Volumetric (manual) image registration procedures are described ([] Evans et al., 1991a). In the context of Pet/mri data analysis, the MRI volume from each subject is aligned with the corresponding PET volume. An orthogonal coordinate frame is then established based on the anterior commissure - posterior commissural (AC-PC) line as identified in the MRI volume ([] Evans et al., 1992). These anatomical frame coordinates are then used to apply a trilinear re-sampling of each matched pair of MRI and PET datasets into a standardized stereotactic coordinate system ([] Fox et al., 1985; [ 13] Talairach et al., 1988). A fully automatic method, based on multi-scale, three dimensional (3D) cross-correlation between features (image intensity and 3D gradient magnitude) of the volume to register and a target model volume (already in the target space) has also been developed in []. When this method fails, we resort to the manual method described previously.