Master's Thesis Defense: Environmental Exposure Patterns and Autism Spectrum Disorder in the ABCD Study: A Nonlinear and Factor-Based Analysis

Autism spectrum disorder (ASD) is a common neurodevelopmental disorder that has become an important public health concern in recent years. This study explores how environmental exposures are related to autism spectrum disorder (ASD) using data from the Adolescent Brain Cognitive Development Study (ABCD) Study. Environmental measures came from the ABCD linked exposure datasets, and information from the Area Deprivation Index (ADI) and socioeconomic status (SES) datasets was used to adjust for confounding. After preprocessing the data and handling missing values with multiple imputation, we used adjusted generalized additive models (GAMs) to examine both linear and nonlinear associations between environmental variables and ASD. This screening step identified a group of variables with strong evidence of nonlinear relationships with ASD. Because these variables were highly correlated, we then used factor analysis on the full environmental dataset to identify broader environmental patterns. These factors were entered into adjusted generalized additive models to evaluate which environmental domains remained associated with ASD. The final results suggest that while several individual exposures showed nonlinear relationships with ASD, the most consistent signals came from broader environmental patterns, especially those related to build and pollution versus greenness, particulate pollution and transport, and local density and development.

Advisor: Bo Li