can align corresponding anatomical features. Even so, cortical anatomy is so complex and variable across subjects that most standard nonlinear registration algorithms cannot match cortical regions very well, as confirmed by a recent large-scale comparison study of 14 registration methods (Klein et al., 2009). Because VBM typically uses a standard automated registration approach and does create explicit models of the cortex, cross-subject registration errors in cortical anatomy can seriously reduce the power to detect group differences and disease effects.
