Toward an accurate multi-fiber assessment strategy for clinical practice

Multi-tensor models provide information about the fiber bundles underlying characteristics and are of great interest for clinical applications. In this work we propose both a novel model fitting procedure and a novel acquisition scheme for multi-tensor assessment. Our fitting procedure includes the estimation of the diffusion of free water and ensures non-degenerate and regularized tensors. Our acquisition scheme combines spherical and cubic sampling. It enables the acquisition of multiple non-zero b-values which has been shown to be required for the full multi-tensor estimation. It enables high b-values to be acquired while achieving the same low echo time as a single-shell HARDI, leading to comparable geometrical and intensity distortion. We evaluate our resulting novel strategy CUSP-MFM (CUbe+SPhere Multi-Fiber Model) on both synthetic and clinical data. We show the ability of CUSP-MFM of assessing complex fiber structures from short duration acquisitions, compatible with clinical routine.