The accuracy and precision of timing of self-paced, repetitive movements in subjects with Parkinson's disease.

In separate experiments, we studied the temporal accuracy and precision of self-paced, repetitive finger-tapping in two groups of 12 patients with Parkinson's disease and a group of 12 controls matched to the patients with respect to age and general cognitive state. One group (I) of patients was studied initially following 12-15 h abstinence from normal levodopa medication ('off') and again, subsequently, approximately 1 h after ingestion of a single normal dose ('on'). A second group (II) of patients, each of whom had bilaterally asymmetrical neurological signs, was tested using 'worse' and 'better' hands separately. Within each session, subjects were tested repeatedly on a tapping task during which they were required to produce a regular series of self-timed inter-tap intervals, the target duration (550 ms) of which had been established previously during an initial period of tapping in synchrony with the beats of a regular metronome. We employed Wing and Kristofferson's (1973) model of control of motor timing to partition the total variance (TV) about the mean inter-response interval (IRI) produced during the self-paced phase of each run into separate components ['clock' variance (CV) and 'motor-delay' variance (MDV)] attributable to hypothetical 'clock' and 'motor-implementation' processes. Although the mean self-paced IRI of parkinsonian patients was generally shorter than that of controls, only during the 'on' medication condition (Group I) was it significantly so. By comparison with control values, and those observed during the 'on' medication condition, values of TV, CV and MDV in Group I were all significantly higher when subjects were 'off' medication. During the 'on' medication condition, only CV was significantly higher than the control value. In Group II, values of TV, CV and MDV associated with use of the 'worse' hand were all significantly higher than both control values and those associated with use of the 'better' hand. Values of these variables when subjects used the 'better' hand did not, however, differ significantly from control values. The theoretical import of these results is discussed in the light of several important procedural, statistical and computational issues and we conclude that TV, CV, and MDV may all vary significantly as a function of the efficacy of dopaminergic transmission in the basal ganglia.