Abstract:
|
Wearable accelerometers are deployed to study patterns of motor activity, sleep, and circadian rhythmicity in health research. Generally, participants wear the device continuously for multiple days. The association of health outcomes with temporal patterns of movement have been assessed using methods for functional data analysis. In the context of scalar outcomes (e.g., disease status), typically individuals’ average movement pattern (averaged across days) is used in regression modelling. However, this averaging loses information contained in day-to-day variability. Here, we discuss a novel method for flexibly modelling the association of daily activity patterns with a scalar outcome. Methods are applied to the National Health and Nutrition Examination Survey (NHANES) 2003-2006 and 2011-2014 cohorts.
|