This paper describes an approach to fine scale non-rigid registration for mapping patterns of tissue volume loss in serial MRI studies of the brain. Specifically it addresses the important confound of diffuse tissue contrast changes which can influence local sub-voxel estimates of volume change. Such changes can be induced by neurodegenerative or neurodevelopmental processes, which not only modify apparent tissue volume, but also modify tissue integrity and its resulting MRI contrast parameters. We derive an approach to the voxel-wise maximization of regional mutual information (RMI) and use this to drive a viscous fluid deformation model between images. This provides a topology preserving map of local changes in volume between time points that is robust to regional changes in tissue contrast. Comparisons with current methodology are included showing that the approach provides a significant reduction in errors when tissue contrast varies locally between MRI acquisitions.