Hi Mike,
that code saves the label in a coordinate system that tkmedit does not use. Try loading the label into it into tksurfer, then saving it, then loading it into tkmedit.
doug
On 4/26/11 6:02 AM, Michael Thomas Smith wrote:
Hello, I'm trying to use the V1 localisation tool provided by Hinds et al[1]. I've found the voxels labelled don't appear to be aligned correctly. The code I ran to do this follows:
First the symbolic link and running the tool:
$ ln -sf ~/V1_average ~/fs_subjects/ $ ~/V1_average/scripts/predict_v1.sh s01
Then I used tkmedit to view the result, comparing to the both T1.mgz or orig.mgz. E.g.
$ tkmedit s01 T1.mgz
then chose File->Label->Load label and selected "lh.v1.predict.label":
http://www.sal.mvm.ed.ac.uk/images/v1_tkmedit.png
I thought maybe this could be because of not realigning the image or some such. I thought I could try looking at the image once converted to nifty format:
$ mri_convert -rl mri/T1.mgz -rt nearest mri/T1.mgz ~/Desktop/temp.nii $ mri_label2vol --label label/rh.v1.prob.label --regheader ~/Desktop/temp.nii --temp ~/Desktop/temp.nii --o ~/Desktop/test.nii
However it still looks wrong:
http://www.sal.mvm.ed.ac.uk/images/v1_nii.png
Any ideas on what I'm doing wrong in the V1 localisation?
Thanks for any advice,
Mike Smith