Regularization of MR Diffusion Tensor Maps for Tracking Brain White Matter Bundles

We propose a new way for tracking brain white matter fiber bundles in diffusion tensor maps. Diffusion maps provide information about mobility of water protons in different directions. Assuming that diffusion is more important along axons, this information could lead to the direction of fiber bundles in white matter. Nevertheless, protocoles for diffusion image acquisition suffer from low resolutions and instrument noise. This paper is essentially dedicated to the design of a Markovian model aiming at the regularization of direction maps, and at the tracking of fiber bundles. Results are presented on synthetic tensor images to confirm the efficiency of the method. Then, white matter regions are regularized in order to enable the tracking of fiber bundles, which is of increasing interest in functional connectivity studies.