Statistical Fiber Tracking on DT-MRI data as a Potential Tool for Morphological Brain Studies

Introduction: The main potential of diffusion weighted MR imaging resides certainly in its capability to give information about nerve fiber tract orientation in the brain [1]. In order to isolate fiber tracts on a 3D tensor field, a specific algorithm is necessary. Deterministic as well as probabilistic approaches are possible. Modeling a nerve fiber tract as a smooth curve following the main directions of diffusion would be an example in the first case [2], whereas in the second case a stochastic process [3] is justified by the fact that the macroscopic description of water diffusion is done in probabilistic terms and the diffusion measured in a voxel is an average of many different sub-compartments. Here a combination of both approaches is implemented and is used to tackle a morphological question. By means of a statistical study we explore the divergence of the nerve fiber tracts radiating from different positions on the sagital midline of the corpus callosum into the hemispheres.