Multi-level Predictors of Depression Symptoms in the Adolescent Brain Cognitive Development (ABCD) Study

ObjectiveTo identify multi-level factors that maximize prediction of depression symptoms in a diverse sample of children in the U.S. participating in the Adolescent Brain and Cognitive Development (ABCD) study. Methods8,507 participants (49.6% female, 75.2% white, ages 9-10) from ABCD provided complete data at baseline and 7,998 of these participants provided one-year follow-up data. Depression symptoms were measured with the Child Behavior Checklist. Predictive features included child demographic, environmental, and structural and resting-state fMRI variables, parental depression symptoms and demographic characteristics, and relevant site and scanner-related covariates. We used linear (elastic net regression, EN) and non-linear (gradient boosted trees, GBT) predictive models to identify which set of features maximized prediction of depression symptoms at baseline and, separately, at one-year follow-up. ResultsBoth linear and non-linear models achieved comparable results for predicting baseline (EN: MAE=3.628; R2=0.232; GBT: MAE=3.555; R2=0.229) and one-year follow-up (EN: MAE=4.116; R2=0.143; GBT: MAE=4.141; R2=0.1400) depression. Parental depression symptoms, family support, and child sleep duration were among the top predictors of concurrent and future child depression symptoms across both models. Although resting-state fMRI features were relatively weaker predictors, connectivity of the right caudate was consistently the strongest neural feature associated with depression symptoms at both timepoints. In contrast, brain features derived from structural MRI did not significantly predict child depression symptoms. Conclusions & RelevanceParental mental health and child sleep quality are potentially modifiable risk factors for youth depression. Functional connectivity of the caudate is a relatively weaker predictor of depression symptoms but may represent a biomarker of depression risk.