Behavioural Sleep State Classification for Research in Newborns

Sleep plays a crucial role in development, starting from right after birth. Newborns cycle between quiet (QS) and active sleep (AS); states that serve unique functions in early brain maturation. Understanding these patterns is essential for advancing our understanding of neural and cognitive development and has been a focus of developmental research for decades. Physiological tools like polysomnography (PSG) offer precision but can be intrusive and impractical for most research settings. Unobtrusive approaches such as Videosomnography (VSG) provide a feasible alternative, but the behavioural expression of sleep states is ambiguous, and attempts to operationalize them have been inconsistent. This thesis presents a framework designed to help researchers classify newborn sleep states through video-based analysis of behavioural sleep state markers. It offers annotation guidelines, a method for systematizing coding data into classifications and flexibility in adapting to different experimental objectives and measurement techniques. Part 2 validates the framework’s use within a research setting, in which it successfully derived quantitative sleep behaviour estimates and detected state-dependent differences in cognitive processing abilities. The framework facilitates unobtrusive newborn sleep assessment and can be a tool to advance new monitoring techniques or to encourage state-sensitive developmental research.