[Mne_analysis] issue with volumetric source estimate morphing

[Eric Larson]

        External Email - Use Caution        

This indeed looks like a bug. Can you open an issue on GitHub and we can
work on it there?

Thanks,
Eric


On Mon, Apr 6, 2020 at 7:19 AM Fleur GAUDFERNAU <fleur.gaudfernau at pasteur.fr>
wrote:

>         External Email - Use Caution
>
> Hello,
>
> I have mixed source estimates that I wish to morph to the fsaverage brain.
> To do so, I separated the mixed source estimates into a surface source
> estimate and a volume source estimate. When trying to morph the lastest
> (using morph_apply), I get the following error:
>
> ValueError: vertices do not match between morph (453) and stc (216) for
> the volume:
>
> (I believe there is a small mistake in morph_apply and what it means is
> actually that vertices do not match between morph (*216*) and stc (*453*)
> for the volume.)
>
>
> Here is my code:
>
>
> *#recuperating volume source estimate from mixed source estimate stc *
> *nb_sources_cortex = stc.vertices[0].shape[0] + stc.vertices[1].shape[0]*
>
>
>
> *vertices_vol = np.concatenate((stc.vertices[2], stc.vertices[3]))
> stc_volume = mne.VolSourceEstimate(stc.data[nb_sources_cortex:], vertices_vol,
>                                     tmin=stc.tmin, tstep=stc.tstep,
>                                     subject = subject_mri)*
>
> *#Maybe not the cleanest way to isolate a volume source estimate from a
> mixed source estimate? *
>
> * #recuperating the source space I want to morph to*
>
> *src_fsaverage_vol = subjects_dir +
> '/fsaverage/bem/fsaverage-vol-5-src.fif' *
> *src_vol  = mne.read_source_spaces(src_fsaverage_vol)*
>
> *#recuperating the source space to morph from*
> *cereb_used = mne.SourceSpaces(inverse_op['src'][2:])*
> *#the source space (inverse_op['src'])*
> * is made of 4 source spaces: 2 surface source spaces and 2 volume source
> spaces *
> *#- here I recuperate the 2 volume source spaces, made of 216 and 237
> vertices (respectively)*
>
>
>
>
>
>
> *morph = mne.compute_source_morph(src = cereb_used,
>                                     subject_from=stc_vol.subject,
>                                     subject_to = 'fsaverage',
>                                     subjects_dir=subjects_dir,
>                                     niter_affine=[10, 10, 5], niter_sdr=[10, 10, 5],  # just for speed
>                                     src_to=src_vol, verbose=True)*
>
> *stc_vol = morph.apply(stc_vol)*
>
> I'm sorry that I cannot provide a replicable code.
>
> I believe the error is coming from the fact that when computing the source
> morph, mne only considers the first of the two volume source spaces in
> cereb_used, hence generating a morph with 216 vertices instead of 216+237
> vertices, but I do not know how to fix it.
>
> King regards,
>
> Fleur
> _______________________________________________
> Mne_analysis mailing list
> Mne_analysis at nmr.mgh.harvard.edu
> https://mail.nmr.mgh.harvard.edu/mailman/listinfo/mne_analysis
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.nmr.mgh.harvard.edu/pipermail/mne_analysis/attachments/20200406/2a39a600/attachment-0001.html