GM volume was obtained from the voxel corresponding to the global maxima of the group difference model which localized to medial OFC (-5, 53, -5, MNI) (Figure 1). Partial correlation between GM volume and behavioral data adjusted for IQ, education, and total GM was considered significant at p<.05, 1-tailed. A 1-tailed test was used because the hypothesis was that a higher GM would correlate with greater avoidance of “bad” cards.
