On the origin ofR₂orientation dependence angle offsets in white matter

Abstract

Purpose: to identify the origin of confounding angle offset ε_0 in R2 orientation dependence and to propose a novel framework for better characterizing anisotropic R2 in brain white matter (WM). Methods: Anisotropic (ε) rather than principal diffusivity direction (Φ) was theorized along axon fiber, with ε determined by all eigenvalues and eigenvectors from diffusion tensor. An extra parameter ε_0 was introduced into generalized magic angle effect function to account for any offset in R2 orientation dependence derived from T2-weighted image (b=0). These dependences referenced by ε were compared to those referenced as usual by Φ at b-values of 1000 and 2000 (s/mm2) on both linear and planar tensor image voxels in WM, based on public domain ultrahigh-resolution (760 μm3) Connectome DTI datasets of a healthy young adult brain. A Student's t-test was used to assess the mean differences and the statistical significance was considered at P < 0.05. Results: Fitted ε_0 became zero if referenced by ε or nonzero if referenced by Φ, signifying the origin of ε_0. Nonzero ε_0 relied on b-values and tensor shapes so did other model parameters, e.g., the amplitude of anisotropic R2 (1/s) significantly increased when using a higher b-value for the linear tensor image voxels, i.e., 3.3 ± 0.1 vs. 1.8 ± 0.1, P < 0.01. Conclusion: The origin of R2 orientation dependence angle offsets has been identified and the combined anisotropic diffusion and transverse relaxation models have fully quantified R2 orientational anisotropies, thus providing novel insights otherwise unattainable on microstructural alterations in WM.