Sleep Validity of a Non-Contact Bedside Movement and Respiration-Sensing Device

Area of Science:

• Biomedical Engineering
• Sleep Medicine
• Wearable Technology

Background:

• Commercial sleep trackers are increasingly popular for monitoring sleep patterns.
• Assessing the validity of these devices against polysomnography (PSG) is crucial for clinical and research applications.
• Bio-motion sensing devices offer a non-contact approach to sleep tracking.

Purpose of the Study:

• To evaluate the sleep detection and staging accuracy of the ResMed S+ bedside bio-motion sensor.
• To assess the impact of algorithm updates on the S+ device's performance.
• To compare the S+ device's validity against actigraphy and polysomnography.

Main Methods:

• A cohort of 27 healthy adults underwent synchronized polysomnography (PSG).
• Data from PSG were epoch-by-epoch compared with synchronized actigraphy and S+ (original V1 and updated V2) recordings.
• Sleep staging and detection accuracy metrics were calculated for each device.

Main Results:

• Both S+ versions demonstrated high sleep detection accuracy (93.3%-93.8%), comparable to actigraphy (96.0%).
• Wake detection accuracy was low for all devices (47.9%-73.1%), though S+ showed higher specificity due to improved wake before sleep onset detection.
• Stage detection accuracy for S+ did not exceed 67.6% for any sleep stage.

Conclusions:

• The S+ device exhibits limitations in detecting wake after sleep onset (WASO), similar to actigraphy.
• The updated S+ algorithm (V2) improved wake before sleep onset (WBSO) detection accuracy compared to V1 and actigraphy.
• Researchers and clinicians must consider the potential impact of algorithm updates on the validity of commercial sleep tracking devices.