Abstract

The globus pallidus external segment (GPe) constitutes part of the indirect pathway of the basal ganglia. Because of inhibitory projections from the striatum, most GPe neurons are expected to reduce activity during movements. However, many GPe neurons in fact display increased activity. We previously found that both excitatory and inhibitory responses were modulated during antisaccades, when eyes were directed away from a visual stimulus. To elucidate the roles of these neurons during antisaccades, we examined neuronal activities as monkeys performed antisaccades, prosaccades, and NoGo tasks under 2 conditions. In the Deliberate condition, the task-rule was instructed by color of the fixation point, while in the Immediate condition, it was given by color of the target. Under both conditions, the increase-type neurons exhibited greater activity during antisaccades compared with the other tasks and neuronal activity negatively correlated with saccade latency. The decrease-type neurons also showed greater modulation during antisaccades but their activity was comparable between NoGo and antisaccade trials in the Immediate condition. These results suggest that the increase-type neurons might play a role in facilitating antisaccades, whereas the decrease-type neurons could mediate signals for reflexive saccade suppression. We propose that these GPe neurons are differently involved in basal ganglia pathways.

Introduction

A widely accepted view regarding the neuronal processes in the basal ganglia assumes the “direct” and “indirect” pathways (Albin et al. 1989; DeLong 1990). Although this hypothesis can explain many pathophysiological aspects of basal ganglia diseases, including Parkinson's disease, the growing body of evidence suggests that this hypothesis might be too simplistic; in addition to the alternative direct and indirect pathways, several other major projections in the basal ganglia have recently been elucidated (for reviews, see Nambu et al. 2002; Wichmann et al. 2011). According to the original hypothesis, the external segment of the globus pallidus (GPe) constitutes part of the indirect pathway, which plays a role in the suppression of actions (Brotchie et al. 1991; for review, see Chan et al. 2005). Because of the inhibitory projection from the striatum, neurons in the GPe are expected to exhibit decreased activity during movements. However, more than half of the neurons in the GPe in fact exhibit increased activity during movements (Mitchell et al. 1987; Turner and Anderson 1997; Yoshida and Tanaka 2009a,b; Shin and Sommer 2010; for review, see Goldberg and Bergman 2011), suggesting that the direct excitatory inputs from the subthalamic nucleus (STN) might regulate neuronal activity in the GPe (Shink et al. 1996) (Fig. 1D).

Figure 1.

Behavioral paradigms and underlying neuronal processes. (A) Sequence of events in the Deliberate condition, in which color of the fixation point indicates the task-rule. A target spot appeared 12° right or left at the time of the fixation point offset. Monkeys were required to make a saccade away from (antisaccade) or toward (prosaccade) the target within 400 ms, or maintain fixation for >400 ms (NoGo). (B) In the Immediate condition, task-rule was indicated by color of the saccade target. This imposed monkeys to select an action immediately, as they were not informed of the trial type in advance. Different trials were randomly interleaved in a block. Anti, antisaccade; Pro, prosaccade. (C) In each task and condition, the brain needs to facilitate or inhibit saccade production. Note that the inhibitory processes during the NoGo trials differ between the conditions. (D) Anatomical connections of the GPe. Neurons in the GPe receive both excitatory (glutamatergic) and inhibitory (GABAergic) projections. Cd, caudate nucleus; GPe, external segment of the globus pallidus; Put, putamen.

Figure 1.

Behavioral paradigms and underlying neuronal processes. (A) Sequence of events in the Deliberate condition, in which color of the fixation point indicates the task-rule. A target spot appeared 12° right or left at the time of the fixation point offset. Monkeys were required to make a saccade away from (antisaccade) or toward (prosaccade) the target within 400 ms, or maintain fixation for >400 ms (NoGo). (B) In the Immediate condition, task-rule was indicated by color of the saccade target. This imposed monkeys to select an action immediately, as they were not informed of the trial type in advance. Different trials were randomly interleaved in a block. Anti, antisaccade; Pro, prosaccade. (C) In each task and condition, the brain needs to facilitate or inhibit saccade production. Note that the inhibitory processes during the NoGo trials differ between the conditions. (D) Anatomical connections of the GPe. Neurons in the GPe receive both excitatory (glutamatergic) and inhibitory (GABAergic) projections. Cd, caudate nucleus; GPe, external segment of the globus pallidus; Put, putamen.

Although the firing properties of 2 types of neurons in the GPe have been well documented, their functional roles remain elusive. Previously we examined GPe neurons responding to saccades, and found that both types of neurons exhibited enhanced firing modulation during antisaccades, when eyes were directed away from the target (Yoshida and Tanaka 2009a). These neurons could have a role in the production of antisaccade commands and/or the suppression of reflexive saccades toward the target, both of which are indispensable neural processes for the generation of antisaccades (Munoz and Everling 2004). Indeed, our previous study showed that inactivation of the GPe with muscimol resulted in the increased number of erroneous antisaccades (Yoshida and Tanaka 2009a).

The purpose of the present study was to elucidate the roles of signals in the GPe in the generation of antisaccades. In addition to the antisaccade and prosaccade trials, we randomly introduced the NoGo trial so as to dissociate the saccade-related movement activity from that related to saccade suppression. Recent studies suggest that there are 2 types of proactive inhibitory controls during antisaccades (Abegg et al. 2012). The “global suppression” withholds any motor response, causing general lengthening of reaction time, whereas the “selective suppression” inhibits a particular response that could be directed to distractive stimuli. These 2 processes appear to be regulated by different pathways in the basal ganglia (Majid et al. 2013; for review, see Aron 2011).

To examine these inhibitory processes, we presented trials in 2 different conditions. In one condition, the task-rule was given by the color of the fixation point in advance of the target appearance (Deliberate condition, Fig. 1A). In this condition, the global suppression may dominate during the NoGo trials in which the subjects should ignore visual stimuli and maintain fixation throughout the trial (Fig. 1A). In the other condition, the task-rule was indicated by the color of the target (Immediate condition, Fig. 1B). Because the subjects need to direct attention to the visual stimuli rather than simply ignoring them even in the NoGo trials, the signals related to selective suppression should be generated in the absence of movements. Thus, using the 3 tasks in 2 different conditions, we were able to examine neuronal activity related to saccade production or suppression, and to dissociate signals related to global suppression from selective suppression (Fig. 1C). Our data show that the increase-type neurons exhibited different firing properties from the decrease-type neurons during the trials, suggesting that the 2 types of GPe neurons might be differently involved in antisaccade generation, one facilitating and the other suppressing saccades.

Materials and Methods

Animal Preparation

Experiments were conducted on 4 Japanese macaques (Macaca fuscata, one 7 kg male and three 4–6 kg females). All experimental protocols were approved by the Animal Care and Use Committee of Hokkaido University. The experimental procedures were similar to those described previously (Tanaka 2005). After initial chair training, a pair of head holders was implanted in the skull using titanium screws and dental acrylic under general isoflurane anesthesia and using sterile procedures. To induce anesthesia, we administered medetomidine (40 µg/kg) and midazolam (0.3 mg/kg) intramuscularly. The animals were intubated to maintain anesthesia using a mixture of isoflurane (1–3%), nitrous oxide, and oxygen. In a separate surgical procedure, a coil of stainless steel wire was implanted under the conjunctiva to record eye movements. During subsequent training and experimental sessions, the monkey's head was fixed to the primate chair, and horizontal and vertical eye position was recorded using the search coil technique. After sufficient training for eye movement tasks, a recording cylinder was installed over a small craniotomy under the same surgical conditions as above. Animals received analgesics (pentazocine or ketoprofen) during each surgery and for several days subsequently. Topical antibiotics were administered around the implant and in the cylinder as necessary. Water intake was controlled daily so that monkeys were motivated to perform behavioral tasks.

Visual Stimulus and Behavioral Paradigms

Experiments were controlled by a Windows-based real-time data acquisition system (TEMPO; Reflective Computing, St. Louis). All events were updated every 5 ms, and visual stimuli were presented on a 24-inch cathode-ray tube monitor (GDM-FW900; Sony, Tokyo; refresh rate: 60 Hz) that was located 38 cm away from the eyes, and subtended visual angle of 64 × 44°. A 0.5° square spot served as a visual stimulus. Targets of different colors (white, red, green, and blue) were used for different means in each trial (see below). Experiments were carried out in a darkened booth. Voltages proportional to horizontal and vertical eye position were calibrated before each experiment by having monkeys fixate on a stationary target spot at known visual angles. Thereafter, visual stimuli were presented in individual trials, and monkeys were rewarded at specific time intervals during each trial with drops of water or apple juice for maintaining eye position within a “window” that surrounded the stimulus position. The size of the window was 2° for initial fixation, 6° (antisaccade task) or 3° (prosaccade task) for the saccade target, and 5° for the maintenance of fixation during target presentation in the NoGo trials. A trial was aborted and followed by a newly selected trial if monkeys failed to maintain eye position within a specified window. Trials of different types and in different directions were presented in a random order within a block. The saccade target was presented either 12° left or right of the fixation point.

The present study used 3 tasks in 2 conditions (Fig. 1). In the Deliberate condition, the task-rule was given by color of the fixation point in advance of the target appearance. In contrast, in the Immediate condition, the task-rule was given by color of the peripheral target so that monkeys were unable to prepare for specific action until the target was presented. We used 4 colors (white, red, green, and blue) for the fixation point and the target. Contrast modulation (Michelson contrast) was calculated for each stimulus using the formula CM = (AB)/(A + B), where A was the luminance of a given stimulus and B was the luminance of the background on the screen (1.59 cd/m2). In all tasks, the fixation point initially appeared at the center of the screen for 1600 ms. In trials with either a green fixation point (CM = 9.18, Deliberate condition) or a green saccade target (Immediate condition), monkeys were required to perform antisaccades 180° away from the target within 400 ms. In trials with a red fixation point or saccade target (CM = 5.10), monkeys were required to perform prosaccades. When either a blue fixation point or a blue saccade target (CM = 3.77) was presented, the animals were required to maintain central fixation for >400 ms following the target onset. The saccade target in the Deliberate condition and the fixation point in the Immediate condition were always white (CM = 14.21, Fig. 1A,B). In both saccade tasks, the peripheral target disappeared after a second fixation interval of 500 ms, and then a liquid reward was given. Introduction of the second fixation interval allowed us to temporally dissociate the neuronal responses to eye movements from those to reward.

Recording Procedures

Tungsten microelectrodes (FHC, Bowdoin, ME) were lowered through a 23-gauge stainless steel tube guided by a grid system (Crist Instrument Co.). The electrodes were advanced using a hydraulic micromanipulator (MO-97S; Narishige, Tokyo) attached to the recording cylinder, which allowed vertical (monkey S) or 38–40° off-vertical penetrations (monkeys U, V, and X) in the coronal plane. Signals obtained from the electrodes were amplified, filtered, and monitored using oscilloscopes and an audio monitor. Once task-related neuronal activity was encountered, spikes of single neurons were isolated using a real-time spike sorter with template-matching algorithms (MSD; Alpha Omega Engineering, Nazareth, Israel). For each experiment, the occurrence of action potentials was time-stamped and saved in files with the data of eye movements, location and timing of visual stimuli.

Data Acquisition and Analysis

Horizontal and vertical eye position signals were directly obtained from the eye coil electronics (MEL-25; Enzanshi Kogyo, Chiba, Japan). Data were digitized and sampled at 1 kHz, and were stored in files during the experiments for further off-line analysis, performed using Matlab (Mathworks, Natick, MA). For each neuron, data were aligned with either the initiation of saccades or the target onset. In some antisaccade trials, monkeys initially generated a saccade toward the target, but redirected their eyes to the location opposite to the target within 400 ms (“turn-around saccade”). Since we detected error trials by analyzing eye position before and 400 ms after the fixation point offset during experiments, monkeys were rewarded in such trials. We detected turn-around saccades off-line and considered them as errors and these data were excluded from the subsequent analyses. Traces of horizontal and vertical eye positions were reviewed with rasters and spike density profiles that were constructed from neuronal data.

Neurons recorded during fewer than 10 trials for each task condition were excluded from subsequent analyses. Neuronal activity was measured during the following 3 time periods: (1) the 300-ms interval immediately before the fixation point onset (baseline period); (2) the 250-ms period after the target onset (target onset period); and (3) the 150-ms period starting from 100 ms before saccades (saccade period). We used a relatively long target onset period to compare neuronal activities during antisaccades with those during NoGo trials, because reaction times of antisaccades were ∼250 ms and the suppression of reflexive saccades should have occurred during this interval in the NoGo trials. We defined neuronal activity as “antisaccade-related” when it differed significantly from the baseline activity during antisaccades (2-tailed paired t test, P < 0.05). These neurons were further classified into 2 groups, “increase-type” or “decrease-type”, according to the direction of firing modulation during antisaccades. The time course of neuronal activity for each condition was examined qualitatively by calculating the spike density function using a Gaussian filter (σ = 15 ms). However, all quantitative measures were performed on the basis of spike counts for specific time windows. Analytical measures are reported in the relevant text in the Results. To evaluate neuronal activity, the mean firing rate was assessed using the analysis of variance (ANOVA) followed by post hoc multiple comparisons (Scheffé's method). Each statistical test that was applied is stated in the relevant text in the Results.

Verification of the Recording Sites

The recording sites were reconstructed based on magnetic resonance images [MRI; GE Signa 1.5 T; 3D T1-weighted images (1 mm thick slice) and 2D T2-weighted images (1.5 mm thick slice)] after implantations of the recoding chamber. Consistent with our previous study (Yoshida and Tanaka 2009a), many antisaccade-related neurons were found in the anterior part of the GPe, close to the anterior commissure (Fig. 2).

Figure 2.

Sites of the task-related neurons are overlaid on MR images shown in the frontal sections (T2-weighted image). Left panel shows the left hemisphere of monkey X, while right panel shows the right hemisphere of monkey S. Shape of symbols indicates response property, and the size indicates the number of recorded neurons. Both MR images are the sections 1 mm posterior to the anterior commissure (AC). Cd, caudate nucleus; GPe, external segment of the globus pallidus; Put, putamen; STN, subthalamic nucleus.

Figure 2.

Sites of the task-related neurons are overlaid on MR images shown in the frontal sections (T2-weighted image). Left panel shows the left hemisphere of monkey X, while right panel shows the right hemisphere of monkey S. Shape of symbols indicates response property, and the size indicates the number of recorded neurons. Both MR images are the sections 1 mm posterior to the anterior commissure (AC). Cd, caudate nucleus; GPe, external segment of the globus pallidus; Put, putamen; STN, subthalamic nucleus.

Results

Database and Recording Sites

Data were collected only after monkeys were well trained on the 3 tasks under both conditions. Table 1 summarizes the behavioral data obtained during 170 recording sessions. All monkeys performed the task correctly in >88% of trials. Error trials in the antisaccade task contained those with reflexive saccades toward the visual stimuli (including turn-around saccades) and those without saccades during 400 ms of target appearance. Overall, a 2-way ANOVA revealed that saccade latencies were statistically different between conditions (192 ± 48 ms and 241 ± 50 ms for the Deliberate and Immediate conditions, respectively, P < 0.05) as well as between pro- (204 ± 56 ms) and antisaccades (230 ± 51 ms, P < 0.05). There was also a significant interaction between the main factors (P < 0.05), indicating that the difference in latency between the 2 saccade tasks was statistically greater in the Deliberate condition than in the Immediate condition.

Table 1

Comparison of behavioral parameters between conditions

 Deliberate
 
Immediate
 
Anti Pro NoGo Anti Pro NoGo 
Monkey S 
N (correct/all) 4539/4573 4548/4549 4570/4751 4549/4574 4569/4581 4581/4617 
 Error rate (%) 0.7 0.0 3.8 0.5 0.3 0.8 
 Latency (ms) 215 ± 30 188 ± 31 — 254 ± 30 247 ± 30 — 
 Error latency 139 ± 51 — — 157 ± 17 224 ± 21 — 
Monkey X 
N (correct/all) 1293/1296 1289/1292 1288/1425 1288/1289 1293/1293 1293/1326 
 Error rate (%) 0.0 0.0 10.6 0.0 0.0 2.50 
 Latency (ms) 300 ± 39 262 ± 22 — 325 ± 30 324 ± 32 — 
 Error latency 213 ± 36 — — — — — 
Monkey V 
N (correct/all) 749/851 848/848 852/853 824/854 851/851 842/842 
 Error rate (%) 11.7 0.0 0.0 3.5% 0.0 0.0 
 Latency (ms) 223 ± 52 182 ± 48 — 295 ± 40 193 ± 59 — 
 Error latency 149 ± 14 — — 206 ± 37 — — 
Monkey U 
N (correct/all) 426/428 425/425 435/440 423/425 427/427 428/434 
 Error rate (%) 0.5 0.0 1.1 0.5 0.0 1.4 
 Latency (ms) 224 ± 32 185 ± 26 — 274 ± 31 260 ± 37 — 
 Error latency 199 ± 44 — — 223 ± 29 — — 
 Deliberate
 
Immediate
 
Anti Pro NoGo Anti Pro NoGo 
Monkey S 
N (correct/all) 4539/4573 4548/4549 4570/4751 4549/4574 4569/4581 4581/4617 
 Error rate (%) 0.7 0.0 3.8 0.5 0.3 0.8 
 Latency (ms) 215 ± 30 188 ± 31 — 254 ± 30 247 ± 30 — 
 Error latency 139 ± 51 — — 157 ± 17 224 ± 21 — 
Monkey X 
N (correct/all) 1293/1296 1289/1292 1288/1425 1288/1289 1293/1293 1293/1326 
 Error rate (%) 0.0 0.0 10.6 0.0 0.0 2.50 
 Latency (ms) 300 ± 39 262 ± 22 — 325 ± 30 324 ± 32 — 
 Error latency 213 ± 36 — — — — — 
Monkey V 
N (correct/all) 749/851 848/848 852/853 824/854 851/851 842/842 
 Error rate (%) 11.7 0.0 0.0 3.5% 0.0 0.0 
 Latency (ms) 223 ± 52 182 ± 48 — 295 ± 40 193 ± 59 — 
 Error latency 149 ± 14 — — 206 ± 37 — — 
Monkey U 
N (correct/all) 426/428 425/425 435/440 423/425 427/427 428/434 
 Error rate (%) 0.5 0.0 1.1 0.5 0.0 1.4 
 Latency (ms) 224 ± 32 185 ± 26 — 274 ± 31 260 ± 37 — 
 Error latency 199 ± 44 — — 223 ± 29 — — 

Values of latency indicate mean ± SD.

We recorded from 170 single GPe neurons in 6 hemispheres of 4 monkeys (n = 102, 34, 20, and 14 for monkeys S, X, U, and V, respectively) that modulated their activity in association with saccades. More than 2 thirds of these neurons increased their firing rates before and during saccades (“increase-type” neurons, n = 117, 70%), whereas the remaining neurons decreased their firing rate (“decrease-type” neurons, n = 53, 30%). For the 4 animals, the proportion of the increase-type neurons ranged from 62–80% (increase/decrease, 70/32, 21/13, 16/4, 10/4, for monkeys S, X, U, and V, respectively).

Figure 2 illustrates the recording sites overlaid on MR images of monkeys X and S. Most saccade-related neurons were found in the anterodorsal part of the GPe, within 2 mm anterior or posterior to the level of the anterior commissure, consistent with our previous study (Yoshida and Tanaka 2009a). Anatomical studies have shown that this part of the GPe receives projections from the associative areas in the striatum (Haber et al. 1993; Francois et al. 1994, 2004; for review, see Parent and Hazrati 1995b). Our recording sites appeared to be somewhat anterior to the sites in other studies that had been examined for saccade-related neuronal activity (Kato and Hikosaka 1995; Shin and Sommer 2010).

Properties of Increase-type Neurons

Figure 3 plots the data of a representative increase-type neuron under 2 different conditions. As the data were aligned to the initiation of saccades, neuronal activity just before antisaccades was greater than that before prosaccades under the Deliberate condition (Fig. 3A, right 2 panels), which is consistent with our previous study (Yoshida and Tanaka 2009a). This neuron also exhibited an enhanced activity during antisaccades under the Immediate condition (Fig. 3B, right 2 panels), while the gradual elevation of firing rate at the time of target onset found in the Deliberate condition disappeared (left panels). To examine whether the GPe was involved in the suppression of reflexive saccades, we randomly introduced the NoGo trials under both conditions. Neuronal activity in the NoGo trials under the Deliberate condition was expected to be related to global saccade suppression, while that under the Immediate condition was likely to be related to selective suppression (Abegg et al. 2012; Fig. 1C). When the data were aligned to the target onset, neuronal activity in the NoGo trials was less than that during antisaccades under both conditions (Fig. 3A,B, left 2 panels), suggesting that this neuron might not be involved in any type of saccade suppression.

Figure 3.

Comparison of the responses between paradigms for representative increase-type neuron. (A) Deliberate condition. (B) Immediate condition. In all panels, data are aligned on the target onset or saccade initiation (vertical lines). Raster lines are sorted according to saccade latency. Traces in different colors indicate spike density for different conditions. The diamond on each raster line indicates the time of saccade initiation (left two panels) or target onset (right 2 panels). Anti, antisaccade; Pro, prosaccade.

Figure 3.

Comparison of the responses between paradigms for representative increase-type neuron. (A) Deliberate condition. (B) Immediate condition. In all panels, data are aligned on the target onset or saccade initiation (vertical lines). Raster lines are sorted according to saccade latency. Traces in different colors indicate spike density for different conditions. The diamond on each raster line indicates the time of saccade initiation (left two panels) or target onset (right 2 panels). Anti, antisaccade; Pro, prosaccade.

To examine the time course of neuronal activity, Figure 4 plots the population activities of the increase-type neurons during different conditions. The traces in Figure 4A,E show that neuronal activities during the NoGo trials were consistently less than those during antisaccades. For individual neurons, Figure 4C,G compare the mean firing rate during 250 ms following the target onset between the antisaccade and NoGo trials. Under the Deliberate condition, 59 (50%) and 55 (47%) out of 117 increase-type neurons displayed activities that were statistically different between the trials for contralateral and ipsilateral targets, respectively (2-way ANOVA, P < 0.05, followed by Scheffé's test, P < 0.05, Fig. 4C, red symbols). Under the Immediate condition, 26 (22%) and 23 (19%) of neurons showed differential activity between the trials for contralateral and ipsilateral targets, respectively (P < 0.05, Fig. 4G). In the whole population, the activity immediately after the target onset in the antisaccade trials was statistically greater than that during the NoGo trials for all combinations of target locations and task conditions (paired t test, P < 10−6), suggesting that these neurons might play only a minor role in saccade suppression. Although many neurons exhibited a gradual elevation of activity prior to target onset during antisaccade trials under the Deliberate condition (Fig. 4A), the baseline activity during stable fixation (500–1000 ms before target onset) did not differ either between conditions or paradigms (2-way ANOVA, P = 0.66), indicating that the variation of neuronal activity during initial fixation did not solely reflect the difference in visual attributes.

Figure 4.

Quantitative analyses of the increase-type neurons. (A) and (B) Time courses of the population activity following the target onset (A) or saccade initiation (B) under the Deliberate condition. The shaded areas indicate ± SEMs. The magenta raster below the traces indicate 25-ms time intervals when neuronal activity during antisaccades differed from that during NoGo trials (A) or prosaccades (B). (C) and (D) Comparison of the firing rate of individual neurons during antisaccades with that during NoGo trials (C) or prosaccades (D) under the Deliberate condition. Neuronal activity was measured for a 250-ms interval following the target onset (black bars in A) or a 150-ms interval starting from 100 ms before saccade initiation (black bars in B). Red symbols indicate the data with a significant difference (P < 0.05). (E) and (F) Time courses of the population activity under the Immediate condition. Conventions are the same as in (A) and (B). (G) and (H) Each data point compares the firing rate of individual neurons during antisaccades with that during the other tasks under the Immediate condition.

Figure 4.

Quantitative analyses of the increase-type neurons. (A) and (B) Time courses of the population activity following the target onset (A) or saccade initiation (B) under the Deliberate condition. The shaded areas indicate ± SEMs. The magenta raster below the traces indicate 25-ms time intervals when neuronal activity during antisaccades differed from that during NoGo trials (A) or prosaccades (B). (C) and (D) Comparison of the firing rate of individual neurons during antisaccades with that during NoGo trials (C) or prosaccades (D) under the Deliberate condition. Neuronal activity was measured for a 250-ms interval following the target onset (black bars in A) or a 150-ms interval starting from 100 ms before saccade initiation (black bars in B). Red symbols indicate the data with a significant difference (P < 0.05). (E) and (F) Time courses of the population activity under the Immediate condition. Conventions are the same as in (A) and (B). (G) and (H) Each data point compares the firing rate of individual neurons during antisaccades with that during the other tasks under the Immediate condition.

To estimate the timing of when neuronal activity during antisaccades differed from that during the NoGo trials, we performed time-series analysis for every 25-ms window starting from 600 ms before and ending at 600 ms following the target onset. The rasters below the traces of population activities in Figure 4A,E summarize the statistical results (Mann Whitney U-test, P < 0.05). When the neuronal differentiation time was defined as the center of the first bin of 10 consecutive intervals with significant difference, the values were −177 ms (contralateral target) and −124 ms (ipsilateral target) for the Deliberate condition, and 27 ms (contra) and 49 ms (ipsi) for the Immediate condition.

The enhancement of neuronal activity during antisaccades suggested a role of the increase-type neurons in the production of volitional saccade commands. To test this, we next examined neuronal activity around the times of saccades. Figure 4B,F show the time courses of the population activity during saccades under the 2 different conditions. The increase-type neurons exhibited enhanced activity during antisaccades compared with that during prosaccades, in particular, under the Deliberate condition (Fig. 4B). To assess the saccade-related activities, the firing rates of individual neurons were measured for a 150-ms interval starting from 100 ms prior to saccades (Fig. 4B,F, black bars). A 2-way ANOVA and the following multiple comparisons revealed that neuronal activity was greater during anti- compared with prosaccades for 30% (n = 35/117, contraversive saccades) and 23% (n = 27, ipsiversive saccades) of neurons under the Deliberate condition, and for 55% (n = 64, contra) and 34% (n = 40, ipsi) of neurons under the Immediate condition (Scheffé's test P < 0.05; Fig 4D,H, red symbols). In the whole population, the magnitude of neuronal activity during antisaccades was greater than that during prosaccades (paired t test, P < 10−4 and P < 10−10 for the Deliberate condition and the Immediate condition, respectively). When we measured the neuronal differentiation time between antisaccade and prosaccade tasks, the values were −325 ms (contralateral saccades) and −263 ms (ipsilateral saccades) in the Deliberate condition, and −21 ms (contra) and −103 ms (ipsi) in the Immediate condition.

To further examine whether the increase-type neurons played a role in saccade generation, trials from each neuron were divided into tertiles according to saccade latencies. Traces in the upper panels of Figure 5A,B show the population activities for the 3 groups aligned to the target onset in contraversive antisaccade trials. Although neuronal activity for trials with shorter latencies was greatest during antisaccades in both conditions, the activity was relatively comparable between groups during prosaccades. When rank correlation between the firing rates (−100 to 50 ms of the target onset) and individual saccade latencies was computed for each neuron, the mean of correlation coefficients was statistically smaller than zero (one sample t test, n = 117, P < 0.05) during antisaccades under all 4 combinations of saccade direction and trial condition. In contrast, the mean of rank correlation coefficients did not differ from zero during prosaccades (P > 0.05). These results suggest that the increase-type neurons might play a role in facilitating antisaccades.

Figure 5.

Correlation between neuronal activity and saccade latency. (A) Time courses of population activity for contraversive saccades in the Deliberate condition. For each neuron, the data were divided into 3 groups with equal number of trials according to saccade latencies. The population activity was computed for each of the 3 groups shown in different line styles. The lower histograms indicate the distributions of correlation coefficients computed between neuronal activities (−100 to 50 ms of target onset, black bar in the upper panel) and individual saccade latency. Arrowhead above each histogram indicates the mean value which was significantly smaller than zero for antisaccades (one-sided t test, P < 10−5) but not for prosaccades. (B) Data for the Immediate condition. Conventions are the same as in (A).

Figure 5.

Correlation between neuronal activity and saccade latency. (A) Time courses of population activity for contraversive saccades in the Deliberate condition. For each neuron, the data were divided into 3 groups with equal number of trials according to saccade latencies. The population activity was computed for each of the 3 groups shown in different line styles. The lower histograms indicate the distributions of correlation coefficients computed between neuronal activities (−100 to 50 ms of target onset, black bar in the upper panel) and individual saccade latency. Arrowhead above each histogram indicates the mean value which was significantly smaller than zero for antisaccades (one-sided t test, P < 10−5) but not for prosaccades. (B) Data for the Immediate condition. Conventions are the same as in (A).

Properties of Decrease-type Neurons

Figure 6 plots the data from a representative decrease-type neuron. This neuron reduced activity in a similar manner during all 3 tasks under the Immediate condition (Fig. 6B), whereas, under the Deliberate condition, the size of the firing modulation was greatest during antisaccades (Fig. 6A). Similarly to previously reported neurons (Yoshida and Tanaka 2009a), this neuron started to modulate activity even during the instruction period when the task-rule was presented, possibly because the monkey prepared for antisaccades even though the target location was still uncertain. For multiple decrease-type neurons (n = 53), Figure 7A,E compare the time courses of the population activities between different trials under both conditions. Unlike the increase-type neurons (Fig. 4E), neuronal activity during the NoGo trials was similar to that during antisaccades under the Immediate condition (Fig. 7E). When the firing rates of individual decrease-type neurons were measured for 250 ms from the target onset (black bars in Fig. 7A,E), 15 (28%, contralateral target) and 10 (19%, ipsilateral target) neurons exhibited differential activity between the antisaccade trials and the NoGo trials under the Deliberate condition, and 5 (9%, contra) and 2 (4%, ipsi) neurons exhibited differential activity under the Immediate condition (2-way ANOVA, P < 0.05, followed by Scheffé's test, P < 0.05, Fig. 7C,G, red dots). In the population as a whole, neuronal activity in the NoGo trials was less than that during antisaccades under the Deliberate condition (paired t test, P < 10−3 and 10−4 for the contralateral and ipsilateral targets, respectively). Meanwhile, neuronal activity was comparable between the trials under the Immediate condition (P = 0.32 and 0.29 for the contralateral and ipsilateral targets, respectively). This suggests that the decrease-type neurons might play a role in “selective” saccade suppression (Fig. 1C). As seen in the increase-type neurons, the decrease-type neurons also altered activity just before the target onset, especially in the antisaccade trials under the Deliberate condition (Fig. 7A). However, neuronal activity during stable fixation (500–1000 ms before target onset) did not vary across tasks and conditions (2-way ANOVA, P = 0.48).

Figure 6.

Comparison of the responses between paradigms for a representative decrease-type neuron. (A) Deliberate condition. (B) Immediate condition. Conventions of figures are the same as in Figure 3.

Figure 6.

Comparison of the responses between paradigms for a representative decrease-type neuron. (A) Deliberate condition. (B) Immediate condition. Conventions of figures are the same as in Figure 3.

Figure 7.

Quantitative analyses of the decrease-type neurons. (AD) Data for the Deliberate condition. (EH) Data for the Immediate condition. Conventions of figures are the same as in Figure 4. Note that the neuronal modulation during the NoGo trials was virtually absent in the Deliberate condition (A), while that in the Immediate condition was comparable with the firing modulation during antisaccades (E).

Figure 7.

Quantitative analyses of the decrease-type neurons. (AD) Data for the Deliberate condition. (EH) Data for the Immediate condition. Conventions of figures are the same as in Figure 4. Note that the neuronal modulation during the NoGo trials was virtually absent in the Deliberate condition (A), while that in the Immediate condition was comparable with the firing modulation during antisaccades (E).

As seen in the example neuron in Figure 6, the decrease-type neurons reduced their firing rate even during prosaccades, but the firing modulation during prosaccades was less than that during antisaccades, at least under the Deliberate condition. For multiple neurons, Figure 7B,F show the time courses of population activities during saccades. Consistent with our previous study (Yoshida and Tanaka 2009a), the firing modulation during antisaccades was greater than that during prosaccades under the Deliberate condition (Fig. 7B). However, there was no significant difference in the magnitude of neuronal modulation between pro- and antisaccades under the Immediate condition (Fig. 7F). When the firing rates of individual decrease-type neurons were measured around the times of saccades (a 150-ms interval starting from 100 ms prior to saccades, Fig. 7B,F, black bars), 25 (47%, contraversive saccades) and 11 (21%, ipsiversive saccades) neurons displayed differential activity between the tasks under the Deliberate condition, and 27 (51%, contra) and 8 (15%, ipsi) neurons showed differential activity under the Immediate condition (unpaired t test, P < 0.05, Fig. 7D,H, red symbols). In the whole population, the magnitude of firing modulation during antisaccades was greater than that during prosaccades in the Deliberate condition (paired t test, P < 10−3 and 0.05 for contraversive and ipsiversive saccades, respectively), while the neuronal modulation was comparable between the trials in the Immediate condition (P = 0.10 and 0.86 for contraversive and ipsiversive saccades, respectively). In contrast to the increase-type neurons (Fig. 5), we failed to find any significant correlation between neuronal activity and saccade latency in any condition, suggesting that the decrease-type neurons might not signal saccade motor commands.

If the decrease-type neurons were involved in saccade suppression, neuronal modulation during erroneous NoGo trials might be reduced as compared with that during correct trials. For a subset of neurons that could be examined during erroneous NoGo trials with targeting saccades under the Immediate condition, the neuronal activities (measured during 250 ms of target onset) averaged 52.8 ± 4.4 (SD, n = 15) and 64.4 ± 5.2 (n = 12) spikes/s for the contralateral and ipsilateral target, respectively. These values did not differ from those during correct trials (51.9 ± 4.1 and 62.2 ± 4.0 spikes/s, paired t test, P > 0.05). We obtained similar results under the Deliberate condition (12 and 10 neurons for contralateral and ipsilateral targets, respectively), likely because the number of error trials in the NoGo trials was too small to detect a significant difference.

Discussion

We found 2 types of GPe neurons that modulated their activity associated with saccades. Similar to our previous study in other animals (Yoshida and Tanaka 2009a), the recording sites were confined to the anterior part of the dorsomedial aspect of the GPe, close to the anterior commissure. According to anatomical studies, our recording sites appear to be within the so-called associative territory, which receives signals from the dorsolateral and medial prefrontal cortices via the dorsal caudate nucleus and/or the ventral STN (Francois et al. 1994; Middleton and Strick 1994; Shink et al. 1996; for review, see Parent and Hazrati 1995b). This part of the GPe is considered to be involved in cognitive processes rather than simple motor executions (Grabli et al. 2004; Saga et al. 2013).

Absence of Confounding Effects of Visual Contrast and Task Difficulty on Neuronal Activity

Most neurons in the GPe exhibited differential activity between the tasks. For example, the magnitude of firing modulation of the increase-type neurons was greatest during antisaccades and was smallest during NoGo trials (Fig. 4). This task-rule-dependent modulation of neuronal activity was not attributed to the difference in stimulus contrast, because the response to the target was similarly affected by the task-rule under both the Deliberate and Immediate conditions, while the target was always white under the former condition (Fig. 1A). Conversely, the neuronal activity during stable fixation (100–500 ms before target onset) under the Deliberate condition did not differ across the tasks, although the color of the fixation point was different.

Also, one might argue that the task-rule-dependent modulation of neuronal activity could reflect the general difficulty of the task. It is practically impossible to adjust difficulty across the 3 paradigms; our data suggest that the task-rule-dependent modulation was not solely due to the general difficulty of the task. For example, the decrease-type neurons exhibited a similar amount of firing modulation between the tasks under the Immediate condition (Fig. 7E). Furthermore, although all tasks under the Immediate condition must be more difficult than those under the Deliberate condition, the magnitude of firing modulation was comparable between the conditions for both types of neurons (Figs. 4 and 7). Thus, the task-rule-dependent modulation of neuronal activity was likely to reflect neuronal signals that were necessary to regulate animals' behavior, rather than the stimulus attributes or general difficulty of the task. In the subsequent sections, we will consider how these signals were generated within the basal ganglia pathways, and how GPe neuronal activity could regulate voluntary movements.

Origin of Saccade Signals in the GPe

The GPe has been viewed as part of the indirect pathway, receiving signals from the striatum and sending inhibitory output to the STN (Fig. 8, the second column). Considering that the striatal neurons display a relatively lower baseline firing rate (Buser et al. 1974; Hikosaka et al. 1989) and send GABAergic projections to the GPe, neurons in the GPe would be expected to exhibit decreased activity during movements, just like our decrease-type neurons. However, we also found many neurons exhibiting increased activity during antisaccades, and in fact, more than half of the task-related neurons showed increased activity. These results are consistent with previous studies of somatic movements (for review, see Goldberg and Bergman 2011).

Figure 8.

Diagrams showing how the signals in the GPe could facilitate antisaccades or suppress reflexive saccades. Black symbol in each rectangle indicates the direction of firing modulation during saccades. The decrease-type GPe neurons could receive direct inputs from the caudate nucleus (Cd) through the indirect pathway, while the increase-type GPe neurons might receive signals through the cortico-STN-GPe pathways. The indirect and hyperdirect pathways may suppress reflexive saccades, while the direct and cortico-STN-GPe pathways might facilitate voluntary saccades: these pathways might be involved in selective and global facilitations, respectively. Cd, caudate nucleus; GPi, internal segment of the globus pallidus; SNr, substantia nigra pars reticulata; STN. Subthalamic nucleus.

Figure 8.

Diagrams showing how the signals in the GPe could facilitate antisaccades or suppress reflexive saccades. Black symbol in each rectangle indicates the direction of firing modulation during saccades. The decrease-type GPe neurons could receive direct inputs from the caudate nucleus (Cd) through the indirect pathway, while the increase-type GPe neurons might receive signals through the cortico-STN-GPe pathways. The indirect and hyperdirect pathways may suppress reflexive saccades, while the direct and cortico-STN-GPe pathways might facilitate voluntary saccades: these pathways might be involved in selective and global facilitations, respectively. Cd, caudate nucleus; GPi, internal segment of the globus pallidus; SNr, substantia nigra pars reticulata; STN. Subthalamic nucleus.

How do neurons in the GPe generate increased activity associated with saccades? Because the GPe receives projections from multiple nuclei in the basal ganglia, there are several possible mechanisms activating GPe neurons. For example, a transient suppression of inhibitory signals from the striatum, a release of lateral inhibition within the GPe, or feedback through the striatal-GPe-STN-GPe pathways might underlie the increased activity. However, previous studies demonstrated that excitation of GPe neurons following cortical stimulation was mediated by the cortico-STN-GPe pathways (Nambu et al. 2000; Kita et al. 2004). Furthermore, a recent study has shown that most increasing signals in GPe neurons during voluntary movements are attributed to glutamatergic inputs, while most of decreasing signals reflect GABAergic projections (Kaneko et al. 2014). Anatomical data also suggest the possibility that the increase-type neurons might receive glutamatergic inputs from the STN (Shink et al. 1996). The STN receives direct projections from the cerebral cortex (Inase et al. 1999; Nambu et al. 1996; Kelly and Strick 2004) and composes the “hyperdirect” pathway that elevates neuronal activity in the substantia nigra pars reticulata (SNr) and the internal segment of the GP (GPi) (Smith et al. 1990; Parent and Hazrati 1995a; Figure 8, the third column). In addition to the hyperdirect pathway, which ultimately suppresses neuronal activity in the superior colliculus (SC) or the thalamus (Nambu et al. 2002), GPe neurons also have direct access to the SNr and GPi (Smith et al. 1990; Sato et al. 2000), which could facilitate signals in the SC and thalamus via the disinhibition mechanisms (Fig. 8, right column). Thus, the different types of signals in the GPe might come from 2 different sources.

Possible Roles of Two Different Pathways Through the GPe

What then are the roles of the pathways mediated by the 2 different types of GPe neurons? Previous studies suggest that each basal ganglia pathway is assigned to the regulation of a different aspect of movement. For example, the direct pathway in the basal ganglia is thought to be involved in the generation of purposeful actions (for review, see Hikosaka et al. 2000). Neurons in the caudate nucleus, the SNr and the SC exhibit greater firing modulation during memory-guided saccades compared with visually guided saccades (Hikosaka and Wurtz 1983a,b; for review, see Shires et al. 2010). The enhancement of neuronal modulation in the direct pathway during antisaccades has also been reported (Ford and Everling 2009; Watanabe and Munoz 2009), and these signals seem to have a causal role in the generation of antisaccades (Watanabe and Munoz 2010b). Based on neuronal activity in the caudate nucleus, Watanabe and Munoz (2009, 2010a) have proposed that neurons with a preference for contraversive saccades may facilitate antisaccades through the direct pathway, while those with an ipsilateral preference may suppress reflexive saccades via the indirect pathway. According to this hypothesis, the decrease-type neurons in the GPe found in the present study might receive volitional signals from the caudate nucleus and suppress contraversive reflexive saccades. On the other hand, we found a significant negative correlation between neuronal activity of the increase-type neurons and antisaccade latency under both conditions (Fig. 5), suggesting that these neurons might play a role in facilitating saccades, in a similar manner to neurons in the caudate nucleus that have a contralateral preference (Hikosaka et al. 1989; Lau and Glimcher 2007). Unfortunately, however, our study could not elucidate how these GPe neurons facilitate antisaccades in cooperation with the volitional neurons in the caudate nucleus. One possibility is that neurons in the SNr and GPi might be suppressed through both the direct and the cortico-STN-GPe-SNr/GPi pathways (Fig. 8, right column), causing disinhibition in the downstream structures when the 2 parallel pathways transmit signals simultaneously.

In relation to this, Isoda and Hikosaka (2008) found 2 types of neurons in the ventral STN as monkeys switched from automatic to volitional saccades. One type of neuron exhibited a transient activity during behavioral switching, while the other was active as the animals suppressed automatic saccades (no-go neuron). Our increase-type neurons in the GPe might be relevant to the former neurons, both of which are likely to facilitate volitional saccades (Isoda and Hikosaka 2008). Consistent with this, many increase-type neurons also exhibited some activity during prosaccades, although the magnitude of firing modulation was smaller than that during antisaccades (Fig. 4). The signals carried by the increase-type GPe neurons might facilitate saccades in general, while the contribution of these signals appears to be dominant for antisaccades. In contrast, the decrease-type neurons in the GPe and the no-go neurons in the STN might play a role in suppressing reflexive saccades. These hypotheses could be tested in future studies by pharmacological manipulation of signals in the different GPe pathways.

Global and Selective Suppression of Movements

To elucidate how signals in the GPe suppress unwanted movements, we introduced the NoGo trial under 2 different conditions. In the Deliberate condition, the animals might simply ignore any visual stimulus during the NoGo trials, whereas in the Immediate condition, they had to pay attention to the target to understand the task being performed. Thus, the suppression of reflexive saccades must be more selective under the Immediate condition compared with the Deliberate condition (Fig. 1C). Our data showed that the magnitude of firing modulation in the decrease-type neurons during NoGo trials was comparable with that during antisaccades under the Immediate condition, whereas the same neurons exhibited only a slight change in activity in the NoGo trials under the Deliberate condition. These results suggest that the indirect pathway mediated by the decrease-type neurons in the GPe might be related to the selective inhibition of unwanted actions. Our findings are in line with the recent proposal that the indirect and hyperdirect pathways mediate signals for selective and global suppression, respectively (Majid et al. 2013; for review, see Aron 2011). To further confirm this hypothesis, future studies need to examine causal roles of each type of GPe signal in the generation of antisaccades.

In conclusion, we found 2 types of neurons in the GPe that exhibited different firing modulation during NoGo and antisaccade trials. We suggest that the increase-type neurons are involved in the process of saccade facilitation, while the decrease-type neurons are relevant to the selective suppression of reflexive saccades. The increase-type neurons could be involved in the cortico-STN-GPe-SNr/GPi pathway, which might operate in parallel with the direct, indirect and hyperdirect pathways in the basal ganglia.

Funding

This work was supported by grants from the Ministry of Education, Culture, Sports, Science and Technology of Japan, the Ministry of Health, Labour and Welfare of Japan, the Japan Society for the Promotion of Science, the Smoking Research Foundation and the Takeda Science Foundation. A.Y. was supported by the Graduate Research Fellowship for Young Scientists from the Japan Society for the Promotion of Science. Animals were provided by the National Bio-Resource Project.

Notes

We are grateful to T. Mori and A. Hironaka for the animal care, M. Suzuki for administrative help, and all laboratory members for comments and discussions. Conflict of Interest: None declared.

References

Abegg
M
Sharma
N
Barton
JJ
.
2012
.
Antisaccades generate two types of saccadic inhibition
.
Biol Psychol
 .
89
:
191
194
.
Albin
RL
Young
AB
Penney
JB
.
1989
.
The functional anatomy of basal ganglia disorders
.
Trends Neurosci
 .
12
:
366
375
.
Aron
AR
.
2011
.
From reactive to proactive and selective control: developing a richer model for stopping inappropriate responses
.
Biol Psychiatry
 .
69
:
e55
e68
.
Brotchie
P
Iansek
R
Horne
MK
.
1991
.
Motor function of the monkey globus pallidus. 1. Neuronal discharge and parameters of movement
.
Brain
 .
114
:
1667
1683
.
Buser
P
Pouderoux
G
Mereaux
J
.
1974
.
Single unit recording in the caudate nucleus during sessions with elaborate movements in the awake monkey
.
Brain Res
 .
71
:
337
344
.
Chan
CS
Surmeier
DJ
Yung
WH
.
2005
.
Striatal information signaling and integration in globus pallidus: timing matters
.
Neurosignals
 .
14
:
281
289
.
DeLong
MR
.
1990
.
Primate models of movement disorders of basal ganglia origin
.
Trends Neurosci
 .
13
:
281
285
.
Ford
KA
Everling
S
.
2009
.
Neural activity in primate caudate nucleus associated with pro- and antisaccades
.
J Neurophysiol
 .
102
:
2334
2341
.
Francois
C
Grabli
D
McCairn
K
Jan
C
Karachi
C
Hirsch
EC
Feger
J
Tremblay
L
.
2004
.
Behavioural disorders induced by external globus pallidus dysfunction in primates II. Anatomical study
.
Brain
 .
127
:
2055
2070
.
Francois
C
Yelnik
J
Percheron
G
Fenelon
G
.
1994
.
Topographic distribution of the axonal endings from the sensorimotor and associative striatum in the macaque pallidum and substantia nigra
.
Exp Brain Res
 .
102
:
305
318
.
Goldberg
JA
Bergman
H
.
2011
.
Computational physiology of the neural networks of the primate globus pallidus: function and dysfunction
.
Neuroscience
 .
198
:
171
192
.
Grabli
D
McCairn
K
Hirsch
EC
Agid
Y
Feger
J
Francois
C
Tremblay
L
.
2004
.
Behavioural disorders induced by external globus pallidus dysfunction in primates I. Behavioural study
.
Brain
 .
127
:
2039
2054
.
Haber
SN
Lynd-Balta
E
Mitchell
SJ
.
1993
.
The organization of the descending ventral pallidal projections in the monkey
.
J Comp Neurol
 .
329
:
111
128
.
Hikosaka
O
Sakamoto
M
Usui
S
.
1989
.
Functional properties of monkey caudate neurons. I. Activities related to saccadic eye movements
.
J Neurophysiol
 .
61
:
780
798
.
Hikosaka
O
Takikawa
Y
Kawagoe
R
.
2000
.
Role of the basal ganglia in the control of purposive saccadic eye movements
.
Physiol Rev
 .
80
:
953
978
.
Hikosaka
O
Wurtz
RH
.
1983a
.
Visual and oculomotor functions of monkey substantia nigra pars reticulata. I. Relation of visual and auditory responses to saccades
.
J Neurophysiol
 .
49
:
1230
1253
.
Hikosaka
O
Wurtz
RH
.
1983b
.
Visual and oculomotor functions of monkey substantia nigra pars reticulata. III. Memory-contingent visual and saccade responses
.
J Neurophysiol
 .
49
:
1268
1284
.
Inase
M
Tokuno
H
Nambu
A
Akazawa
T
Takada
M
.
1999
.
Corticostriatal and corticosubthalamic input zones from the presupplementary motor area in the macaque monkey: comparison with the input zones from the supplementary motor area
.
Brain Res
 .
833
:
191
201
.
Isoda
M
Hikosaka
O
.
2008
.
Role for subthalamic nucleus neurons in switching from automatic to controlled eye movement
.
J Neurosci
 .
28
:
7209
7218
.
Kaneko
N
Hatanaka
N
Takara
S
Takada
M
Nambu
A
.
2014
.
Glutamatergic and GABAergic control of monkey pallidal activity during performance of a motor task
.
Japan Neurosci Soc Abstr
 .
37
:
1.141
.
Kato
M
Hikosaka
O
.
1995
.
Functional of the indirect pathway in the basal ganglia oculomotor system: visuo-oculomotor activities of external pallidum neurons
. In:
Segawa
M
Nomura
Y
, editors.
Age-related dopamine-deficient disorders
 .
Basal
:
Karger
. pp.
178
187
.
Kelly
RM
Strick
PL
.
2004
.
Macro-architecture of basal ganglia loops with the cerebral cortex: use of rabies virus to reveal multisynaptic circuits
.
Prog Brain Res
 .
143
:
449
459
.
Kita
H
Nambu
A
Kaneda
K
Tachibana
Y
Takada
M
.
2004
.
Role of ionotropic glutamatergic and GABAergic inputs on the firing activity of neurons in the external pallidum in awake monkeys
.
J Neurophysiol
 .
92
:
3069
3084
.
Lau
B
Glimcher
PW
.
2007
.
Action and outcome encoding in the primate caudate nucleus
.
J Neurosci
 .
27
:
14502
14514
.
Majid
DS
Cai
W
Corey-Bloom
J
Aron
AR
.
2013
.
Proactive selective response suppression is implemented via the basal ganglia
.
J Neurosci
 .
33
:
13259
13269
.
Middleton
FA
Strick
PL
.
1994
.
Anatomical evidence for cerebellar and basal ganglia involvement in higher cognitive function
.
Science
 .
266
:
458
461
.
Mitchell
SJ
Richardson
RT
Baker
FH
DeLong
MR
.
1987
.
The primate globus pallidus: neuronal activity related to direction of movement
.
Exp Brain Res
 .
68
:
491
505
.
Munoz
DP
Everling
S
.
2004
.
Look away: the anti-saccade task and the voluntary control of eye movement
.
Nat Rev Neurosci
 .
5
:
218
228
.
Nambu
A
Takada
M
Inase
M
Tokuno
H
.
1996
.
Dual somatotopical representations in the primate subthalamic nucleus: evidence for ordered but reversed body-map transformations from the primary motor cortex and the supplementary motor area
.
J Neurosci
 .
16
:
2671
2683
.
Nambu
A
Tokuno
M
Hamada
I
Kita
H
Imanishi
M
Akazawa
T
Ikeuchi
Y
Hasegawa
N
.
2000
.
Excitatory cortical inputs to pallidal neurons via the subthalamic nucleus in the monkey
.
J Neurophysiol
 .
84
:
289
300
.
Nambu
A
Tokuno
H
Takada
M
.
2002
.
Functional significance of the cortico-subthalamo-pallidal ‘hyperdirect’ pathway
.
Neurosci Res
 .
43
:
111
117
.
Parent
A
Hazrati
LN
.
1995a
.
Functional anatomy of the basal ganglia. I. The cortico-basal ganglia-thalamo-cortical loop
.
Brain Res Brain Res Rev
 .
20
:
91
127
.
Parent
A
Hazrati
LN
.
1995b
.
Functional anatomy of the basal ganglia. II. The place of subthalamic nucleus and external pallidum in basal ganglia circuitry
.
Brain Res Brain Res Rev
 .
20
:
128
154
.
Saga
Y
Hashimoto
M
Tremblay
L
Tanji
J
Hoshi
E
.
2013
.
Representation of spatial- and object-specific behavioral goals in the dorsal globus pallidus of monkeys during reaching movement
.
J Neurosci
 .
33
:
16360
16371
.
Sato
F
Lavallee
P
Levesque
M
Parent
A
.
2000
.
Single-axon tracing study of neurons of the external segment of the globus pallidus in primate
.
J Comp Neurol
 .
417
:
17
31
.
Shin
S
Sommer
MA
.
2010
.
Activity of neurons in monkey globus pallidus during oculomotor behavior compared with that in substantia nigra pars reticulata
.
J Neurophysiol
 .
103
:
1874
1887
.
Shink
E
Bevan
MD
Bolam
JP
Smith
Y
.
1996
.
The subthalamic nucleus and the external pallidum: two tightly interconnected structures that control the output of the basal ganglia in the monkey
.
Neuroscience
 .
73
:
335
357
.
Shires
J
Joshi
S
Basso
MA
.
2010
.
Shedding new light on the role of the basal ganglia-superior colliculus pathway in eye movements
.
Curr Opin Neurobiol
 .
20
:
717
725
.
Smith
Y
Hazrati
LN
Parent
A
.
1990
.
Efferent projections of the subthalamic nucleus in the squirrel monkey as studied by the PHA-L anterograde tracing method
.
J Comp Neurol
 .
294
:
306
323
.
Tanaka
M
.
2005
.
Involvement of the central thalamus in the control of smooth pursuit eye movements
.
J Neurosci
 .
25
:
5866
5876
.
Turner
RS
Anderson
ME
.
1997
.
Pallidal discharge related to the kinematics of reaching movements in two dimensions
.
J Neurophysiol
 .
77
:
1051
1074
.
Watanabe
M
Munoz
DP
.
2009
.
Neural correlates of conflict resolution between automatic and volitional actions by basal ganglia
.
Eur J Neurosci
 .
30
:
2165
2176
.
Watanabe
M
Munoz
DP
.
2010a
.
Presetting basal ganglia for volitional actions
.
J Neurosci
 .
30
:
10144
10157
.
Watanabe
M
Munoz
DP
.
2010b
.
Saccade suppression by electrical microstimulation in monkey caudate nucleus
.
J Neurosci
 .
30
:
2700
2709
.
Wichmann
T
DeLong
MR
Guridi
J
Obeso
JA
.
2011
.
Milestones in research on the pathophysiology of Parkinson's disease
.
Mov Disord
 .
26
:
1032
1041
.
Yoshida
A
Tanaka
M
.
2009a
.
Enhanced modulation of neuronal activity during antisaccades in the primate globus pallidus
.
Cereb Cortex
 .
19
:
206
217
.
Yoshida
A
Tanaka
M
.
2009b
.
Neuronal activity in the primate globus pallidus during smooth pursuit eye movements
.
Neuroreport
 .
20
:
121
125
.