The substantial reduction in group differences when comparing alcohol dependence with controls once PC1 was removed suggested that the originally observed group differences were largely driven by PC1. Given the overlap between PC1 and the numerous other psychiatric and neurological effect maps, it follows that the original alcohol dependence effects were mostly not disease-specific. Although we were not able to directly examine the impact of removing PC1 on group differences in other disorders, the high correlations between PC1 and the other published effects suggest that the global abnormalities captured by PC1 may well dominate the group differences in other disorders. Therefore, a potential extension of the present findings would be to use the PC1 to first identify and statistically set aside global, non-specific disease effects before examining whether more disease-specific, cortical effects remain. In this scenario, group differences demonstrating greater similarity with PC1 would be consistent with distributed effects along the ‘global’ axis while less similarity would indicate more disease- or region-specific differences. For instance, the weaker and non-significant correlation between MDD and Combined-PC1 might represent more disease- or region-specific
