Tract-of-interest segmentation following FBA pipeline completion

1

Hi,

I am attempting to derive average FBA metrics from a tract of interest and I have realized there are a couple of ways of going about this. I have just completed whole-brain tractography from 61 subjects using the FBA pipeline for MSMT-CSD(Fibre density and cross-section - Multi-tissue CSD — MRtrix3 3.0 documentation) I am wondering which technique is the most efficient and powerful.

So far, I believe my two options are either:

  1. performing tractography (tckgen) on my FOD template using inclusion and exclusion regions and paramters and then mapping the tractogram of my bundle of interest to my fixel mask.
  2. using TractSeg to derive tract masks by segmenting my FOD template and then applying tck2fixel to the streamlines created from the TOMs (the TOMs created by tractseg are changed into FODs right?)

In short, I am looking for some guidance on what my next steps should be now that I have my whole-brain tractography and wish to complete ROI analysis on specific tracts.

2

Hi,

Those choices might depend on the tracts you are interested in and on the precision you’re looking for, or how many you want to analyse. TractSeg might be the simplest and most convenient if it already provides the tracts you are looking for with successful segmentation. But it may not let you segment all tracts/fibres you’d like to look at. TractSeg’s Tracking function will get tracks from the TOMs and tck2fixel will return a fixel mask in the same space as the FOD template.

If you’re looking to deal with many tracts and have an atlas, you could also warp it to your FOD template and play with tck2connectome/connectome2tck, which allows you to extract tracks connecting specific combinations of regions.

@rsmith did a summary of various ways to build bundle of interest and I think they should also work for a FOD template:

3

Thank you! Much appreciated