how to build reliable and nice tracks

Dear Mrtrix Experts,
I would like to build tracks ifof CST OR etc based upon some ROIs I have manually made. I have tried 2 ways: with and without -act options
these are the actual commands I am using in the first case

dwi2response dhollander $basname_dti outresponse.txt out2.txt out3.txt -shell $shell -fslgrad $basname_bvec $basname_bval -mask mask.nii -force

dwi2fod msmt_csd $basname_dti outresponse.txt FOD_2.nii -fslgrad $basname_bvec $basname_bval -mask mask.nii -debug -force

tckgen FOD_2.nii CSD_{track_string}_{numberoftracks}.tck -seed_image ROI0 -mask mask.nii -angle 40 -select $numberoftracks -force

tckedit -include ROI2 -exclude ROI3 CSD_{track_string}_{numberoftracks}.tck CSD_{track_string}_{numberoftracks}_roi_new_refined.tck -force

in this case I am getting too disperse tracks which are not good enough for a paper figure. any hints to make less scattered fibers?
In the second way these are the commands:

tt=5tt.mif

if [ -e {tt} ] then echo "file {tt} exists: I do nothing"
else
echo “file ${tt} does not exist: computing it”
flirt -in $basname_mprage -ref diffusion_b0_mean.nii -dof 6 -omat T12dwi.mat
transformconvert T12dwi.mat basname_mprage diffusion_b0_mean.nii flirt_import T12dwi_matrix.txt mrtransform -linear T12dwi_matrix.txt {basname_mprage} {basname_mprage}_registered.nii 5ttgen fsl -mask mask.nii {basname_mprage}_registered.nii 5tt.mif
fi

dwi2response dhollander $basname_dti outresponse.txt out2.txt out3.txt -shell $shell -fslgrad $basname_bvec $basname_bval -mask mask.nii -force
dwi2fod msmt_csd $basname_dti outresponse.txt FOD_2.nii -fslgrad $basname_bvec $basname_bval -mask mask.nii -debug -force

tckgen FOD_2.nii CSDACT_{track_string}_{numberoftracks}.tck -seed_image ${outputin[0]} -mask mask.nii -angle 40 -act 5tt.mif -backtrack -crop_at_gmwmi -force

tckedit stringin CSDACT_{track_string}{numberoftracks}.tck CSDACT_{track_string}${numberoftracks}_roi_new_refined.tck -force

I am getting better tracks but still too scattered.
is there any way I can refine them?

Thanks a lot for your help!

Antonio

Hi @Antonio ,

You could add more include or exclude masks if you know where these tracts should or shouldn’t pass.

Cheers,
Nick

Dear Antonio,

Can you please try using a higher cutoff for the tracts, while generating the tracts with tckgen use -cutoff with all regular options, default of this is 0.1 but increasing it to 0.20 might give a more restrictive result.