High-Resolution Whole-Brain Diffusion Tensor Imaging Exploiting Rapid Single-Slab 3D EPI Strategy.

Abstract

The purpose of this work is to investigate the feasibility of high-resolution whole-brain diffusion tensor imaging (DTI) using a rapid single-slab 3D pseudo-random EPI encoding strategy with physical constraints. A spin-echo-based diffusion-weighted imaging was modified to incorporate both single-slab 3D segmented EPI for high-resolution diffusion imaging and unsegmented EPI with short readouts for segment-specific motion-induced phase navigation. A physically constrained, segment-wise grouped phase encoding strategy is introduced, yielding a rapid, pseudo-random traversal of $k$ -space with smooth signal transition in local neighborhood even in the presence of magnetic field inhomogeneities. Numerical simulations and in vivo studies were performed to validate the feasibility of the proposed method for high-resolution whole-brain DTI. The proposed method exhibits a robust point spread function (PSF) even in the presence of magnetic field inhomogeneities and produces a clear depiction of DTI parameter maps from highly incomplete measurements (reduction factor = 5.5). Furthermore, the proposed method outperforms the conventional 2D single-shot EPI and the conventional simultaneous multislice EPI due to its robust PSF, high encoding efficiency, and high signal gain. We successfully demonstrated the rapid single-slab 3D pseudo-random EPI encoding strategy with physical constraints, which makes it possible to achieve high-resolution (1.0 mm$^{3}$) single-slab 3D DTI roughly in 14 minutes without apparent artifacts and noise. This is the first work that prospectively demonstrates a rapid, physically constrained pseudo-random EPI strategy for high-resolution single-slab whole-brain DTI.