Running title:
Simulation of Diffusion in the CortexPurpose: Diffusion MRI of the cortical gray matter has shown to be sensitive to diseases such as Alzheimer's disease and multiple sclerosis. Additionally, there is a growing interest in remodelling, characterization and improvement of the understanding of different brain areas using a combination of quantitative MRI or 3D microscopy. The purpose of this study was to quantitatively characterize the cytoarchitecture of cortical tissue from fluorescence microscopy in order to simulate diffusion in the cortex and to better understand its diffusion signal characteristics Methods: Cellular volume fractions, and sizes of glial nuclei, and neurons were measured from 3D microscopy by volume segmentation and non-linear least squares ellipsoid fitting.Diffusion of water molecules in the cortex was simulated for variable permeability levels, free diffusion coefficients, cell (neuron, glia, dendrite or axon) densities, and sizes and diffusion times. Some parameters such as axonal density were extracted from the available literature. In an alternative method, the simulation geometry was directly extracted from available microscopy data of the cortex. The simulations were compared with fractional anisotropy, apparent diffusion and kurtosis values reported in vivo in the healthy brain.
Results:Assuming a permeability of around 20-30 µms -1 and a free diffusion coefficient of around 1.6-1.7 µm 2 ms -1 , the simulated fractional anisotropy, diffusion coefficient and kurtosis of the cortical gray matter were on average around 0.2-0.4, 0.7-1.3 μ m 2 ms −1 , and 0.1-1.0, respectively, which are in ranges in agreement with previous in vivo reports.
Conclusions:The Monte Carlo simulations of different layers or areas of the healthy cortex, and extrapolation of these simulations to other cell sizes or densities is an additional step in better quantitative characterization or remodelling of the healthy or diseased cortex from diffusion MRI.