Hi DTI fsl experts,
being quite new to DWI proceesing, I am surprised not to find much information about running probtrackx on HCP data.
What I am interested in is to use the 32k surfaces and the subcortical ROIs to build a connectivity matrix using cortical parcellation.
I created new L/R gifti surfaces combining fsaverage32k surfaces (T1w/fsaverage_LR32k/*.{L,R}.white.32k_fs_LR.surf.gii) with a label which excludes the medial wall of freesurfer (MNINonLinear/fsaverage_LR32k/*.R.atlasroi.32k_fs_LR.shape.gii).
I also created a ROIs mask in the DWI space including the ROIs used in fMRI in HCP.
This 3 files were listed in file areas2.txt and I ran probtrackx with the following commandline:
probtrackx2 -s merged -m ../Diffusion/nodif_brain_mask.nii.gz -x areas2.txt --targetmasks=areas2.txt --opd --s2tastext -V 1 --dir=prob_test4 --forcedir --omatrix1 -P 50 -S 2000 -l --onewaycondition --os2t --stop=./aparc+aseg_wm_dwi_inv.nii.gz
My stop mask is everything not in white matter.
Does it seems reasonable way to go?
It seems to run well, there is just a problem with the distribution because of the brainstem being one of my ROI.
My question is about the output:
fdt_matrix1.dot is a 3 column file which supposedibly describes the seed points indices pairs with a count.
coords_for_fdt_matrix1.nii.gz is supposed to help matching which seed points is used in the matrix but:
it is of shape 24034x5 (in docs it is supposed to be 3 columns?), with example values:
[[ -6.21752501 -14.92267704 28.0871563 90. -11.75583458]
[-15.32361126 -14.51436615 28.7354393 91. -11.47147846]
[-45.92995834 -14.80008698 29.36710739 92. -10.26763535]]
how can I match it back to different seeds mask/surfaces?
why is that only 24034 long as the seeds number amounts to about 90k = 3*30k + 30k (subcortical)?
Many thanks for your insight on this.
Cheers.
basile
