Student’s t-tests were computed to determine group differences in age, education, and alcohol/substance use variables (Table 1). Linear model (LM) fits were performed to determine group differences for the cortical and subcortical region volumetric estimates, and the FA, AD, RD values in the 15 clusters identified in the voxelwise analysis of the DTI maps. Intracranial volume (ICV) used as covariates for the volumes of cortical and subcortical regions whereas only age was used as a covariate for all DTI measures. To determine the correct p-value threshold for a familywise error rate (FWER) of 5% for all group comparisons, permutation tests (10,000 iterations) of all the LMs were performed to determine the expected FWER when the hypothesis of no group difference was true. The p-value for a single comparison was selected to guarantee that the FWER (type I error α) was controlled at < 0.05 level. Paired t-test was computed to determine group-independent difference between bilateral putamen. Partial correlation was performed to analyze the relationship between bilateral HV while controlling for age and ICV. To analyze the relationship between length of
