You Are How You Sleep: Personalized Sleep Monitoring Based on Wrist Temperature and Accelerometer Data

Abstract

Good sleep is a key component of good health, and as such, how to obtain quality sleep is of concern to many people. Circadian rhythms vary between individuals and play an important role in regulating sleep, however, they are currently not monitored by commercially available wearables. Previous work has shown that circadian rhythm is reflected in changes of wrist temperature. In this work, we present a prototype wristband that measures motion and temperature at the wrist. We developed an algorithm to detect wrist temperature increase onset, which is an indicator of the body preparing for sleep. Our results demonstrate that our algorithm is able to detect wrist temperature increase onset, which appears to occur at the same time for the same person. We also show that temperature increase onset varies between people as does overall temperature patterns between people. The detection of wrist temperature patterns gives us a deeper understanding of the mechanisms underlying sleep and could be a valuable component of a personalized sleep monitoring algorithm.