When you use --regmethod volume, it assumes that the label coordinates are in native anatomical space and then applies the mni305 registration (talairach.xfm). In this case, your source subject is fsaverage so the native space is assumed to be mni305 space (and talairach.xfm is the identity). So if you don't start with mni305, the the output surely won't be right. Second, when you use --regmethod volume, any surface information will be lost (or inappropriate), which is probably why you are seeing such chaos on the surface. I think by far the best solution is to use --regmethod surface. Talairach is so inaccurate that 0-4mm of misalignment is meaningless.
On 7/25/16 9:23 AM, Agoston Mihalik wrote:
Hi Freesurfer Developers,
I would like to ask for your expertise in the following problem I encountered:
I want to create subject specific labels based on a label I created in fsaverage. The transformation needs to be based on Talairach intermediate registration, since my label was drawn based on Talairach coordinates.
I ran the following command:
mri_label2label --srclabel test.label --srcsubject fsaverage --trglabel lh.test2.label --trgsubject
subj222 --regmethod volume
The problem is that the output label is scattered and all around the place when I look at it in tksurfer - see the image attached (lh.regmethod_volume.tif).
When I did the registration based on surface, it worked fine (see the image also attached - lh.regmethod_volume.tif), however, it resulted in 0-4 mm misalignment when I looked at the Talairach coordinates of the label borders. I guess it is due to the surface registration that optimises the surfaces to be aligned, not the Talairach coordinates.
Please find some information attached in case it is any help – content of files, output etc.
Here is my freesurfer version: freesurfer-Darwin-lion-stable-pub-v5.3.0
Your help would be really appreciated.
Thanks and best,
Agoston
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer