Fast Monte Carlo simulations replace analytical tissue models in diffusion MRI

Introduction Characteristics of tissue micro-structure, such as axonal diameters (d) and membrane permeability (p), can be estimated from diffusion MRI data obtained with different diffusion times [1,2]. Micro-structural parameters are currently extracted using analytical models, based on approximations of the tissue geometry and of the measurement on that geometry [3,4]. For example, the intracellular signal is modelled using either the short-gradient-pulse (SGP) or the Gaussian-phase-distribution (GPD) approximation [5]. As a consequence, analytical models suffer from crude approximations. In this work, we propose a novel approach to the extraction of microstructural parameters from diffusion MRI data without the use of analytical models, by using Monte Carlo simulations of particles diffusing in tissue geometries.