Background: Conventional diffusion MRI employs diffusion encoding along three orthogonal directions in 3D space using a set of motion probing gradients. Multidimensional diffusion encoding (MDE) extends this concept by encoding diffusion by using non-orthogonal gradient directions or additional gradient directions beyond the three spatial dimensions. Recently developed MDE technique, diffusional variance decomposition (DIVIDE), may provide more detailed insights into tissue microstructure by capturing diffusion anisotropy and heterogeneity in a more comprehensive manner in complex tissue environments, such as crossing or kissing fiber configurations in the brain. This study aimed to evaluate the feasibility of the DIVIDE imaging for the human brain.
Materials and Methods: Ten healthy subjects (8 men; mean age 29 years, range 23-37 years) were included in this study. During March 2021 to June 2021, whole-brain diffusion-weighted imaging was performed using MDE twice with an interval of more than 1 week using a 3T scanner with a 48-channel head coil. The DIVIDE sequence is a modification of the spin-echo 2D echo-planar imaging (EPI). Imaging parameters were as follows: TR/TE = 5200/96.6 ms, FA = 90°, 60 slices without gaps, FOV = 256×256 mm2, spatial resolution = 2 × 2 × 2.5 mm2, bandwidth = 3906 Hz/pixel, parallel imaging factor = 2 (anterior-posterior), partial-Fourier factor = 0.75 . MDE encoding was performed with 29 linear and 26 spherical b-tensors in an interleaved fashion using optimized gradient waveforms for b = 100, 1000, 2000 s/mm² with 4, 10, 15 directions, respectively, for linear encodings and with 6, 10, 10 directions, respectively, for spherical encodings, giving a total scan time about 5 min. Automated extraction of 20 major regions of interest (ROI) based on JHU DTI-based white-matter atlas was performed to obtain DIVIDE metrics including fractional anisotropy (FA), mean diffusivity (MD), axial diffusivity (AD), radial diffusivity (RD), anisotropic mean kurtosis (MK_A), isotropic mean kurtosis (MK_I), total mean kurtosis (MK_T), and microscopic fractional anisotropy (_μFA). The coefficient of variation (CV) and interclass correlation coefficient (ICC) were calculated to assess the intra-subject (scan-rescan) and inter-subject reproducibilities of the regional DIVIDE metrics.
Results: The intra-subject CV was less than 5% in 20/20 ROIs for FA, 20/20 for MD, 20/20 for AD, 19/20 for RD, 18/20 for MK_A, 13/20 for MK_I, 19/20 for MK_T, 20/20 for _μFA. The intra-subject CV was lower than that of inter-subject CV in all the metrics of all the regions. The ICC showed almost perfect agreements for all of the regional values of FA, MD, AD, RD, MK_A, MK_T, _μFA. The ICC showed substantial to almost perfect agreement in 17/20 ROIs for MK_I.
Conclusion: Scan-rescan repeatability was acceptable in all the metrics. DIVIDE may be reliably used for measuring diffusion metrics.

Acknowledgements

No acknowledgement found.