Full-brain Q-Ball Imaging in a Clinically Acceptable Time : Application to White Matter Fibre Tractography

Introduction Q-Ball Imaging (QBI) has been presented by Tuch et al. as a model-free method for measuring the diffusion orientation distribution function (ODF) [5]. QBI yields a surface ( ,φ) that reflects the underlying fibre structure, even in the presence of subvoxel partial volume averaging of fibre directions. This surface can potentially be utilized for fibre tract reconstruction without imposing a limit on the number of fibre directions in a voxel, performing peak extraction, or imposing a model for the shape of the diffusion displacement pdf. The downfall of the protocol suggested by Tuch et al. is that it requires measurement of 484 diffusion-weighted images, and therefore requires a discouragingly long acquisition time for the full brain (approximately one hour). In this study we investigate the feasibility of using fewer (90) diffusion weighted images to obtain full-brain high angular resolution images of the function , and use this surface to perform fibre tracking via a surface evolution-based tractography algorithm [1]. Methods MRI data were acquired on a Siemens 1.5T Sonata MR scanner (Siemens Medical Systems, Erlangen, Germany) using an 8-channel phased-array head coil. Diffusion encoding was achieved using a single-shot spin-echo echo planar sequence with twice-refocused balanced gradients. Four coregistered datasets were acquired, each consisting of 90 diffusion weighted images using isotropically spaced gradient directions, 2.8mm isotropic resolution, b=3000 s/mm, q=0.35 m, TR=8s, TE=110ms, and 30 slices. All diffusion scans were cardiac gated to reduce pulsatile motion artifacts. Each acquisition of 90 diffusion weighted images (DWIs) took approximately 12 minutes, for a total scan time of 48 minutes for the four datasets. The angular resolution achieved with this protocol is approximately 19. A 1mm isotropic resolution T1 weighted anatomical scan was also acquired (TR=22ms, TE=9.2ms, α=30). For each separate 90-DWI dataset, we calculated ( ,φ) using the Funk-Radon transform, as described in [5]. For visualization and for fibre tracking, we normalized the values of to have a mean of unity in each voxel. We defined an anisotropy index based on the fractional anisotropy (FA) [4] commonly used in diffusion tensor (DT) experiments. Our generalized fractional anisotropy (GFA) is the normalized standard deviation of the squared diffusion ODF measurements,