Evaluation of Personally Worn and Ceiling-based Sensors in Circadian Rhythm Monitoring

In the recent years, we saw a rapidly increasing interest in monitoring the human circadian rhythm. Yet, accurate measurement of an individual's retinal light exposure appears to be the Achilles heel of many applications based on circadian physiology. In fact, available body-worn and infrastructure-based devices typically exhibit very large inaccuracies and discrepancies, depending for instance on the sensor position. In this paper, we study the feasibility of available sensors for circadian rhythm monitoring and whether the reported light exposure inaccuracies are reflected in the estimations of the actual circadian state. We applied the dynamic circadian oscillator model by Kronauer to compare the phase shift resulting from body-worn and ceiling-based light measurements to reference retinal light exposure measurements. Our results stress out that despite their low accuracy, the efficacy of body worn and infrastructure-based solutions is adequate for circadian rhythm monitoring as no significant phase-shift errors were observed. In fact, the maximum daily phase shift error was-12 ± 3 min.