Shining Light on Degeneracies and Uncertainties in the NEXI and SANDIX Models with µGUIDE

Motivation: New biophysical models of gray matter microstructure have been introduced, with a particular focus on exchange time and soma size estimations. However, the fitting quality of these models has not been studied. Goal(s): Our goal is to study the feasibility of estimating both exchange time and soma size in a clinical setting. Approach: We applied µGUIDE, a Bayesian inference framework, to quantify the quality of the fitting of two models from the literature, NEXI and SANDIX, using an extensive protocol, and a clinical one. Results: Estimations of both exchange time and soma size in clinical setting shows high uncertainty. Impact: For the first time, we are using µGUiDE on two microstructure models that take into account the exchange between neurites and the extra-cellular space. We applied it to both synthetic and clinical data.