Smartwatch-based Activity Analysis During Sleep for Early Parkinson's Disease Detection

Area of Science:

• Neuroscience
• Biomedical Engineering
• Sleep Medicine

Background:

• Parkinson's Disease (PD) is a common neurodegenerative disorder where non-motor symptoms often precede motor diagnosis.
• Accurate early detection of PD is crucial for timely intervention and management.
• Current diagnostic methods may not fully capture the spectrum of early, non-motor symptoms like sleep disturbances.

Purpose of the Study:

• To propose and validate a novel angle-based analysis of smartwatch activity data for quantifying sleep quality.
• To assess the utility of this method in discriminating between early Parkinson's Disease patients and healthy controls.
• To correlate extracted sleep metrics with established clinical measures of PD-related sleep issues.

Main Methods:

• Utilized smartwatch triaxial accelerometry data to capture arm angle changes during sleep.
• Developed an algorithm to estimate sleep/wake states and compute key sleep metrics (sleep efficiency, total sleep time, fragmentation index, etc.).
• Validated the approach against polysomnography (PSG) and tested on a cohort of early PD patients and healthy controls in a real-world setting.

Main Results:

• The proposed method demonstrated comparable sleep state estimation to PSG in a clinical setting.
• Analysis of sleep metrics showed up to 0.77 AUC for classifying early PD patients versus controls.
• Significant correlations (up to 0.46) were found between computed sleep metrics and the PD Sleep Scale 2.

Conclusions:

• Smartwatch-based nocturnal activity analysis can effectively quantify sleep quality.
• This approach shows potential for detecting non-motor symptoms associated with early-stage Parkinson's Disease.
• The method offers a non-invasive, accessible tool for early PD detection and monitoring.