Switched manual pursuit tracking to measure motor performance in Parkinson's disease

Abstract

Control theoretic measures are proposed to assess motor performance in Parkinson's disease (PD), a neuro-degenerative disorder that impairs motor skills, speech, and aspects of cognition. Ten normal and 14 PD subjects performed a series of manual pursuit tracking tasks: three tasks were first performed separately, then as a merged sequence with sudden, unenunciated task changes. The tasks differed in whether the tracking errors appeared amplified, attenuated or unaltered. From the discrete block experiments, subject- and task-specific second-order, linear time invariant models were derived, with the trajectory subjects are asked to track as input and the subject's motor response as output. Multiple model adaptive estimation was employed on the merged sequences to determine whether, and with what delay, each subject modified their performance after a task change. Although all normal subjects detected the task change, less than one-third of PD subjects did (and with longer delay). Further, those PD subjects who detected the task change had estimators with higher damping ratio than those PD subjects who did not. Since cerebellar structures may affect damping ratio, and the basal ganglia are often associated with switching behaviour, the proposed method provides a comprehensive assessment of motor structures that may be affected in PD.