Introduction

Early blindness (EB) affects the optic radiation (OR), the component of the visual pathway conveying signals from the lateral geniculate nucleus (LGN) of the thalamus to the occipital cortex. Previous diffusion MRI studies^1,2 showed that the OR of EB individuals presents lower fractional anisotropy (FA) values compared to matched sighted controls. However, these studies either (i) averaged FA along the OR tract, or (ii) considered voxel-based analyses. As a result, the spatial distribution of FA changes along the OR in EB remains not well characterized. Here we investigate whether the FA profile along the OR is affected by blindness relative to sighted controls. The possible detection of localised patterns of changes might provide new insights into the mechanisms of plasticity in sensory-deprived individuals.

Methods

A group of 16 early-blind subjects (8 males, mean age 33.9, onset ≤ 3 years old) was investigated in comparison with a gender- and age-matched group of healthy sighted controls. A 4T Bruker Medspec MRI scanner was used to acquire diffusion (2.3mm isotropic voxel, 60 DW and 10 b0 volumes, b = 1500s/mm2 , TE = 99ms) and T1 anatomical (MPRAGE, 1 mm isotropic) data. Diffusion data were corrected for eddy currents and head motion (FSL) followed by bias field correction and global intensity normalization in MrTrix3³. Diffusion profiles (FODs) were estimated on upsampled DWI data from the group-specific average response function. Anatomically-constrained probabilistic tractography (ACT)⁴ was performed at following parameters: 0.5 mm step size, 45° maximum angle threshold, 1000 seeds/voxel. Tracking was randomly initiated from each voxel in the LGN and the primary visual cortex (V1) was used as target. The LGN was extracted from the Juelich atlas (threshold=25%), inflated (2mm) to account for inter-subject variability, and non-linearly registered to the subject’s diffusion space in FSL(FNIRT). V1 was automatically parcellated in Freesurfer(v5.3.0). Parcellations were visually inspected and, manually edited by an expert operator (author H.T.). Resulting tractograms were then filtered and transformed into TrackVis file format using nipype. The Along-Tract-Statistic Toolkit⁵ was used to estimate FA at multiple locations along the tract length: each tract streamlines were reoriented to a common origin and re-parameterized with cubic B-splines, resampled to new vertices in such a way that each streamline’s point closest to the midpoint vertex of the calculated mean tract streamline was identified, and each half of each streamline’s restant vertices fell respectively on their side of the midpoint vertex. The number of new vertices was calculated by dividing the length of each tract mean streamline by the voxel resolution, and then averaging it across subjects. Each subject’s FA map was then sampled at these new vertices and FA values were collapsed across streamlines at each group of correspondent vertices in order to get mean (M) and standard deviation (SD) FA estimates along the length of the tract. FA was also extracted by collapsing each individual’s along-tract FA values to compare Hedges’s g in this condition with Hedges’s g for each vertex showing significant FA differences between groups.

Results

For both hemispheres, significant effects of group (left: p=0.004, right: p=0.005) and vertex (left: p<0.001, right: p<0.001) were observed across all the tract length. Significant interactions of vertex and group were found for the right hemisphere (p<0.001, figure 1). Post-hoc t-tests show that verteces presenting significant group interactions (p<0.05, corr.) are segregated in two distinct segments along the tract course, and that they are localized respectively in the postero-central and posterior regions of the OR (figure 2, 3).

Discussion and Conclusion

These results confirm previous findings for decreased FA values within the OR of early-blind people, and shed new light on the spatial pattern characterizing such changes. Interestingly, FA changes appear to mainly affect two distinct segments of the tract course, and this might point to the occurrence of focal patterns of white matter reorganization occurring in specific tract positions. The localization of such phenomenon might provide new understandings of mechanisms of plasticity in sensory-deprived individuals, and might represent new valuable opportunities in the assessment of suitability of sight restoration.

Acknowledgements

We acknowledge Maxime Descoteaux and the SCIL for sharing the code.