Even without physiological changes, multi-modal fusion is a made more difficult by non-rigid spatial transforms inherent to some of the modes (e.g. CT quite accurate spatially, but MR is certainly not, and can vary with pulse sequence also)