Microscopic Propagator Imaging with diffusion MRI

Microscopic Propagator Imaging (MPI) is a novel diffusion MRI technique that estimates properties, referred to as indices, of the microscopic propagator. This is the probability distribution of water displacements within tissue microstructures. Unlike conventional mean apparent propagator methods, MPI is designed to minimize the sensitivity of indices to mesoscopic confounds such as axonal orientation dispersion, yielding diagnostic maps that more directly reflect presence, integrity, and shape of microstructures rather than their directional arrangement within the voxel. The method is implemented as a machine learning framework that exploits zonal relationships among spherical harmonic coefficients of multi-shell diffusion data to map such data to the microscopic propagator indices. Applied to human brain data, MPI yields reliable voxelwise estimates, with resulting maps exhibiting expected spatial patterns and systematic differences relative to the corresponding mean apparent propagator indices. These findings suggest that MPI provides microscopic-specific and complementary information beyond classical propagator methods, with potential to improve the characterization of brain tissue microstructure.