Statistical Modeling of Cross-Systems Biomarkers

• Activity Number: 635 - Monitoring Health Behaviors with Multi-Sensor Mobile Technology
• Type: Invited
• Date/Time: Thursday, August 1, 2019 : 10:30 AM to 12:20 PM
• Sponsor: ENAR
• Author(s): Vadim Zipunnikov* and Mike Xiao and Kathleen Merikangas
• Companies: Johns Hopkins University and National Institute of Mental Health and National Institute of Mental Health
• Keywords: electronic diary; wearables; EMA

Abstract

Daily electronic diaries available through smart-phones/watches are now extensively used for ecological momentary sampling that taps patterns of many homeostatic systems including sleep, emotional states, dietary intake and others to assess behavioral components of human homeostatic systems. We develop a statistical framework that focuses on joint modelling the complex interaction of multiple biomarkers of sleep, physical activity, and circadian system coupled with the context from electronic diaries to characterize multi-modal behavioral phenotypes. We apply the approach to the data from Family Study of Mood Disorders that collected EMA and actigraphy data on 315 subjects. We will illustrate how the approach provides a deeper understanding of the dynamic interplay of multiple brain-body systems and will facilitate the development of personalized in-time prevention strategies.