The characterization of shape in the brain is of great importance for understanding differences in structure and the relationship to function. Structural differences have been associated with, for example, age, sex, handedness, cognitive abilities and many neurologic and psychiatric conditions. Nonrigid registration methods enable the characterization of shape differences between images based on the transformation that relates them. Unlike methods which characterize shape in terms of geometric features computed from individual structures, transformation-based deformation description characterizes the entire space and therefore may better reflect the interrelationships between structures, as well as changes within and near structure. The transformation, as characterized by the local Jacobian, can yield an expressive description of local shape differences.