It is important to consider the limitations of this study. First, while the study of alcohol and substance naïve children allowed us to preclude that associations may be attributable to alcohol exposure, we were unable to test whether brain structure correlates of genomic risk are associated with onsets of alcohol use and other drinking milestones. As the longitudinal ABCD study continues to collect data, it will be particularly interesting to examine how the interplay between polygenic risk and exposure influences brain structure trajectories, and vice versa (i.e., how malleability of brain development modulates escalation or desistance in alcohol use). Second, PRS typically have low cross-trait (i.e., from GWAS phenotype to a related phenotype) predictive utility. While our findings suggest that variability in brain structure may represent predispositional biomarkers for alcohol involvement, the small effect sizes (i.e., maximum R2=0.003) and resource intensive nature of neuroimaging limit their clinical utility. Further, these small effects combined with multiple testing burden may have contributed to false negatives; as such null effects should not be interpreted to suggest that an association observed in adults may
