Computer-implemented techniques for mapping of the connectivity of the nervous system of a mammal use data derived from diffusion-weighted magnetic resonance imaging. Diffusion data (5) representing a probability density function for the diffusion parameters at respective voxels in a subject region of the nervous system is derived and used to derive a connectivity pattern data set (7) describing the anatomical connectivity of said nervous system to a seed voxel. Seed voxels are classified on the basis of their respective connectivity pattern data sets (7) according to a predetermined criterion such as the probability of connection of the seed voxels to portions of a target region. The techniques allow analysis of the anatomy of the nervous system because the unique connectivity pattern of a brain region determines the type of information available to it and therefore influences its function. The use of probabilistic techniques provides for robust and effective analysis and imaging.
