Towards Passive and Continuous Assessment of Subjective Sleep Quality Using Smartwatches
Towards Passive and Continuous Assessment of Subjective Sleep Quality Using Smartwatches
Omer Hausner,M.Sc student at the School of Industrial Engineering
Advisor: Prof. Erez Shmueli
Abstract:
Poor sleep quality poses serious risks to both mental and physical health, having been linked to cardiovascular disease, elevated cortisol levels, anxiety, depression, inflammation, impaired cognitive performance, and even increased mortality. As such, continuous monitoring of sleep quality is critical for health assessment and the promotion of well-being. Sleep quality is widely recognized as comprising both Objective Sleep Quality (OSQ) and Subjective Sleep Quality (SSQ), which are distinct yet complementary indicators, each independently associated with mental and physical health outcomes. While advances in wearable technology now allow OSQ to be monitored passively and continuously, SSQ still relies exclusively on self-reported questionnaires, leaving no suitable alternative for its passive and continuous assessment in the current research landscape. In this study, we propose a machine learning approach for passive and continuous assessment of SSQ by integrating objective sleep metrics, physical activity patterns, and physiological signals collected via commercial smartwatches. To improve predictive accuracy, we also incorporate sociodemographic factors, personality traits, and health-related characteristics. Using the ongoing PerMed study dataset comprising over 5,000 participants, more than 900,000 self-reported questionnaires, and over 2 million days of smartwatch data collected over three years, we developed a multi-layered predictive model that achieves an AUC of 0.88. Our findings highlight the potential of wearable data to reliably estimate subjective sleep quality in real-world settings, enabling scalable, passive, and continuous health monitoring.
Bio:
Omer Hausner is an Industrial Engineer currently pursuing an M.Sc. in Data Science at Tel Aviv University, where he researches the use of commercial smartwatches to predict and detect well-being-related events. In addition to his academic work, Omer is also a musician — a composer and pianist — and a graduate of the Rimon School of Music
