Streamlines produced by ACT expected to start/end in the GM?

,
1

Hi,
I am trying to understand the effect of tckgen (called with iFDOD2+ACT) and tckedit.

Am I understanding correctly that tckgen called with iFDOD2+ACT, e.g.

tckgen \
  -algorithm ifod2 \
  -act "${5ttgen_fname}" \
  -backtrack \
  -seed_grid_per_voxel "${gmwmi_fname}" ${grid_size} \
  -nthreads ${n_threads} \
  -minlength ${min_length} \
  -maxlength ${max_length} \
  -select ${strml_count} \
  ${fod_fname} \
  ${tractogram_fname}

should produce a tractogram with streamlines ending/starting in the GM, regardless of the step size employed for propagation?

I have computed the binary mask resulting from the union of the cortical and subcortical GM from the result of the 5ttgen voume, and visualizing cgm_sgm_mask.mif seems to have correctly included all cortical and subcortical GM.

I have then called tckedit on the tractogram with that mask and not all of the streamlines are kept:

mrconvert 5ttgen.mif -coord 3 0 tmp_cgm.mif
mrconvert 5ttgen.mif -coord 3 1 tmp_sgm.mif
mrcalc tmp_cgm.mif tmp_sgm.mif -add | mrthreshold - cgm_sgm_mask.mif

tckedit ${in_tractogram_fname} ${out_tractogram_fname} -include cgm_sgm_mask.mif

What am I missing here?

Thank you.