Acoustic methods in diagnosis of Parkinson’s disease

The article discusses the usefulness of speech signals in the context of screening or objectively monitoring the progress of Parkinson's disease (PD) treatment. The AGH team developed an application for a mobile device (smartphone), which uses the built-in microphone to record the acoustic signal during a phonetic test performed by the patient. The research task was to identify parameters estimated from the time courses of the speech signal that objectively and statistically significantly differentiate the states of Parkinson's disease. Ultimately, nine different acoustic parameters were selected, and their usefulness was statistically examined to determine the significant difference between the studied groups. Voice parameters of individuals with Parkinson's disease (27 people) were compared with a control group of healthy individuals (15 people). The difference in voice between patients with low and high levels of disease severity on the Hoehn-Yahr scale was also examined. Inference was conducted using the Mann-Whitney U test, suitable for the given sample size and group imbalance. The obtained research results and conducted analyses show a high potential for diagnostics based on the speech signal, which can be combined with upper limb tremor diagnostics in the recognition of early-stage PD, thus providing the possibility of facilitating doctors' work in the context of objectively monitoring the adopted therapy.