Movement accuracy constraints in Parkinson’s disease patients

Abstract

This study examined the hypothesis that the kinematics of movements performed by PD (Parkinson’s disease) patients are differentially affected depending on whether or not the aiming movement has an accuracy constraint. The aiming movements required elbow extension in the horizontal plane on a digitizer. There were two movement conditions: (1) one having a spatial accuracy requirement in which the subjects moved to the defined target and stopped on it; and (2) one requiring the subjects to move toward the defined target without stopping precisely on it. Subjects were instructed to make their movements as fast and as accurate as possible in response to the auditory imperative signal.

PD patients modified the movement speed and kinematics depending on the two accuracy conditions. However, when the accuracy constraint was imposed, movement slowness observed in the patients was much more pronounced. The most revealing result was localized to the deceleration phase, particularly as the target was approached. The patients also were found to make a higher number of acceleration zero crossings from negative to positive to reach the target, indicating that the movements were more irregular. For the patients, the first acceleration zero crossing from negative to positive occurred much earlier in the movement than that for the controls. In addition, when movement accuracy was constrained, the number of zero crossings was accentuated. These data show that when PD patients make aiming movements to a target, their deceleration phase becomes longer and more variable.

Introduction

It is known that the spatial accuracy requirement for terminating a movement causes pronounced slowing in Parkinson’s disease (PD) patients compared to age-matched control subjects [32], [39], [56]. For example, Sanes [39] demonstrated pronounced slowing of movements in PD patients related to a reduction of target size and an increase in movement amplitude. However, Teasdale et al. [46] showed that PD patients could regulate their movement speeds as well as normal subjects when no spatial target constraint was imposed. It was also observed that when the movement did not require spatial accuracy, PD patients could saturate their initial EMG burst. The influence of target accuracy on control and coordination of movements performed by PD patients was interpreted as evidence that basal ganglia impairment reduces the capability to control and regulate force parameters because PD patients were less able to regulate velocity and acceleration magnitudes of movements, particularly when accuracy constraints were imposed [7], [56], [57].

Rapid aiming tasks have been used extensively to examine whether there are abnormalities in muscle activation patterns during movements performed by PD patients [2], [15], [18], [19], [41], [46]. A recent animal study using aiming tasks demonstrated that inactivating the putamen or globus pallidus produced abnormal braking of the arm movements and difficulty in stopping at a target [26]. A failure to achieve the correct balance of activation between agonist and antagonist muscles was also observed during cooling of these same structures [24]. In a fast aiming task to a pre-determined target location, tri-phasic (agonist-antagonist-agonist) EMG activation patterns were observed for normal subjects [4], [28], [29]. However, Hallett and Khoshbin [19] demonstrated that Parkinson patients did not show a typical tri-phasic muscle activation pattern during voluntary movement. These abnormal muscle activation patterns in PD patients suggest that basal ganglia impairment also reduces the capability to regulate and coordinate agonist and antagonist muscle activities.

For the movements made to a defined target, the known failure to saturate initial EMG burst could be one of the reasons for movement slowing in PD patients [19], creating a need to produce additional agonist EMG bursts to reach a target. Consequently, the movement duration and trajectory variability are likely to increase. Bradykinetic movements performed by PD patients have been further characterized in abnormal kinematic profiles of movements [15], [35], [42], [51], [56]. Normally, a single, optimally performed smooth movement is characterized by smooth bell-shaped velocity profiles with single acceleration and deceleration components between the movement initiation and termination [21]. PD patients exhibited asymmetrical velocity profiles with a prolonged deceleration phase [15], [56]. Furthermore, PD patients showed several alternating phases of acceleration and deceleration during movements [15], [35], indicating that their movements were segmented.

In contrast to the results of the aiming movements to a defined target by Hallett and Khoshbin [19], our laboratory has previously shown [46] that when accuracy requirements were minimized, PD patients were able to scale their movement speeds as instructed and saturate their initial EMG burst. Thus, it appears that activation of the initial EMG burst is not a fundamental impairment in the PD patients. Such selective impairments of the muscle activation should consequently influence the movement kinematics of the PD patients. We, therefore, hypothesized that the kinematics of movements should be different depending on whether the terminal accuracy is constrained. To our knowledge, no study comparing kinematics of movements between conditions with/without accuracy constraint has been carried out previously in PD patients. It is known that increased accuracy demand, typically causes normal subjects to slow their movements by shortening the acceleration phase and lengthening the deceleration phase. The decelerating phase is lengthened presumably to allow for more error corrective adjustments [30], [52]. These corrections are typically observed in the form of a zero crossing in the acceleration curve. A zero crossing is when the acceleration curve changes from negative to positive and typically reflects changes in the control and regulation of the movement.

This experiment examined the hypothesis that the kinematics of the movements in PD patients are selectively altered depending on whether the aiming movement has a terminal accuracy requirement. In one condition subjects terminated their movements at a defined target location. In the second condition they made a movement to the same target but were not required to stop the movement on the target. The crucial difference between the two conditions is whether the movements had to be precisely terminated. Kinematic profiles were compared between the PD patients and age-matched controls in the two accuracy conditions. Based on our earlier findings that PD patients do saturate their initial EMG bursts when a spatial accuracy is removed [46], we postulated that the kinematic profiles of these aiming movements would be different between the two groups in the accuracy condition but that the differences would be minimized in the no accuracy condition. The results support the prediction that the slower movement observed in PD patients was more pronounced when the terminal accuracy constraint was imposed. In addition, the kinematics of PD patients were abnormal, characterized by a higher number of acceleration zero crossings and that the first zero crossing occurred much earlier in the movement. Preliminary studies were reported elsewhere [37], [38].