Latent Variable Poisson Models for Assessing the Regularity of Circadian Patterns over Time

Area of Science:

• Biological Rhythms
• Biostatistics
• Chronobiology

Background:

• Biological rhythms, particularly circadian patterns, are crucial in biology and medicine.
• Disturbances in circadian rhythms may signal future health issues.
• Existing methods lack robust analysis for irregular circadian patterns in population count data.

Purpose of the Study:

• To develop novel statistical methodology for assessing circadian pattern disturbance in longitudinal count data.
• To model individual variations in circadian and short-term trend components.
• To incorporate covariate dependencies into circadian rhythm analysis.

Main Methods:

• Latent variable Poisson modeling approach.
• Inclusion of circadian and autoregressive latent process components.
• Bayesian posterior computation using Markov chain Monte Carlo sampling.
• Deviance Information Criterion for model comparison.

Main Results:

• A flexible statistical framework for analyzing circadian rhythm irregularity.
• Demonstrated ability to model individual variability and covariate effects.
• Successful application to longitudinal physical activity count data in adolescents.

Conclusions:

• The proposed methodology provides a robust tool for quantifying circadian rhythm disturbances.
• This approach can aid in understanding the link between biological rhythms and disease.
• The findings are applicable to analyzing longitudinal count data in various biological and medical research areas.