Objective: Development of an automated audio-analysis software framework for aiding in biomarker identification to objectively quantify ED severity.

Background: ED, the second most common phenotype of Musician’s Dystonia, affects the muscles involved in blowing air into wind instruments. ED diagnosis and assessment of severity often relies on subjective rating of the sound quality produced by musicians, having low reliability intra and inter subjects, which makes a long-term study difficult. Regardless of the evident need, only two studies have quantified sound characteristics in ED subjects during musical performance [1,2]. There is clinical need for a more comprehensive automated approach to acoustic analysis of ED.

Method: A protocol of musical performance exercises was designed, including sustained and sequenced notes played at multiple combinations of volume, pitch and tempo. This encompasses musicians’ full playing range. Audio recordings were made and a custom-developed software program was developed to extract acoustic features from such recordings. Fundamental frequency and amplitude perturbations were analysed by Power Spectral Density and Time Domain analyses. Rhythmic abilities were evaluated from sequences by metronome and inter-note comparisons.

Results: The proposed protocol (30 min duration) is optimised to enable subsequent automation of the analysis. The main output measures include: shimmer and jitter for sustained notes; inter-notes consistency parameters (rhythm, volume and pitch) for sequences. The protocol and software analysis were tested with recordings from four professional wind instrumentalists. For example, mean jitter was 0.18% for the low register with a coefficient of variation across higher pitches of 0.47%.

Conclusion: The protocol and software developed allow automatic extraction of acoustic features and form a baseline to compare to ED patients. A major strength of this methodology is the potential for remote data recording and processing, facilitating its use in a larger ED population.

References: 1. A Morris, S Norris, J Perlmutter, J Mink, Quantitative, clinically relevant acoustic measurements of focal embouchure dystonia, Mov. Disord., 33(3):449–458, 2018. 2. A Lee, S Furuya, M Morise, P Iltis, E Altenmüller, Quantification of instability of tone production in embouchure dystonia, Park. Relat. Disord., 20(11):1161–1164, 2014.

« Back to 2019 International Congress

MDS Abstracts - https://www.mdsabstracts.org/abstract/a-software-framework-for-objective-and-quantitative-audio-analysis-of-embouchure-dystonia-ed/