Associations between alcohol involvement PRS (i.e., PAU-PRS and DPW-PRS) and brain structure phenotypes were estimated using mixed-effects models in the lme4 (29) package in R. Family ID and scanner MRI serial number were included as random effects to account for non-independence of measurement associated with relatedness and scanner/site. We controlled for the first 20 ancestry principal components, mean of each modality (separately when predicting regions of that modality, e.g., mean thickness predicting regional thickness), prenatal exposure to alcohol (i.e., retrospective report of maternal use of alcohol during pregnancy prior to [no/yes] or after [no/yes] maternal knowledge of the pregnancy (30)), age, sex, age by sex, socioeconomic status proxies (i.e., caregiver education and combined household income), genotyping batch as fixed effects in each model. We ran a test of association separately for each brain region within each modality (e.g., cortical thickness) with each PRS. False-discovery rate (FDR) correction within modality (e.g., cortical thickness) was used to adjust for multiple testing. (Gray matter cortex = 34 regions, white matter = 36 tracts, subcortex = 26 regions). Results surviving FDR correction are discussed
