Developing an SNR-efficient tensor-valued diffusion encoding protocol for studying brain microstructural changes in...

Tensor-valued diffusion encoding (TDE) is an emerging diffusion MRI technique that uses advanced diffusion-encoding waveforms and modeling to provide enhanced specificity to brain microstructure compared with conventional Stejskal-Tanner pulsed-gradient encoding. However, the diffusion encoding duration to achieve the same b-value is substantially increased in TDE, resulting in relatively low SNR and spatial resolution (>2 mm isotropic). In this study, we developed an SNR-efficient, high-resolution TDE protocol with 1.8-mm isotropic resolution and clinically feasible scan time and evaluated its reproducibility and sensitivity to age-related microstructural differences. Eleven cognitively normal older adults (5F/6M, 62-71 years) and seven younger adults (3F/4M, 22-31 years) were scanned at 3T using an in-house TDE sequence incorporating an SNR-efficient multi-band multi-shot EPI readout and reconstruction. Quantitative diffusion metrics, including mean diffusivity (MD), fractional anisotropy (FA), microscopic FA (μFA), anisotropic mean kurtosis (MKA), isotropic mean kurtosis (MKI), and total mean kurtosis (MKT), were evaluated in the bilateral temporal parts of the cingulum bundles (CBT), bilateral fornix (FX), and global white matter (WM). Older participants underwent two scans to assess reproducibility using intraclass correlation coefficients (ICCs) and Bland-Altman analysis, and diffusion metrics were compared between younger and older groups using Wilcoxon signed-rank tests with false discovery rate (FDR) correction. All metrics except MKI demonstrate moderate-to-excellent reproducibility across all WM regions (ICCs≥0.64). Exploratory comparisons show age-related differences in FA, μFA, and MKA with p<0.05 before FDR correction, particularly in the fornix, although none remains significant after correction. These findings demonstrate the feasibility of SNR-efficient TDE at 1.8-mm isotropic resolution with approximately 70-mm brain coverage in approximately 12 minutes and establish its reproducibility for high-resolution brain microstructural imaging.