High-resolution diffusion-weighted imaging at 7 Tesla: single-shot readout trajectories and their impact on signal-to-noise ratio, spatial resolution and accuracy

Diffusion MRI (dMRI) is a valuable imaging technique to study the brain in vivo. However, the resolution of dMRI is limited by the low signal-to-noise ratio (SNR) of this technique. Various acquisition strategies have been developed to achieve high resolutions, but they require long scan times. Imaging at ultra-high fields (UHF) could further increase the SNR of single-shot dMRI; however, the shorter T2* and the greater field non-uniformities will degrade image quality. In this study, we investigated the trade-off between the SNR and resolution of different k-space trajectories, including echo planar imaging (EPI), partial Fourier EPI, and spiral, over a range of resolutions at 7T. The effective resolution, spatial specificity and sharpening effect were measured from the point spread function (PSF) of the simulated diffusion sequences for a nominal resolution range of 0.6-1.8 mm. In-vivo scans were acquired using the three readout trajectories. Field probes were used to measure dynamic magnetic fields up to the 3rd order of spherical harmonics. Using a static field map and the measured trajectories image artifacts were corrected, leaving T2* effects as the primary source of blurring. The effective resolution was examined in fractional anisotropy (FA) maps. In-vivo scans were acquired to calculate the SNR. EPI trajectories had the highest specificity, effective resolution, and image sharpening effect, but also had substantially lower SNR. Spirals had significantly higher SNR, but lower specificity. Line plots of the in-vivo scans in phase and frequency encode directions showed ~0.2 units difference in FA values between the different trajectories. The difference between the effective and nominal resolution is greater for spirals than for EPI. However, the higher SNR of spiral trajectories at UHFs allows us to achieve higher effective resolutions compared to EPI and PF-EPI trajectories.