Group analyses were conducted on contrast images transformed into standard space. Z-statistic images were thresholded with cluster-based corrections for multiple comparisons based on the theory of Gaussian Random Fields with a cluster-forming threshold of Z>1.96, 2.3, or 3.7, depending on the analysis (see Tables and Figures for details), and a probability threshold of p<0.05 (Worsley, 2001). In particular, the more stringent thresholds were used to refine the localization of the clusters of activation thus informing the interpretation of the findings. Specifically, the voxel height threshold (i.e., “cluster-forming threshold”) of the Z map (Gaussianized t-map) was first specified followed by a corrected cluster significance threshold of p=0.05. Alcoholism severity factor scores were modeled as the explanatory variable on the whole-brain contrast maps. Anatomical localization within each cluster (maximum Z statistics and MNI coordinates) was obtained by searching within maximum likelihood regions from the FSL Harvard-Oxford probabilistic atlas. To test the primary aim, OPRM1 genotype (i.e., AA and AG/GG) was entered as a predictor variable and examined in relation to the computed contrasts using a whole-brain approach.
