Bayesian Covariate-Dependent Circadian Modeling of Rest-Activity Rhythms

TL;DR

This paper introduces a Bayesian covariate-dependent circadian model that analyzes activity data from wearables, improving interpretability and identifying key predictors of rest-activity rhythms for personalized healthcare.

Contribution

It presents a novel covariate-dependent anti-logistic circadian model with an l_1-ball prior for sparsity, enhancing analysis of activity data with covariate integration.

Findings

Effective in uncovering relationships among demographic, psychological, and medical factors.

Demonstrates improved interpretability and flexibility over existing models.

Provides insights for personalized clinical assessments.

Abstract

We propose a Bayesian covariate-dependent anti-logistic circadian model for analyzing activity data collected via wrist-worn wearable devices. The proposed approach integrates covariates into the modeling of the amplitude and phase parameters, facilitating cohort-level analysis with enhanced flexibility and interpretability. To promote model sparsity, we employ an l_1-ball projection prior, enabling precise control over complexity while identifying significant predictors. We assess performances on simulated data and then apply the method to real-world actigraphy data from people with epilepsy. Our results demonstrate the model's effectiveness in uncovering complex relationships among demographic, psychological, and medical factors influencing rest-activity rhythms, offering insights for personalized clinical assessments and healthcare interventions.