Quantitative Analysis of Voice in Parkinson Disease Compared to Motor Performance: A Pilot Study

Alice K. Silbergleit, Henry Ford Health System
Peter A. LeWitt, Henry Ford Health System
Edward L. Peterson, Henry Ford Health System
Glendon M. Gardner, Henry Ford Health System

Abstract

BACKGROUND: Characteristic features of hypokinetic dysarthria develop in Parkinson disease (PD). We hypothesized that quantified acoustic changes of voice might provide a correlate of disease severity.

OBJECTIVE: To determine if there are significant differences in acoustic measures of voice between mild and moderate PD; 2) To evaluate correlations between acoustic parameters of voice and subtests of the UPDRS in mild and moderate PD.

METHODS: Twenty six participants with PD underwent vocal acoustic testing while off PD medication, for comparison to 22 healthy controls. Participants with PD were divided into two groups based upon UPDRS activities of daily living (ADL) ratings: summed scores were used to define mild and moderate PD. Participants voiced /i/ ("ee") at comfort, high, and low pitch (3 trials/pitch). The CSpeech Waveform Analysis Program was used to analyze cycle-to-cycle frequency ("jitter") and amplitude ("shimmer") irregularities of the vocal signal, signal-to-noise ratio, and maximum phonation frequency range converted to semitones. Sections of UPDRS scores were correlated to acoustic variables of voice.

RESULTS: Key findings included a significant difference between the semitone range of the control subjects and the moderate PD group (p = 0.036). Further analyses revealed significant differences in semitone range for males between the controls vs. mild PD (p = 0.014), and controls vs. moderate PD (p = 0.005). Significant correlations were also found between acoustic findings and both the ADL and motor portions of the UPDRS.

CONCLUSIONS: Acoustic analysis of voice, particularly frequency range, may provide a quantifiable correlate of disease progression in PD.