Dear Cristian & SPMers, I've read in one of the emails of this list that vbm8 does not require covariating for GM or intracranial volumes. If that is the case, could you explain me why it doesn't and whether there is some literature supporting this statement? Many thanks, Joana