Probabilistic normative maps of the corticospinal and cortico-reticulospinal tracts in streamline and volumetric formats

Abstract

Background

Normative maps of brain tracts are key tools for assessing the extent of tract injury and plasticity after brain lesions such as stroke. The conventional procedure for tract mapping reduces diffusion tractography streamlines to a volumetric image before warping them to a standard coordinate space. Unfortunately, this volumetric reduction discards information about tract connectivity across voxels, complicating estimation of tract injury from the resulting map. This issue could be addressed by using a streamline-based tract map, but prior efforts to do so using group-based tractography have not accounted for individual variability, and thus are at risk of underestimating tract extent.

New Method

Direct streamline normalization (DSN) directly warps individual tractography streamlines to a standard coordinate space, which can be pooled across individuals to account for individual variability. Here, DSN was used to create normative tract maps for the first time.

Results

Novel probabilistic maps of the corticospinal tract (CST) and cortico-reticulospinal tract (CRST) were generated in streamline format using DSN.

Comparison with Existing Methods

Compared with group-based tractography, tract maps generated with DSN had 3.1–5.2 times greater volume, 1.0–1.5 times greater mean diameter and 7.8–15.3 times greater cortical origin area.

Conclusions

DSN produces tract maps with more accurate representation of axonal geometry than the conventional procedure and greater incorporation of individual variability than group-based tractography, likely resulting in larger maps. Future studies can use the CST and CRST maps generated here to optimize tract injury and plasticity measurements, which may provide an improved understanding of motor impairment and recovery.