Inconsistent surfaces extracted using watershed

✉️ Mailing List Archive (read-only)
1

External Email - Use Caution

Dear MNE expert,

I am working with a EEG dataset and I wish to run source on that.
Unfortunately the T1w are not FLASH and I know that in the absence of FLASH
images, MNE offers a somewhat less accurate
solution based on the watershed algorithm (see Jas et al., 2018).

Indeed for some subjects in my dataset (see attached 1) I got the following
error when I am going to run make_bem_model.

RuntimeError: Surface inner skull is not completely inside surface outer
skull

For other subject instead the surfaces are simply wrong (see attached 2),
even though make_bem_model hadn't give me any error

I was wondering whether if you have some tips on how to fix them.

Thanks

Dave
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.nmr.mgh.harvard.edu/pipermail/mne_analysis/attachments/20200121/4c3f1047/attachment-0001.html
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 1.png
Type: image/png
Size: 75536 bytes
Desc: not available
Url : http://mail.nmr.mgh.harvard.edu/pipermail/mne_analysis/attachments/20200121/4c3f1047/attachment-0002.png
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 72752038-34c66200-3b8f-11ea-9b5b-5620ce8cd13f.png
Type: image/png
Size: 74634 bytes
Desc: not available
Url : http://mail.nmr.mgh.harvard.edu/pipermail/mne_analysis/attachments/20200121/4c3f1047/attachment-0003.png

2

External Email - Use Caution

Can you see if any of the tips in these old posts help? This looks like a
Freesurfer reconstruction / function problem.

https://mail.nmr.mgh.harvard.edu/pipermail//mne_analysis/2019-February/005561.html
https://mail.nmr.mgh.harvard.edu/pipermail//mne_analysis/2019-April/005805.html
https://mail.nmr.mgh.harvard.edu/pipermail/mne_analysis/2019-November/006317.html

In this particular case I'm guessing that your `brainmask.mgz` does not
look correct. I would tweak the freesurfer reconstruction before working on
BEM mesh extraction.

HTH,
Eric

3

External Email - Use Caution

I found this issue often when I started with MNE/Freesurfer. My personal choice as a solution is to use the method specified by ezemikulan to manually edit the meshes.

They have created a very nice straightforward guide for this here: https://github.com/ezemikulan/blender_freesurfer

I had no experience with Blender in the past, but their guide made it very clear to use and this solution means that there is no trial and error required with tweaking parameters. That said, the links that Eric posted may also provide a solution that is preferable to you.

Cheers,
Joel