The relationship between long-term sleep profiles and challenging behavior in individuals with low-functioning autism

Autism spectrum disorder (or autism) is associated with a high prevalence of sleep and behavioral difficulties. Prior research indicates that between 40-80% of children experience problems with sleep and approximately 64-93% have at least one challenging behavior (i.e., behaviors that are physically dangerous and impact learning; for example, aggression, self injury or tantrums). Although the association between sleep and behavior has been investigated in adults and children with high- functioning autism, these relationships are not fully understood, and have yet to be studied at all in children with low-functioning autism (i.e., individuals with severe intellectual impairment, IQ < 70). It is well known that individuals low-functioning autism have a higher likelihood of sleep and behavior problems compared to individuals with high-functioning autism, however the relationship between sleep and behavior has yet to be explored in this population. Furthermore, as previous research has focused only on broad associations between sleep and behavior across individuals, it remains unclear whether these relationships translate in real-time for a given individual, or whether prior sleep is predictive of subsequent daytime behavior. The overarching aim of this research was to systematically identify the nature of sleep difficulties and the relationship between sleep and behavior in an understudied population of children with low functioning autism. To achieve this overall aim this research examines an unprecedented dataset of nightly sleep-awake recordings and daily behavioural recordings across a 6 month- 3 year time range obtained from a cohort of 179 individuals with low functioning autism living in two residential facilities in Boston, USA. The first chapter, reviews the existing evidence for the relationship between sleep and behavior in autism and highlights the need to study these relationships in individuals with low functioning autism. The second chapter, provides an understanding of the dataset and the machine learning techniques y used to uncover patterns of sleep and behavior in the unprecedented large dataset. The third chapter examines the different sleep phenotypes in low functioning autism and their relationship to autism clinical symptoms. The final chapters four and five, examine the predictive real-time predictive relationship between sleep and behavior in children with low functioning autism. This thesis builds on the research to date by proposing a study of sleep, and sleep and behaviour in children with low functioning autism, delving beyond traditional cross-sectional designs. The results of this thesis pave the way for future work in developing a real-time monitoring tool to predict problems and facilitate prophylactic treatment for individuals with autism.