Fixel processing in marmoset brain images

Mrtrix developers,

I have in-vivo, multishell marmoset data that I want to perform an FBA on. I have manual segmentations of the WM and GM voxels in all specimens and have done msmt FOD estimation. The WM FODs in the population template are very strong in the cortex (see screenshot).

After I do streamlines+sift+connectivity and then fixelcfestats the white matter fixels are filtered out of the results (presumably due to the lack of streamlines connectivity warning).

The cortex and GM structures are not of interest, can I perform streamlines tractography on just the WM voxels masked with a template tissue mask?

Related question (since I am having a similar issue with human data) how can I debug the “no streamlines connectivity” warning message in fixelcfestats? What criteria are used to filter fixels for stats?

Thanks in advance,

Chris Adamson.