Envelope_correlation

💬 Support & Discussions
1

Hi!
I want to analyse resting data using the envelope correlation of the hilbert transformed data for different frequency bands.
In the mne version v.0.22.0. there are two examples, one in source space using dSPM the other one in volume space using a lcmv beamformer.
I am a little bit confused how the noise covariance matrix is calculated in these examples. Before looking at these examples, I used empty room data for calculating the noise covariance instead of the broadband raw signal. Which option is better? And in case of using empty room data for calculating the noise covariance data, should one filter it before or use the broadband signal as well?
Thank you very much for your help!
Cheers, Hannah

2

hi,

I would try empty room first. And for your source imaging you must use the same preprocessing for empty room and subject data (same temporal filter, same ICA rejection etc.)

HTH
Alex

3

How do we calculate the envelope correlation across two channels?
@agramfort

4

@BarryLiu97 did you search the docs before asking? If you’re still not sure after reading the docs, please open a new question thread with details of which parts of the docs don’t make sense.

  1. mne.connectivity.envelope_correlation — MNE 0.22.0 documentation
  2. Compute envelope correlations in source space — MNE 0.22.0 documentation
  3. Compute envelope correlations in volume source space — MNE 0.22.0 documentation