Hi,

I think I got the problem, just updating, so other people won't do the same mistake:

The warp stored in the 'talairach.m3z' file should be applied on a conformed volume, meaning the raw image after the the following command was cried out:
"mri_convert img.nii.gz img_conformed.nii.gz --conform" (conform to 1mm voxel size in coronal)

then the warp can simply applied by the following command:
'mri_convert img_conformed.nii.gz --apply_transform talairach.m3z -oc 0 0 0 img_warped.nii' (works for me!)

I also noticed that the '--conform' option convert the img data type to integer (if it was float before, like in my case, where it turned all the values to 0). So you should be aware to that.

Good luck!
gidon l.


2017-10-19 16:40 GMT+03:00 Gidon Levakov <gidonle@post.bgu.ac.il>:
Hi Freesurfer experts,

I have several nifti images (created with another software) in the subjects original anatomical space (native space) that I want to transform to a common space (freesurfer version of MNI305 should be fine) using non-linear transformation.

Reading a previous answer regarding this issue (https://mail.nmr.mgh.harvard.edu/pipermail//freesurfer/2013-September/033556.html), I understand I should use the mri_vol2vol command to apply the transformation stored in the file mri/transforms/talairach.m3z (I already run recon all). But after checking the result, the images are still not align with each other and do not resemble the mni305.cor.mgz file.

I used the following command:

mri_vol2vol --m3z my_subjects_dir/sub_X/mri/transforms/talairach.m3z --noDefM3zPath --reg_header --mov file_loc/img.nii.gz --targ /usr/local/freesurfer/average/mni305.cor.mgz --o file_loc/img_warped.nii.gz

I also tried without the --reg_header option with no success.

Any idea what I am doing wrong? can you give me an example that works?

Thanks in advance!

Gidon l.