Fully Automated Detection of Isolated Rapid-Eye-Movement Sleep Behavior Disorder Using Actigraphy.

Area of Science:

• Neurology
• Sleep Medicine
• Biomedical Engineering

Background:

• Isolated rapid-eye-movement (REM) sleep behavior disorder (iRBD) is a strong early predictor of Parkinson's disease.
• Current screening methods for iRBD lack specificity due to symptom overlap with other sleep disorders.
• Nocturnal wrist actigraphy shows potential for iRBD detection but often requires sleep diary data.

Purpose of the Study:

• To develop and validate a precise, actigraphy-only method for detecting iRBD.
• To combine abnormal nighttime activity and 24-hour rhythm disruption markers for improved iRBD detection.
• To establish a fully automated approach for large-scale iRBD screening in the general population.

Main Methods:

• Optimized nighttime actigraphy models using automated sleep period detection in iRBD patients and controls.
• Optimized 24-hour rhythm actigraphy models using data from iRBD patients and matched controls from the UK Biobank.
• Fused nighttime and 24-hour rhythm classifiers using logistic regression for enhanced iRBD detection accuracy.

Main Results:

• Both nighttime and 24-hour rhythm actigraphy features effectively distinguished iRBD from control groups.
• The fused classifier achieved high diagnostic performance: 78.9% sensitivity, 96.4% specificity, and 0.954 AUC.
• The developed method preliminarily validates a fully automated approach for iRBD detection using actigraphy.

Conclusions:

• Actigraphy alone can accurately detect iRBD by analyzing nighttime activity and circadian rhythm disruption.
• This automated method offers a promising tool for widespread, non-invasive screening of iRBD.
• Early detection of iRBD via actigraphy may facilitate timely intervention for individuals at risk of Parkinson's disease.