add mne based bem script for forward solutions

[maedoc]

In order to complete the present anatomical pipeline for complete
simulated modalities in TVB, forward solutions for MEG and EEG
require that the head model be formed and solved. This script
uses the FreeSurfer segmentation and MNE command line tools to
build the BEM and solve it. It is still necessary to manually
align sensors to MRI space, but after this step, forward solutions
are trivial to obtain using the mne-python library.

[maedoc]

This script adds a dependency on the MNE tools package, which is freely downloadable and primarily serves to build and solve the head model. Other pipelines exist, e.g. for MATLAB, but I haven't used them and they pull in tools from SPM to do segmentation, which is redundant with FreeSurfer.

If you're OK with all that, still please do not accept immediately. This script should be adapted to reuse an existing segmentation before merginng. Please suggest modifications.

[timpx]

In the concept, this can be easily integrated at the end of script.
One could also add an option if sensor positions are available and in this case compute mne_setup_forward_model.
For the modification, mainly:

• bash variables are already set
• mne and freesurfer are sourced at the beginning of the main script
• recon-all is already performed
• I try to not cd, so I use complete path using bash variables
• results should be stored somewhere in the $PRD directory
• I also usually add check to not redo a operations already performed

Conclusion: I can easily insert an adapted version into scripts if that's ok for you

[maedoc]

Ok if you are comfortable adapting the script then feel free to do so.

For the sensors I agree but not sure what are the best to use. We can ask the mne people about that. The difficulty is to have automagically the meg-mri alignment transform which is usually done visuallly.

timpx notifications@github.com wrote:

In the concept, this can be easily integrated at the end of script.
One could also add an option if sensor positions are available and in this case compute mne_setup_forward_model.
For the modification, mainly:

bash variables are already set
mne and freesurfer are sourced at the beginning of the main script
recon-all is already performed
I try to not cd, so I use complete path using bash variables
results should be stored somewhere in the $PRD directory
I also usually add check to not redo a operations already performed
Conclusion: I can easily insert that into scripts if that's ok for you

—
Reply to this email directly or view it on GitHub.

[maedoc]

closing in deference to #36