Inferring the brain connectivity from MR diffusion tensor data

Abstract

Magnetic resonance diffusion tensor imaging (DTI) has become an established research tool for the investigation of tissue structure and orientation. In this paper, we are interested in the connectivity induced by white matter fibers in the living human brain. Considering the data resolution, this question has to be addressed at the level of fascicles made up by a bunch of parallel fibers. We propose first an algorithm dedicated to the fascicle tracking in a direction map inferred from diffusion data. This algorithm takes into account the usual fan shaped merge of several fascicles in one larger bundle. Then, we propose a way of inferring a regularized direction map from diffusion data in order to get a robust tracking. The regularization stems from a priori knowledge on the white matter geometry embedded in a model of the bending energy of an equivalent spaghetti plate. Finally, we propose a study of the tracking behaviour according to the weight given to the regularization and some examples of the tracking results with real data.