I don't know what might have happened (under the hood, mni152reg uses the FSL flirt program). If you want a non-linear registration, then you'll need to use CVS (mri_cvs_register) and choose the mni target. If you just want the aparc of the aparc+aseg, then don't go into the volume at all.
On 04/17/2018 01:57 PM, Zimmerman, Jared wrote:
Hi All,
I’m trying to register the aparc+aseg volumes from subject T1 space to MNI152 space and running into problems. I’ve taken the following steps
1.mri152reg –s Subj241
2.mri_vol2vol --targ $FSLDIR/data/standard/MNI152_T1_2mm_brain.nii.gz --mov $SUBJECTS_DIR/Subj241/mri/aparc+aseg.mgz --o Subj241_aparc+aseg_mni2mm.nii.gz --reg $SUBJECTS_DIR/Subj241/mri/transforms/reg.mni152.2mm.dat
This yields an image that is rotated and shrunken relative to the MNI152_T1_2mm template as shown in the attached image.
I’m wondering if this is the right way to do what I’d like to do. The reg.mni152.2mm.dat file is only a linear transform, but I’d like to apply a non-linear transformation to the subject aparc+aseg to get it well matched to the MNI152 template space. Some of my subjects have atypical head shapes and I really think a non-linear warp will be necessary to achieve a good fit.
I’ve run the MNI152 1mm template through recon-all so I have that as a reference. Is there a way to map volumes from one subject space to another via fsaverage or sphere sort of like mri_label2label does for labels?
Alternatively, I could just calculate the appropriate T1 to MNI warps with FNIRT or ANTs and apply those to the aparc+aseg, but I figured it would be possible to do with FreeSurfer without calculating any new transforms.
Thanks,
Jared
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer