. Annu Rev Neurosci 9: 357–381, 1986.[CrossRef][ISI][Medline]
Allum JHJ, Anner-Baratti REC, and Hepp-Raymond MC. Activity of neurons in the motor thalamus and globus pallidus during the control of isometric finger force in the monkey. Exp Brain Res 7, Suppl, 194–203, 1983.
Anderson ME and Horak FB. Influence of the globus pallidus on arm movements in monkeys. III. Timing of movement-related information. J Neurophysiol 54: 433–448, 1985.
Anderson ME and Turner RS. A quantitative analysis of pallidal discharge druing targeted reaching movement in the monkey. Exp Brain Res 86: 623–632, 1991.[ISI][Medline]
Apicella P, Ljungberg T, Scarnati E, and Schultz W. Responses to reward in monkey dorsal and ventral striatum. Exp Brain Res 85: 491–500, 1991.[ISI][Medline]
Baron MS, Wichmann T, Ma D, and DeLong MR. Effects of transient focal inactivation of the basal ganglia in parkinsonian primates. J Neurosci 22: 592–599, 2002.
Boussaoud D and Kermadi I. The primate striatum: neuronal activity in relation to spatial attention versus motor preparation. Eur J Neurosci 9: 2152–2168, 1997.[CrossRef][ISI][Medline]
Brotchie P, Iansek R, and Horne MK. Motor function of the monkey globus pallidus. I. Neuronal discharge and parameters of movement. Brain 114: 1667–1683, 1991a.
Brotchie P, Iansek R, and Horne MK. Motor function of the monkey globus pallidus. II. Cognitive aspects of movement and phasic neuronal activity. Brain 114: 1685–1702, 1991b.
Burbaud P, Bonnet B, Guehl D, Lagueny A, and Bioulac B. Movement disorders induced by gamma-aminobutyric agonist and antagonist injections into the internal globus pallidus and substantia nigra pars reticulata of the monkey. Brain Res 780: 102–107, 1998.[CrossRef][ISI][Medline]
Clower DM, Dum RP, and Strick PL. Substantia nigra pars reticulata provides input to area 7b of parietal cortex. Society for Neuroscience 32nd annual meeting, Orlando, FL, 2002, 460.461.
DeLong M. Activity of pallidal neurons in the monkey during movement and sleep. Physiologist 12: 207, 1969.
DeLong MR, Crutcher MD, and Georgopoulos AP. Relations between movement and single cell discharge in the substantia nigra of the behaving monkey. J Neurosci 3: 1599–1606, 1983.[ISI][Medline]
DeLong MR and Georgopoulos AP. Motor functions of the basal ganglia. In: Handbook of Physiology. The Nervous System. Motor Control. Bethesda: Am. Physiol. Soc., 1981, sect. 1, vol. II, part 2, p. 1017–1061.
Francois C, Yelnik J, and Percheron G. Golgi study of the primate substantia nigra. II. Spatial organization of dendritic arboriations in relation to the cytoarchitectonic boundaries and to the striatonigral bundle. J Comp Neurol 265: 473–493, 1987.[CrossRef][ISI][Medline]
Gdowski MJ, Miller LE, Parrish T, Nenonene EK, and Houk JC. Context dependency in the globus pallidus internal segment during targeted arm movements. J Neurophysiol 85: 998–1004, 2001.
Georgopoulos AP, DeLong MR, and Crutcher MD. Relations between parameters of step-tracking movements and single cell discharge in the globus pallidus and subthalamic nucleus of the behaving monkey. J Neurosci 3: 1586–1598, 1983.[Abstract]
Gilman S, Dauth GW, Frey KA, and Penney JBJ. Experimental hemiplegia in the monkey: basal ganglia glucose activity during recovery. Ann Neurol 22: 370–376, 1987.[CrossRef][ISI][Medline]
Haber SN, Kunishio K, Mizobuchi M, and Lynd-Balta E. The orbital and medial prefrontal circuit through the primate basal ganglia. J Neurosci 15: 4851–4867, 1995.[Abstract]
Haber SN, Lynd-Balta E, and Spooren WPJM. Integrative aspects of basal ganglia circuitry. In: The Basal Ganglia IV. New Ideas and Data on Structure and Function, edited by Percheron G, McKenzie JS, and Feger J. New York: Press, 1994, p. 71–80.
Hamada I, DeLong MR, and Mano NI. Activity of identified wrist-related pallidal neurons during step and ramp wrist movements in the monkey. J Neurophysiol 64: 1892–1906, 1990.
Hedreen JC and DeLong MR. Organization of striatopallidal, striatonigral and nigrostriatal projections in the macaque. J Comp Neurol 304: 569–595, 1991.[CrossRef][ISI][Medline]
Hikosaka O, Sakamoto M, and Usui S. Functional properties of monkey caudate neurons. III. Activities related to expectation of target and reward. J Neurophysiol 61: 814–832, 1989.
Hikosaka O and Wurtz RH. Visual and oculomotor functions of monkey substantia nigra pars reticulata. III. Memory-contingent visual and saccade responses. J Neurophysiol 49: 1268–1284, 1983.
Hollerman JR and Schultz W. Dopamine neurons report an error in the temporal prediction of reward during learning. Nat Neurosci 1: 304–309, 1998.[CrossRef][ISI][Medline]
Hollerman JR, Tremblay L, and Schultz W. Influence of reward expectation on behavior-related neuronal activity in primate striatum. J Neurophysiol 80: 947–963, 1998.
Hoover JE and Strick PL. Multiple output channels in the basal ganglia. Science 259: 819–821, 1993.
Hoover JE and Strick PL. The organization of cerebellar and basal ganglia outputs to primary motor cortex as revealed by retrograde transneuronal transport of herpes simplex virus type 1. J Neurosci 19: 1446–1463, 1999.
Hosokawa S, Kato M, and Kuroiwa Y. Topographical distribution of propagation of seizure activity in the basal ganglia during focal motor seizures in the monkey. Neurosci Lett 38: 29–33, 1983.[CrossRef][ISI][Medline]
Ilinsky IA, Jouandet ML, and Goldman-Rakic PS. Organization of the nigrothalamocortical system in the rhesus monkey. J Comp Neurol 236: 315–330, 1985.[CrossRef][ISI][Medline]
Ilinsky IA, Tourtellotte WG, and Kultas-Ilinsky K. Anatomical distinctions between the two basal ganglia afferent territories in the primate motor thalamus. Stereotactic Funct Neurosurg 60: 62–69, 1993.
Inase M, Buford JA, and Anderson ME. Changes in the control of arm position, movement, and thalamic discharge during local inactivation in the globus pallidus of the monkey. J Neurophysiol 75: 1087–1104, 1996.
Jaeger D, Gilman S, and Aldridge JW. Primate basal ganglia activity in a precued reaching task: preparation for movement. Exp Brain Res 95: 51–64, 1993.[ISI][Medline]
Kermadi I and Joseph JP. Activity in the caudate nucleus of monkey during spatial sequencing. J Neurophysiol 74: 911–933, 1995.
Kermadi I, Jurquet Y, Arzi M, and Joseph JP. Neural activity in the caudate nucleus of monkeys during spatial sequencing. Exp Brain Res 94: 352–356, 1993.[ISI][Medline]
Lauwereyns J, Takikawa Y, Kawagoe R, Kobayashi S, Koizumi M, Coe B, Sakagami M, and Hikosaka O. Feature-based anticipation of cues that predict reward in monkey caudate nucleus. Neuron 33: 463–473, 2002.[CrossRef][ISI][Medline]
Lestienne F and Caillier P. Role of the monkey substantia nigra pars reticulata in orienting behavior and visually triggered arm movements. Neurosci Lett 64: 109–115, 1986.[CrossRef][ISI][Medline]
Lestienne FG and Thullier F. Performance of visually triggered wrist movements task in monkey: an application of information theory to evaluate deficits following unilateral substantia nigra pars reticulata lesion. Neurosci Lett 251: 177–180, 1998.[CrossRef][ISI][Medline]
Lynd-Balta E and Haber SN. Primate striatonigral projections: a comparison of the sensorimotor-related striatum and the ventral striatum. J Comp Neurol 345: 562–578, 1994.[CrossRef][ISI][Medline]
Magarinos-Ascone C, Buno W, and Garcia-Austt E. Activity in monkey substantial nigra neurons related to a simple learned movement. Exp Brain Res 88: 283–291, 1992.[CrossRef][ISI][Medline]
, responses to arm movements; 
, leg responses; 
, orofacial responses; x, responses to trunk examination; 
, saccade and visual responses. Nonresponding neurons are shown (—), while neurons that were encountered but not tested are shown as ·.
