I am currently trying to look at structural underpinnings of some functional connectivity findings between two regions that are indirectly connected, i.e. a polysynaptic pathway. Is there a way to probe these paths with tractography? In other words, is it possible to do ROI to ROI tractography between two regions that are not directly connected? I brought this issue up to several people at OHBM and the only answer I got was tracer studies.
My second question is along the same lines. The only idea I have had to solve this problem is to look at graph space and calculate something like path length between the two ROIs. For example, if two regions have a path length of 3, looking at which 3 connections make up this path and then bringing this back into anatomical space to looks at the tracts involved. Is there a good way to do this with MRtrix (maybe along with the BCT)?
Any suggestions or guidance is much appreciated. This is an issue I’ve been mulling over for a long time now.
Thanks so much for your time,
Looking at the connectome matrix is certainly one way you could go about it. Though I would avoid thinking about the situation as having an integer path length and a fixed set of connections. With tracking as implemented in MRtrix3, the connectome matrix is almost fully connected; therefore one must instead think about all possible “polysynaptic” paths, and the strengths of the connections along each path. Precisely how to interrogate / describe / quantify this is however left as an exercise to the reader.
An alternative approach I’ve used once before that you’re free to try (not sure if it’s been described elsewhere before, so give me credit if it hasn’t!), is:
Seed from your starting region, generating some number of tracks.
tckmap on the resulting tracks with the
-ends_only option, to get a map of the density of streamlines terminations (this includes both start and end points, which isn’t ideal for this particular experiment, but it’ll suffice for a proof of concept). Also use the
-crop_at_gmwmi option if using ACT.
Provide the resulting termination density map in
tckgen using the
-seed_rejection option. This will seed new streamlines in those locations where streamlines from the previous iteration terminated. By using a rejection seeder, a voxel in which a large number of streamlines terminated in the previous iteration will be used to seed streamlines more frequently than a voxel in which few streamlines terminated.
Go back to step 2.