[Mne_analysis] cross_spectral_density_matrix

[Edoardo Pinzuti]

        External Email - Use Caution        

Dear MNE team,

I needed to implement a csd estimation that return the single trials data.
For creating the common filter I used the pooled data (avg csd) but then I
apply the filter to single condition (with single trials).  Keeping  single
trials allow me to have a first level statistic condition vs baseline.

To do that:
- I don 't allow the function to average the csd at the end and I keep the
trials as second dimension ( for the function this work treating them as
different frequencies bin)
-I apply the filter at each trials, now the function instead of applying
the filter at each frequency it      does it at each trials

I get a source power n vertices x n trials which is what I need, people
more experienced with the beamforming do they see any issue with this
approach ? In few words I used the same function csd_morlet and
apply_dics_csd  but instead of the second dimension frequency I put trials.

Thanks you
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.nmr.mgh.harvard.edu/pipermail/mne_analysis/attachments/20201104/c35037c8/attachment.html