single-layer BEM model for EEG

✉️ Mailing List Archive (read-only)
1

Hi,

Would it be possible to calculate an EEG forward solution with single-layer
BEM in MNE?

I know it's generally a bad idea. I just want to see how bad the source
localization is with the single-layer BEM.

I have the MRI scans (Flash5 and Flash30) and the 3-layer BEM for the
subject. The source localization of auditory N1 works great with the
3-layer BEM. I generated a single-layer BEM and tried to use it to compute
a forward solution, but MNE do_forward_solution returned with error.

Thanks!

Best
Le

2

hi Le,

indeed you cannot use a homogeneous model in EEG forward calculations.

it is not supported.

Best,
Alex

3

Hi, Alex,

Thank you very much for the confirmation. I will try to get a 3-layer model
from just the MPRAGE (no FLASH for future subjects due to time
constraints).

Best
Le

4

Hi Le,

You can create decent BEM models from a single MEFLASH at 5 degrees
(some users have even reported good results from a single echo). Which
would save you half the time.

HTH
D

5

Thanks you! This is very useful info to know. I will try to convince the
MRI people to run the FLASH at just 5 degrees.

Best
Le

6

Thanks again! Which FLASH protocol do you recommend for creating BME models
for EEG source localization?

So I've talked with our (f)MRI people about adding a single FLASH scan in
their protocol. They are mainly concerned about the time requirement.

I found a FLASH protocol on
https://surfer.nmr.mgh.harvard.edu/fswiki/FreeSurferWiki?action=AttachFile&do=get&target=FreeSurfer_Suggested_Morphometry_Protocols.pdf,
which takes about 8.5 min.

I don't know if this is the most up-to-date FLASH protocol at Martinos --
Is there a newer FLASH protocol that can produce decent skull boundaries
and takes less time to run?

Another question is whether there's a FLASH protocol already imported on
Bay 1 (Skyra scanner) at Martinos (where our MRI team run MPRAGE). If not,
is it ok to run FLASH and MPRAGE on different scanners for the same subject?

Thanks!

7

Hi Le,

The one I know of on bay 1 takes 8:25. If you want to do some further
inquiries offline (email me @ dwakeman at nmr.mgh.harvard.edu ) and we
can inquire together from the sequence programmers here at Martinos.

HTH
D