Non-contact home-adapted device estimates sleep stages in middle-aged men: A preliminary study

Abstract

BACKGROUND:

Sleep monitoring is essential to maintain a healthy life, especially for the elderly who want to age well. Among various forms of sleep devices, the non-wearable and home-adapted device might be preferred because it can be easily used.

OBJECTIVE:

In this study, we evaluated the performance of a non-contact home-adapted device compared to polysomnography (PSG), a gold standard method.

METHODS:

As a preliminary study, eight subjects were recorded over fourteen nights through PSG. The non-contact home-adapted device comprised a microwave sensor, passive infrared sensor, and smartphone application. Through the device, heart rate, respiratory rate, and body movement were collected and used to estimate sleep stages. Sleep structure was labeled in four classes: wake, REM, light, and deep sleep, and were classified using a weighted k-nearest neighbor algorithm.

RESULTS:

The device correctly estimated sleep structures with an overall epoch-by-epoch accuracy of 98.65% $\pm$ 0.05% based on leave-one-out cross-validation. The device showed significantly positive correlations with PSG in sleep indices including total sleep time, sleep efficiency, and wake after sleep onset.

CONCLUSIONS:

Our findings demonstrate a good performance of this non-contact and home-adapted device and suggest its suitability for sleep monitoring.