A Single-Channel EEG Approach for Sleep Stage-Independent Automatic Detection of REM Sleep Behavior Disorder

Area of Science:

• Neurology
• Sleep Medicine
• Artificial Intelligence

Background:

• Rapid Eye Movement (REM) Sleep Behavior Disorder (RBD) is a sleep disorder linked to neurodegenerative diseases.
• Current diagnosis via polysomnography (PSG) is time-consuming and labor-intensive.
• Idiopathic RBD is a prodromal sign for α-synucleinopathies, with high conversion rates.

Purpose of the Study:

• To develop a Machine Learning (ML) framework for automatic RBD detection.
• To utilize unstaged, single-channel EEG sleep data for efficient analysis.
• To create a scalable, lightweight clinical decision support system.

Main Methods:

• A stage-agnostic ML framework was applied to EEG data from 58 subjects (32 with RBD).
• The model analyzed whole-night, single-channel EEG recordings.
• Performance was evaluated using accuracy, sensitivity, and F-1 score.

Main Results:

• The best ML model achieved 86.21% accuracy.
• Sensitivity reached 90.6%, and the F-1 score was 87.9%.
• The framework demonstrated strong predictive power for RBD detection.

Conclusions:

• This study introduces the first whole-night EEG-based ML approach for RBD detection.
• The developed tool offers a scalable and lightweight solution for early neurodegenerative disease screening.
• This facilitates timely interventions and improved patient management.