SPECTRE-A novel dMRI visualization technique for the display of cerebral connectivity

The visualization of diffusion MRI related properties in a comprehensive way is still a challenging problem. We propose a simple visualization technique to give neuroradiologists and neurosurgeons a more direct and personalized view of relevant connectivity patterns estimated from clinically feasible diffusion MRI. The approach, named SPECTRE (Subject sPEcific brain Connectivity display in the Target REgion), is based on tract-weighted imaging, where diffusion MRI streamlines are used to aggregate information from a different MRI contrast. Instead of using native MRI contrasts, we propose to use continuous template information as the underlying contrast for aggregation. In this respect, the SPECTRE approach is complementary to normative approaches where connectivity information is warped from the group level to subject space by anatomical registration. For the purpose of demonstration, we focus the presentation of the SPECTRE approach on the visualization of connectivity patterns in the midbrain regions at the level of subthalamic nucleus due to its importance for deep brain stimulation. The proposed SPECTRE maps are investigated with respect to plausibility, robustness, and test-retest reproducibility. Clear dependencies of reliability measures with respect to the underlying tracking algorithms are observed.

Automated summary

The study introduces a simple visualization technique named SPECTRE for estimating connectivity patterns from diffusion MRI. Based on tract-weighted imaging, it uses continuous template information as the underlying contrast for aggregation. The research focuses on visualization of connectivity patterns in midbrain regions essential for deep brain stimulation.