What radio waves tell us about sleep!

Area of Science:

• Biomedical Engineering
• Artificial Intelligence in Healthcare
• Sleep Medicine

Background:

• Assessing sleep and breathing disorders at home is challenging.
• Current methods often require uncomfortable on-body sensors.
• Longitudinal sleep data is crucial for understanding disease interactions.

Purpose of the Study:

• To develop an advanced machine learning algorithm for passive sleep and nocturnal breathing monitoring.
• To validate the algorithm's performance against the gold standard (polysomnography).
• To explore the relationship between sleep stages and various diseases.

Main Methods:

• Utilized radio wave analysis reflecting off sleeping individuals.
• Developed a machine learning algorithm for sleep stage and apnea detection.
• Validated the model with 880 participants against polysomnography.

Main Results:

• Achieved 80.5% accuracy in sleep hypnogram classification (wake, light, deep, REM).
• Demonstrated high accuracy in detecting sleep apnea (AUROC = 0.89) and measuring Apnea-Hypopnea Index (ICC = 0.90).
• Showcased equitable performance across diverse demographics and revealed disease-sleep stage interactions.

Conclusions:

• The developed algorithm offers a powerful, non-contact method for sleep and breathing assessment.
• This technology has significant potential for clinical trials, routine care, and disease research.
• Highlights the critical role of sleep in understanding and managing neurological, psychiatric, cardiovascular, and immunological disorders.