Retinal tractography: Influence of diffusion acquisition, ocular laterality and ocular dominance in healthy subjects

Abstract

Purpose Exploration of the human retina has never been described with MRI tractography, supposedly because of ocular movement-related artifacts and imaging distortions with EPI diffusion acquisition. In this work, we used track-weighting imaging to assess retinal fascicles. Diffusion acquisition parameters influence was studied and the relationship between variability and ocular laterality and ocular dominance was explored. Materials and Methods Twenty healthy subjects underwent MRI, with retinal tractography calculated with the constrained spherical deconvolution model. A DWI protocol with b value of 1000 s/mm 2 and 60 diffusion weighting directions was structured in 4 scans of 15 directions, and b = 0 s/mm 2 value repeated at each scan to ease motion estimation and correction. At post-processing, we merged subsets of 15 directions according to varied combinations. We generated whole-brain tractography with 10 million streamlines based on 30, 45 and 60 diffusion weighting directions. Ocular dominance was determined using a hole-in-a-card test. Results Retinal nerve fiber layer appeared to have a heterogeneous distribution, with a fascicle-like display on tractography. Diffusion acquisition required at least 45 directions to display retinal fascicles. There was no significant difference on the number of fascicles displayed between 45 and 60 direction acquisitions, whereas there were significantly more fascicles using 60 and 45 directions than with 30 directions ( P ≤0.001). Regarding ocular dominance (OD), we observed there were more superior fascicles on the right OD group than on the left one ( P ≤0.05), whereas there were more inferior fascicles on the left OD group ( P ≤0.05). Conclusion Although the distribution of reconstructed retinal fascicles depends on the number of diffusion weighting directions used, assessment of ocular laterality and dominance–related variability seems possible.