A method for characterizing daily physiology from widely used wearables

Area of Science:

• Physiological monitoring
• Chronobiology
• Data science

Background:

• Wearable devices generate vast amounts of heart rate (HR) and activity data.
• Existing analyses often overlook complex physiological rhythms and individual variations.

Purpose of the Study:

• To develop a statistical method for extracting six key physiological parameters from wearable data.
• To track circadian rhythms in HR (CRHR) and their modulations by activity, meals, posture, and stress.
• To create personalized models of CRHR dynamics and responses to stimuli.

Main Methods:

• A novel statistical approach applied to over 130,000 days of real-world data from medical interns.
• Analysis of HR and activity data to identify underlying circadian rhythms and external influences.
• Estimation of individual-specific phase-response curves of CRHR to activity.

Main Results:

• Circadian rhythm in HR (CRHR) dynamics are distinct from sleep-wake and activity patterns.
• CRHR patterns exhibit significant inter-individual variability.
• Personalized phase-response curves quantify how real-world stimuli alter circadian timekeeping.

Conclusions:

• The developed method effectively extracts and tracks key physiological parameters from wearable data.
• Individualized CRHR dynamics and responses to stimuli can be accurately modeled.
• The findings highlight the complexity and personalization of human circadian rhythms.