search

mne-tools / mne-python

MNE: Magnetoencephalography (MEG) and Electroencephalography (EEG) in Python
https://mne.tools
BSD 3-Clause "New" or "Revised" License
2.76k stars 1.33k forks source link

add pick_normal= True or False for the lcmv beamformer #571

Closed NicZ closed 11 years ago

NicZ commented 11 years ago

I would like to use the lcmv reconstruction to compare the results I obtained with the MNE reconstruction when setting pick_normal= True.

It seems however that the lcmv beamformer only returns the norm (i.e. equivalent to pick_normal= False). It would be nice to add this feature too in lcmv.

Especially, I think it would be useful for those who want to do (or compare) some frequency / time-frequency analysis with lcmv.

Cheers, Nicolas

agramfort commented 11 years ago

that would indeed be a nice addition.

Altough beamformers commonly use PCA rather than pick_normal to have 1 time-series per location. maybe an API change so pick_normal=True becomes pick=None | "normal" | "pca" ?

WDYT?

agramfort commented 11 years ago

cc/ @Eric89GXL @mluessi @dengemann ...

dengemann commented 11 years ago

maybe an API change so pick_normal=True becomes pick=None | "normal" | "pca" ?

Yes, this would be nice + elegant. Maybe the next challenge for @rgoj ;-)

On Wed, Apr 17, 2013 at 2:53 PM, Alexandre Gramfort < notifications@github.com> wrote:

cc/ @Eric89GXL https://github.com/Eric89GXL @mluessihttps://github.com/mluessi @dengemann https://github.com/dengemann ...

— Reply to this email directly or view it on GitHubhttps://github.com/mne-tools/mne-python/issues/571#issuecomment-16503599 .

mluessi commented 11 years ago

I think this would be nice, the only downside I can think of is that it may be computationally quite costly, as you need to compute an SVD for each source space location. This API change could later easily be expanded to return all orientations.

agramfort commented 11 years ago

I think this would be nice, the only downside I can think of is that it may be computationally quite costly, as you need to compute an SVD for each source space location. This API change could later easily be expanded to return all orientations.

indeed.

@rgoj do you want to take a stab at it?

I'll review you're other PR asap.

rgoj commented 11 years ago

I think this would be nice, the only downside I can think of is that it may be computationally quite costly, as you need to compute an SVD for each source space location. This API change could later easily be expanded to return all orientations.

indeed.

@rgoj do you want to take a stab at it?

This does look like a good way to start working with the beamforming code--I'll take a look. But I don't know the code well enough to give an estimate of how long this could take me.

But I'll take a stab at it and let you know how I got on.