| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
J Neurophysiol (February 1, 2003). 10.1152/jn.00114.2002
Submitted on Submitted 15 February 2002; accepted in final form 15 October 2002
Department of Anatomy and Neurobiology, Washington University School of Medicine, St. Louis, Missouri 63110
 |
ABSTRACT |
|---|
 TOP
 ABSTRACT
 INTRODUCTION
METHODS
RESULTS
DISCUSSION
APPENDIX
REFERENCES
|
|---|
Goodkin, H. P. and W. T. Thach. Cerebellar Control of Constrained and Unconstrained Movements. I. Nuclear Inactivation. J. Neurophysiol. 89: 884-895, 2003. The aim of this study was to determine in monkeys if inactivation of dentate and lateral interposed deep cerebellar nuclei preferentially impairs certain movements relative to others. Constrained movements of the digits were trained with digits, hand, and elbow constrained in a cast. Simple movements were flexion of Thumb or Index. A compound movement was simultaneous flexion of Thumb+Index. An unconstrained movement consisted of a reach to, pinch of, and retrieval of a small food reward (Reach+Pinch). In two monkeys we mapped the dentate and interpositus with 66 injections of muscimol (3 µl of 5 µg/µl). Thirty-two percent of the injections resulted in increased reaction times of Thumb, Index, and Thumb+Index (mean = 24, 24, 28 + 26, respectively). Fifty percent of the injections impaired Reach+Pinch, producing target overshoot, curved reach trajectory, missed target (X and Y errors), and clumsy pinch with dropped fruit bits. Inactivation impaired each and all of Thumb, Index, Thumb+Index, and Reach+Pinch in 27%, only Reach+Pinch in 23%, and only Thumb, Index, Thumb+Index in 5% of injections. In sum, all types of movement were impaired. Thumb+Index was no more impaired than Thumb or Index alone, suggesting that the lateral cerebellar nuclei are not specifically required for combining movements of the two digits when constrained. Reach+Pinch appeared so greatly impaired and Thumb, Index, Thumb+Index so little as to be consistent with the hypothesis that a principal role of the cerebellum is to control those interactions that occur between body segments in natural unconstrained movements. However, the fact that all movements were impaired shows that the cerebellum contributes to the control of all movements.
| |
INTRODUCTION |
|---|
|
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
APPENDIX
REFERENCES
|
|---|
"Simple"
movements, caused by relatively few muscles acting at one or a few
joints, may have an increased reaction time after cerebellar lesions
(Holmes 1917
; Meyer-Lohmann et al. 1977).
"Compound" movements, caused by more muscles acting over more
joints, have seemed disproportionately impaired after cerebellar
lesions, showing the classic errors of curved trajectory, overshoot of
target with vertical and lateral misses, and trial-to-trial variability
(Gilman et al. 1976; Holmes 1939). The
argument has been made that the greater impairment of compound
movements may result from summation of the effects of the greater
number of moving parts (Holmes 1939).
Another argument is that since compound movements appear more severely
impaired than simple movements, compound movements may be actively
combined from simple movement components by the cerebellum. To
distinguish it from the "summation" hypothesis, this has been
called the "combination" hypothesis (Babinski 1899; Fluorens 1824; cf. Goodkin et al. 1993;
Thach et al. 1992).
Prior studies have compared constrained simple movements with
unconstrained compound movements. The question thus remains as to
whether the greater deficits seen in the unconstrained compound movements are more a function of their being unconstrained than of
their being compound. Bernstein (1967) discussed the
problem of dealing with the added complexity of dynamic interactions
generated by one moving segment on another in a free multi-segment
movement. Bastian et al. (1996), Cooper et al.
(2000), and Topka et al. (1998) have shown that
the cerebellum may control dynamic interactions generated in
multi-segment unconstrained limb movements. This gives rise to a third
hypothesis of cerebellar control of movement coordination, an
"interaction" hypothesis.
To distinguish among these three hypotheses, we have studied the effects of cerebellar nuclear inactivation on movements that were constrained, unconstrained, simple, and compound. We trained Rhesus monkeys (Macaca mulatta) to perform constrained simple (Thumb or Index) and constrained compound (Thumb+Index) digit movements for a water or food reward. No interaction was required between digits in the constrained Thumb+Index task. An unconstrained compound movement consisted of a natural reach to, precision pinch of, and retrieval of fruit bits (Reach+Pinch). Interaction occurred between arm and digit segments in this task. Inactivation of portions of the dentate and interposed nuclei was produced by mapping them with microinjections of the GABA agonist, muscimol. The deficits in Thumb, Index, Thumb+Index, and Reach+Pinch were then compared to test the three hypotheses.
| |
METHODS |
|---|
|
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
APPENDIX
REFERENCES
|
|---|
All aspects of this study were approved by the Internal Review Board of Washington University School of Medicine and performed in compliance with Animal Welfare guidelines.
Behavioral paradigms
CONSTRAINED THUMB, INDEX AND THUMB+INDEX TASKS.
Two monkeys (M. mulatta) sat in a primate chair that
constrained movement of the neck, trunk, shoulder, and elbow (Fig.
1A). The right hand was held
in a Plexiglas glove fixed to the chair with all digits constrained
(Schieber 1991). Thumb and index were held so as to
approximate the configuration of a pinch. The ventral surface of the
apical pad of each of thumb and index rested on the ends of lever arms
(25 mm) of snap-action microswitches that were closed by flexion of a
30 g force exerted over a range of 3 mm. Force was measured
using strain gauges mounted on the microswitch lever arms. Each gauge
was in a separate bridge circuit whose output was amplified, low-pass
filtered (5 KHz), biased, and sampled at 100 Hz using a Cambridge
Electronic Design (CED, UK) 1401 interface.
|
UNTRAINED UNCONSTRAINED COMPOUND MOVEMENT TASKS. Reaching for, pinching, and retrieving fruit bits were evaluated before and after muscimol injections in the following two conditions: 1) To measure overshoot and reach path trajectory, the monkey stood in the open room and reached for, pinched, and retrieved fruit bits held between thumb and forefinger of the hand of one of the investigators and was videotaped from the side; 2) To measure x-y accuracy and the incidence of dropped fruit, the monkey was seated in the chair and the freed arm reached to, pinched, and retrieved fruit bits impaled on a pin in a food well and was videotaped from behind. Wrist, elbow, and shoulder were marked using white nontoxic bond correction fluid. A Peak Performance Technologies Inc. system was used to digitize videotaped reaches (60 fields/s).
As the monkey stood in the open room and reached for fruit bits, the reach was videotaped from the right side perpendicular to the direction of the reach, before and after injections.OVERSHOOT (Z-AXIS ERROR). Overshoot was defined as the digit tips and/or palm striking the fruit or the investigator's hand and was subsequently rated independently by two observers viewing videotapes. Reaches in which no visible overshoot occurred were counted as correct. Those reaches in which a visible overshoot occurred were counted as "incorrect." Concordance was nearly 100%; the rare disputed trials were counted as correct. The number of overshoots before and after injections was compared using the Fisher exact test (CoStat Software, P < 0.05).
A wrist path length ratio was used as the measure of the straightness of the wrist path. The wrist path length ratio was defined as the ratio of the actual path length (distance traveled by the wrist during the initial segment of the reach prior to any corrective movement or terminal oscillations) to the ideal path length (straight line distance from the position of the wrist at the beginning of the reach to the position at the end of the initial segment of the reach). Wrist path length ratios before and after cerebellar injections were compared using the Mann-Whitney nonparametric test (P < 0.05). As the monkeys sat in the primate chair used for the constrained digit tasks, in front of the monkey a deep, narrow food well was mounted at shoulder height less than an arm's length away (Fig. 1C). A fruit reward was placed on a cantilever pin centered in the middle of the food well. Strain gauges attached to the cantilever detected when the monkey first touched the fruit reward. The monkey had to reach 20-25 cm to pinch and retrieve the fruit bit. The well's depth was the length of the opposed thumb and index digit; its width was twice that of the monkey's thumb, encouraging the monkey to use a precision pinch strategy.REACH ENDPOINT X-Y-AXIS ERROR AND DROPPED FRUIT. Reach+Pinch was videotaped from behind the seated monkey, its head held parallel to the direction of the reach. Each trial was searched field-by-field to determine the tangential (x-y) endpoint of the initial segment of the reach prior to any corrective movements or terminal tremor. Once this field was found, the position of the metacarpophalangeal (mcp) joint of the index digit, the corners of the food well, and the target (fruit bit) were digitized. Differences (P < 0.05) in variability of the tangential xy position of the index digit's metacarpophalangeal joint at the end of the initial segment of reach before and after muscimol were determined by analysis of variance (ANOVA).
Trials of the Reach+Pinch were also rated as to whether the fruit was successfully placed in the monkey's mouth or was dropped prior to placement. Differences (P < 0.05) in the numbers of successful and unsuccessful (dropped fruit) trials before and after muscimol injection were compared using the Fisher exact text (CoStat Software).RANGE AND FORCE OF MOVEMENTS. In Thumb, Index, and Thumb+Index, the range of movement was 3 mm, and the torques were estimated to be 0.007 Nm for thumb and 0.013 for index at mcp joints (see APPENDIX). In Reach+Pinch, the range of movement was 20-25 cm, and the torque was estimated to be 0.87 Nm at the shoulder (see APPENDIX).
Task control and data acquisition
A Rockwell AIM 65 microprocessor was programmed to control the LED instructional display, monitor the thumb and index digit microswitches (open or closed), and dispense the fruit drink or water reward for correct trials. The CED 1401 interface and Spike2 Data Capture software controlled all aspects of data acquisition. Spike train data [single unit and voltage-to-frequency converted electromyogram (EMG)], digitized analog signals (strain gauge and EMG), and event codes (generated by the AIM microprocessor), were buffered by the CED 1401 and then sent to a PC for permanent storage. Data analysis was also performed on the PC using the Spike2 Data Analysis software (CED) and custom-designed raster analysis software.
Surgery
Ketamine HCl 10-25 mg/kg im was given with atropine sulfate 0.04 mg/kg im prior to surgical anesthesia. Thiopental sodium 12.5 mg/kg iv was then given to effect and maintained throughout the procedure. The monkeys were placed into a stereotaxic apparatus; the scalp and temporalis muscles were reflected, and the periosteum removed. Six stainless steel screws were placed in holes drilled into the calvarium. A cap of dental acrylic that covered the exposed bone was molded over the screws. Bolts placed in the rostral and caudal borders of the cap attached to the primate chair to restrict head movements.
One month later, when performance with the head restrained returned to
the preoperative level, a second procedure under anesthesia was
performed. A 16-mm hole, stereotaxically centered at P7 L10 (Winters and Kado 1969), was trephined through the cap
and the underlying skull. The dura mater was left intact. A Lucite
chamber was mounted in the acrylic over the trephined hole. The chamber was filled with isotonic saline and covered with a Lucite cap. The
chamber was cleansed daily with 3% hydrogen peroxide.
After recording EMG and single-unit activity in the deep cerebellar nuclei (see companion paper), the GABA agonist muscimol was used to study the effect of inactivation on the performance of 1) Thumb, Index, Thumb+ Index, and 2) Reach+Pinch while standing (reach trajectory and z-axis endpoint accuracy) and seated (reach x-y endpoint accuracy and frequency of dropped fruit bits).
Temporary pharmacological focal inactivation of the dentate nucleus with muscimol
Injections of muscimol were made with a 10 µl GlenCo syringe through a 4 in., 26-gauge needle. The syringe was mounted on the same x,y,z coordinate system used previously for single-unit recordings in these monkeys. This allowed for stereotaxic injections of muscimol into the cerebellar dentate nucleus, the boundaries of which had been determined by single-unit recordings. The location of 45 injections into the cerebellar nuclei of monkey 1 and 22 injections in monkey 2 are shown in Fig. 2.
|
Muscimol injection sessions were carried out according to the following
protocol: The monkey was videotaped in the observation room sitting,
standing, and walking. Reaches and pinches of fruit bits held in the
investigator's hand were videotaped from the side to determine
trajectory and z-axis endpoint accuracy. The monkey was then
placed in the primate chair; the head was restrained, and the syringe,
filled with the muscimol solution, was mounted on the chamber. The
syringe needle was advanced to the intended site of the injection.
Unconstrained Reach+Pinch (10 correct trials) was performed and
videotaped from behind to determine x-y axes endpoint
accuracy and number of fruit-bit drops. Thumb, Index, and Thumb+Index
(20 correct trials each) were performed and the output of the
microswitch strain gauges recorded. Muscimol, 3 µl of the 5 µg/µl
in physiological saline (pH 7.3) or 3 µl of sterile saline (control
injection 22 in monkey 2), was injected over a vertical
range of 4 mm (1 µl at 2-mm intervals) over a 15-min interval.
Diffusion for muscimol has been calculated from behavioral effects in
monkey (Thach et al. 1998) and cat (Martin et al.
2000) cerebellar nuclei to be approximately 1 mm. Injection sites were spaced at 2-mm intervals so that the injected material would
meet/overlap and form a continuum. Thus injections would theoretically
inactivate a vertical cylinder of cerebellar nuclear cells 
6 mm deep
(maximum depth of the nucleus) and 2 mm wide. Injections on a boundary
between dentate and interpositus or interpositus and fastigius would
inactivate each for a millimeter to either side. A 15- to 30-min wait
period allowed for muscimol diffusion (during the injection interval
and wait period, the monkey sat in the primate chair with the arm still
placed in the elbow cast and the digits held). A postinjection set of
Thumb, Index, and Thumb+Index (20 correct trials each) was performed
and recorded. Postinjection Reach+Pinch (10 correct trials) was
performed and videotaped. The syringe was withdrawn; the head was
freed, and the monkey was removed from the primate chair and brought
back into the observation room. Postinjection sitting, standing,
walking, reaching, and pinching were performed and videotaped as at the beginning of the session.
At least 1 day intervened between muscimol injections to insure no residual effect of the previous injection.
Histology
After each cerebellum was sectioned and stained with thionin, the injection sites were mapped and drawn on tracings of the deep cerebellar nuclei. The injection sites were identified from the tracks caused by the insertion of the syringe needle into the cerebellum or interpolated when these were not visible.
| |
RESULTS |
|---|
|
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
APPENDIX
REFERENCES
|
|---|
Constrained digit tasks
Constrained digit movements Thumb, Index, Thumb+Index were examined for differences before and after the injections. Performance was assessed for changes in the percentage of correct trials, reaction time, and peak force exerted against the microswitches.
In Fig. 3, the analog outputs of the strain gauges attached to the thumb and index digit microswitches are displayed for correct trials of the constrained digit tasks before and after muscimol injection 10 in the first monkey. Following the injection, the monkey was still able to perform Thumb, Index, Thumb+Index. There was no significant change (G test, P > 0.05, CoStat Software) in the percentage of correct trials after the injection as compared with before.
|
There was an increase in reaction time (P < 0.05 by the Mann-Whitney nonparametric test) for both digits during both single and combined movements after injection 10 in monkey 1 shown in Fig. 3. Increases in mean reaction time were as follows: for the Thumb task, 19.95 ms; for the Index task, 27.27 ms; and for the Thumb+Index(2) task, 22.86 ms for the thumb and 23.35 ms for the index digit.
The largest increases in reaction time across all tasks occurred in the first monkey following injection 15: Thumb, 42.1 ms; Index, 45.7 ms; and Thumb+Index task, 48.5 ms for the thumb and 49.8 ms for the index. For all injections in which there was a significant increase in reaction time, the mean increases are as follows: Thumb, 24.3 ms; Index, 24.2 ms; and Thumb+Index, 27.8 ms for the thumb and 25.9 ms for the index.
Figure 4, top, shows a plot of the increases in reaction time for the thumb during the Thumb+Index against the thumb during the Thumb. Figure 4, bottom, plots the corresponding graph for the index digit. The increases in reaction time for both digits in the Thumb+Index task are positively correlated with the increases during the corresponding single-digit tasks.
|
For two injections
no. 44 in the first monkey and no. 19 in the
secondthere was a significant increase in reaction time measured for
one but not the other tasks. In both cases, only the index digit during
the Index task had increased reaction time (marked by the open circle
in Fig. 5). For all the other 22 injections that produced an increase in reaction time, the increase was
for both digits during all the tasks (marked by the filled diamonds in
Fig. 5).
|
The mean peak force produced during these constrained digit tasks was examined before and after muscimol injection. In Fig. 3, for injection 10 (first monkey), the mean peak force exerted by both thumb and index during both the single digit and the two digit tasks actually increased more than was necessary to close the switches. Significant increases and decreases in mean peak force were measured during these tasks after injection. The changes were not consistent across injections or within an injection and occurred even after a control injection of normal saline at injection site 22 in the second monkey.
LOCATION OF SITES WHERE INJECTIONS INCREASED REACTION TIME IN CONSTRAINED DIGIT MOVEMENTS. Figure 5 shows the locations of 23/66 (35%) of injections in which there was an increase in constrained movement reaction time. For injections in or within 1 mm of the dentate nucleus, there was an increase in reaction time in 23/55 (42%); for the 11 injections outside this boundary, there was no such increase.
Unconstrained compound tasks (Reach+Pinch)
Following the injections into the dentate and lateral interpositus nuclei, no deficits were observed in sitting or standing.
REACH OVERSHOOT. Prior to injection of muscimol, reaches ended on target without overshoot. Figure 6A (top) shows a single reach trial before injection 18 in the first monkey. The position of the target is marked (+). The path of the tip of the index digit (index finger) and the sequential positions of wrist (W), elbow (E), and shoulder (S) are displayed. The shoulder and elbow angles are plotted against time at the bottom of Fig. 6A. The angles change continuously and smoothly throughout each reach.
|
REACH WRIST PATH LENGTH RATIO. The wrist path for the initial segment of five reaches before and after muscimol injection 11 into the dentate nucleus of monkey 2 are displayed in Fig. 7A. Prior to the injection, the wrist path was slightly curved. The mean wrist path length ratio was 1.05 ± 0.05. After the injection (Fig. 7B), the initial segment was more curved and the monkey overshot the target. The mean wrist path length ratio after injection was 1.24 ± 0.19.
|
movement at one joint only (elbow or shoulder) followed by
movement at the other joint onlythen one would expect to see
horizontal and vertical lines. These lines would represent a segment of
the reach during which the elbow moved by itself (with shoulder fixed)
or the shoulder moved by itself (with elbow fixed). After inactivation there was increased trial-to-trial variability in the relation between
the shoulder angle and the elbow angle (e.g., Fig. 7B). Nevertheless, there was no appreciable decomposition of
movement as defined above. The principal change that increased the path curvature was at the end of movement in the elbow angle.
The location of the injections for which there was an increase
(P < 0.05) in the wrist path length ratio is marked in
Fig. 8.
|
REACH TERMINAL TANGENTIAL ERROR (XY ACCURACY). Figure 9 shows before and after an injection the positions of the index digit's mcp joint at the end of the initial reach segments (prior to any corrective movements or terminal tremor). The target fruit bit is represented by a plus sign Before muscimol is represented by open squares and after muscimol is represented by filled circles. The outer rectangle marks the perimeter of the Lucite block (5.5 × 9 cm); the inner rectangle marks the opening to the food well (1.5 × 3 cm). For trials before the injection, the monkey was able to insert its thumb and index digit into the food well opening with little or no corrective movements.
|
LOCATION OF SITES OF INJECTIONS THAT CAUSED ERRORS DURING REACH. Figure 8 shows the sites in which muscimol injection caused errors (P < 0.05) of overshoot, wrist trajectory, and tangential inaccuracy, and their co-localization. Following 35/66 (53%) of the injections in monkey 1 and monkey 2, there was an increase (P < 0.05) in the error of at least one of the parameters of reach. In monkey 1, 4/44 injections (11%) affected all parameters; 5/44 (11%), z-axis accuracy (overshoot) only; 2/44 (5%) xy accuracy only; 7/44 (16%), z-axis accuracy (overshoot) and wrist path length ratio only; and 1/44 (2%), z-axis accuracy (overshoot) and xy accuracy only. In monkey 2, 10/22 injections (45%) affected all parameters; 2/22 (9%) z-axis accuracy (overshoot) only; 1/22 (5%) z-axis accuracy (overshoot) and wrist path length ration only; 1/22 (5%) z-axis (overshoot) and xy accuracy only; and 1/22 (5%) xy accuracy and wrist path length ratio only.
PINCHING. After dentate injections, the monkey's use of the digits appeared "clumsy" and "awkward." Kinematic analysis showed that the monkey did not pinch but instead used an index single digit "winkling" strategy. For the first monkey (prior to modifying the food well) the winkling strategy usually proved successful: the monkey would pull the fruit off the cantilever with the index digit and into the palm, holding it there with the index or by closing the fist. Figure 10 shows EMG in the first monkey for three trials before (A, B, and C) and three trials after (D, E, and F) injection 45. EMG traces are shown for the flexor pollicus brevis (top, FPB, thumb flexor) and flexor digitorum profundus (middle, FDP, index flexor), aligned on first detected touch of fruit. The bottom trace (force) represents output of the cantilever strain gauges; downward deflection indicates first touch of the fruit by the index digit. The index digit slid the fruit off the cantilever. As the fruit left the cantilever, the force trace returned to baseline. EMG traces show that before injection, both the thumb FPB and the index FDP were active during part of the reach and the entire pinch of the trial. Although coordination between the two digits was necessary, there was no apparent constancy of onset timing for each. After the injection, index FDP was active. However, thumb FPB was no longer active, consistent with the sole use of the index in a single digit winkling strategy.
|
LOCATION OF INJECTION SITES CAUSING ERRORS DURING PINCH. In Fig. 11, 9/66 (14%) injections for which an increase (P < 0.05) in the number of dropped fruit trials occurred are represented by an open square. Of these, 2/44 (5%) were in monkey 1, and 7/22 (32%) were in monkey 2. Comparison of Figs. 11 and 8 shows that injections causing pinch error and dropped fruit also caused at least one abnormal parameter of reach: in monkey 1, all reach parameters (injection 15), and tangential xy inaccuracy only (injection 22); in monkey 2, all reach parameters (injections 19, 11, 4, 1, 5, and 18), and z-axis accuracy (overshoot) only (injection 9). Comparison of Figs. 11 and 5 shows that injections causing pinch error and dropped fruit sometimes but not always also caused reaction time delays in the constrained digit tasks (6/9 injections total; 1/2 in monkey 1; 5/7 in monkey 2). Similarly, injections causing reaction time delays in the constrained digit tasks sometimes but not always also caused pinch errors and dropped fruit (5/23 injections total; 1/14 in monkey 1; 4/9 in monkey 2). In sum, though errors of reaction time in the constrained digit task and of reach and pinch in the unconstrained task were often caused by a single injection, in other injections each error could occur without being accompanied by the others.
|
| |
DISCUSSION |
|---|
|
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
APPENDIX
REFERENCES
|
|---|
In this study we tried to determine in monkeys if inactivation of
lateral cerebellar output preferentially impairs certain movements
relative to others. Our goal was to test three hypotheses of cerebellar
control of movement:
1)
Summation hypothesis: All movements are controlled equally, those involving few muscles and those involving many, and the latter apparently more so only insofar as the errors of the simpler movement components summate in the more complex movements.
2)
Combination hypothesis: Many-muscled movements are controlled preferentially or exclusively, whether unconstrained or constrained.
3)
Interaction hypothesis: Unconstrained, many muscled movements are preferentially controlled to correct for interaction torques between body segments.
Designing this experiment to isolate the key variables has been only partially successful. Ideally, a simple movement would occur at one joint caused by prime movers: the limb would be supported (constrained), and other muscles would be inactive. In our study, this was not achieved, as many muscles other than prime movers were also active. Our constrained, compound movement consisted of simultaneous flexion of the index and thumb: the many other nonprime muscles were also active. Our unconstrained, compound movement consisted of a precision pinch in which the two digits worked together interactively toward a specific goal. However, our unconstrained precision pinch was combined with an unconstrained reach. Ideal tasks would have involved movement of the same muscles and joints in both tasks, but would have eliminated the interaction torques in the constrained task and involved them in the unconstrained task. Inadvertently, this may have been partially achieved. In the constrained tasks we attempted to restrict movements to the prime mover muscles of the digits; however, other muscles were active throughout the fore- and upper limb, shoulder girdle, and trunk. The pattern of muscle activity in the constrained digit tasks overlapped the pattern of muscle activity in the unconstrained Reach+Pinch task (see companion paper). Thus fortuitously, the pattern of muscle activity in the task with interaction torques minimized was more similar than not to the pattern of muscle activity in the task with the interaction torques maximized.
Deficits in the performance of the constrained digit tasks
For about half the injections in or within 1 mm of dentate,
reaction time was increased, usually for both digits during the single-digit movements and for both digits during the compound movement. The magnitude of the increase in reaction time was not significantly different for Thumb+Index than for Thumb or Index tasks.
This finding agrees with most studies of simple movements which have
reported an increase in reaction time following inactivation of the
dentate nucleus (Meyer-Lohmann et al. 1977; Mink
and Thach 1991; Thach et al. 1992;
Trouche and Beaubaton 1980). There was no observed
weakness after inactivation. For the similarly constrained movements
Thumb, Index, Thumb+Index, the fact that cerebellar inactivation
produced no significant additional increase in reaction time
in Thumb+Index would appear to go against the combination hypothesis in
this condition.
It has long been known that large inactivations of the cerebellum may
cause an increase in reaction time of a variety of movements. Holmes (1917) showed large increases in reaction time of
200 ms in unconstrained manual grasp and squeeze of a bulb
manipulandum. Similarly, in unconstrained movements in the monkey,
large increases in reaction time have been reported: moving a gripped
lever by flexion/extension at elbow of 100 ms (Meyer-Lohmann et
al. 1977) and 50-70 ms (Spidalieri et al. 1983)
flexion/extension at wrist of the mittened hand of 30-50 ms
(Mink and Thach 1991). Our attempts to constrain and
restrict movement further to the digits still resulted in reaction time
increases of 50 ms, with a mean of half that.
Deficits in the performance of the unconstrained compound task
Reaches had excessively curved trajectories, target overshoot, and
terminal errors in the tangential plane. These findings concord with
previous observations (Bastian and Thach 1995;
Bastian et al. 1996; Beaubaton and Trouche
1982; Cooper et al. 2000; Gilman et al.
1976; Goodkin et al. 1993; Holmes
1939; Thach et al. 1992).
Pinches were clumsy, with frequently dropped fruit. A strategy of
single-digit winkling when permitted was often successful. These
observations also agree with previous observations (Bastian and
Thach 1995; Goodkin et al. 1993; Thach et
al. 1992).
Co-localization of deficits of constrained digit movements and unconstrained reach and pinch
The spatial distributions of those injections that caused deficits of constrained and unconstrained movements overlapped. Injections sometimes impaired all measured parameters of all movements, consistent with the hypothesis that complex movements are the sum of their simpler components. However, often injections caused errors in one task and not in the others, which is inconsistent with this hypothesis.
The region midway between anterior and posterior in dentate and
interpositus has been shown to represent the forelimb in the monkey in
anatomical tracing (Asanuma et al. 1983; Kalil
1982), electrical stimulation (Rispal-Padel et al. 1982,
1983), and unit recording-behavioral correlation (Thach
et al. 1993) studies. Mason et al. (1998)
proposed a different scheme in which coordination of fingers (in grasp)
is represented in central interpositus and shoulder (in reach) in
posterior interpositus. This interpretation was based on muscimol
injections (8 in one monkey, 7 in a second) centered along the boundary
dividing interpositus and dentate. Their search planwith its grid
narrow in the medio-lateral dimension, centering on the
interpositus/dentate border, and the muscimol spreading from there to
involve both nucleiseems unlikely to have distinguished between
interpositus and dentate contributions. As to a separation of reach and
pinch in the rostro-caudal dimension, we have not seen it in this or
other monkey studies (Mink and Thach 1991;
Thach et al. 1992). In the cat, Cooper et al.
(2000) reported differences in posterior and anterior
interpositus in the ability to control different aspects of trajectory,
with little effect from inactivation of dentate, similar to what
Mason et al. (1998) reported in the monkey. How much
these discrepancies are due to species differences and how much to
technique remains to be seen.
Preferential cerebellar control of unconstrained movements?
Estimated torques and range of movement at mcp joints of thumb and index were many times smaller than those estimated at the shoulder at the onset of reach. It is conceivable that the greater deficits in unconstrained movements could have been due to a general weakness in the face of the greater torques required by the task. However, we do not think so, for the following reasons: 1) digit force was not impaired; 2) while reach force was not measured, shoulder-upper arm EMG amplitudes were comparable before and after injections, as were reach movement times and acceleration; 3) reaches forcibly overshot, and were above, as well as below, the target; and finally, and 4) pinches forcibly held the reward, if once acquired and not dropped.
These and previous results are theoretically consistent with a
summation hypothesis. Yet a linear summation hypothesis does not easily
explain the qualitative difference in the deficits in constrained and
unconstrained movements. Why might dentate and interpositus seem
preferentially to control the unconstrained compound movements? One
possibility is that the CNS must account for the complexity of the
movement and the mechanical interactions (Bernstein
1967; Macpherson 1991).
Reaching requires control of interaction torques generated by one limb
segment on another. During constrained simple movements, the path and
trajectory of the moving segment is determined solely by the kinematics
at a single joint. The torque acting on the moving joint arises only
from the muscles acting on that joint and gravity. During unconstrained
compound movements, temporal, kinematic, and kinetic variables multiply
across limb segments and joints greatly determining the trajectory of
the moving limb segments (Bernstein 1967; Hasan
1991; Hollerbach and Flash 1982; Kaminski
and Gentile 1989; Sainburg et al. 1993). The
normally straight reach trajectory (Morasso 1981;
Soechting and Lacquaniti 1981) becomes curved and the
reaches become inaccurate. This pattern has been observed after
cerebellar ablation in this study and by Gilman et al.
(1976). Bastian et al. (1996) and Cooper
et al. (2000) have shown that the abnormal trajectory and the
overshoot of the reach after lateral cerebellar nuclear lesions and
cortical atrophy can be largely explained by a specific inability to
anticipate and permit/counteract interaction torques.
Pinching also requires control of interactions between thumb, index,
and object. Each digit must respond quickly and precisely to any
changes in force produced by the other digit to maintain the proper net
force across the object (Westling and Johansson 1984).
For precision pinch, redundant degrees of freedom would include the
distance the thumb and index digit need to travel, the paths of the
thumb and index digit, the tangential velocities of the digits,
individual joint rotation magnitudes and velocity profiles, and the
spatial position of the digit tips at contact. Cole and Abbs
(1986) showed that, for the precision pinch, the CNS reduces
redundant multiple degrees of freedom by specifying the contact site
between the thumb and index digit. Individual digit trajectories, joint
angular position, and muscle contractions systematically co-varied
across the digits to minimize the variation in the point of contact.
Mai et al. (1988) have shown that patients with
cerebellar disease may be impaired under certain conditions in control
of isometric pinch force, although maximum grip force was within the
normal range.
The cerebellum may control simple movement components and the interactions between them
Cerebellar dentate neurons fire at high, maintained rates and
would seem likely to exert an excitatory bias on their
targets, the downstream movement generators. Inactivation of a part of the tonic excitatory output of cerebellar nuclei might therefore seem
sufficient to have caused the movement onset delays seen in the
constrained movements. Nevertheless, experimental removal of the tonic
cerebellar nuclear neuron discharge by permanent or reversible lesions
has been shown not to lessen tonic motor cortex neuron
discharge (Hore and Flament 1988; Lamarre and
Jacks 1978; Lamarre et al. 1978;
Meyer-Lohmann et al. 1975, 1977;
Spidalieri et al. 1983). However, the phasic discharge
of motor cortex neurons is delayed: there are thus changes
in the phasic but not the tonic discharge. This is more compatible with
loss of a phasic patterned signal from cerebellum rather than loss of a
supportive role in maintaining tonic background dischargeunless the
tonic background discharge supports a "phasic generator" in
thalamus or motor cortex.
The greater deficits in the unconstrained compound movements may
reflect their complexity. Movement of one body part results in
interaction torques on other body parts; the torques must be anticipated and either used in the movement or counteracted by muscle
action. As the number of body parts that must be moved in synergy (or
kept still) (Bastian et al. 2000) increases, the number
of the combinations of various interactions between the constituent
simple movement components increases nonlinearly in relation to the
number of components. During the performance of these complex
movements, the cerebellum may control the constituent simple movements
and their interactions. Following cerebellar injury, the
inability to control the great number of interactions will give the
accurate impression that unconstrained compound movements are more
greatly impaired than are their simple movement components.
| |
APPENDIX |
|---|
|
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
APPENDIX
REFERENCES
|
|---|
1) To find the inertial acceleration torque
(It) of the arm moving from rest, we
used the moment of inertia (Iprox)
regression equation for the entire upper extremity (from shoulder,
Vilensky et al. 1979)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
ACKNOWLEDGMENTS |
|---|
We thank S. Norris, A. Bastian, J. Mink, and T. Martin for help analyzing these data and preparing the manuscript, and we thank anonymous reviewers for constructive criticisms.
This work was supported by National Institute of Neurological Disorders and Stroke Grants NS-12777 and NS-15070.
| |
FOOTNOTES |
|---|
Address for reprint requests: W. T. Thach, Department of Anatomy and Neurobiology, Washington University School of Medicine, 660 S. Euclid Ave. St. Louis, MO 63110 (E-mail: thachw{at}pcg.wustl.edu).
| |
REFERENCES |
|---|
|
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
APPENDIX
REFERENCES
|
|---|
) mark sites where reaction time was increased
for the Index task only. Filled diamonds (
) mark sites
where reaction time was increased for all 3 tasks. Small dots mark
sites where no significant differences were observed before and after
muscimol injection. In no injection were reaction times increased for
the Thumb only or Thumb+Index task only.
)
Injection sites producing an increase in the number of overshoots.
(×) Injection sites producing an increase in the wrist
path length. (
) the injection. The food well is represented by the
two concentric rectangles.
