BEM construction - FLAIR vs FLASH?

For BEM construction, can MNE take advantage of Freesurfer's multispectral
imaging pipeline (using a single FLAIR scan), or is it still better to
collect the two FLASH scans? I grepped through the linux & python scripts
but didn't see anything promising. Cutting a scan would free up some MRI
time on a new project, but I don't want to hamstring myself.

Thanks,

Jon
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.nmr.mgh.harvard.edu/pipermail/mne_analysis/attachments/20140227/62d676e0/attachment.html

hi Jon,

For BEM construction, can MNE take advantage of Freesurfer's multispectral
imaging pipeline (using a single FLAIR scan), or is it still better to
collect the two FLASH scans? I grepped through the linux & python scripts
but didn't see anything promising. Cutting a scan would free up some MRI
time on a new project, but I don't want to hamstring myself.

I am not aware of a solution using FLAIR. I guess it could work should you
dig deep enough in the mne_flash_bem shell script.

Let us know if you find a way. That would be a nice contribution for MNE users.

Best,
Alex

Given the contrast in FLAIR images IMO it would not be possible to
use the standard mne_flash_bem script tools to generate useful bem
surfaces (you could obviously write your own). I can help you out by
saying that in my experience, you can get away with only utilizing the
FLASH 5 data (although this assumes you have similar image quality).
If you are getting very good contrast between the critical tissues, in
your FLASH 5 alone. This should save you half the time.

D

Dear Dan, Jon and Alex,

Having recently received the sequence parameters from Alex (thx!) for the multi-echo FLASH, I can second Dan's observation of practically identical surfaces using 5 deg alone vs. 5 and 30 degs. This is on a 3T Skyra, anyone care to weigh in with experiences from other scanners?

My naive take on FLAIR is that it might help getting the pia-dura separation right, i.e., improve cortex thickness estimates, but probably not that critical for M/EEG forward models?

/Chris
/CFIN, Aarhus, Denmark

To speak to the second point. In MNE the pial surface is not used in
the forward model. Furthermore, I would expect (read I have no direct
tests) that you would get more benefit (relative to effort) from
correctly measuring the inner skull boundary then from adding in
another layer for the pial surface. And FLAIR (Fluid Attenuation
Inversion Recovery) will unfortunately not help you with that
measurement, due to its intentional insensitivity to CSF (hence being
beneficial for pial measurements). However, for cortical thickness
measurements, this is supposed to be very useful.

D