Define the scientific needs to extract LF phase

[oliche]

The instantaneous phase of the LF (and possible also AP), and especially the difference in such instantaneous phase across channels will be very informative of brain region and boundaries between brain regions respectively, as seen on the raw data.

It is not clear how to compute the instantaneous phase as the Hilbert transform is slow, and STFT does not work (Slack).
One other option is complex coherence, but the standard Python SciPy implementation throws away the phase.

If looking only at boundaries between brain regions, another approach is to not compute the difference of the instantaneous phase across channels, but rather compute the crosscorrelations of adjacent channels directly.

This requires further identification of the scientific needs.

Cf also GC retrospective --> Peter Dayan's note

[oliche]

I would do a review of methods and tools to compute phase precession and features relating the LF to the AP band