[Nipy-devel] fif IO in Python

Alexandre Gramfort alexandre.gramfort@inria...
Fri Jan 7 08:45:58 CST 2011

> On Thu, Dec 30, 2010 at 09:29:58AM -0500, Alexandre Gramfort wrote:
>> Thanks Stephan
>> I've been working on this since monday. I can know read most fif files
>> with pure Python. I need to do the writing routines now.
>> I still have to negotiate with Matti the license I can use for this code.
>> Hopefully it will be BSD so I can benefit from your help for testing
>> and setting up a nice documentation.
> If you can get this code in BSD, you'll contribute it to nibabel, right?

Yes ideally. The problem for me now is that MNE has a customized
version of fif libraries and It's not clear to me what is standard and what is
MNE specific. For example MNE supports fif files with 3D complex
matrices which is not standard. It means for nibabel either sticking
to the neuromag standard or having an explicit reference to MNE.

Also on my side it means clearly separating what is pure IO and
simple but useful examples for processing MEG data.

> If nibabel keeps growing this way, it will be a killer in the
> neuroscience labs: I can see so many people struggling with IO around me.

me too ...


More information about the Nipy-devel mailing list