Applying mathematical models of human circadian rhythms for experimental design and data analysis

Abstract

The adult human circadian pacemaker has an intrinsic period just over 24 h that entrains to a 24 h day by environmental cues such as light, eating, and exercise. Dynamic phenomenological mathematical models of the human circadian pacemaker have been developed to simulate the pacemaker's response to light. These models have been widely used for applications such as minimizing jet lag and optimizing experimental protocol design. This thesis is focused on applying mathematical models to account for the interindividual differences that are inherently present in humans and therefore also in circadian data. We optimized an experimental protocol to ensure robust performance across individuals with varying intrinsic circadian periods. Next, we developed novel MCMC-based methodology to use phase shift data to determine the mean intrinsic circadian period of a group of study participants. Finally, we investigated parameter sensitivity of a circadian pacemaker model and applied this knowledge to an adolescent data set where distinct behavior within the cohort was observed. This thesis highlights the utility of human circadian pacemaker models in a variety of contexts and establishes new insights into properties of these models and their influence on behavior.