Impaired Reaching and Grasping After Focal Inactivation of Globus
Pallidus Pars Interna in the Monkey
KRISTIN K. WENGER,1 KRYSTINA L. MUSCH,2 AND JONATHAN W. MINK1,3
1
Department of Neurology and 2Program in Physical Therapy, Washington University School of Medicine; and 3St. Louis
Children’s Hospital, St. Louis, Missouri 63110
INTRODUCTION
It is generally agreed that the basal ganglia play a role in the
control of movement. However, the specific contribution of the
basal ganglia to normal movement remains controversial. Inactivation or ablation of basal ganglia output neurons for limb
movement in globus pallidus pars interna (GPi) has been used
to study the contribution of basal ganglia circuits to an otherwise intact motor system. Although many previous studies
have described clear and consistent deficits after GPi lesions,
others have revealed little or no effect on movement (see
The costs of publication of this article were defrayed in part by the payment
of page charges. The article must therefore be hereby marked “advertisement”
in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
DeLong and Georgopoulos 1981; Mink 1996a for review).
Discrepancy among findings has been attributed to factors such
as lesion placement, lack of quantitative behavioral measures,
or task differences. An additional confounding factor is the
observation that lesions of the GPi can improve movement in
Parkinson’s disease (Baron et al. 1996; Lang et al. 1997).
The most consistent finding in primates after GPi lesion is
slowing of movement, muscular co-contraction, and a flexor
positional bias (DeLong and Coyle 1979; Hore and Vilis 1980;
Inase et al. 1996; Mink and Thach 1991). In concordance with
known properties of GPi discharge during movement, we have
hypothesized that the basal ganglia output acts during voluntary movement to inhibit potentially competing motor patterns
(Mink 1996a; Mink and Thach 1993; Thach et al. 1993).
During any given movement, the desired movement must be
planned and executed, but also a multitude of potentially
competing motor mechanisms must be inhibited to prevent
them from interfering with the desired movement. It follows
that removal of GPi output would lead to unwanted competition resulting in slowness of movement, muscular co-contraction, and abnormal postures.
The purpose of this study was to test the general hypothesis
that the GPi output contributes to inhibition of competing
motor patterns in monkeys performing a natural reach and
grasp task before and after reversible inactivation of GPi neurons. For the primate sitting upright, the typical posture is one
of flexed arms, legs, and neck. In that posture, tonic neck,
vestibular, and other postural mechanisms favor flexion
(Denny-Brown 1967; Magnus 1924; Mink 1996a). Movement
of any limb out of the initial position requires selective inhibition of those postural mechanisms, for that limb only, in
addition to activation of mechanisms involved in producing the
desired movement. Inability to inhibit those posture-holding
mechanisms for the moving limb would be expected to slow
movement and to bias movement toward that initial flexed
posture. We hypothesized specifically that GPi inactivation in
the monkey would result in an impairment in moving a limb
out of the initial posture, but little or no impairment in returning the limb to the initial posture.
To test this hypothesis, two monkeys were studied while
they performed a reaching task before and after GPi inactivation with the GABAA agonist muscimol. The task consisted of
a reach-grasp-retrieval sequence in which the monkey’s target
(and reward) was an apple bit. The monkeys, already familiar
with the experimenter and with sitting in a primate chair,
readily learned this task over the course of only five training
sessions. No limits on reaction time, movement time, or reach
0022-3077/99 $5.00 Copyright © 1999 The American Physiological Society
2049
Downloaded from http://jn.physiology.org/ by 10.220.33.1 on June 16, 2017
Wenger, Kristin K., Krystina L. Musch, and Jonathan W. Mink.
Impaired reaching and grasping after focal inactivation of globus
pallidus pars interna in the monkey. J. Neurophysiol. 82: 2049 –2060,
1999. The purpose of this study was to test the hypothesis that the
basal ganglia output from globus pallidus pars interna (GPi) contributes to inhibition of competing motor patterns to prevent them from
interfering with a volitional movement. To test this hypothesis, the
kinematics of a natural reach, grasp, and retrieval task were measured
in the monkey before and after focal inactivation in GPi with the
GABAA agonist muscimol. Two rhesus monkeys were trained to
reach in a parasagittal plane to grasp a 1-cm cube of apple and retrieve
it. Reflective markers were applied to the shoulder, elbow, wrist, and
index finger. Movements were videotaped at 60 fields/s, digitized, and
analyzed off-line. In each session the monkey performed 12–15
reaches before and 12–15 reaches after injection of 0.5 ml of 8.8 mM
muscimol. Muscimol was injected into 22 separate locations in the
“arm” area of GPi. Inactivation of the GPi with muscimol produced
movement deficits in a reach-grasp-retrieve task that can be summarized as follows: 1) decreased peak wrist velocity during the reach to
target; 2) decreased elbow and shoulder angular velocities, with elbow
angular velocity relatively more impaired than shoulder angular velocity; resulting in 3) higher maximum vertical wrist and index finger
positions at the apex of the reach; 4) prolonged latency from the end
of the reach to the completion of grasp; and 5) less impairment of
retrieval than reach, with inactivation at the majority of sites causing
no impairment and some actually speeding up retrieval despite slow
reaching. The results of this study show that reaching movements are
impaired in a specific way after focal inactivation of GPi in previously
normal monkeys. The slowing of the reach with normal (or fast)
retrieval suggests that there is difficulty inhibiting the posture holding
mechanisms that were active before the reach, but that assist the
retrieval. The nature of the impairment supports the hypothesis that
GPi lesions disrupt the ability to inhibit competing motor mechanisms
to prevent them from interfering with desired voluntary movement.
2050
K. K. WENGER, K. L. MUSCH, AND J. W. MINK
path were imposed by instrumentation. Very small injections
of muscimol were used to avoid spread to nearby structures.
In the present study, inactivation at many sites in GPi caused
a specific impairment of the reach-grasp-retrieval task. Reaching to the target was slow and performed with a flexor bias.
Grasping was impaired after some injections due to excessive
flexion of the fingers. After most injections, the retrieval (return to flexion) was unimpaired, but in others it was slow.
Results from injection at several different sites in GPi sites
suggested a topographic basis for the two different deficit
patterns. These results have been presented previously in abstract form (Thach et al. 1996).
METHODS
Subjects and apparatus
Surgery was performed in a stereotaxic frame under general anesthesia with Isoflurane. Using sterile technique, the scalp, the left
temporalis muscle, and the periosteum were reflected. Six stainless
steel screws were inserted through the outer table of the skull and
covered with methyl methacrylate. The exposed skull was also covered with methyl methacrylate, and bolts were imbedded in the acrylic
to allow fixation of the head during recording. After the animal
recovered and the behavior returned to baseline, a second procedure
was performed to place a 20 3 20 mm square Lucite chamber
centered over the left GPi at 50° from vertical in the coronal plane
[stereotaxic coordinates A12.0, L9.0, H6.0 (Snider and Lee 1961)].
The animal was given analgesia with buprenorphine and acetaminophen after each procedure.
The borders of GPi were localized with single-unit recording during
a wrist movement task and with microstimulation. The animal was
seated in a primate chair with attached wrist manipulandum as described by Mink and Thach (1991), and the head was held by bolts
attached to the acrylic cap. Multiple penetrations were made at 1-mm
intervals while recording neuronal discharge patterns at rest and
during wrist movement (Mink 1996b). GPi was identified by the
characteristic high-frequency tonic discharge rate of neurons in this
structure (DeLong 1971). The adjacent internal capsule was identified
by electrical activity characteristic of fibers and with microstimulation. The optic tract was identified by change in discharge evoked by
light. Injection sites were chosen to be at sites within GPi where
neurons discharged in relation to right upper extremity movements
and a minimum of 1 mm from the dorsolateral border to avoid spread
of muscimol into GPe.
Before injection, a series of 12–15 reaches was performed by the
monkey, and data were collected as described below. A 10-ml syringe
with a 24-gauge needle was filled with 8.8 mM (1 mg/ml) muscimol
(Sigma). In some sessions, a tungsten microelectrode (impedence 300
kV to 1 MV) was attached to the syringe needle with the electrode tip
extending ;300 mm beyond the needle tip. The head was held, and
the syringe was attached to a microdrive and the Lucite chamber.
After the monkey performed a series of wrist movements (Mink
1996b), the syringe was lowered to the predetermined target coordinates. A total of 0.5 ml of muscimol was injected in 0.1– 0.2 ml
increments over a 5-min period. The monkey performed a series of
wrist movement tasks again, the injection needle was withdrawn, and
a second series of 12–15 reaches was performed 30 – 45 min after the
muscimol injection was completed. At least 2 days elapsed between
injections to assure that there was no residual effect of muscimol. In
four sessions, 0.5 ml of saline was injected as a control for nonspecific
effects of injection.
Data collection and kinematic analysis
FIG. 1. A: schematic representation of a monkey reaching to an apple cube
target and retrieving it. ●, joints marked for kinematic analysis. Arrows
indicate movement direction. B: stick figure representation of the arm during
reach and retrieval. The shoulder, elbow, wrist, and index finger are indicated
with the letters S, E, W, and I, respectively. Arrows indicate the direction of
movement.
To identify the index finger, wrist, elbow, and shoulder joints for
video-based motion analysis, bright white circular markers, 1.5 cm
diam, were painted on the skin overlying the acromion (shoulder),
lateral epicondyl (elbow), and extensor aspect of the wrist. The tip of
the right index finger was also painted white. The white markers were
outlined in black to increase the contrast. The subjects were videotaped at 60 fields per second. The video camera recorded the movements in the sagittal plane and was zoomed-in to provide the largest
image possible. Marker positions were tracked and digitized using a
Peak Performance Motion Measurement System, and the data were
smoothed using a fourth-order Butterworth filter at 6 Hz. Angular
displacements were calculated from the marker positions; angular and
linear displacement data were numerically differentiated to calculate
velocity and acceleration. The joint angle conventions used for data
analyses were as follows: 1) for the elbow, 180° was defined as full
extension (arm and forearm in line) and decreasing angle was referred
to as flexion; and 2) for the shoulder, 180° was defined with the arm
Downloaded from http://jn.physiology.org/ by 10.220.33.1 on June 16, 2017
Two male rhesus monkeys (monkeys T and L) weighing 7–10 kg
were trained to sit in a primate chair and reach for a bit of apple. The
monkeys were restrained only by a collar-like aluminum plate attached to the chair into which they inserted their necks. The monkeys
were conditioned to sit upright in the chair with their legs flexed,
upper arms vertical, elbows flexed 80 –90°, and each hand lightly
grasping the top of a 1-in. diam pole (Fig. 1). An apple bit, ;1 cm3,
was presented on the end of a small flexible stick (mounted on a fixed
pole) at shoulder height in front of the right shoulder and at a distance
such that it was aligned with the metacarpal-phalangeal joint with the
arm outstretched. The straight path distance of the wrist from the
starting position to the target was ;24 cm for monkey T and 28 cm for
monkey L. The apple bit was presented as a target only if the monkey
assumed the required initial posture. Once the apple bit was attached
to the stick, the monkey could reach with the right arm to grasp,
retrieve, and eat the apple bit. If the monkey began to reach with the
left arm, he was prevented from obtaining the apple bit. The monkeys
had been trained previously to perform a wrist movement task similar
to those described by Mink and Thach (1991).
Localization of GPi and injection protocol
IMPAIRED REACHING AFTER GLOBUS PALLIDUS INACTIVATION
Histology
To verify the accuracy of localization using physiological methods,
the brain of one monkey was processed for histology. After being
given a lethal dose of pentobarbital sodium, the monkey was perfused
transcardially with normal saline followed by 4% paraformaldehyde.
The brain was removed, fixed for 4 wk, and subsequently saturated
with 4% paraformaldehyde in a 40% sucrose solution. The brain was
blocked in the coronal plane, frozen, cut in 50-mm sections with a
microtome, and stained with thionin. Evaluation of the stained sections in comparison with the physiologically identified landmarks
revealed that the physiological mapping was accurate to within 0.5
mm. The injection locations from the second monkey were based on
physiological mapping.
RESULTS
The normal reach in our paradigm consisted of a slightly
curved wrist path that was made by flexing the shoulder and
slightly flexing and then extending the elbow (Fig. 2). As the
apple bit target was reached, the monkey grasped the target and
then retrieved the apple bit to his mouth by flexing the elbow
and extending the shoulder. As described in METHODS, the
sequence was divided into three phases for analysis: 1) Reach
to target, 2) Grasp of target, and 3) Retrieval. Figure 2 shows
wrist position, and elbow and shoulder angles over the course
of a typical reach. In the normal state, both the reach to target
and retrieval were performed with a single movement, i.e., with
no corrective sub-movements.
Muscimol was injected at a total of 22 sites in the GPi of 2
monkeys: 10 in monkey T, 12 in monkey L. Using peak tangential wrist velocity during the reach to target phase as the
dependent variable, significant deficits (P , 0.05, t-test) were
seen after injection into nine sites in monkey T and six sites in
monkey L (Tables 1 and 2).
Slowing of reach to target after GPi inactivation
Inactivation in GPi caused slowing of peak wrist velocity
during the reach at the majority of injection sites. Average peak
wrist velocities for the effective injection sites are shown in
Table 1. When averaged across injections, peak tangential
wrist velocity was decreased to a similar degree in the two
monkeys (Fig. 3). In monkey T, peak wrist velocity decreased
from 87 to 74 cm/s (t 5 8.76, P , 0.001) and in monkey L,
peak wrist velocity decreased from 100 cm/s to 85 cm/s (t 5
4.55, P , 0.01). Accompanying the decreased peak tangential
wrist velocity was a prolongation of movement time from 435
to 562 ms (t 5 24.53, P , 0.01) in monkey T and from 440 to
528 ms (t 5 22.92, P , 0.05) in monkey L. Correlation
between injection number and preinjection (control) wrist velocity in each monkey was not significant (monkey T, r 5 0.42;
monkey L, r 5 20.59), hence the deficits were not due to injury
caused by repeated penetrations.
Injections that produced slowing of peak wrist velocity
during the reach also decreased elbow and shoulder angular
velocities. In both monkeys, the percentage decrease of elbow
angular velocity was significantly greater than the percentage
decrease of shoulder angular velocity. Average peak elbow and
shoulder angular velocities for the effective injection sites are
shown in Table 2. In monkey T, elbow angular velocity decreased from 263 to 182 deg/s (t 5 14.58, P , 0.001), and
shoulder angular velocity decreased from 330 to 273 deg/s (t 5
6.91, P , 0.001). The percentage elbow angular velocity
decrease (29.4%) was greater than the shoulder angular velocity decrease (16.9%; t 5 9.06, P , 0.001). In monkey L, elbow
angular velocity decreased from 296 to 240 deg/s (t 5 10.41,
P , 0.001), and shoulder angular velocity decreased from 302
to 253 deg/s (t 5 8.24, P , 0.001). The elbow angular velocity
decrease (19.3%) was greater than the shoulder angular velocity decrease (16.4%; t 5 3.97, P , 0.05).
Slow reaching due to GPi inactivation was accompanied by
an increase in the maximum vertical wrist position during the
reach. During the normal reach to target, the monkey’s wrist
typically followed a curvilinear path in which the wrist acquired its maximum vertical position when the target was
grasped. After GPi inactivation, the monkey tended to follow a
more curved wrist path in which the maximum vertical position
of the wrist was greater than the final vertical position and
occurred before the target was grasped. In monkey T, the mean
maximum vertical wrist position increased from 2.9 cm below
target to 2.3 cm below target (t 5 3.21, P , 0.05). In monkey
L, maximum vertical wrist position increased from 1.1 cm
below target to 0.8 cm below target (t 5 4.78, P , 0.01).
Reach endpoint accuracy
GPi inactivation caused slight yet consistent impairment of
index finger endpoint accuracy in the vertical (Y) direction but
not in the horizontal (X) direction (Fig. 4). When averaged
across injections, the Y endpoint of the index finger was increased to a similar degree in the two monkeys. In monkey T,
the Y endpoint increased from 0.7 6 0.17 cm above the target
to 1.3 6 0.46 cm above the target (t 5 24.18, P , 0.01) and
in monkey L, the Y endpoint increased from 0.9 6 0.27 cm
above target to 1.2 6 0.31 cm above target but did not reach
significance. These results are comparable to those shown
Downloaded from http://jn.physiology.org/ by 10.220.33.1 on June 16, 2017
vertical at the monkey’s side and decreasing angle was referred to as
flexion.
For analysis, the movement was broken into three phases: 1) reach
to target, 2) grasp of target, 3) retrieval of target. Using custom
software, the following events were defined and marked: 1) “Start of
movement” (reach or retrieval) was defined as the time and position at
which the tangential wrist velocity exceeded 10% of its peak; 2) “End
of movement” (reach or retrieval) was the time and position at which
the wrist velocity dropped below 10% of the peak or (in a few trials)
when the tangential velocity reached a minimum prior to a subsequent
corrective movement; 3) “Grasp” was defined as the time at which the
hand closed around the apple cube target. Kinematic data for the reach
and retrieval phases were analyzed from start of movement to end of
movement, before any corrective movements. Endpoint error was the
distance between the tip of the index finger and the target in horizontal
(X) and vertical (Y) dimensions at the end of the first phase of the
movement.
Data were analyzed from 10 trials before injection and 10 trials
after injection. The 1st 10 trials in which the monkey satisfied the
above specified criteria (i.e., waiting for the target to be in position
before beginning reach, reaching with right arm) were chosen for
analysis. For comparisons between control and muscimol within each
injection, the kinematic data were analyzed using unpaired t-tests. If
the peak tangential wrist velocity after muscimol injection was significantly different from control, the data were included for further
analysis across injections using paired t-tests. Unless otherwise indicated, the results are expressed as means 6 SD.
2051
2052
K. K. WENGER, K. L. MUSCH, AND J. W. MINK
above for maximum vertical wrist position during the reach
and reflect a slight impairment of the reach path.
Retrieval phase of movement
For injections that caused slowing of the reach to target,
retrieval speed was normal after most injections (n 5 11), but
was slow after others (n 5 4). We grouped the data into two
categories for further analysis: injections causing slow reach
but normal retrieval, and injections causing slow reach and
slow retrieval (Fig. 5).
Injections causing slow reach but normal retrieval
Inactivation in GPi did not cause slowing of peak tangential
wrist velocity during the retrieval of the apple bit target for five
injections in monkey T and for six injections in monkey L. In
fact, the velocity of retrieval was slightly faster on average
after muscimol injection in monkey L. Figure 2 shows the
kinematics for a injection with slow reaching but normal retrieval. Averaged across those injections in monkey T, peak
wrist velocity was 61 6 3 cm/s before and 57 6 3 cm/s after
muscimol (NS) and in monkey L, peak wrist velocity was
slightly increased from 100 6 12 cm/s before to 106 6 9 cm/s
after muscimol (t 5 22.94, P , 0.05). Injections that produced
no slowing of tangential wrist velocity during retrieval likewise
did not cause significant slowing of elbow or shoulder angular
velocities in either monkey.
Injections causing slow reach and slow retrieval
Inactivation in GPi caused slowing of mean peak wrist
velocity during the retrieval phase at four injection sites in
monkey T, but none in monkey L. Figure 6 shows the kinematics for an injection that caused slow reaching and slow retrieval. Mean peak wrist velocity was decreased from 69 6 10
cm/s to 56 6 3 cm/s (t 5 3.78, P , 0.05). The slowing was
accompanied by decreased peak angular velocities at both
elbow and shoulder. Averaged across the four injections, elbow
angular velocity decreased from 348 6 32 deg/s to 294 6 10
deg/s (214.7%; t 5 4.67, P , 0.05), and shoulder angular
velocity decreased from 189 6 11 deg/s to 167 6 7 deg/s
(210.8%; t 5 3.80, P , 0.05). The slowing of retrieval after
Downloaded from http://jn.physiology.org/ by 10.220.33.1 on June 16, 2017
FIG. 2. Movement kinematics before and after muscimol for an injection (injection 7, monkey T) causing slow reaching but
normal retrieval. Individual graphs are as follows: wrist position (A and G), tangential wrist velocity (D and J), elbow angle (E and
H), elbow angular velocity (E and K), shoulder angle (C and I), and shoulder angular velocity (F and L). Data are shown for the
last 5 of 10 trials analyzed in each condition. z z z, data before injection of muscimol; —, data after injection. For the position traces,
the target is located at (0,0). Data are time-aligned on the start of wrist movement and displayed from start to end of movement
(see METHODS). In B, the time between the 2 arrows delimits (for control trials) initial elbow flexion. After muscimol, there is
decreased wrist velocity, elbow angular velocity, and shoulder angular velocity during the reach, but no impairment during the
retrieval.
IMPAIRED REACHING AFTER GLOBUS PALLIDUS INACTIVATION
TABLE
1.
Peak wrist velocity and movement time before and after GPi inactivation with muscimol
Peak Wrist Velocity, cm/s
Injection
2053
Control
Muscimol
Movement Time, ms
% Change
Control
Muscimol
% Change
402 6 30
485 6 40
482 6 34
448 6 24
420 6 38
390 6 22
457 6 74
430 6 22
397 6 13
434 6 36
468 6 34
688 6 156
720 6 153
535 6 53
492 6 113
535 6 74
495 6 45
683 6 60
443 6 23
562 6 106
14.2
29.6
33.1
16.2
14.5
27.1
7.7
37.1
10.5
21.1
433 6 36
460 6 28
468 6 29
449 6 68
423 6 18
409 6 23
440 6 23
577 6 85
513 6 137
655 6 85
572 6 109
423 6 18
430 6 28
528 6 91
24.9
10.4
28.5
21.5
0
5.0
15.1
Monkey T
92 6 4
79 6 7
77 6 5
88 6 10
87 6 3
92 6 5
92 6 9
86 6 4
91 6 5
87 6 6
3
5
7
9
10
12‡
13‡
14‡
15‡
Mean
79 6 4*
62 6 9*
54 6 9*
76 6 7†
79 6 8†
82 6 7*
79 6 10†
69 6 6*
82 6 5*
74 6 10*
214.8
221.3
230.0
212.7
210.1
211.6
214.0
220.3
29.6
216.0
Monkey L
84 6 9§
74 6 12*
77 6 7*
84 6 15§
95 6 4*
97 6 8
85 6 9†
212.8
227.9
216.8
214.6
27.4
27.3
214.5
Values are means 6 SD. GPi, globus pallidus pars interna.
* Differences are significant at P , 0.001. † Differences are significant at P , 0.01. ‡ Slow retrievals. § Differences are significant at P , 0.05.
some injections but not others could not be attributed to a more
severe movement deficit during the reach to target. Peak velocity during the reach to target was not significantly different
for the two groups (normal retrievals vs. slow retrievals).
To further explore the differential effect of GPi inactivation
on elbow angular velocity in normal as opposed to slow retrievals, we examined the elbow flexion that occurred at the
beginning of the reach (see Fig. 2) to see whether injections
would also affect the angular velocity of that initial elbow
TABLE
2.
flexion (Fig. 7). Injections in monkey T causing slow retrievals
also caused slowing of elbow flexion angular velocity at the
beginning of the reach from 120 6 11 deg/s to 99 6 13 deg/s
(t 5 25.70, P , 0.05). Injections that produced normal retrievals did not cause slowing of the initial elbow flexion
during the reach, despite significant slowing of the elbow
extension. In monkey T, angular velocity of the initial elbow
flexion was 118 6 19 deg/s in control measurements and
107 6 33 deg/s after muscimol (NS). In monkey L, angular
Kinematics of the elbow and shoulder before and after GPi inactivation with muscimol
Peak Elbow Angular Velocity, deg/s
Injection
Control
Muscimol
Peak Shoulder Angular Velocity, deg/s
% Change
Control
Muscimol
% Change
362 6 24
291 6 29
318 6 28
343 6 17
323 6 18
329 6 15
317 6 28
323 6 22
364 6 22
330 6 23
306 6 17*
249 6 31†
206 6 40*
290 6 20*
261 6 29*
287 6 23*
292 6 33
250 6 19*
320 6 17*
273 6 35*
215.3
213.1
235
215.1
219.3
212.7
27.4
222.1
212
216.9
282 6 39
290 6 16
286 6 13
304 6 18
330 6 10
319 6 11
302 6 20
239 6 24†
224 6 39*
222 6 20*
251 6 41†
295 6 12*
288 6 15*
253 6 32*
215.4
222.8
222.3
217.2
210.6
29.9
216.4
Monkey T
3
5
7
9
10
12
13
14
15
Mean
291 6 26
283 6 34
223 6 5
263 6 24
230 6 16
244 6 18
259 6 25
253 6 23
322 6 17
263 6 31
218 6 14*
181 6 34*
128 6 29*
188 6 25*
147 6 14*
166 6 26*
209 6 30*
152 6 18*
250 6 19*
182 6 39*
224.4
222.4
242.3
227.4
235.9
231.7
218.9
239.4
222.2
229.4
Monkey L
2
9
12
15
17
18
Mean
296 6 11
249 6 16
272 6 19
293 6 29
336 6 16
328 6 35
296 6 33
234 6 24*
184 6 31*
200 6 24*
241 6 33†
296 6 21*
287 6 17†
240 6 45*
220.9
226.1
226.6
217.8
211.9
212.7
219.3
Values are means 6 SD. GPi, globus pallidus pars interna.
* Differences are significant at P , 0.001. † Differences are significant at P , 0.01.
Downloaded from http://jn.physiology.org/ by 10.220.33.1 on June 16, 2017
96 6 14
103 6 5
93 6 4
99 6 7
103 6 3
105 6 3
100 6 5
2
9
12
15
17
18
Mean
2054
K. K. WENGER, K. L. MUSCH, AND J. W. MINK
velocity of the initial elbow flexion was 86 6 14 deg/s in
control measurements and 75 6 17 deg/s after muscimol (NS).
Thus when retrieval was normal, elbow flexion angular velocity was normal in both the reach and retrieval, despite marked
slowing of elbow extension angular velocity during the reach.
Timing of grasp components
Inactivation in GPi with muscimol caused an increased
latency from the end of the reach to the start of retrieval for five
injections in monkey T and for three injections in monkey L
(Fig. 8). Averaged across these injections in monkey T, the
latency from the end of the reach to the start of retrieval
increased from 163 6 32 ms to 488 6 153 ms (t 5 24.77, P ,
FIG. 4. Index finger position at the end of the reach, before any corrective
submovements. Data from all injections that caused slowing of the reach in
both monkeys are included. The point (0,0) is the center of the target,
represented by the 1 3 1 cm box. Arrow indicates the direction of the reach.
■, data from preinjection controls; E, data after inactivation with muscimol.
Large open square represents the average final position in controls; the open
circle represents the average final position after muscimol. After GPi inactivation, the endpoint position of the index finger is higher in the vertical plane.
FIG. 5. Effect of GPi inactivation on peak tangential wrist velocity during
retrieval. A: average peak velocity 6SE from the 5 injections in monkey T that
caused no impairment of retrieval. B: average peak velocity 6SE from the 4
injections in monkey T that caused slowing of retrieval. *P , 0.05
0.01). In monkey L, the latency from the end of reach to the
start of retrieval increased from 51 6 48 ms to 167 6 101 ms,
but did not reach significance due to the small N (P 5 0.06).
We divided the grasp period into two components for further
analysis: the latency from the end of the reach to completion of
the grasp, and the latency from the completion of grasp to the
start of retrieval. GPi inactivation prolonged the latency from
end of reach to grasp substantially more than the latency from
grasp to start of retrieval in monkey T, but not in monkey L. In
monkey T, the time from end of reach to grasp increased from
61 6 26 ms (37% of total time) to 303 6 103 ms (62% of total
time) (t 5 24.64, P , 0.01), whereas the time from grasp to
the start of retrieval increased from 102 6 27 ms (63% of total
time) to 184 6 60 ms (38% of total time) (t 5 23.42, P ,
0.05). In monkey L, time from the end of reach to grasp
increased from 41 6 8 ms (81% of total time) to 106 6 75 ms
(63% of total time), whereas the time from grasp to the start of
return increased from 9 6 40 ms (19% of total time) to 61 6
30 ms (37% of total time).
GPi injections sites
All injections in both monkeys were in the posteroventral
portion of GPi (Fig. 9) and were placed at sites where previous
Downloaded from http://jn.physiology.org/ by 10.220.33.1 on June 16, 2017
FIG. 3. Effect of globus pallidus pars interna (GPi) inactivation on peak
tangential wrist velocity during reaching. Bars represent the average of peak
velocity 6SE after the 9 injections in monkey T and the 6 injections in monkey
L that caused slowing of movement. **P , 0.01; ***P , 0.001
IMPAIRED REACHING AFTER GLOBUS PALLIDUS INACTIVATION
2055
recording had demonstrated activity related to movements of the
right upper extremity. In monkey T, injections that slowed both the
reach and retrieval were located, on average, anterior to those that
slowed the reach but not the retrieval. As described above, no
injection in monkey L slowed the retrieval. There was no simple
topographic difference between injections that impaired grasp and
those that did not. Injections that did not cause slowing during the
reach to target did not have consistent significant effects on
retrieval. There was no impairment of the grasp in any injection
that did not cause slowing of the reach.
DISCUSSION
Inactivation at GPi sites with muscimol produced movement
deficits in a reach-grasp-retrieve task that can be summarized
as follows: 1) decreased peak wrist velocity during the reach to
target; 2) decreased elbow and shoulder angular velocities,
with elbow angular velocity relatively more impaired than
shoulder angular velocity; resulting in 3) higher maximum
vertical wrist and index finger positions at the apex of the
reach; 4) prolonged latency from the end of the reach to the
completion of grasp, and 5) less impairment of retrieval than
reach, with inactivation at the majority of sites causing no
impairment despite slow reaching.
Kinematics of reach and retrieval
Reaching was slowed after GPi inactivation, compared with
preinjection control. Peak tangential wrist velocity, elbow angular velocity, and shoulder angular velocity were slowed, with
elbow velocity relatively slower than shoulder angular velocity. This disproportionate slowing of elbow extension relative
to shoulder flexion produced a more curved reach path. Further, the position of the index finger at the end of the reach was
significantly higher in the vertical plane Thus after muscimol
injection into GPi, there was a flexor bias at the elbow that
affected both movement velocity and endpoint position.
The muscimol injections that slowed reaching were subdivided into two groups based on peak velocity of retrieval: slow
retrieval and normal retrieval. Inactivation at the majority of
Downloaded from http://jn.physiology.org/ by 10.220.33.1 on June 16, 2017
FIG. 6. Movement kinematics before and after muscimol for an injection (injection 14, monkey T) causing both slow reaching
and slow retrieval. Individual graphs are as follows: wrist position (A and G), tangential wrist velocity (D and J), elbow angle (E
and H), elbow angular velocity (E and K), shoulder angle (C and I), and shoulder angular velocity (F and L). Conventions are the
same as in Fig. 2. After muscimol, there is markedly decreased wrist velocity, elbow angular velocity, and shoulder angular velocity
during the reach, with slight decreases during the retrieval. Note that the wrist path during the reach to target is more curved and
reaches a higher vertical position after muscimol.
2056
K. K. WENGER, K. L. MUSCH, AND J. W. MINK
GPi sites had no effect on retrieval speed despite significant
slowing of the reach. For those sites where muscimol did cause
slowing of retrieval, the reach and retrieval were slowed to a
similar degree. In our paradigm, there was a slight elbow
flexion at the beginning of the reach as the monkey lifted its
hand from the initial position. For those injections that did not
impair the retrieval, the velocity of the initial elbow flexion
were also unimpaired. Thus for most injections, extension was
impaired, but not flexion.
Slowing of movement after GPi inactivation
FIG. 8. Timing of phases during the transition from reach to retrieval. The
latency from end of reach to start of retrieval was divided into 2 components:
end of reach to grasp, and grasp to start of retrieval. A: averages 6SE from the
5 injections in monkey T that prolonged grasp latency. B: averages 6SE from
the 3 injections in monkey L that prolonged grasp latency. *P , 0.05; **P ,
0.01.
FIG. 7. Average peak elbow angular velocity during reach and retrieval.
Bars represent mean 6 SE of elbow angular velocity during the elbow flexion
and elbow extension phases of the reach and elbow flexion during retrieval (see
Fig. 2 for kinematics). A: data from injections causing slow reaching but
normal retrievals in monkey T. B: data from injections causing slow reaching
and slow retrieval in monkey T. *P , 0.05; **P , 0.01; ***P , 0.001.
and Thach 1991) or cooling (Hore and Vilis 1980) impaired
both visually guided and self-paced movements. In reaching
tasks, kainic acid lesions, cooling, or muscimol inactivation of
GP prolonged movement time but had no effect on reaction
time (Horak and Anderson 1984a; Inase et al. 1996; Trouche et
al. 1979). Thus similar deficits have been reported for singlejoint and multijoint movements.
A criticism of previous studies has been that several of the
lesions were large and involved GPe, and thus it could not be
concluded whether damage to GPe or GPi, or both, caused the
movement deficits. Based in part on the model of Alexander
and Crutcher (1990) and DeLong (1990), in part on the apparent lack of deficits after GP lesions in older nonquantitative
studies (reviewed in DeLong and Georgopoulos 1981), and in
part on the observation that GPi lesions improve movements in
Parkinson’s disease (Baron et al. 1996; Laitinen et al. 1992;
Lang et al. 1997), it has been suggested that lesions restricted
to GPi should not impair voluntary movement (Wichmann and
DeLong 1997; Wichmann et al. 1995). However, lesions involving just GPi had effects similar to lesions involving both
GPe and GPi (Mink and Thach 1991). In the present study, our
goal was to selectively inactivate relatively small groups of
Downloaded from http://jn.physiology.org/ by 10.220.33.1 on June 16, 2017
Both slowing of movement and a flexor bias after GPi
inactivation or ablation are consistent with previous work on
movement deficits after pallidal lesions. Most prior studies
described slowing of movement with a flexor positional bias
(DeLong and Coyle 1979; Hore et al. 1977; Hore and Vilis
1980; Inase et al. 1996; Mink and Thach 1991) and little or no
effect on reaction time after cooling, muscimol injection, or
kainic acid lesions. Inactivation of GP with muscimol (Mink
IMPAIRED REACHING AFTER GLOBUS PALLIDUS INACTIVATION
2057
FIG. 9. Site of all 0.5-ml muscimol injections into GPi.
Injection sites were mapped onto drawings of GPe and GPi
from the atlas of Snider and Lee (1961). E, injections that
slowed reaching but not retrieval. ■, injections that slowed
both reaching and retrieval; 3, injections that had no significant effect on the task.
primary deficits and that the slowing does not represent a
compensatory strategy. Despite the task differences between
the present study and that of Inase et al. (1996), the findings are
consistent. This strengthens the conclusion that the described
deficits are directly due to GPi inactivation.
Is reaching impaired after GPi inactivation because of
impaired inhibition of competing posture-holding
mechanisms?
We have hypothesized that the basal ganglia act to facilitate
desired motor patterns and inhibit potentially competing motor
patterns (Mink 1996a; Mink and Thach 1993; Thach et al.
1993). Specifically, it was hypothesized that during a voluntary
movement, a small subset of GPi neurons decrease discharge
(under inhibitory influence from striatum) to remove tonic
inhibition from thalamic and brain stem mechanisms and allow
the desired movement to proceed. Simultaneously and in parallel, the majority of GPi neurons increase discharge (under
excitatory influence from the subthalamic nucleus) to increase
inhibition of mechanisms such as those involved in postureholding and prevent them competing with the desired movement. Potentially competing posture-holding mechanisms were
hypothesized to include tonic neck reflexes, contact righting
and placing mechanisms, long-loop stretch reflexes, and vestibulospinal and reticulospinal mechanisms (Mink 1996a). For
the primate sitting upright, the typical posture is one of flexed
arms, legs, and neck. In that posture, tonic neck and vestibular
mechanisms favor flexion (Magnus 1924), and movement of
any one limb out of the initial position requires inhibition of
those postural mechanisms in addition to activation of mech-
Downloaded from http://jn.physiology.org/ by 10.220.33.1 on June 16, 2017
neurons restricted to GPi by 1) using a small injection volume
of muscimol, 2) identifying the GPi borders physiologically
before injection, and 3) injecting into areas of GPi where
neurons were related to arm movements [in a separate task
(Mink 1996b)].
The present results and those of another recent study of
reaching after GPi inactivation (Inase et al. 1996) indicate that
movement is indeed slow after focal inactivation of GPi with
small injections of muscimol. Muscimol injections restricted to
GPi in monkeys performing a two-dimensional arm movement
in the horizontal plane caused slowing of movement with
cocontraction and a flexor bias (Inase et al. 1996). The task
used in those studies was a center-out planar reaching task to
eight different targets requiring movements of up to ;7.5 cm
and holding of both initial and final positions. The authors
proposed that the decrease in peak movement velocity was
likely the result of an attempt to stabilize the limb against the
flexor drift.
Although the results of the present study are consistent with
those from other studies of reaching or single joint movements
after GPi lesions, there are several distinguishing features of
the present task. Our task required relatively large amplitude
arm movements, comparable to those of Horak and Anderson
(1984a), but much larger than those of Inase et al. (1996).
Furthermore, there was no required hold period and no imposed limitation on movement time or accuracy to obtain the
reward. The only limitation was that the monkey had to grasp
and retrieve the apple bit without dropping it to eat it. Finally,
our paradigm involved reaching without any instrumentation.
Because of the few accuracy requirements in our task, we
suggest that the slowing of movement and flexor bias are
2058
K. K. WENGER, K. L. MUSCH, AND J. W. MINK
overwhelming evidence that GPi lesions do not impair movement initiation or generation (Mink 1996a). Furthermore, the
component of the reach that was relatively most impaired was
elbow extension. During reaching in the vertical plane, the
forces that extend the elbow come primarily from gravity and
interaction torques produced by shoulder movement (Bastian et
al. 1996). Elbow extension is accompanied by gradual reduction of torque produced by the elbow flexor muscles and not by
active force generation by elbow extensors. Thus the impairment of elbow extension is consistent with the findings of Mink
and Thach (1991) that GPi lesions cause greater disability
turning off previously active muscles than turning on muscles.
An alternative hypothesis is that the basal ganglia are involved
in movement sequencing. This hypothesis will be discussed
below.
Comparison to pallidotomy in Parkinson’s disease (PD)
Our results confirm those of many previous studies showing
that GPi lesions cause movement abnormalities in normal
monkeys. However, GPi lesions (pallidotomy) improve many
movement abnormalities in Parkinson’s disease (PD) (Baron et
al. 1996; Laitinen et al. 1992; Lang et al. 1997). Because of the
improved bradykinesia in PD and apparent lack of adverse
effects, it has been argued that GPi lesions cause no long-term
deficits and that deficits seen after GP lesions in monkeys are
due to involvement of GPe or even of the internal capsule.
(Wichmann and DeLong 1997). However, in the present study
and in others where the lesion or inactivation site was restricted
to GPi, the deficits are unmistakable. There may be a fundamental reorganization of basal ganglia circuits in PD that
explains the opposite effects of GPi lesions in PD and in
normal animals.
Prolonged grasp latency
In the present study, GPi inactivation at several sites prolonged the time required to grasp the apple bit as measured by
the latency from the end of reach to the start of retrieval. When
broken into components, the time from end of reach to completion of grasp was relatively more prolonged than the time
from completion of grasp to start of retrieval in monkey T but
not in monkey L. We attribute some of the discrepancy in these
results between the two monkeys T and L to their individual
reaching and grasping styles. Monkey T usually grasped the
apple bit with the thumb and index finger, sometimes using the
third finger also. In contrast, monkey L usually grasped the
apple bit by contacting with the palm and curling all fingers
around the bit. As a result, monkey T’s grasp required more
precision of the fingers than did monkey L’s, and the greater
effect of GPi inactivation in monkey T may have been due to
that factor.
We observed, but were unable to quantify with the present
methodology, a bias toward finger flexion after GPi inactivation.
In the normal reach to grasp in human subjects, it has been shown
that grip aperture is formed relatively early during the reach
(Jeannerod 1984). In our monkeys, we also observed that formation of the grip aperture began early in the reach. However, after
muscimol injection it appeared that the aperture was too small
(i.e., the fingers were too flexed), so that when the target was
reached the monkey had to extend the fingers to widen the
Downloaded from http://jn.physiology.org/ by 10.220.33.1 on June 16, 2017
anisms involved in producing the desired movement. Inability
to inhibit those posture-holding mechanisms for the moving
limb would be expected to slow movement and to bias movement toward that initial position, i.e., flexion.
The data presented here support our hypothesis. Sitting in
the initial posture before reaching, posture-holding mechanisms are active to maintain position of the body. These
mechanisms would include tonic neck reflexes, labyrinthine
reflexes, contact righting responses, optic righting responses,
and possibly others. Movement away from the initial posture
would be opposed by those posture-holding mechanisms, but
movement back toward the initial posture would be assisted.
That is what was seen after the majority of injections into GPi.
However, not all potentially competing mechanisms favor flexion. For example, long-latency stretch reflexes are present in
both flexors and extensors. Thus one might expect that both
flexion and extension would be slowed by disinhibition of
these reflexes. Indeed, after a few injections in GPi in one
monkey, all components of the reach-grasp-retrieval task were
slow.
Injection sites causing slow retrievals were topographically
different from sites that did not affect retrieval. This is consistent with the idea that for a given movement, certain motor
mechanisms must be expressed and others must be inhibited.
By changing the location of injection, these mechanisms may
be affected to different degrees, in some cases to favor flexion
and in others not. Support for topographic representation of
different mechanisms in GPi comes from three lines of evidence. First, there is anatomic evidence for parallel, functionally different circuits through the basal ganglia (see Alexander
et al. 1986 for review) and for separation of GPi neurons that
project via thalamus to different cortical areas (Hoover and
Strick 1993; Middleton and Strick 1994). Second, microstimulation of some sites in GPi slows reaching, but microstimulation of others speeds reaching (Horak and Anderson 1984b).
Third, neuronal discharge patterns indicate a somatotopic organization in GPi (DeLong et al. 1985). In light of these
observations, different effects of inactivation at different sites
in GPi is not surprising. It may seem more surprising that
inactivation at most sites did not impair the retrieval movement
in spite of significant slowing of the reach. However, this was
predicted by the hypothesis (Mink 1996a).
Although we favor the hypothesis of selective facilitation
and inhibition of competing motor patterns, other hypotheses
of basal ganglia function have been proposed and are relevant
to the present study. One such hypothesis is that the basal
ganglia scale the magnitude of muscle activity to produce a
movement of desired amplitude and velocity (Hallett and
Khoshbin 1980). This hypothesis was supported by Horak and
Anderson (1984a) and is inherent in the model of Alexander
and Crutcher (1990). Although the present task did not look at
scaling in a controlled manner, the slow reaching but normal
retrieval after most injections suggests that the deficit is not one
of scaling generally. The finding that elbow extension was
relatively more impaired than was shoulder flexion during the
reach could be interpreted as incorrect scaling of activity in one
muscle to another, but this interpretation would not explain the
similar impairment of elbow and shoulder during the retrievals.
Likewise, it is difficult to interpret the deficits after GPi inactivation as impaired movement generation. Although reaction
time was not measured in the present experiments, there is
IMPAIRED REACHING AFTER GLOBUS PALLIDUS INACTIVATION
aperture before flexing around the apple bit. The prolonged latency from end of reach to grasp appeared to reflect the need to
reextend the fingers from their overly flexed posture before grasping the apple bit. Thus the flexor bias slowed reaching and
impaired the grasp, but had less or no effect on the retrieval that
involved returning to a flexed posture.
Present results do not support a primary role for the basal
ganglia output in the generation of movement sequences
movement sequences and suggest that the two hypotheses are
compatible (Mink 1996a).
The results of this study show that reaching movements are
impaired specifically after focal inactivation of GPi in previously normal monkeys. The nature of the impairment supports
the hypothesis that GPi lesions disrupt the ability to inhibit
competing motor mechanisms to prevent them from interfering
with desired voluntary movement. Further study is required to
specify exactly which motor mechanisms contribute to the
described deficits and how they are represented topographically in GPi.
The authors thank W. T. Thach and A. J. Bastian for helpful advice during
the development of these experiments. We also thank A. J. Bastian for
development of some of the analysis software.
This work was supported by National Institute of Neurological Disorders
and Stroke Grants K08 NS-01808 and T32 NS-07027, The American Parkinson Disease Association, The McDonnell Center for Higher Brain Function,
and the Department of Neurology at Washington University School of Medicine.
Address for reprint requests: J. W. Mink, Dept. of Neurology, Washington
University School of Medicine, Box 8111, 660 S. Euclid Ave., St. Louis, MO
63110.
Received 15 March 1999; accepted in final form 22 June 1999.
REFERENCES
ALDRIDGE, J. AND BERRIDGE, K. Coding of serial order by neostriatal neurons:
a “natural action” approach to movement sequences. J. Neurosci. 18: 2777–
2787, 1998.
ALEXANDER, G. E. AND CRUTCHER, M. D. Functional architecture of basal
ganglia circuits: neural substrates of parallel processing. Trends Neurosci.
13: 266 –271, 1990.
ALEXANDER, G. E., DELONG, M. R., AND STRICK, P. L. Parallel organization of
functionally segregated circuits linking basal ganglia and cortex. Ann. Rev.
Neurosci. 9: 357–381, 1986.
BARON, M. S., VITEK, J. L., BAKAY, R.A.E., GREEN, J., KANEOKE, Y., HASHIMOTO, T., TURNER, R. S., WOODARD, J. L., COLE, S. A., MCDONALD, W. M.,
AND DELONG, M. R. Treatment of advanced Parkinson’s disease by posterior
GPi pallidotomy: 1-year results of a pilot study. Ann. Neurol. 40: 355–366,
1996.
BASTIAN, A. J., MARTIN, T. A., KEATING, J. G., AND THACH, W. T. Cerebellar
ataxia: abnormal control of interaction torques across multiple joints. J. Neurophysiol. 76: 492–509, 1996.
BENECKE, R., ROTHWELL, J. C., DICK, J.P.R., DAY, B. L., AND MARSDEN, C. D.
Performance of simultaneous movements in patients with Parkinson’s disease. Brain 109: 739 –757, 1986.
BENECKE, R., ROTHWELL, J. C., DICK, J.P.R., DAY, B. L., AND MARSDEN, C. D.
Simple and complex movements off and on treatment in patients with
Parkinson’s disease. J. Neurol. Neurosurg. Psychiatry 50: 296 –303, 1987.
BHATIA, K. P. AND MARSDEN, C. D. The behavioural and motor consequences
of focal lesions of the basal ganglia in man. Brain 117: 859 – 876, 1994.
BROTCHIE, P., IANSEK, R., AND HORNE, M. K. Motor function of the monkey
globus pallidus. 2. Cognitive aspects of movement and phasic neuronal
activity. Brain 114: 1685–1702, 1991.
CROMWELL, H. C. AND BERRIDGE, K. C. Implementation of action sequences by
a neostriatal site: a lesion mapping study of grooming syntax. J. Neurosci.
16: 3444 –3458, 1996.
DELONG, M. R. Activity of pallidal neurons during movement. J. Neurophysiol. 34: 414 – 427, 1971.
DELONG, M. R. Primate models of movement disorders of basal ganglia origin.
Trends Neurosci. 13: 281–285, 1990.
DELONG, M. R. AND COYLE, J. T. Globus pallidus lesions in the monkey
produced by kainic acid: histologic and behavioral effects. Appl. Neurophysiol. 42: 95–97, 1979.
DELONG, M. R., CRUTCHER, M. D., AND GEORGOPOULOS, A. P. Primate globus
pallidus and subthalamic nucleus: functional organization. J. Neurophysiol.
53: 530 –543, 1985.
DELONG, M. R. AND GEORGOPOULOS, A. P. Motor functions of the basal
ganglia. In: Handbook of Physiology. The Nervous System. Motor Control.
Bethesda, MD: Am. Physiol. Soc., 1981, sect. 1, vol. 2, p. 1017–1062.
Downloaded from http://jn.physiology.org/ by 10.220.33.1 on June 16, 2017
A hypothesized role of the basal ganglia is that they contribute
to the automatic execution of movement sequences (Bhatia and
Marsden 1994; Brotchie et al. 1991; Houk et al. 1995; Marsden
1987). Specifically, it has been proposed that the basal ganglia
contribute to the generation of the second and later components of
a movement sequence, but not to the initial component (Brotchie
et al. 1991; Marsden 1987). Supporting this hypothesis is the
characteristic micrographia of PD where initial letters of a written
sentence are substantially larger than subsequent letters. Some GP
neurons change activity after the onset of the first movement in a
sequence of two movements, but before the second (Brotchie et al.
1991) and some neurons in striatum and pallidum has been shown
to correlate with movement sequence (Aldridge and Berridge
1998; Mushiake and Strick 1995). People with PD are more
impaired when performing simultaneous or sequential movement
than when performing each component separately (Benecke et al.
1986, 1987; Marsden and Obeso 1994). Furthermore, inactivation
in putamen can disrupt the learning or execution of movement
sequences (Miyachi et al. 1997). However, it should be noted that
simple (nonsequential) movements are also impaired in PD
(Evarts et al. 1981; Hallett and Khoshbin 1980) and after striatal
or pallidal lesions (Kato and Kimura 1992; Mink and Thach
1991).
Several aspects of our results are not consistent with a
specific basal ganglia role in the programming of movement
sequence components subsequent to the initial one. First, the
retrieval phase was never more impaired than the reach phase
and in most cases was normal. If programming of latter sequence components was a primary function of the basal ganglia, one would expect the latter components of the sequence to
be more impaired than the earlier ones. The greater prolongation of the latency from the end of reach to grasp compared
with the latency from the grasp to start of retrieval in monkey
T also argues against a specific deficit of sequence generation,
per se. Studies of reach to grasp movements in normal human
subjects suggest that the reach and grasp are programmed as
one pattern rather than simultaneous execution of two separate
patterns (Jeannerod 1984). Thus the movements involved in
our task may be viewed as a sequence of two rather than three
movements. Nonetheless, the latter component of the sequence
was less affected than was the former. Although the present
results do not support a primary role of the basal ganglia output
in the programming of movement sequences, they do not
exclude the basal ganglia as an important contributor to the
control of movement sequences. The areas of GPi that were
inactivated in the present study are in the “motor circuit” of the
basal ganglia, and injections in more anterior portions may
have disrupted sequencing. Furthermore, basal ganglia may
play a role in some aspects of sequence control that were not
required in our task. We favor the idea that dynamic inhibition
of competing motor patterns is a critical feature in many
2059
2060
K. K. WENGER, K. L. MUSCH, AND J. W. MINK
MIDDLETON, F. A. AND STRICK, P. L. Anatomical evidence for cerebellar and
basal ganglia involvement in higher cognitive function. Science 266: 458 –
461, 1994.
MINK, J., KARPITSKAYA, Y., WENGER, K., AND FLEKAL, J. The pallidotomy
paradox: different motor effects of globus pallidus lesions in normal and
parkinson’s disease brains. Soc. Neurosci. Abstr. 24: 409, 1998.
MINK, J. W. The basal ganglia: Focused selection and inhibition of competing
motor programs. Prog. Neurobiol. 50: 381– 425, 1996a.
MINK, J. W. Slowing of wrist movement and impaired ability to turn off active
muscles following inactivation of globus pallidus pars interna (GPi): localization of effective sites. Soc. Neurosci. Abstr. 22: 415, 1996b.
MINK, J. W. AND THACH, W. T. Basal ganglia motor control. III. Pallidal
ablation: normal reaction time, muscle cocontraction, and slow movement.
J. Neurophysiol. 65: 330 –351, 1991.
MINK, J. W. AND THACH, W. T. Basal ganglia intrinsic circuits and their role
in behavior. Curr. Opin. Neurobiol. 3: 950 –957, 1993.
MIYACHI, S., HIKOSAKA, O., MIYASHITA, K., KARADI, Z., AND RAND, M. K.
Differential roles of monkey striatum in learning of sequential hand movement. Exp. Brain Res. 115: 1–5, 1997.
MUSHIAKE, H. AND STRICK, P. L. Pallidal neuron activity during sequential arm
movements. J. Neurophysiol. 74: 2754 –2758, 1995.
SNIDER, R. S. AND LEE, J. C. A stereotaxic atlas of the monkey brain (Macaca
mulatta). Chicago, IL: Univ. of Chicago Press, 1961.
THACH, W. T., BASTIAN, A. J., AND MINK, J. W. Impairment of reaching after
muscimol inactivation of globus pallidus pars interna in the monkey. Soc.
Neurosci. Abstr. 22: 1085, 1996.
THACH, W. T., MINK, J. W., GOODKIN, H. P., AND KEATING, J. G. Combining
versus gating motor programs: differential roles for cerebellum and basal
ganglia. In: Role of the Cerebellum and Basal Ganglia in Voluntary Movement, edited by N. Mano, I. Hamada, and M. R. DeLong. Amsterdam:
Elsevier, 1993, p. 235–245.
TROUCHE, E., BEAUBATON, D., AMATO, G., LEGALLET, E., AND ZENATTI, A. The
role of the internal pallidal segment on the execution of a goal directed
movement. Brain Res. 175: 362–365, 1979.
WICHMANN, T. AND DELONG, M. R. Physiology of the basal ganglia and
pathophysiology of movement disorders of basal ganglia origin. In: Movement Disorders: Neurologic Principles and Practice, edited by R. Watts and
W. Koller. New York: McGraw-Hill, 1997, p. 87–97.
WICHMANN, T., VITEK, J. L., AND DELONG, M. R. Parkinson’s disease and the
basal ganglia: lessons from the laboratory and from neurosurgery. The
Neuroscientist 1: 236 –244, 1995.
Downloaded from http://jn.physiology.org/ by 10.220.33.1 on June 16, 2017
DENNY-BROWN, D. The fundamental organization of motor behavior. In: Neurophysiological Basis of Normal and Abnormal Motor Activities, edited by
M. Yahr and D. Purpura. New York: Raven, 1967, p. 415– 442.
EVARTS, E. V., TERAVAINEN, H., AND CALNE, D. B. Reaction time in Parkinson’s disease. Brain 104: 167–186, 1981.
HALLETT, M. AND KHOSHBIN, S. A physiological mechanism of bradykinesia.
Brain 103: 301–314, 1980.
HOOVER, J. E. AND STRICK, P. L. Multiple output channels in the basal ganglia.
Science 259: 819 – 821, 1993.
HORAK, F. B. AND ANDERSON, M. E. Influence of globus pallidus on arm
movements in monkeys. I. Effects of kainic acid-induced lesions. J. Neurophysiol. 52: 290 –304, 1984a.
HORAK, F. B. AND ANDERSON, M. E. Influence of globus pallidus on arm
movements in monkeys. II. Effects of stimulation. J. Neurophysiol. 50:
305–322, 1984b.
HORE, J., MEYER-LOHMANN, J., AND BROOKS, V. B. Basal ganglia cooling
disables learned arm movements of monkeys in the absence of visual
guidance. Science 195: 584 –586, 1977.
HORE, J. AND VILIS, T. Arm movement performance during reversible basal
ganglia lesions in the monkey. Exp. Brain Res. 39: 217–228, 1980.
HOUK, J. C., DAVIS, J. L., AND BEISER, D. G. Models of information processing
in the basal ganglia. In: Computational Neuroscience, edited by T. J.
Sejnowski and T. A. Poggio. Cambridge, MA: MIT Press, 1995.
INASE, M., BUFORD, J. A., AND ANDERSON, M. E. 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.
JEANNEROD, M. The timing of natural prehension movements. J. Mot. Behav.
16: 235–254, 1984.
KATO, M. AND KIMURA, M. Effects of reversible blockade of basal ganglia on
a voluntary arm movement. J. Neurophysiol. 68: 1516 –1534, 1992.
LAITINEN, L. V., BERGENHEIM, A. T., AND HARIZ, M. I. Leksell’s posteroventral
pallidotomy in the treatment of Parkinson’s disease. J. Neurosurg. 76:
53– 61, 1992.
LANG, A. E., LOZANO, A. M., MONTGOMERY, E., DUFF, J., TASKER, R., AND
HUTCHINSON, W. Posteroventral medial pallidotomy in advanced Parkinson’s disease. N. Engl. J. Med. 337: 1036 –1042, 1997.
MAGNUS, R. Korperstellung. Berlin: Julius Springer, 1924.
MARSDEN, C. D. What do the basal ganglia tell premotor cortical areas? Ciba
Found. Symp. 132: 282–300, 1987.
MARSDEN, C. D. AND OBESO, J. A. The functions of the basal ganglia and the
paradox of stereotaxic surgery in Parkinson’s disease. Brain 117: 877– 897,
1994.