Incorporating the effect of white matter microstructure in the estimation of magnetic susceptibility in ex-vivo mouse brain

Abstract

Accurate estimation of microscopic magnetic field variations induced in biological tissue can be valuable for mapping tissue composition in health and disease. Here, we present an extension to Quantitative susceptibility mapping (QSM) to account for local white matter (WM) magnetic microstructure by using our previously presented model for solid cylinders with arbitrary orientations to describe axons in terms of concentric cylinders. Through computer simulations, we find that our model improves susceptibility estimation compared to QSM, and Susceptibility Tensor Imaging (STI) are substantially biased by the unaccounted-for structural anisotropy due to the mesoscopic frequency contribution, indicating the observed STI tensor might not represent susceptibility anisotropy as expected. Experiments in mouse brains acquired at ultrahigh field shows the mesoscopic contribution due to WM microstructure to be substantial. This in turn changed estimated susceptibility values up to 25% root-mean-squared-difference in WM compared to standard QSM. Our work underscores how microstructural field effects impact susceptibility estimates, and should not be neglected when imaging anisotropic tissue such as brain WM.