[Mne_analysis] Spectral analysis with LCMV

Eric Larson larson.eric.d at gmail.com
Mon Jun 10 13:01:54 EDT 2019
Search archives:

        External Email - Use Caution        

You should be able to compute the LCMV source values using
apply_lcmv_epochs, then loop over the resulting STCs, passing stc.data to
psd_multitaper:

https://martinos.org/mne/stable/generated/mne.time_frequency.psd_multitaper.html

If you want it to be more efficient, you could write more complicated code
to compute the single-epoch spectral estimates (linearly, not computing
power or absolute value), apply the LCMV filters to these, then take the
power (or abs). But it would involve more complicated code.

Eric


On Wed, Jun 5, 2019 at 6:26 AM Виктория Манюхина <mvo96 at inbox.ru> wrote:

>         External Email - Use Caution
>
> Dear MNE developers,
>
> We would like to use MNE/LCMV for spectral analysis on virtual sensors.
>
> Previously we used Fieldtrip for this purpose.
> Function ft_sourceanalysis  created spatial filters, that we multiplied
> to each data trial  to create
> VERTUALSENSORS structure with field .trial{Ntrials}.{Nsources x
> Ntimepoints}.
> ft_freqanalysis then produced power spectrum of dimensions: Ntrials x
> Nsources x Nfrequencies
>
> What would be most efficient way to do the similar analysis in the MNE?
>
> Best regards,
> Viktoria
>
> _______________________________________________
> 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/20190610/0a0d08a4/attachment.html 


More information about the Mne_analysis mailing list