Question about Atlas File Integration in TI-Toolbox GUI

[LawsOfForm]

Subject: Question about Atlas File Integration in TI-Toolbox GUI

Dear TI-Toolbox team,

I recently installed the latest version of the TI-Toolbox and had great success running the pre-processing module via the Docker GUI.

I am now trying to use the Optimizer with a predefined atlas region. In the TI-Toolbox/resources/atlas folder, I can see several atlas files, including:

• HarvardOxford-sub-maxprob-thr0-1mm.nii.gz
• HOS-thr0-1mm.nii.gz
• MNI_Glasser_HCP_v1.0.nii.gz
  These are accompanied by corresponding .txt files.

However, when I select “Cortical” as the ROI Definition Method in the Optimizer tab, the “Atlas” dropdown menu remains empty and does not show these atlases as options.

Could you please advise:

Where should the atlas files be placed so they appear in the GUI dropdown?
Or is there an additional step I might be missing to enable them?
Thank you in advance for your support.

Best regards,
Filip

[idossha]

Hi @LawsOfForm,

I see the problem. Could you please convert this to its own discussion post and not a comment to the welcome message? I just want it to be available on the main discussion feed page.

Regarding the actual problem, the flex-search does not actual use those NIfTI files, those are used later in an optional MNI voxel analysis. The flex-search uses .annot files created during the preprocessing of the m2m folder We have recently refactor that preprocessing module, and may have introduced a potential bug as we did not set up a robust testing suite yet. In the meantime there are two ways around it,

Place a spherical target with appropriate size in your region of interest.
If the .annot files indeed were not created during preprocessing, you can manually generate them if you have a fully formed m2m folder. The command you can use is the following: subject_atlas -m "$m2m_folder" -a "$atlas" -o "$output_dir" Your outpur dir should be you m2m-subID/segmentation/. This should then be available to you for atlas based targeting in the flex-search.

Update

I have identified the problem and have fixed it in the codebase. I have incorporated the changes in the current version (v2.2.4). For it to take effect, please log out of the TI-Toolbox and your next login will incorporate the update.

Now, every m2m step will automatically create the necessary .annot file. However, if you already have a number of head models and do not want to repeat that complete process, run this subject_atlas -m "$m2m_folder" -a "$atlas" -o "$output_dir" where Your outpur dir should be you m2m-subID/segmentation/ and atlas sould be either a2009s , DK40 , HCP_MMP1.

Let me know how it goes !

Ido

[LawsOfForm]

Atlas Configuration Notes

Hi Ido,

I did as you recommended, and the .annot files where created, thanks for your help. If I know choose the subject in the Basic Parameters tab I can see all 3 atlas options.

For everyone else who is not totally familiar how to use the command above in a docker environment, this will create the annot. files in your example folder

• Run the bash ./loader.sh to open docker
• Example path:
  When prompted with: "Press Enter to use this directory or enter a new path":
  /media/Data03/Studies/Flowmodulation/code/Prestudy/ExampleT1/Example_DICOM
• This path in the docker is now /mnt/Example_DICOM so to use subject_atlas command inside the docker it has to look like this:
• subject_atlas -m /mnt/Example_DICOM/derivatives/SimNIBS/sub-001/m2m_001 -a "DK40" -o /mnt/Example_DICOM/derivatives/SimNIBS/sub-001/m2m_001/segmentation

[idossha]

This is correct, thank you for documenting your process. However, to clarify, this manual approach is not necessary, as the preprocessing step of creating an m2m folder should do it for you.