Contactless Camera-Based Sleep Staging: The HealthBed Study

Area of Science:

• Biomedical Engineering
• Sleep Medicine
• Signal Processing

Background:

• Polysomnography (PSG) is the gold standard for sleep monitoring but is obtrusive.
• Advances in contactless sensing and data analysis enable remote physiological monitoring.
• Heart rate variability (HRV) metrics indirectly reflect sleep stages.

Purpose of the Study:

• To evaluate a camera-based remote photoplethysmography (PPG) setup for automated sleep stage classification in near darkness.
• To assess the feasibility of contactless sleep monitoring using HRV analysis.
• To compare the performance of the remote PPG system against polysomnography (PSG) and electrocardiography (ECG).

Main Methods:

• Utilized a previously developed HRV-based algorithm for 3 and 4-class sleep stage classification.
• Collected simultaneous overnight data from 46 healthy participants using PSG and a camera-based remote PPG setup.
• Validated performance using ECG data for benchmarking.

Main Results:

• The remote PPG-based algorithm achieved moderate agreement with manually scored PSG for 3-class (Wake-N1/N2/N3-REM) and 4-class (Wake-N1/N2-N3-REM) classification.
• Average kappa (κ) values were 0.58 for 3-class and 0.49 for 4-class classification.
• Accuracy reached 81% for 3-class and 68% for 4-class classification, comparable to wearable sleep staging technologies.

Conclusions:

• Camera-based remote PPG shows potential for non-contact, automated sleep stage classification.
• The system demonstrates performance in range with existing wearable sleep staging technologies.
• This technology could enable more accessible and comfortable remote sleep monitoring.