Dear Alfonso / CONN community,

I would like to use subject-specific ROIs for certain 'sensitive/small' areas in the MTL (hipp. subfields, entorhinal c.)
Up to now I have done a first analysis using the data at the MNI-space.

I imported directly/manually via conn batch the MNI-space preprocessed images (preproc was done at fMRIPrep).

I saw in other posts that I could add via batch the raw fMRI in addition to the preproc data I have already imported & specify to CONN to do things separately for some ROIs on native-space and the rest keep them at MNI-space.
But it is unclear to me how to do this. 

In a previous thread:
https://www.nitrc.org/forum/message.php?... 
You mention: 
"...(if your functional volumes defined in batch.Setup.functionals are in MNI space and are named 'swauFunctionals.nii', for example, this rule will consider instead the files named 'auFunctionals.nii' -in subject-space- when CONN needs to extract BOLD-signals from your ROIs; see help conn_batch for more details)"

------
Can you elaborate on what you mention with "swau" and "au"?
how I could deal with that in my case using the batch? (fMRIprep preproc data manually imported via batch)

2) does it make sense later to use the subject-specific ROIs (e.g. entorhinal c.) to see connectivity with other mni-space ROIS? (MNI-space ROIs default from CONN) e.g. FPN.lateral PFC

Would that work well with CONN ? What would you recommend?


Thank you very much for the help :)

Dear Panagiotis,

Yes, it is perfectly fine to combine subject-specific ROIs with MNI-space ROIs. Once the BOLD data is imported from each corresponding ROI CONN will no longer "care" about how this ROI was defined when performing seed-to-voxel and ROI-to-ROI analyses. 

To do this you only need to have several "versions" of your functional data, each in the corresponding space (e.g. one version of the functional data in subject-space and another version in MNI-space). One way to do that, for example, is in CONN preprocessing window to click 'Load' and select a pipeline named DefaultMNISSCombined (this pipeline combines the "subject-space" and "MNI-space" default preprocessing pipelines in one). After running your data through this pipeline, you will already have these two different versions of the functional data for each subject, so all you have to do now is associate each ROI with the corresponding version of the functional data (i.e. MNI-space ROIs associated with MNI-space functional data, and subject-space ROIs associated with subject-space functional data). To do that, in CONN's Setup.ROIs tab, simply click on each ROI and select the appropriate secondary dataset (e.g. select the data labeled as 'MNI-space data' for MNI-space ROIs and the dataset labeled 'subject-space data' for the subject-space ROIs).

Hope this helps
Alfonso 
Originally posted by Panagiotis Iliopoulos:

Thank you very much Alfonso.

How can I do this using the conn batch, by importing data already preprocessed outside of the CONN toolbox (e.g. fMRIprep)?

I must use the setup.ROI field I assume. An example would help :)
I am wondering how I can still use the default ROIs imported (when not specifying the setup.ROIs field), but adding the custom ROIs (a new atlas) and the subject-space ROIs as well.

I have trouble ensuring this would work smoothly (use default ROIs, + custom atlas, + subj. ROIs, all via batch).


Thank you a lot in advance! :)