Problem with coregistration of MRI/MEG

Dear all,

I am a newbie to Brainstorm (and MEG) in general, so I may just be missing something entirely obvious (but, since I’ve been stuck for quite a while now, I’d be grateful for your help).

I’ve preprocessed my MEG data with fieldtrip and would now like to do source-localization analyses in brainstorm. To this extent, I have segmented my MRI scans with freesurfer and imported those into Brainstorm using a script (wmp_bst_import_mri.m) that allows for the manual identification of the relevant coordinate points (nasal, preauricular, etc.). From what I can tell, this step seems to work fine. I then use another script (wmp_bst_import_recordings.m) to import the preprocessed MEG data as a function of condition and create a channel file for each of these conditions. I don’t know what/where exactly, but, apparently, something goes wrong here because the registration between the MRI and the recordings does not seem to work at all for any of my subjects (see attached example): Usually, about half of the digitized electrodes are inside the skull and the position of the head in the helmet is far from perfect (sometimes, as in the example, the head is even outside the helmet - even after applying Brainstorm’s refinement algorithm ). I suppose that I could manually adjust the registration, but I am a bit worried that it doesn’t work for any of my subjects. Do any of you have an idea of what might be going on and what I could do to improve the registration?

Many thanks in advance!
Darinka

Hi Darinka,

I’m not sure to understand what your problem is.
To me the screen capture you posted here looks really good: the electrodes are really close to the head surface, the head of the subject is correctly positioned and touching the helmet (it looks like they are intersecting, but they are rather tangent, we’re not working at precisions below the millimeters), all the points look correctly placed.

If those yellow points we see are EEG electrodes, you might want to just project them on the scalp surface for calculating a more a accurate forward model for the EEG.
Right-click on the channel file > Edit: EEG > Click on the button [Project electrodes on surface].
From a script, you can use the process version of it: Import recordings > Project electrodes on scalp.

Cheers,
Francois