Estimation of sleep posture using a patch-type accelerometer based device.

Area of Science:

• Biomedical Engineering
• Sleep Science
• Wearable Technology

Background:

• Accurate sleep posture tracking is crucial for understanding sleep quality and related health conditions.
• Current methods for sleep posture monitoring can be invasive or lack precision.
• Wearable sensors offer a promising non-invasive approach for continuous physiological monitoring.

Purpose of the Study:

• To develop and validate a sleep posture estimation algorithm using 3-axis accelerometer signals.
• To enable accurate identification of multiple sleep postures, including supine, lateral, and prone positions.
• To assess the algorithm's performance against established polysomnography (PSG) and video recording methods.

Main Methods:

• Utilized a patch-type sensor to collect 3-axis accelerometer data during sleep.
• Analyzed accelerometer signal characteristics for distinct sleep postures (supine, left/right lateral, prone, non-sleep).
• Developed decision rules for a posture estimation algorithm and validated it using data from 13 subjects during PSG.

Main Results:

• The algorithm achieved an average agreement of 99.16% in estimating sleep postures.
• Demonstrated high reliability with a Cohen's kappa of 0.98 compared to reference methods.
• Successfully differentiated between five sleep postures and non-sleep states.

Conclusions:

• The developed algorithm accurately estimates sleep postures using wearable accelerometer data.
• This technology can serve as a valuable supportive tool in clinical polysomnography (PSG) studies.
• The system is suitable for out-of-hospital sleep monitoring and research applications.