I’m trying a simple creation of tract in the white matter mask, but for some reason it looks weird…
I used multishell data with filed map correction, and the function: tckgen -act ${SUB}_5tt_coreg.mif -backtrack -seed_gmwmi ${SUB}_gmwmSeed_coreg.mif -nthreads 8 -maxlength 250 -cutoff 0.06 -select 10000000 ${SUB}_wmfod_norm.mif ${SUB}_tracks_10M.tck
image attenched
My best guess based on the screenshots is that you have a gross misalignment between the DWI and anatomical images. In the coronal image in top left, it looks like there are red streamlines attempting to traverse the corpus callosum but are instead within the cingulum bundles relative to the anatomical segmentation. Try e.g. loading your DWI image as the main image in mrview, and the 5TT image in the Overlay tool.
If the alignment looks correct, then try the dwigradcheck tool in case it’s a permutation of the X and Y components; but from the sagittal image at top left I don’t think that’s the case.