[Mne_analysis] "fsaverage" surface generation

[Leonard Varghese]

Hi,

I'm attempting to scale the "fsaverage" head based on subject-specific
digitization points, and to create and scale 3 layer BEM that will be
used to localize EEG data (only). The machines I'm using have Freesurfer
6.0, and I'm running mne-python in a Python 2 virtual Environment. 

When I first load the coregistration GUI, I'm given the option to
transfer the fsaverage head to my subjects directory, which also copies
some MNE-specific files over
(https://github.com/mne-tools/mne-python/tree/master/mne/data/fsaverage).

If I exit the GUI, and then run:

mne make_scalp_surfaces -s fsaverage -n

It seems like it should generate a high-resolution copy of the head. But
when I load the coreg GUI back up, and toggle the "high resolution head
option", I see the following:

Low-res head:
https://drive.google.com/file/d/10k4PDpLwvERYSAbrLUtum0vDRvXJYLe4/view?usp=sharing

Hi-res head:
 https://drive.google.com/file/d/11hzLgWYmQXtoeU1J013lRhLsNFnQuzeF/view?usp=sharing

The shapes of the head seem different moreso than just the the face
distortion, and the fiducial markers, by eye, don't seem to be quite at
the same anatomical landmarks...even though both heads correspond to
"fsaverage". 

Is it "safe" to ignore this discrepancy, and simply run the watershed
algorithm and 3-layer BEM construction for fsaverage?
 
-Lenny

--
Leonard Varghese
Auditory Neuroscience Laboratory
Department of Biomedical Engineering
Boston University
610 Commonwealth Avenue, 9th floor
Boston, MA 02215