Multi-Fascicle Model Reconstruction from Acquisitions at a Single b-value with a Population-Informed Prior

Purpose. Diffusion tensor images (DTI) have been widely used to characterize the white matter microstructure. The incapacity of DTI to represent crossing pathways has motivated the developement of novel diffusion models, such as the multi-tensor models. However, to estimate a multi-tensor model, one needs to acquire the diffusion weighted images (DWI) at multiple b-values, unlike the common single-shell HARDI acqusition that has been used for years. This incompatibility comes with a high financial burden since studies based on single-shell HARDI acquisitions should be renewed and completed with novel acquisitions. In this study, we propose a method to retrospectively reconstruct a multi-tensor model from single-shell HARDI data, based on a multi-tensor atlas.