Research and Development of a Novel Approach for Accurate Personal Sleep Tracking
Description: This study aimed to develop accurate home sleep tracking technology. Our main findings are listed below. (1) Consumer sleep trackers agreed well to medical devices in measuring total sleep duration and sleep efficiency, but classifying sleep stages remains challenging for both devices. (2) In terms of user experience with sleep tracking devices, people had incomplete and faulty mental models of sleep, and there are two blind spots in people’s mental models of sleep tracking devices. In addition, they encountered challenges in reconciling device differences for assessing their sleep quality. (3) We developed a new sleep staging algorithm based on Generalized Linear Mixed Model to count in fixed and random effects during sleep tracking. Compared to the proprietary algorithm of Fitbit, our new algorithm achieved similar specificity, better sensitivity and thus higher balanced accuracy. This study provided rich implications to the design of accurate home sleep tracking technologies.
Amount: ¥2,990,000
Funding type: fellowship
Funder scheme: Grant-in-Aid for Research Activity Start-up
Start date: 2016-08-26
End date: 2018-03-31
Investigator: ZILU LIANG
Institution: The University of Tokyo
Funded outputs count: 13
Funder award ID: 16H07469
OpenAlex ID: G2999300560