Along-Tract Parameterization of White Matter Microstructure using Medial Tractography Analysis (MeTA)

Diffusion MRI tractography is a noninvasive method to estimate the structural connectivity of white matter (WM) bundles (tracts) in the human brain, which can help us understand brain function and neurodegenerative diseases. Existing techniques for analyzing WM microstructure along the length of bundles often require registering all individuals into a common space that may ignore potentially key differences in the shape and alignment of the tracts. We propose the Medial Tractography Analysis (MeTA) method to reduce partial voluming and microstructural heterogeneity in dMRI metrics while retaining bundle shape and capturing the regional variation within bundles. We performed reliability, compatibility, and disease-based validations. MeTA showed moderate to good overall overlap for most bundles in a test-retest dataset and preserved regional compatibility when applied to a dataset of subjects scanned with both high and low angular resolution protocols. Diffusion tensor imaging (DTI) metrics along the length of MeTA bundles had strong associations with cognitive impairment in ADNI. MeTA may be a reliable approach to identify regional abnormalities in clinical populations across multiple diffusion acquisitions.