Wakefulness evaluation during sleep for healthy subjects and OSA patients using a patch-type device.

Area of Science:

• Biomedical Engineering
• Sleep Medicine
• Wearable Technology

Background:

• Obstructive sleep apnea (OSA) is a prevalent sleep disorder linked to fatigue and accidents.
• Effective long-term sleep monitoring is crucial for OSA management.
• Current ambulatory sleep monitoring methods lack the reliability of polysomnography (PSG).

Purpose of the Study:

• To develop and validate an algorithm for distinguishing wakefulness from sleep using a novel patch-type device.
• To assess the algorithm's applicability for both healthy individuals and OSA patients.
• To provide objective sleep data for improved OSA symptom prevention and management.

Main Methods:

• Utilized electrocardiogram (ECG) and 3-axis accelerometer signals from a single patch-type device.
• Developed six parallel methods based on movement and autonomic nervous activity to determine wakefulness.
• Conducted a five-fold cross-validation on data from 25 subjects (15 low RDI, 10 high RDI) and compared extracted sleep parameters (TST, SE, SOL, WASO) with PSG.

Main Results:

• Achieved an average Cohen's kappa of 0.60, 91.24% accuracy, 64.12% sensitivity, and 95.73% specificity for wakefulness detection.
• Demonstrated significant correlations (p < 0.001) between the algorithm's extracted TST, SE, SOL, and WASO values and those obtained from PSG.
• The algorithm successfully provided wakefulness-related information.

Conclusions:

• The proposed algorithm, using data from a patch-type device, effectively provides wakefulness information.
• The algorithm's performance in wakefulness detection is comparable to existing studies.
• This system presents a viable solution for objective, long-term sleep monitoring in both healthy individuals and OSA patients.