Dear Freesurfer community,
I would like to be able to use any preexisting Nifti atlas with Freesurfer’s parcellation methods. As a proof-of-concept, I am attempting to use the parcellation scheme produced by Finn and colleagues (2015) with mris_ca_train and mris_ca_label. I have some questions about this.
I wonder whether what I am trying to do makes sense. To train mris_ca_train, I would feed it the Collin MNI template space and Finn’s parcellation scheme (which is registered to Collin MNI as far as I can tell). This is not ideal for two reasons. First, it would be better to train mris_ca_train on a group of subjects. Second, the anatomical data and parcellation data come from two separate data sets. Is my understanding correct so far? On the other hand, the other way I can think of to parcellate an individual’s data based on an atlas is to register the data and the atlas and assign labels from the atlas to the individual where they overlap. This doesn’t seem any better to me. Are there other ways of using atlases to parcellate individual data that I am not aware of?
Right now, I am stuck trying to produce an annot file to give to mris_ca_train. I ran mri_vol2surf --src shen_1mm_268_parc.mgz --hemi lh --mni152reg --out_type paint --o ./lh-surf-parc.mgz where shen_1mm_268_parc.mgz is the result of converting the parcellation scheme from nifti to mgz with mri_convert. The resulting mgz file seems reasonable. However, I don’t know how to make an annot file from it. Is there a way to do this?
I also attempted the following: mris_sample_parc -ct ./ctab.txt -sdir .. -file ./ctab.txt FS lh shen_1mm_268_parc.mgz lh-parc.annot The results don’t make sense to me and I don’t understand how mris_sample_parc works. What is the purpose of the translation file (-file option)? From what I’ve read on the wiki, mris_sample_parc is used to sample a manual parcellation done on an individual subject’s volume to that subject’s surface. In my case, the “subject” (Collin MNI brain) happens to be aligned with the parcellation. Does it make sense to run mris_sample_parc?
Thank you. Best regards, Ruy