To find out cross frequency coupling with MVL can we directly give signal and do the following steps, simulate-simulate PAC signals then frequency-phase amplitude coupling ? or do we need to filter the frequency bands first and then do we need to take the hilbert transform ?
If you have your own signal, there is no need for simulating PAC, you can easily do the analysis on it.
(The algorithm will take care of all filtering and Hilbert transforms).
You can also try time-resolved PAC (available in the same menu)!
Thank you. In time resolved PAC, will it show the maximum PAC strength after processing signal?
Sure,
The value is stored in the output structure (sPAC.maxPAC). You can easily extract it from the results in time-fA domain (initial tPAC output), or transfer the results to comodulogram (fP_fA) map which show the maximum value even in the graphical illustration (using: Run > Frequency > tPAC > Extract comodulogram).
When using the MLV pipeline for PAC (non-time resolved) is there a way to concatenate the trials together prior to processing or is it only possible to run on each individual trial epoch and then average the the values across trials?
Thanks
(If your question was technical and not theoretical: you can use process "Standardize > Contactenate time")
Perfect! It was technical actually. Thanks
Sorry for the premature "Perfect!" It looks like that option doesn't work for source localized signals.
I'm looking to apply an option similar to what's available for a variety of the connectivity analyses ("Concatenate input files before processing (one file)"). Doesn't look like it's an option available for PAC analysis, but can probably apply as a separate step if there's a way to concatenate source resolved epoched data.
Indeed... There is another solution, with the process "Extract > Extract values", with the option "Concatenate time".
Note that this may create gigantic files, that would days hours or days to process.
You might be interested in working with regions of interest instead.