To investigate whether the reduction in global GMV associated with alcohol intake stems from relationships in specific regions, we estimate regression models to quantify the association of alcohol intake with a total of 139 regional GMV IDPs. These IDPs were derived using parcellations from the Harvard–Oxford cortical and subcortical atlases and Diedrichsen cerebellar atlas. Of the 139 GMV IDPs, 125 (88.9%) are significantly associated with log alcohol intake (see Supplementary Table 3). We observe the strongest associations in frontal, parietal, and insular cortices, temporal and cingulate regions, putamen, amygdala, and the brain stem. In these regions, alcohol intake explains between 0.3 and 0.4% of the variance in local GMV above the other covariates. Supplementary Figure 2 illustrates the marginal effect of increasing daily alcohol units on regional GMV IDPs, grouped by lobe. All associations are negative, except that involving the right pallidum—where the effect size is positive but very small (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\triangle {R}^{2}$$\end{document}△R2 = 0.0005). Importantly, the largest regional association was less than half the size of the association between alcohol consumption
