Computing phase coherence between two regions using electrode sets

Hi all,

I’m attempting to estimate frontoparietal theta phase coherence using an EEG dataset. I would like to estimate coherence between a set of frontal (e.g., AFz, Fz, FPz, AF3, and AF4) and parietal (e.g., PO8, P8, and P10) electrodes. I’m unsure what the best approach to achieve this is:

  1. Average the time domain signal over the electrode sets first. Then estimate coherence between the averaged ‘frontal’ and ‘parietal’ region.
  2. Extract the phase angle time series for each electrode separately, average phase angles over the electrode sets, and then estimate phase coherence.
  3. Compute coherence between all pairs of electrodes and then average coherence over channels.

Just to add, I have already applied a Laplacian filter to the data.

Any thoughts? TIA!

Hi Samantha. Your question is a common – and difficult – one: When and how to average to boost SNR? A few thoughts:

  1. This is a good approach, and something I’ve done in the past. Just keep in mind that you don’t want to average too many electrodes together given that you’ve applied the Laplacian.
  2. This is not a good idea, unless you are averaging the vectors in the complex domain. Otherwise, you run into the issue that mean([0 2*pi])=pi, even though 0 and 2pi are the same angle.
  3. This is what I’ve done most often. The advantage here is that you don’t need to worry about phase differences across electrodes.

So overall I recommend option 3.

Hi Mike, that’s great, thanks for such a quick and helpful response :slight_smile: