Please offer the customized .txt file and clean.py etc, and check the reproducibility of all code on data preparing, thank you.

[ShoujunYu]

Main Issues:

1. Missing txt file
   

   DeepMultiConnectome/data/tractography.sh

   Line 475 in b40c398

   labelconvert "${parcellation}" ./txt_files/FreeSurferColorLUT.txt ./txt_files/fs_${parc}.txt "${parcellation_converted}" 2>&1 | tee -a "${log_file}"

2. Registration error
   

   DeepMultiConnectome/data/tractography.sh

   Line 334 in b40c398

   mrregister "${dwi_meanbzero_brain}" "${T1_brain_nii}" \

3. Missing .py file
   

   DeepMultiConnectome/data/tractography.sh

   Line 594 in b40c398

   python3 ./clean_log.py "${log_file}"

4. Please clearly describe the source of each file. For example, in the HCP directory there are two versions of aparc.a2009s+aseg.nii.gz and aparc+aseg.nii.gz: one under the T1w folder and another under the MNINonLinear folder.

Other:
Based on my understanding and the code logic you provided, I made some modifications to parts of /data/tractography.sh.

transform_T1_to_DWI="${dmri_dir}/struct2diff_mrtrix.txt"

mrregister "${T1_brain}" "${dwi_meanbzero}" \
          -type rigid \
          -rigid "${transform_T1_to_DWI}" \
          -transformed "${dmri_dir}/T1_brain_dwi_from_mrregister.mif" \
          -mask2 "${dwi_meanbzero_brain_mask}" \
          -rigid_init_translation geometric \
          ${threading} -info 2>&1 | tee -a "${log_file}"

mrtransform "${seg_5tt_T1}" "${seg_5tt}" \
          -linear "${transform_T1_to_DWI}" \
          -template "${dwi_meanbzero}" -interp nearest \
          ${threading} -info 2>&1 | tee -a "${log_file}"

mrtransform "${T1_brain}" "${T1_brain_dwi}" \
          -linear "${transform_T1_to_DWI}" \
          -template "${dwi_meanbzero}" -interp linear \
          ${threading} -info 2>&1 | tee -a "${log_file}"

mrtransform "${gmwm_seed_T1}" "${gmwm_seed}" \
          -linear "${transform_T1_to_DWI}" \
          -template "${dwi_meanbzero}" -interp nearest \
          ${threading} -info 2>&1 | tee -a "${log_file}"

At present, the connectome generated with aparc+aseg looks fine, but the connectome matrix generated with aparc.a2009s+aseg appears to have the following issue. Could you please let me know how to resolve it? The connectome figure is attached.

Thank you again, looking forward to your reply.

[MarcusVroemen]

Dear @ShoujunYu, thanks for the detailed report and for sharing your helpful modifications.

We've pushed fixes to main addressing all issues:

The missing .txt file (FreeSurferColorLUT.txt) needs to be downloaded from https://surfer.nmr.mgh.harvard.edu/fswiki/FsTutorial/AnatomicalROI/FreeSurferColorLUT. For the label conversion files, we replaced the missing custom files with MRtrix3's built-in fs_default.txt (aparc+aseg) and fs_a2009s.txt (aparc.a2009s+aseg). This should also fix the aparc.a2009s+aseg connectome issue, as the missing fs_a2009s.txt caused streamlines to be assigned to the wrong or background nodes. Your changes to the registration were incorporated, and the clean_log.py file was unnecessary and therefore removed.

We also added a note to the README clarifying that we used the MNINonLinear/ folder for the parcellation files, but either version works.

Could you test with the updated script and let us know if the aparc.a2009s+aseg connectome looks correct?

[ShoujunYu]

Dear @MarcusVroemen,

I have implemented the plot_connectome.py script with the assistance of GPT, following the logic of the shell pipeline.

The script is for visualizing connectome matrices generated from different parcellation schemes, including aparc+aseg and aparc.a2009s+aseg.

Could you please check whether this is correct?

plot_connectome.py

[MarcusVroemen]

Dear @ShoujunYu,

Thank you for the help, that script had indeed been missing. I pushed a change with this script added. Can you verify if the connectome generation works now?