Generalised Hierarchical Bayesian Microstructure Modelling for Diffusion MRI

Microstructure imaging combines tailored diffusion MRI acquisition protocols with a mathematical model to give insights into subvoxel tissue features. The is typically fit voxel-by-voxel the image least squares minimisation voxelwise maps of parameters relating microstructural features, such as diffusivities and compartment fractions. However, this fitting approach susceptible noise, which can lead erroneous values in parameter maps. Data-driven Bayesian hierarchical modelling defines prior distributions on learns them from data, hence reduce noise effects. has been demonstrated for microstructure MRI, but only few, relatively simple, models. In paper, we generalise wide range multi-compartment models, models Markov chain Monte Carlo (MCMC) algorithm. We implement our method by utilising Dmipy, software package data. Our code available at github.com/PaddySlator/dmipy-bayesian.