Objective: We aimed to design and test the sensitivity of a fully-automated and noise-resistant smartphone-based system that can unobtrusively screen for prodromal parkinsonian speech disorder in subjects with isolated rapid eye movement sleep behaviour disorder (iRBD) in a real-world scenario.
Background: Smartphones allow non-invasive detection and tracking of early signs of Parkinson’s disease (PD). However, most available tests rely on the active involvement of investigated subjects, while an ideal digital biomarker can be measured passively. Speech dysfunction represents one of the first motor manifestations that develop in PD.
Method: This cross-sectional study assessed regular, everyday voice call data from individuals with iRBD compared to early PD and healthy controls via a developed smartphone application [1]. In addition, the participants performed an active, regular reading of a short passage via smartphone. The smartphone data was continuously collected for up to three months after the standard in-person assessments at the clinic, where the speech was recorded using a high-quality microphone.
Results: A total of 3525 calls leading to 5990 minutes of preprocessed speech were extracted from 73 participants, including 23 iRBD, 25 PD, and 25 controls. With a high area under curve of 0.85 between iRBD and controls, the combination of passive and active smartphone data provided comparable or even more sensitive evaluation than laboratory examination using a high-quality microphone. Eighteen minutes of speech corresponding to approximately nine calls were optimal to obtain the best sensitivity for the screening. The most sensitive features to elicit prodromal neurodegeneration in iRBD were imprecise vowel articulation in phone calls (p = 003) and monopitch in reading (p = 0.05). Between calls and laboratory monologue, a high correlation coefficient was achieved only in imprecise vowels (r = 0.67, p < 0.001). Between reading tasks, the correlations were generally stronger, with a high correlation coefficient demonstrated in monopitch (r = 0.70, p < 0.001), voice quality (r = 0.66, p < 0.001), and articulation rate (r = 0.70, p < 0.001).
Conclusion: We consider the developed tool widely applicable to dense, longitudinal digital phenotyping data with future applications in neuroprotective trials, deep brain stimulation optimization, neuropsychiatry, speech therapy, population screening, and beyond.
References: Kouba T, Illner V, Rusz J. Study protocol for using a smartphone application to investigate speech biomarkers of Parkinson’s disease and other synucleinopathies: SMARTSPEECH. BMJ Open 2022; 12:e059871.
To cite this abstract in AMA style:
V. Illner, M. Novotny, T. Kouba, T. Tykalova, M. Simek, P. Sovka, J. Svihlik, E. Ruzicka, K. Sonka, P. Dusek, J. Rusz. Passive smartphone speech monitoring in isolated rapid eye movement sleep behaviour disorder [abstract]. Mov Disord. 2024; 39 (suppl 1). https://www.mdsabstracts.org/abstract/passive-smartphone-speech-monitoring-in-isolated-rapid-eye-movement-sleep-behaviour-disorder/. Accessed October 5, 2024.« Back to 2024 International Congress
MDS Abstracts - https://www.mdsabstracts.org/abstract/passive-smartphone-speech-monitoring-in-isolated-rapid-eye-movement-sleep-behaviour-disorder/