An Anisotropic Material Model for Image Guided Neurosurgery

In order to combine preoperative data with intraoperative scans for image-guided neurosurgery visualization, accurate registration is necessary. It has been determined previously that a suitable way to model the non-rigid deformations due to brain shift is via a biomechanical model that treats the brain as a homogeneous, isotropic, linear elastic solid. This work extends that model-based non-rigid registration algorithm to take into account the underlying white matter structure, derived from diffusion tensor MRI, to more accurately model the brain. Experiments performed on retrospective surgical cases were used to evaluate the results of the registration algorithm in comparison to the earlier model.