Hi Richard, See comments below. Lilla ________________________________ From: freesurfer-bounces@nmr.mgh.harvard.edu freesurfer-bounces@nmr.mgh.harvard.edu on behalf of Jones, Richard richard.jones@choa.org Sent: Wednesday, April 7, 2021 12:38 PM To: freesurfer@nmr.mgh.harvard.edu freesurfer@nmr.mgh.harvard.edu Subject: [Freesurfer] CVS register
External Email - Use Caution
I have a couple of questions regarding CVS register. I’m using Freesurfer 7.1 and I want to align some QSM images in order to analyze some sub-cortical structures that are not defined by Freesurfer (red nucleus, substantia nigra). I ran CVS register to register a subject to MNI152 space and the “final_CVSmorph_tocvs_avg35_inMNI152_norm.mgz file was in excellent registration with the mni152 brain, but when I looked at the corresponding aseg file (final_CVSmorphed_tocvs_avg35_inMNI152_aseg.mgz) the general registration was good but it appears to be slightly smaller than the MNI152 brain (see image below), is that to be expected?
Is the background the MNI152 template or the CVS_template in MNI152 space? The subject was nonlinearly aligned to the latter. At atlas creation time the CVS-template was affinely registered to the MNI space though and then reconned, so there might be some inaccuracies.
[cid:dbcb2edf-6f51-46cc-a156-cb1b622a7980][cid:4f893058-7979-4b87-bd92-5a8670af6424]
To move the QSM image into MNI space I can use one of the echoes used as the input for the QSM processing to derive the registration between the QSM image and the T1 volume using BBR register to give me a transformation file “qsm2anat.dat”. Looking at the documentation it seems I have to use CreateMorph to recreate the morph required to shift the QSM data from freesurfer to MNI space (and I’m a little confused about exactly what the morph final_CVSmorph_tocvs_avg35_inMNI152.m3z actually does).
I thought that you has used the CVS_template in MNI space, did not you?
final_CVSmorph_tocvs_avg35_inMNI152.m3z maps the correspondence between the native structural (recon) space of your subject to the CVS_template in MNI152 space.
You should be able to just use mri_vol2vol include in --m3z with the above output of cvs_register and --lta that specifies the QSM->T1 affine registration. Thetarget should be the norm from /subjects/cvs_avg35_inMNI152/mri
For CreateMorph I’m a little uncertain which of the files shown above I should use for the –in option to CreateMorph and for the –template option do I just specify the brain.mgz from /subjects/cvs_avg35_inMNI152/mri directory?
Having created the morph file “my _morph_file.m3z” I could use apply_morph having first transformed the QSM image into freesurfer space, but would it be easier to use mri_vol2vol with the following options:-
The below looks good, but you do not need a new m3z file, just use the output from mri_cvs_register
mri_vol2vol --targ $FREESURFER_HOME/subjects/cvs_avg35/mri/norm.mgz --m3z $SUBJECTS_DIR/my_subject/cvs/my _morph_file.m3z --noDefM3zPath \
-reg qsm2anat.dat --mov my_qsm_image.nii –out my_qsm_image_in_mni152.nii --interp trilin --no-save-reg
Thanks,
Richard
