Creating a Connectome of Regions Between Injections and Atlas ROIs


First off, I want to say thank you for making such comprehensive software! I’m new to comp neuro, so it was really helpful to have a package that can do most of the things I needed, and give guides to where else I can go :slight_smile:

I’m currently working on with a monkey dataset, where I have tracts generated from tract-tracing in .tck files, and injection sites that specify where the tract-tracing was done from in a Nifti .nii.gz format. I also have an atlas for different regions of the brain I want to explore, which is in a .nii file. My question is, if I want to study the connection strength from each injection site to each ROI site, is there any way for me to do this? I know the tck2connectome command exists, but that’s typically used with a single atlas to make a square connectivity matrix, and I’m more so wondering if there was any way or command I can use to pass in both an atlas and the injection sites (as a mask with all of the injection sites, or something), and get the connectivity strength between injection site ↔ ROI rather than between atlas ROI ↔ ROI.

I hope this makes sense! The only way I can think of doing this is to maybe create a mask of each injection site with each atlas ROI and pass that to the tck2connectome command, and then repeat that for every combination (which would take a long time), or make a mask of all injection sites and the atlas, then passing that to tck2connectome (which would also take a long time, and I believe give a much bigger matrix than I want).

Thank you for help!

Hi Hana,

I think tckedit can help you. :grinning:

  1. Generate a whole-brain tractography.
  2. Run tckedit on the whole-brain tractography to select streamlines that pass the site and the assigned ROI with two -include xxx.
  3. Repeat this step to get results between the site and different ROIs.
  4. Use tckinfo to check the number of streamlines between the site and different ROIs, which measures the connective strength.

Besides, I think your description is somewhat similar to lesion network mapping, maybe you can take a look. :stuck_out_tongue_winking_eye: