Multimodal Monitoring Application for Parkinson’s Disease

Parkinson’s disease is a representative neurodegenerative disorder that primarily affects older adults, and since it is incurable, appropriate pharmacological treatment based on continuous monitoring constitutes the core therapeutic strategy. However, frequent examinations are often difficult due to high medical costs and shortages of medical personnel, which can result in missed opportunities for timely medication adjustments. To address this limitation, we propose a multimodal monitoring application that tracks the rate of change in disease severity in a home environment at low cost and issues clinical recommendation alerts when the changes exceed predefined thresholds. Data are collected through eye-response, gait, and speech modules, and the proposed algorithm analyzes the temporal progression of disease severity and shares the results with both the patient’s mobile application and a simulated environment–based clinician-facing patient management interface. Prototype evaluation results indicate that the eye-response and gait modules demonstrated stable real-time performance, while the speech module requires further improvement in terms of severity estimation accuracy. This system aims to enable early detection of rapid deterioration based on patient-specific cumulative data trends and to bridge gaps between in-person clinical visits, with the expectation that continued data accumulation will lead to improved accuracy of the learning model over time.