Quantifying Psychological Resilience and Susceptibility to Predict Childhood Outcomes - This proposal aims to quantify the link between social determinants of health (SDoH) and psychopathology in adolescence. As social and environmental factors are robust predictors of positive life outcomes, it is important to understand the variable ways in which they affect development. Factors such as low socioeconomic status and environmental adversity have been linked to increased risk for psychopathology, but individual responses to these challenges differ, suggesting that the effect is not immutable. Here, I propose to identify factors associated with deviations in expected psychopathology using childhood behavioral problems as a metric. To this end, I will apply machine learning models to an existing large and longitudinal dataset of children, the Adolescent Brain and Cognitive Development (ABCD) Study, to estimate expected psychopathology based on SDoH. Next, I will use the difference between predicted and reported psychopathology to generate a psychological resilience/susceptibility estimate, which is the extent to which one’s actual psychopathology deviates, for better or worse, from their model prediction. Estimating this gap will allow for the identification of neural and genetic factors associated with better- than-expected and worse-than-expected psychopathology. Aim 1 involves generating an individualized framework for accurately estimating psychological resilience/susceptibility. I will train and optimize a machine learning model to predict a behavioral problems score based on broad SDoH data, encompassing prenatal, interpersonal, and community-wide factors. Once a model of sufficient accuracy is achieved, model predictions for each individual can be compared to actual psychopathology, and the difference in scores will comprise psychological resilience/susceptibility (RS-Gap). The RS-Gap will be validated using psychiatric diagnoses at a follow-up timepoint, and associations with longitudinal quality of life measures (e.g., sports participation, grades, and substance use) will also be assessed. The goal of Aim 2 is to identify genetic correlations between RS-Gap and traits of interest such as impulsivity and neuroticism within the ABCD cohort. I anticipate a shared genetic basis between RS-Gap and predisposition to psychiatric disorders. Finally, Aim 3 concerns neuroimaging measures associated with psychological resilience/susceptibility. I hypothesize that the RS-Gap will be associated with structural and functional differences in brain areas associated with stress reactivity and emotion circuits. Together, these results will identify risk factors for the development of psychopathology and serve as a new avenue to study the pre- onset stages of neuropsychiatric disease. The activities in this proposal will be conducted in an exceptional training environment to help my development into a productive and independent researcher. I will learn to conduct science that is rigorous, reproducible, and statistically sound, to effectively communicate my findings to a range of audiences, and to mentor young scientists to do the same.
