Hi all, I have used eConnectome for computing connectivity in individual participants but I would like to have the connectivity in a group. Have someone done this before? Any advice which is the best way to do it?
As eConnectome deals with a simple data set at the time, I am sure which it is the best/correct way to enter the data.
- averaging all participant EEG data and then computing the DTF?
- averaging the individual DTF?
- ??
Advice or suggestions more than welcome!
Valia
Averaging individual DTF will give you more meaningful results
Hi!
I assume you use Matlab or Octave, so you might try EEGLAB's extension SIFT.
EEGLAB and SIFT work together to allow cohort level observations on different features including connectivity. You will want to have a decently powerful workstation.
Let me know how it goes.
Cheers!
Hi. This is a more general "how to compute group connectivity" issue rather than just EConnectome.
So, it is advisable that you estimate individual participant interchannel DTFs. Then test if the distribution of DTF values for all the subjects is similar or if there is great intersubject variability in data.
If there is a large variance in DTF values across the subjects, then you may have to normalize the data.
However, with or without normalization, you will have multiple individual subject DTF matrices. You can then calculate group-level statistics using methods published in some of the connectivity papers correcting for multiple comparisons and handling the direction component in the data (could be based on PLI/cross-corr/PDC/DTF, etc etc).
A word of caution, do not average without testing for intersubject variance. DTF is a fairly sensitive measure particularly if its scalp data. It may work better on invasive data (not sure which one is being analyzed in your case).
GL!
CG.
- Badges
- Science topic