Performance Evaluation of the Circadia Contactless Breathing Monitor and Sleep Analysis Algorithm for Sleep Stage...

Area of Science:

• Sleep science and technology
• Biomedical engineering
• Artificial intelligence in healthcare

Background:

• Polysomnography (PSG) is the gold standard for sleep studies but is lab-based.
• Wearable and contactless monitors offer potential for home-based, long-term sleep monitoring.
• Validation of novel sleep monitoring technologies against PSG is crucial.

Purpose of the Study:

• To evaluate the sleep staging performance of the radar-based Circadia Contactless Breathing Monitor (C100) and its Sleep Analysis Algorithm.
• To assess the device's accuracy in both home and sleep lab environments using healthy sleepers.
• To compare the C100's performance against existing wrist-worn sleep monitoring devices.

Main Methods:

• The C100 device recorded sleep data from participants alongside PSG.
• Respiration and body movement features were extracted and used to train a machine learning algorithm for sleep stage prediction.
• The algorithm was trained on 17 nights of data and validated on 24 independent nights using leave-one-subject-out cross-validation.

Main Results:

• Epoch-by-epoch recall for 'Deep', 'Light', 'REM', and 'Wake' stages were 75.0%, 59.9%, 74.8%, and 57.1% respectively.
• Similar performance was observed in the independent validation dataset, indicating model robustness.
• The C100 outperformed consumer and medical-grade wrist-worn devices in sleep metric estimation accuracy.

Conclusions:

• The radar-based non-contact monitor is a viable alternative to current wrist-worn sleep monitoring methods.
• The C100 demonstrates feasibility for longitudinal sleep stage monitoring in a home environment.
• This technology advances the potential for accessible, long-term sleep health assessment.