Info on Nifti format for CAB-NP Atlas

[prabhjot kaur]

Hello Tim

I want to apply CAB-NP atlas in CONN for static FC. Where to get NIFTI volumetric version or how to convert CIFTI to NIFTI.

Thanks

Prabhjot Kaur

INMAS, DRDO, New Delhi

India

[Ruben Sanchez]

CAB-NP atlas is originally defined in cifti space (see the Ji et al. paper and repository, https://github.com/ColeLab/ColeAnticevicNetPartition) No need to use the volumetric space.

--
You received this message because you are subscribed to the Google Groups "HCP-Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to hcp-users+...@humanconnectome.org.
To view this discussion visit https://groups.google.com/a/humanconnectome.org/d/msgid/hcp-users/CAAAPaW9sc2597JgKuQvPY2SQAWev6ogCRxzScFK9n-YStBRWsg%40mail.gmail.com.

[prabhjot kaur]

I am aware of the CIFTI format. 

Thanks, but got me confused. Are you saying that CONN can use CIFTI or CIFTI cannot be converted to NIFTI, because to my knowledge, CONN uses NIFTI data. Please clarify and guide me on this.

[Tim Coalson]

Group volume alignment techniques don't work well enough to faithfully represent detailed functional parcellations of human cortex as a volume file, because the folding patterns have too much subject-specific variability in most of the brain, and volume alignments are often based only on the shape of the brain, while functional areas can vary in position even with respect to some of the more stable folds.  Please see our paper https://www.pnas.org/doi/10.1073/pnas.1801582115 .

It is possible to make a subject-specific volumetric version of a parcellation by using the subject's own surface files, but it will be noticeably different for each subject, due to following the subject's cortical ribbon.  However, we instead do cortical connectivity analysis on the surface, where the cross-subject correspondence is much better.  If the CONN toobox can't do group analysis using surface-based data, then we'd recommend looking for another way to do your analysis.

Tim

--

You received this message because you are subscribed to the Google Groups "HCP-Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to hcp-users+...@humanconnectome.org.

To view this discussion visit https://groups.google.com/a/humanconnectome.org/d/msgid/hcp-users/CAAAPaW_mNM5p7AfHV_3RPXnBFGtPrVdAaHE7nsZQFcwF3r5fOA%40mail.gmail.com.

[Tim Coalson]

The CONN toolbox website does mention that they make use of FreeSurfer, and even has an example command of saving a random both-surface data array to...1-D nifti format.  A lot of the images on their website appear to be using group volume methods (displaying big blobs of activation along with a surface mesh for reference, but no pictures that show actually using surface-mapped cortical data), so you'd have to look closely through their documentation to see if they support cifti or other surface-based data or methods, but it isn't looking promising to me.

Tim