shift in freeview

Mailing List Archive (read-only)
1

Dear experts,

I am perplexed how to fix the problem which is shown in the pictures.
The volume is the T1.mgz. The surfaces are computed by MNE (red -
inner_skull, yellow - outer_skull, green - outer_skin). In TkMedit the inner
and outer skull are aligned to the T1.mgz, respectively.

I used <mne_watershed_bem> and <mne_setup_forward_model> to get the
surfaces.

I get the same shift with all surfaces I compute with mne_write_surface, as
well.

Since I would like to visualize 3-4 surfaces with a mri volume I would like
to use freeview instead of TkMedit.

Do you have any hint how to fix that shift problem?

Thanks so much,
Alexander

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.nmr.mgh.harvard.edu/pipermail/mne_analysis/attachments/20101110/12583aaf/attachment.html
-------------- next part --------------
A non-text attachment was scrubbed...
Name: T1_bem_sagittal_freeview.jpeg
Type: image/jpeg
Size: 209238 bytes
Desc: not available
Url : http://mail.nmr.mgh.harvard.edu/pipermail/mne_analysis/attachments/20101110/12583aaf/attachment.jpeg
-------------- next part --------------
A non-text attachment was scrubbed...
Name: T1_bem_tkmedit.jpeg
Type: image/jpeg
Size: 31672 bytes
Desc: not available
Url : http://mail.nmr.mgh.harvard.edu/pipermail/mne_analysis/attachments/20101110/12583aaf/attachment-0001.jpeg

2

Dear Alexander,

This question is probably better off answered on the freesurfer list than the MNE list, since it is a freesurfer question (mne_watershed_bem simply uses freesurfer to generate the surfaces (as described in the manual)). My suspicion is that freeview (which I believe is still in beta), is utilizing the nifti origin, while by default freesurfer calculates surfaces with it's own origin (http://surfer.nmr.mgh.harvard.edu/fswiki/CoordinateSystems).

If you only care about the visualization, the easy thing to do would be to

first find out what coordinate frame it does use
second read in the surfaces in matlab (for example) and perform the translation (plus any rotations if necessary, though it doesn't appear as though that is).
Save these new surfaces under different names (so as not to cause problems with your MNE calculations). e.g. inner_skull.freeview
Then use these for visualization.

D