How to incorporate ACT framework when including small brainstem substructures?

1

Hi MRtrix team,

I’m working on a connectomics pipeline that includes small brainstem substructures (e.g., reticular nuclei, arousal-related nodes), and I’m wondering whether there is a feasible way to incorporate the ACT framework in this context.

My understanding is that the standard 5TT-based ACT segmentation is suboptimal for complex tracking involving subcortical regions, as streamlines are allowed to propagate into subcortical GM, but terminate within and cannot exit.

My questions are:

  1. Is there a recommended way to adapt or extend the ACT framework to accommodate small brainstem ROIs? For instance, is it possible to edit the 5TT image to reclassify certain voxels as white matter (WM) or to assign custom labels to preserve tracking through them?

  2. Alternatively, can I partially use ACT (for example, for cortical and subcortical boundaries) while relaxing constraints in specific brainstem regions?

  3. Would combining ACT with custom WM masks or seeding masks that include these small regions help preserve tracking fidelity without violating ACT constraints?

I’d love to use ACT for its anatomical precision, but I’m struggling with how to make it compatible with brainstem-focused tractography. Any guidance, workarounds, or examples would be incredibly helpful!

Thank you again for all your support and for building such a flexible toolset!

— Monica

2

Hi Monica,

Here are some ideas…

  1. Appropriate modelling the 5TT image is highly dependant on the research question, and yes there is a way to modify the 5TT image (check out 5ttedit). mrgrid might come in handy here if the small brainstem substructure has different dimensions to the 5TT image (along with the -template option)
  2. Not sure how this would work, but (1) will permit streamline behaviour manipulation
  3. Our lab has previously done something like this - check out the following links

Hope this helps!

Cheers,
Arkiev