= 7,858), we trained lasso models to perform variable selection on 325 factors hypothesized to be associated with development of DD. We separately evaluated predictors of baseline DD when children were 10-11 years of age and change in DD between baseline and 2-year follow-up. The lasso identified 33 predictors that explained 7% of the variance in baseline DD. Baseline DD explained 40% of the variance in change in DD and 52 predictors explained an additional 4%. These results were comparable to other studies that have taken a machine-learning approach the evaluating individual-level predictors in the ABCD. Although no single variable demonstrated large explanatory power, the overall set of environmental factors together appears to be as important as individual factors in the development of DD and underscores that DD likely develops as the result of many interacting factors with small individual effects. (PsycInfo Database Record (c) 2026 APA, all rights reserved).
Kahn et al. (Thu,) studied this question.