Ch Khodaverdian
H Soltanian-Zadeh

Estimation of multiple fiber tracts with different orientations within a voxel has been an important problem in white matter fiber tractography using Diffusion Tensor Magnetic Resonance Imaging (DT-MRI) data. In order to solve this problem, based on the fact that DTI signals have super-Gaussian pdf's, fast Independent Component Analysis (ICA) is applied to decompose the original DTI signal from voxels containing multiple fibers into its non-Gaussian components. In this paper, we present a new method for tractography, which considers the eigenvectors of the independent components simultaneously. The voxels with crossing fibers are identified based on the orientation difference, number of fibers coexisting in a voxel, and the high parallel anisotropy value. Results show more reconstructed fibers by the proposed method compared to the previous methods.

Fiber tracking, Diffusion Tensor Imaging, Independent Component Analysis, FACT