Hey all - I'm starting to get a bit into the world of MRI preprocessing for
an ECoG project. We have a number of tools in the lab that do this in
Matlab, but I'm wondering if such tools exist in python as well.
For example, a post-doc in our lab is one of the developers of Fieldtrip,
and he's been using their calls to SPM (or FSL) to do CT-MRI
coregistrations in Matlab. Does something like this exist in python, or do
most people go to an external package / use Matlab?
Chris
hi Chris,
you should ask this question on the neuroimaging at python.org mailing
(former nipy list).
anatomical registration is so far out of the scope of MNE but maybe
someone here can help using freesurfer/fsl tools I personally don't
master.
Alex
Good idea - will do. Don't MEG people do lots of MRI stuff? In tandem with
source localization etc?
Good idea - will do. Don't MEG people do lots of MRI stuff? In tandem with
source localization etc?
Actually less so, speaking for people I know. You don't need fMRI to source
localize MEG.
Hi Chris,
You might find SimNIBS (http://www.simnibs.de) interesting. They put together a nice pipeline using Freesurfer, FSL and others to pull of good BEM-segmentations (see also their documentation for suggestions on MR sequences to run). A major part of their work is then related to tessellation for FEM, which may be relevant, depending on how ?deep? you need to go when modelling your ECoG data.
Cheers, Chris
Actually less so, speaking for people I know. You don't need fMRI to
source localize MEG.
You generally DO need a structural MRI, for calculation of the inner
skull surface used in generating theoretical magnetic fields. Without
an MRI you can only approximate the shape of the brain using a sphere.
Multisphere or Nolte is typically used given a real brain shape. AFNI's
3dSkullStrip is fine for generating the surfaces.
No ambiguities here, of course you need a structural scan. I might hav
misread Chirs' question. We commonly use Freesurfer here. But we don't need
functional scans.
Yep - I'm only talking about structural, not functional. We have a similar
problem in ECoG, but in our case we need to localize electrodes on the
cortical surface. We don't have an MRI of these electrodes, but we do
usually have CT scans. So I'm trying to find a python pipeline for
co-registering the CT (where the electrodes are) to the MR (where the
anatomy is). E.g., I asked on the neuroimaging listserv and someone has put
together a GUI for something like this:
https://github.com/aestrivex/gselu/
The reason I'd like to do this is in order to use Freesurfer for the MR
processing / surface extraction, and then to do electrode plotting on top
of that.
Chris
Hey Chris,
We've developed the "grids and strips electrode localization utility" tool
in our lab, so please let me know if you have any questions.
For electrodes plotting, we developed a visualization tool in python which
is based on Blender, a free source 3D visualization tool.
You can take a look on a quick demo here (It's better to watch it on a
desktop computer and not on a smartphone, otherwise you won't be able to
see the annotations):
https://www.youtube.com/watch?v=tJzoRWbseKw
The tool is for multi-modality neuroimaging data visualization, like fMRI,
MEG, invasive electrodes, etc. 5 minutes into the demo you can find how to
visualize depth electrodes.
Let me know if you want to use it.
Best,
Noam
Hey Noam - it looks like a really cool tool. I've been chatting with Roan
on the repository w/ some issues I had doing the basic co-registration, and
I'm going to keep working through it to try and make it work. Maybe I can
help out with sharing my experience for documentation purposes etc. It
would be great if something like this existed for python, as all of our lab
uses some combination of SPM/FSL/custom code for this.
Chris
Hey Chris,
Sure, any help with the documentation will be a blessing 