A framework for modelling the regional variation of white matter microstructure

Introduction We present a natural framework for group-wise comparison of the regional variation of tissue parameters estimated from diffusion-weighted MRI (DW-MRI) over an anatomical region of interest (ROI). The regional variation model (RVM) fits curves of parameter variation directly to the DWMRI signals which captures the spatial changes in a natural way as well as reducing the effects of noise. Using statistical tests, we are able to localise regions of significant difference along the curves. Unlike traditional ROI analysis, we obtain a continuous representation of the parameter variation. Unlike VBM or TBSS [9], we detect differences in global trends of parameters and exploit the local coherence of parameter values in a natural way. We demonstrate the technique on the mid-sagittal corpus callosum (MSCC) in two distinct age groups and localise significant differences in the genu of the MSCC.