Automatic sleep-stage scoring based on photoplethysmographic signals.

Area of Science:

• Biomedical Engineering
• Signal Processing
• Sleep Medicine

Background:

• Accurate sleep-stage scoring is crucial for evaluating sleep quality and diagnosing sleep disorders.
• Current methods often require specialized equipment and trained personnel.
• Photoplethysmographic (PPG) signals offer a non-invasive alternative for physiological monitoring.

Purpose of the Study:

• To develop and evaluate an automatic sleep-stage scoring method utilizing PPG signals.
• To assess the feasibility of using PPG for sleep monitoring in diverse sleep quality conditions.
• To explore the potential of PPG-derived features for classifying sleep stages.

Main Methods:

• Extracted time-domain, frequency-domain, and pulse rate variability (PRV) features from PPG signals.
• Employed an artificial neural network integrating multiple support vector machine classifiers for sleep-stage classification.
• Utilized leave-one-subject-out cross-validation for robust model evaluation.

Main Results:

• The model achieved up to 78% accuracy (κ=0.54) for three-stage sleep classification (wake, NREM, REM) in subjects with high sleep quality.
• Performance for three-stage classification was 75% (κ=0.52) in subjects with poor sleep quality.
• The model demonstrated consistent performance across different sleep stage groupings (three, four, and five stages).

Conclusions:

• PPG signals show significant potential for automatic sleep-stage scoring.
• The proposed method offers a promising approach for non-invasive, home-based sleep monitoring.
• Further research can refine PPG-based sleep analysis for clinical applications.