Customized head/forward model

External Email - Use Caution

Dear MNE developers,

I'm starting to look into the MNE python package and I really like what I experienced so far, e.g. in terms of data I/O, quick analysis facilities, plotting routines and documentation. Great effort you are taking here!

I am wondering if it would be possible to use a customized head and forward model with MNE. As far as I understand (from the documentation inhttp://martinos.org/mne/stable/auto_tutorials/plot_forward.html) MNE builds upon FreeSurfer utilities for segmentation and BEM generation which is very reasonable and convenient.

However, would it be possible to use my own customized forward model? Let's say I have all information I'd need for an appropriate forward model (e.g. my own segmentation in form of a triangulated mesh with different compartments for different tissue classes, dipole coordinates and normals, the lead field matrix, the MRI volume, electrode positions and registration matrices to MRI space etc.) could I use it with MNE for subsequent source analysis and visualization? If yes: Is there a documentation on how to format and provide the data in ASCII format? I understand that there are methods for I/O forward operators (e.g. mne.read_forward_solution) but I assume these need to be previously saved with MNE.

Thanks a lot!

Best, Arne

External Email - Use Caution

hi Arne,

thanks for the kind words.

regarding your question, yes technically everything is possible as you
can modify
the in memory python objects. However not much attempt has been done in this
direction. I just start to see how to use OpenMEEG:

https://github.com/mne-tools/mne-openmeeg/blob/master/demo_openmeeg.py

note that this is a 2h hack and still quite far from prime time.

If you have time to make this happen, we'll be happy to offer some guidance.

Alex

External Email - Use Caution

Hi Alex,

thanks for your quick reply. Sounds good. Important to know that subsequent operations like inverse calculation and plotting could be used without building upon the FreeSurfer data structure or other internal assumptions. I have to discuss it internally but can very well image to look into this in details after finishing my current project. I'll get back to you for guidance then - much appreciated!

Thanks again! Cheers, Arne