kinsonsyndrom und ihre beeinflussung durch L-dopa. Fortschr Neurol Psychiat 1981;49:187-192. 14. Barbeau A. Six years of high-level levodopa therapy in severely akinetic parkinsonian patients. Arch Neurol 1976;33:333-338. 15. Rajput AH, Offord KP, Beard CM, Kurland LT. Epidemiology of parkinsonism:incidence, classification and mortality. Ann Neurol 1984;16:278-282. 16. Curtis L, Lees AJ, Stern GM. Effect of L-dopa on course of Parkinson’s disease. Lancet 1984;2:211-212. 17. Hoehn MM. Parkinson’s disease: progression and mortality. In:
Yahr MD, Bergmann KJ, eds. Advances in neurology,vol45. New York Raven Press, 1986:457-461. 18. Istituto Centrale di Statistica: Annuario Statistic0 Italiano. Roma: ISTAT, 198753344. 19. Kaplan EL, Meier P. Non-parametricestimation from incomplete observations. J Am Stat Assoc 1958;53:457-481. 20. Pet0 R, Peto J. Asintottically efficient rank invariant test procedures. J R Stat SOCA 1972;135:185-198. 21. Cox DR. Regression models and life tables (with discussion by J ROY).Stat SOCB 1972;34187-220.
The movement speed/accuracy operator in Parkinson’s disease Erwin B. Montgomery, Jr., MD, and Julie Nuessen, MHS, PT
Article abstract-This study examined the effects of Parkinson’s disease (PD) and its treatment on the speed of rapid arm movements to different target sizes in a simple reaction time task in 8 PD and 9 normal subjects. Testing in PD subjects occurred after an overnight fast from their medications and 1 hour after their usual 1st dose. They had longer reaction and movement times, both of which shortened followingmedication. Like normal subjects, PD subjects had decreased movement times to larger targets, but without drug were unable to maximize speed within the range of target sizes tested whereas PD subjects taking drug and normal subjects could. These findings suggest that a speed/accuracy operator is functioning in PD but at reduced initial gain and at an offset towards longer movement times. NEUROLOGY 1990:40:269-272
Past theories of Parkinson’s disease (PD) pathophysiology emphasize problems in movement initiation, as evidenced by prolonged reaction times. However, reaction time abnormalities have been Studies have more consistently shown slowing of movement (ie, bradykinesia) as measured by prolonged movement times although the mechanisms are not understood.2p6Recent studies in PD humans6 and single neuron recordings in the globus pallidus7and putamen of monkeysa suggest that programming target acquisition may be an important basal ganglia function. This function may be disturbed in PD and account for bradykinesia. One aspect of target acquisition programming may be determining movement speed in relationship to required accuracy. Movement time for rapid tasks in normal subjects is inversely proportional to target size.g Since these movements may occur too fast for peripheral feedback to control, the patterns of muscular activity (ie, motor programs) are specific to target size and probably established prior to movement initiation by some speed/accuracy operator, possibly involving the basal ganglia. This operator may be abnormal in PD. Previous studies in PD subjects showed differences
in motor performance with different target conditions.’*1° However, medications were not controlled which may affect performan~e.~.~ The present study controlled testing relative to the timing of medications in an effort to separate disease from drug effects. Furthermore, this study simplified the tasks to a single movement towards a single target of varying size. Originally, alternating movements between 2 targets were used by Fittsgin normal subjects and by Sanes’O in PD subjects. Rapidly alternating movements may be qualitatively different from the simple single movements used here. Therefore, previous results may not be generalizable. Methods. Nine normal (mean age, 63.3 years) and 8 PD (mean age, 59 years) subjects were studied. Subjects sat a t a table holding a stylus as they would a pencil in their dominant hand. The stylus was held on a 2.5 X 2.5 cm square home base in front of the subjects to the side of their nondominant hand (figure 1).Following an auditory tone “go” signal, subjects moved the stylus to a target 60 cm from the home base (measured from the center of the home base to the center of the target). The time from the go signal to lifting the stylus off the home base was measured as the reaction time. The time from the stylus lifting off the home base to placement on the target
From the Department of Neurology and Neurological Surgery (Neurology),Washington University School of Medicine, St. Louis, MO. Supported by a grant from the American Parkinson Disease Association and the Jane K. Pelton Fund for Movement Disorders Research. Received June 7,1989. Accepted for publication in final form July 14, 1989. Address correspondence and reprint requests to Dr. Erwin B. Montgomery, Jr., Department of Neurology and Neurological Surgery (Neurology), Box 8111, Washington University SchooI of Medicine, 660 South Euclid Avenue, St. Louis, MO 63110.
February 1990 NEUROLOGY 40 269
was measured as the movement time. Targets were circular measuring 3.8,2.5, 1.9, and 1.2 cm in diameter. Target sizes were small relative to the distance to target, thus minimizing the effect of landing at different points within the target. All subjects were repeatedly encouraged to react and move as fast as possible. All were allowed to practice until proficient. Trials were presented in blocks of 10 of a single target size beginning with the largest and ending with the smallest. Any training effect would most improve performance to the smallest target. Therefore, any decrease in movement time to large compared with small targets could not be the result of training. PD subjects had bilateral disease stage I1 or I11 when assessed without medications, according to the Hoehn and Yahr scale.” They were tested following an overnight fast from medication (at least 8 hours) and 1hour following their usual 1st dose of medication (table).
Results. Median values for reaction and movement times for each subject for all target sizes are depicted in
Home Base
Target
60 cm
Figure 1. The experimental apparatus. The subject holds a stylus on the home base and, following an auditory tone, moves the stylus to the target.
figures 2 and 3. Reaction and movement times for normal subjects were compared with PD subjects with and without drug using an analysis of variance based on a split plot factorial design.12Normal subjects had significantly shorter movement times than PD subjects without drug (p < 0.02) over all target sizes but not compared with the same PD subjects taking drug. Each group showed significant increases in movement times with decreasing target size (p < 0.001). One-way analysis of variance for individual subjects showed a target size effect on movement times for 7 of 9 normal subjects, 7 of 8 PD subjects without drug, and 6 of 8 with drug (p < 0.05).13 Thus, performance can be operationally described by the gain and offset of a hypothetical speed/ accuracy operator. Gain is defined as the change in movement time with the change in target size, and offset is defined as the difference from normal subjects over all target sizes. The overall gain, as measured by the average reduction in movement time between the smallest and largest targets, was not significantly different between normal subjects (0.20 second), PD subjects when not taking drug (0.20 second), and PD subjects taking drug (0.15 second). Therefore, the offset accounts for most of the PD slowing. However, the character of the operator’s gain was different between normal and PD subjects. Normal subjects, and PD subjects taking drug, were able to reach maximum speed with the 2nd iargest target size, whereas PD subjects without drug could not. No significant movement time differences for the 3.8 cm (largest) target compared with the 2.5 cm (2nd largest) target were found using a 1-sample t test where the dependent variable was the difference of the natural log of the movement times seen for normal subjects ( t = 1 . 2 1 , ~> 0.05, degrees of freedom = 8) and PD subjects taking drug (t = 2.26, p > 0.05, degrees of freedom =
Table. Description of normal and parkinsonian subjects Normal subjects Pt no. Sex Age 1
F
70
2
F
66
3
F F
49
4
5
F
57
6
F
53
7
M
74
8 9
F F
80 60
Pt no.
Parkinsonian subjects Hoehn & Yahr scale without drugs
Sex
Age
1
M
55
I11
2
M
45
I11
3
M
67
I1
4
F
56
I1
5
F
68
I1
6
M
65
111
7
M
64
I1
8
M
52
I1
61
270 NEUROLOGY 40 February 1990
Medications used in testing Bromocriptine 1.25 mg, L-dopa 100 mg, carbidopa 25 mg L-Dopa 300 mg, carbidopa 75 mg L-Dopa 200 mg, carbidopa 50 mg L-Dopa 50 mg, carbidopa 50 mg L-Dopa 100 mg, carbidopa 25 mg L-Dopa 450 mg, carbidopa 112.5 mg L-Dopa 250 mg, carbidopa 62.5 mg L-Dopa 150 mg, carbidopa 37.5 mg, benztropine mesylate 1 mg
F
0.38 B
Parkinsonians off
w
0.38
c
Parkinsonians 1 hr after medications
:::m o*22w -
0*38 0.34 A
0.34
0.30 ""
0.30 0.26
0.18
0.18
0'141.2 1.6 2.0 2.4 2.8 3.2 3.6 4.0
I
111111111111111
141.2 1.6 2.0 2.4 2.8 3.2 3.6 4.0 TARGET SIZE (cm)
Figure 2. T h e median reaction times, for each target size, for (A) normal subjects, (B) parkinsonian subjects when not taking drug, and (C) parkinsonian subjects after taking drug.
1 .1
A
Normals
*2c\B
Parkinsonions off medications
1.1
1.01
\
1.1
c Parkinsonians 1 hr after medications
1 .ol
tllll.IIIIIIIIl
0.41.2 1.6 2.0 2.4 2.8 3.2 3.6 4.0 TARGET SIZE (cm) Figure 3. The median movement times, for each target size, for (A) normal subjects, (B) parkinsonian subjects when not taking drug, and (C) parkinsonian subjects after taking drug. 7).12 PD subjects without drug were significantly different ( t = 3.94, p < 0.01, degrees of freedom = 7). Reaction times were not different between normal subjects and PD subjects with or without drugs over all target sizes (split plot factorial analysis of variance).l* Neither normal subjects nor PD subjects with or without drug showed any effect of target size on reaction time. Reaction and movement times for PD subjects when not taking drug were compared with times when taking drug using a general linear models procedure.12Movement times were shorter after medication over all target sizes (p < 0.001) with target size having an effect 0, < 0.001). Reaction times were shorter following medication (p < 0.01) but did not show any effect of target size.
Discussion. Performance in these tasks can be described by the offset and gain of a hypothetical speed/
accuracy operator, which programs movement time programming as a function of target size. This operator is abnormal in PD subjects. There is an offset towards longer movement times that is improved with medication. Although the overall gain is normal, the initial gain is less and the gain does not maximize or saturate as in normal subjects or PD subjects taking medication. Thus, the gain normalized with medication in PD subjects. These abnormalities may account for at least 1 component of bradykinesia. These results support a basal ganglia role in the speed/accuracy operator and, therefore, in programming target acquisition. This description is not to suggest that the speed/ accuracy operator is necessarily represented in a unique localizable neuronal structure. However, findings of putamen neurons related to and preceding target attainment suggest that the basal ganglia may be involved in programming target acquisition.8 A speed/accuracy February 1990 NEUROLOGY 40 271
operator may be a special case of target acquisition programming. In a previous study: PD subjects, in contrast to normal subjects, were unable to move faster to a target bounded by mechanical stops fie,bounded target) eompared with an unbounded target of a narrow range. The bounded target would appear to the subject as unlimited in size. According to the findings of the present study, movement times to the bounded targets should have been less than to the unbounded targets if bounded tasks are generated by the same mechanisms as unbounded tasks. Rather, combining the results of both studies supports the hypothesis that movement to unbounded versus bounded targets are qualitatively different apart from target size. These movements may have different underlying physiologic mechanisms that are affected differently in PD. Reaction times in PD subjects shortened following medication, implying that reaction times without drug are abnormal. In this case, each subject served as his or her own control. Comparison with normal subjects failed to demonstrate a difference, probably due to the intersubject variability. However, a previous study which used each PD subject taking drug as the subject’s own control did not show any difference in reaction times.6 This may be due to many factors, one of which may be a difference in the go signal. In the previous study, the appearance of a new target to which the subject moved was the go signal. This may require more cognitive operations and, therefore, may be more difficult compared with the present study in which the target did not change and an auditory cue was used. This increased difficulty for both normal and PD subjects could have masked any differences in the times needed for actual movement generation. Comparison with Sanes’ studylois problematic. That study used rapidly alternating movements which may be qualitatively different from movements used in this study and may be differentially affected by PD. Therefore, direct comparison of Sanes’ study to the present
272 NEUROLOGY 40 February 1990
study may not be possible. Furthermore, medications were not controlled in Sanes’ study.1°
Acknowledgments The authors thank Mr. Irving Ward for developing the test equipment, the Memory and Aging Project of the Alzheimer’s Disease Research Center (Dr. Leonard Berg, Director) for providing some of the normal subjects, and Mr. Jack Baty of the Division of Biostatistics.
References 1. Flowers K. Ballistic and corrective movements in an aiming task
intention tremor and parkinsonian movement disorders compared. Neurology 1975;25:413-421. 2. Evarts EV, Teravianen H, Calne DB. Reaction time in Parkinson’s disease. Brain 1981;104167-186. 3. Pullman SL, Watts RL, Juncos JL, Chase TN, Sanes JN. Dopaminergic effects on simple and choice reaction time performance in Parkinson’s disease. Neurology 1988;38:249-254. 4.Rogers MW, Chan CW. Motor planning is impaired in Parkinson’s disease. Brain Res 1988;438271-276. 5. &fa1 RD, Friedman JH, Lannon MC. Preparation of manual movements in hemiparkinsonism. J Neurol Neurosurg Psychiatry iga9;52:399-402. 6. Montgomery EB, Hunt L, Mandir A, Watts RL. Effects of two target conditions on motor performance in normal and parkinsonian subjects. Submitted. 7. Mink JW, Thach WT. Timing of EMG, globus pallidus and cerebellar neuron activity in relation to initiation and termination of set-tracking wrist movement [Abstract]. Abstr SOCNeurosci 1986;12:208. 8. Montgomery EB Jr, Buchholz SR. Changes in putamen activity in Neurosci 1988;14:720. behaving parkinsonian monkeys. Abstr SOC 9. Fitts PM. The information capacity of the human motor system in controlling the amplitude of movement. J Exp Psycho1 1954;47:381-391. 10. Sanes JN. Information processing deficits in Parkinson’s disease during movement. Neuropsychologia 1985;23:381-392. 11. Hoehn MH, Yahr MD. Parkinsonism: onset, progression and mortality. Neurology 1967;17:427-442. 12. SAS Institute Inc. SAS user’s guide: basics, version 5 edition. Cary, NC: SAS Institute Inc, 1985. 13. BBN Software Products. RS/1 user’s guide release 3.0. Cambridge, MA: BBN Software Products Corp, 1987.
