Multi-tissue constrained spherical deconvolution with spherical mean density estimation

Motivation: Multi-tissue constrained spherical deconvolution (MT-CSD) average densities depend on the local fibre arrangement, as it relies on the non-negativity of the fibre orientation density functions (fODF). Goal(s): We investigate whether MT-CSD can be performed without influence of the orientation structure of fibre ODF. Approach: Using the spherical mean of each b-value shell to factor out the angular structure, we rely purely on the different tissue types' b-value dependence, with the angular part of the fODF estimated subsequently . Results: The method produces plausible, but subtly different results from 'regular' MT-CSD, which appear to be less affected by fibre arrangement. It is also substantially faster. Impact: The ability to decompose the signal based purely on its spherical mean b-value dependence may provide more reliable density estimates, though this needs further investigation. The computational efficiency of this approach may also make it more suitable for real-time use.