[Mne_analysis] A Question about mne_setup_forward_model

[Matti Hamalainen]

On Feb 22, 2010, at 8:22 PM, supra at nmr.mgh.harvard.edu wrote:

> Hi MNE Users
>
>    I have a question about the --ico option in the script
> mne_setup_forward_model. Does the ico number have to be the same as
> the one which we wish to use for the source space in computing the
> forward solution
> or is this independent? It creates a file
> structural_mri-5120-5120-5120-bem-sol.fif, which I think corresponds  
> to
> ico 4.
>
> Now is it OK to use this BEM file for computing forward solution if  
> I wish
> to have ico 5 source space corresponding to (10242*2sources)?
>
> mne_forward_solution --src
> $SUBJECTS_DIR/$SUBJECT/bem/structural_mri-ico-5-src.fif --mri
> $SUBJECTS_DIR/$SUBJECT/mri/T1-neuromag/sets/COR- 
> asalazar-090406-142236.fif
> --meas $SUBJECTS_DIR/$SUBJECT/eeg_mri/base/dec_base_raw-ave.fif --bem
> $SUBJECTS_DIR/$SUBJECT/bem/structural_mri-5120-5120-5120-bem-sol.fif  
> --fwd
> dec_base_raw-ave-eeg-all-fwd.fif --eeg --all

They are independent. The BEM model resolution essentially determines  
how close to the inner skull surface the sources can be for the field  
computation to be accurate.
The ico 4 choice for the BEM should be sufficient. It is practically  
not possible to go to higher resolution due to memory limitations and  
accumulating numerical errors in the computation.

- Matti