Sleep Duration as a Neural Survival Model
Johns Hopkins University
Circadian rhythms govern almost every daily activity, including the duration of sleep. In the literature, several medical models and mathematical models have been published that illustrate disparate perspectives of sleep-wake dynamics. This thesis focuses on modeling sleep duration with the most prominent medical model for sleep, the two-process model, in conjunction with a novel mathematical architecture that aspires to capture both the circular nature and the complexity of daily schedules and habits. Specifically, the two aims of this thesis are (1) to develop a theoretical mathematical framework to describe cyclical sleep and wake patterns and (2) to test this framework computationally with empirical patient data. The theoretical framework shows how a variety of mathematical techniques, including neural implementations of Cox proportional hazards models, could provide insights into the physiology of sleep in new ways. The subsequent computational implementation investigates how aspects of the theoretical mathematics presented can also be demonstrable in practice with medical records. With more research, survival models constructed with the methods utilized in this thesis could potentially be used in a clinical setting by physicians to prescribe more effective treatments for patients that will better align with their circadian activities.
Survival Analysis, Circadian Rhythms, Neural Networks, Machine Learning