Finding tracks between ROIs: tckedit on WB tractogram or targeted tracking

tractography

#1

Dear MRtrix experts,

I am trying to find a method to extract tracks starting from one gray matter region, traversing a white matter ROI, to another gray matter region.

I’ve read a few posts related to this question, but it seems there is no one best way to do it.

Basically, I’ve read about three ways to do this:

  1. do whole brain tractography with dynamic seeding; then do SIFT; then do tckedit with include regions. (Finding tracts from one region to another region)
  2. use tckgen to do targeted tractography with ROI induced constraints. (Tckedit or tckgen for tract dissection)
  3. Merge targeted tracking pathways with whole brain tractogram; then do SIFT; then do tckedit with include regions. (Tckedit after targeted tracking)

I am wondering what are the differences among these methods, especially between method 1 and method 3 and which one I should use in my case.

Sorry if my questions are too basic and thank you in advance for your kind help!

Best,
Xiaoxiao


#2

So there’s a couple of different points to consider here: one is in the selection of streamlines, the other is the quantification of connection strength.

For selection of streamlines corresponding to a bundle, I tend to recommend the tck2connectome - connectome2tck approach that’s been described in other threads, as it’s more likely to result in the “expected” behaviour when selecting streamlines based on grey matter regions of interest (where ACT-terminated streamlines may not actually intersect a ROI if it is defined only inside grey matter voxels).

For quantification, only 1 and 3 permit quantitative interpretation of connectivity using the SIFT model. Though note that 3 only works with SIFT2: SIFT would simply remove the large number of streamlines within your pathway of interest as part of its optimisation process. If your pathway is reasonably large, I would suggest sticking with approach 1 for simplicity.

Rob


#3

Dear Rob,

Thank you for you kind reply.

Xiaoxiao