Quanti cation of movement regularity during internally generated and externally cued repetitive movements in patients with Parkinson's disease

Clinical tests used to rate the severity of bradykinesia in patients with Parkinson's disease generally involve self-initiated and self-maintained repetitive movements (e.g. ngertapping). This is based on evidence that movement impairment is exacerbated by the removal of external timing cues or performance feedback (i.e. internally generated movements). Typically performance is rated subjectively by a clinician and is based on observation. Currently no clear means of quantifying the performance of these types of repetitive movements exist. Here we evaluate candidate measures that assess the amount and structure of the variability in the movements respectively. These measures are the sample entropy, which quanti es the degree of irregularity in a signal and the standard deviation which quanti es the spread or variability in the signal. This was tested under three different conditions: (1) slow versus fast repetitive movements; (2) externally cued versus internally cued movements; and (3) movements performed before and after the administration of antiparkinsonian medications. Entropy and variability were calculated for both acceleration and EMG signals. The results suggest that internally cued movements are more irregular than externally cued movements and that both regularity and variability in amplitude increase with antiparkinsonian medications. We conclude that the sample entropy is an effective means of objectively quantifying the performance of repetitive movements.