[Mne_analysis] computing forward solutions separately for MEG and EEG

[Baptiste Gauthier]

Dear mne users,

at first and before all: Happy New Year to the list!

Then, back to work ;) :

I recently tried to compute separate forward models for MEG, EEG and
combined EEG/MEG with the mne suite and mne_do_forward_solution

here is the bash code I used for the MEG-only computations:

arr1=($ListSubj)
>
> Dir=/neurospin/meg/meg_tmp/MTT_MEG_Baptiste/MEG/
>
> for i in {0..19}; do
>     echo ${arr1[i]}
>
> export SUBJECTS_DIR=/neurospin/meg/meg_tmp/MTT_MEG_Baptiste/mri/
> export SUBJECT=${arr1[i]}
> export
> MEG_DIR=/neurospin/meg/meg_tmp/MTT_MEG_Baptiste/MEG/${arr1[i]}/mne_python
>
> # for MEG
> mne_setup_source_space --ico 5 -overwrite
> mne_setup_forward_model --homog --surf --ico 4
>
> mne_do_forward_solution --mindist 5 --spacing ico-5 --mricoord
> $Dir${arr1[i]}/raw_sss/run3_DG_trans_sss-trans.fif --meas
> $Dir${arr1[i]}/raw_sss/run3_DG_trans_sss.fif --bem ${arr1[i]}-5120.fif
> --megonly --overwrite --fwd $MEG_DIR/run3_ico-5_megonly_-fwd.fif
>
> done
>

and it worked fine.

For EEG only, I modified the code to use a 3 layer BEM (remove --homog
option in mne_setup_forward_model) and use the --eegonly option in
mne-do-forward solution.

arr1=($ListSubj)
>
> Dir=/neurospin/meg/meg_tmp/MTT_MEG_Baptiste/MEG/
>
> for i in {0..1}; do
>     echo ${arr1[i]}
>
> export SUBJECTS_DIR=/neurospin/meg/meg_tmp/MTT_MEG_Baptiste/mri/
> export SUBJECT=${arr1[i]}
> export
> MEG_DIR=/neurospin/meg/meg_tmp/MTT_MEG_Baptiste/MEG/${arr1[i]}/mne_python
>
> # for EEG
>
> mne_setup_source_space --ico 5 -overwrite
> mne_setup_forward_model --surf --ico 4
>
> mne_do_forward_solution --mindist 5 --spacing ico-5 --mricoord
> $Dir${arr1[i]}/raw_sss/run3_DG_trans_sss-trans.fif --meas
> $Dir${arr1[i]}/raw_sss/run3_DG_trans_sss.fif --bem ${arr1[i]}-5120.fif
> --eegonly --overwrite --fwd $MEG_DIR/run3_ico-5_eegonly_-fwd.fif
>
> done
>

Then I got this error:

Read 306 MEG channels from
> /neurospin/meg/meg_tmp/MTT_MEG_Baptiste/MEG/cb130477/raw_sss/run3_DG_trans_sss.fif
> Read  61 EEG channels from
> /neurospin/meg/meg_tmp/MTT_MEG_Baptiste/MEG/cb130477/raw_sss/run3_DG_trans_sss.fif
> MEG not requested. MEG channels omitted.
> Some EEG channels do not have locations assigned.
> Forward computation failed (see above)
>

As I understood, if EEG channel number >60, it could raise an error and
prevent mne to retrieve channel location.
So I tried the dedicated fix:  mne_check_eeg_locations

mne_check_eeg_locations --file
> $Dir${arr1[i]}/raw_sss/run3_DG_trans_sss-trans.fif --fix
>

and got:

mne_check_eeg_locations version 1.2 compiled at Jan  5 2015 04:22:49
>
> input file          :
> /neurospin/meg/meg_tmp/MTT_MEG_Baptiste/MEG/sd130343/raw_sss/run3_DG_trans_sss-trans.fif
> digitizer data file :
> /neurospin/meg/meg_tmp/MTT_MEG_Baptiste/MEG/wl130316/raw_sss/run3_DG.fif
>
> Checking...
> Could not find the channel information.
>

I suspect that the EEG064 reference is the problem so I try to modify
directly the EEG status of this channel with mne_rename_channels and the
following alias: "EEG064:MISC666:502" in a text file, trying to transform
this problematic channel into a harmless MISC.

mne_rename_channels --fif
> $Dir${arr1[i]}/raw_sss/run3_DG_trans_sss-trans.fif --alias
> /neurospin/meg/meg_tmp/MTT_MEG_Baptiste/SCRIPTS/define_aliases.txt
>

and got:

mne_rename_channels version 1.5 compiled at Jan  5 2015 04:17:51
>
> fif file   :
> /neurospin/meg/meg_tmp/MTT_MEG_Baptiste/MEG/sd130343/raw_sss/run3_DG_trans_sss-trans.fif
> alias file :
> /neurospin/meg/meg_tmp/MTT_MEG_Baptiste/SCRIPTS/define_aliases.txt
> 1 aliases read from
> /neurospin/meg/meg_tmp/MTT_MEG_Baptiste/SCRIPTS/define_aliases.txt
> 0 changes done
>

So it appears that it didn't change anything.
I checked that this EEG064 is in the fiff file and I confirmed it exists
from fiedtrip header content of the same file.

I'm a bit stuck at this step and would appreciate any feedback on my
unlucky attempts :).

Best regards,

Baptiste

-- 
Baptiste Gauthier
Postdoctoral Research Fellow

INSERM-CEA Cognitive Neuroimaging unit
CEA/SAC/DSV/DRM/Neurospin center
Bât 145, Point Courier 156
F-91191 Gif-sur-Yvette Cedex FRANCE
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.nmr.mgh.harvard.edu/pipermail/mne_analysis/attachments/20150107/6c815490/attachment.html