Tckedit after targeted tracking

5

I assume that means I’d also need to obtain denser sampling for the WB tractogram, again by smaller step size or upsampling? … Any suggestions there?

Perhaps better would be:

  1. Upsample streamlines data corresponding to pathway of interest after tckgen, or run tckgen for that pathway using a smaller step size.

  2. Run tckedit to refine selection of the pathway of interest.

  3. Downsample the streamlines data for the pathway of interest, so that the step size matches that of the whole-brain data.

  4. Concatenate pathway with whole-brain fibre-tracking data.

  5. Run SIFT2.

When concatenating track files, it’s preferable for them to have the same step size, as otherwise the output file from the concatenation process will flag that the step size is ‘variable’, preventing automated calculations of certain parameters down the line. Over and above that, the calculated length of a streamline (and the length of its intersection within each voxel traversed) can in fact change depending on the step size, hence it would be preferable for it to be equivalent between the targeted pathway and the whole-brain tractogram.

Regarding a “precise” option for tckgen/tckedit, I like the idea.

I’ve added it to GitHub issues. But it’s probably not likely to make its way to the top of the priority list any time soon. The list of issues is much longer than the list of people addressing them :ocean: :raising_hand_man:

Any plans for GPU-enabled functions?

As much as @jdtournier and I would love to nerd out by porting various MRtrix3 capabilities to GPU, it simply fails to take priority over fixing bugs / providing new features. So unless a third party chooses to take this on, it’s unlikely.