| dc.description.abstract | Musician s Dystonia (MD) is the most common movement disorder affecting musicians. Being a specific phenotype of Dystonia, MD is a focal, task-specific and painless disorder that affects motor control during musical performance, characterised by abnormal muscle contractions. With an estimate of 1-2% in professional musicians affected, MD has been reported in virtually every instrument, including keyboard, strings, plectrum, woodwind, brass and drums. The main phenotypes associated with MD are Hand Musician s Dystonia (HD) and Embouchure Dystonia (ED). Nevertheless, the disorder has been reported to manifest in several body regions (e.g. arm, hand, lower cranial structure, cervical muscles, breathing muscle, vocal cords, legs).
There exists a lack of research in MD and this is reflected in the absence of objective methodologies for its assessment, specifically those that meet the requirements of ease of use in clinical environments. There is clinical need for a tool that ensures such requirements and that would allow MD assessment alongside the Dystonia-triggering musical instrument. Such a tool would also provide a standardised methodology for longitudinal studies of MD.
The aim of this research was to address these challenges by developing a software platform for objective and quantitative acoustic-analysis of Musician s Dystonia. The main outcomes of the research presented in this thesis are as follows:
An experimental set-up for audio-recordings and a protocol of musical exercises for ED evaluation. The experimental set-up enables subsequent automated analysis of the acquired recordings. The protocol permits studying ED across the disorder s task-specific domains. ED musical technique-specificity is evaluated by means of sequenced and sustained notes. Register-specificity is tested by notes encompassing complete playing and dynamic ranges. Sequenced notes played at slow, medium and fast tempos enable speed-specificity study.
Software analysis for Embouchure Dystonia s severity assessment. The software platform allows automated evaluation across the following specific acoustic domains by extracting key acoustic-features: loudness and pitch instability, rhythmic abilities, note attack precision, loudness consistency and pitch consistency.
Healthy wind instrumentalist s normative acoustic data and feature set. Seven healthy wind instrumentalists followed the experimental protocol and generated a series of audio recordings, which formed a baseline dataset and following application of the software platform a reference output feature sets This data and feature set provide a comparison for data acquired from ED musicians.
A comparative study between audio-based features extracted from ED and control musicians acoustic data. A freely available database of six ED subjects recorded during musical performance was analysed by the developed software platform. Inter-group comparisons revealed greater sound instability in notes played by ED than in those recorded from healthy musicians.
Software analysis for Hand Musician s Dystonia severity assessment. The software suite allows analysis of audio-recordings acquired from HD subjects during performance of protocol-guided musical exercises. Loudness and pitch instability, and tone quality within the acoustic data are evaluated and key audio-based features extracted.
Analysis of audio-recordings acquired from HD musicians before and after a sensory-motor rehabilitation therapy. This is, to our best knowledge, the first study attempting to perform acoustic analysis of HD. Analysed data consisted of audio-recordings of seven HD musicians performing protocol-guided musical exercises, acquired pre and post a rehabilitation designed intervention. Improvement or reduction in improvement of musical performance was quantified for each HD subject. The analysis software demonstrated an ability to resolve changes in impairments during musical performance pre- and post-intervention.
In summary, a software suite that allows objective analysis of Musician s Dystonia is presented in this thesis. MD s severity is quantitatively assessed by analysis of audio-recordings acquired from subjects during the musical performance of protocol-guided exercises. The presented methodology requires simple, low-cost and mobile equipment. No staff expertise is required for its implementation and it provides fast and automated analysis. A major strength of the presented analytical approach is the potential for remote data recording and processing, facilitating its use at scale and in large MD populations. Future research may combine this acoustic approach with complementary quantitative analysis (e.g. clinical, kinematic, neurophysiologic and neuroimaging) to provide further characterisation of symptomatology, assessment of efficacy of new treatments and also longitudinal analysis of Musician s Dystonia. | en |
