The effect size for the case-control differences between participants with alcohol dependence and non-dependent controls was examined in the ENIGMA-Addiction and UKB separately. Specifically, an LME model with diagnosis, age, sex, and ICV included as fixed effects and site included as a random effect was applied to each regional brain measure. The normality and homoscedasticity of residuals were visually checked. The effect size of diagnosis was calculated as [30]: Cohen′sd=t×(n1+n2)(n1×n2)×df where n1 and n2 represent the numbers of cases and controls, respectively, and df is degrees of freedom. The case-control analysis with alcohol dependent samples comprised a total of 1,234 cases and 1,084 controls for ENIGMA-Addiction and UKB. The control group (n=296) for UKB was selected from participants in UKB with AUDIT<1 and other demographic variables matched with the case group. Diagnosis of alcohol dependence is described in Supplementary Methods.
