Examining Heterogeneity in Neurodevelopmental Conditions– A Data-Driven Network-Based Approach
Year: 2024
Type: dissertation
Study design: Observational Study
Abstract: Neurodevelopmental conditions including autism spectrum disorder (ASD; autism), attention deficit/hyperactivity disorder (ADHD), and obsessive-compulsive disorder (OCD) are highly heterogeneous and overlapping. Given this heterogeneity, data-driven approaches, such as clustering, have gained attention for discovering transdiagnostic groups of individuals who share similar neurobiological features that can ultimately inform personalization of interventions and supports. However, existing approaches in this field are critically limited by the lack of replicability and generalizability of the discovered subgroups in unseen datasets. This presents a significant limitation to the clinical translation of data-driven approaches. To address these gaps, this thesis advances our understanding of neurobiological heterogeneity in neurodevelopmental conditions through the application of data-driven network-based approaches to discover replicable subgroups of individuals with shared patterns of brain structure. The specific contributions of this thesis are three-fold: (1) Using a novel graph network analytic framework, we discovered transdiagnostic subgroups of children who shared similar patterns of brain structure; (2) We examined the replicability of clustering structures across four measures of brain structure, male and female subgroups, and independent datasets; and (3) identified clusters of neurodivergent children whose brain structure deviated significantly from a reference model, demonstrating the generalizability of the findings through out-of-sample prediction. Across these studies, neurobiological homogeneity did not align with existing labels of autism, ADHD, OCD, and neurotypical development or individual behavioural domains as the diagnostic criteria for these conditions. Instead, homogeneity aligned with constellations of differences across multiple phenotypic domains. Overall, the results of this thesis enhance our understanding of neurodevelopmental diversity, setting the stage for personalized interventions based on neurobiological profiles.
Source: TSpace (University of Toronto)
Author: Sadat-Nejad, Seyed-Youns
Institution: University of Toronto
Language: English
Subfield: Cognitive Neuroscience
Field: Neuroscience
Domain: Life Sciences
Open Access status: green
OpenAlex ID: W7226055320
