I want to calculate the connectivity of EEG signals. The signals are complex, since I have to do some special preprocessing, and this gives me complex values back. I still have the signal shape of channel*time though.
My question is: Can I directly use the complex signals as input in mne_connectivity.spectral_connectivity_time (or mne_connectivity.spectral_connectivity_epochs)? Are the results still correct?
All the methods in spectral_connectivity_time() and spectral_connectivity_epochs() are based on computing a cross-spectral density from real-valued time series data.
I can’t say if an error will be raised (might be a casting warning for discarding the imaginary part), but the results would not be valid.
Off the top of my head, I’m not aware of spectral connectivity methods designed for complex signals.