Histologically Informed Periodic Axon Substrate Generator for Time-Dependent Diffusion MRI

Motivation: Existing methods for constructing white matter numerical substrates suffer from geometric discontinuities at substrate boundaries. Goal(s): Develop a framework for realistic white matter substrates for diffusion MRI simulations that capture microstructure continuity across substrate boundaries. Approach: This study introduces a novel framework for ensuring microstructure continuity across boundaries by periodically adding reciprocal axons for items that cross or close to boundaries during substrate generation. Results: The method successfully constructs complex, biomimetic white matter substrates with geometric continuities in 3D, while achieving high packing densities (~74%) and controlled diameter distribution, orientation dispersion, and beading. Impact: The framework developed can facilitate more accurate Monte Carlo simulations without confounding effects of boundary artifacts for investigating microstructural contributions to diffusion MRI.