Group level nets

[Malthe Brændholt Sørensen]

Hi all,

First of all, I want to express my gratitude for the IDTxl software package.

I am trying to use IDTxl for a resting state analysis of two different conditions in MEG. I have only one (but long, 8 minutes at 1000 Hz) replication per condition per subject.

To compare the conditions at the group level, I first estimate each subject's network under each condition and then use the "dependent" setting in the "compare_between" method of "NetworkComparison" to compare the two sets of networks.

This approach gives me exactly what I want; however, I don't have a group-level quantification of the networks under each condition separately.

Is there a way to estimate common network patterns in a group of subjects?

Since I am not interested in the local TE, I have been considering estimating one network per condition by adding subjects as replications. Is this meaningful, or is there a better way to make group-level inference?

If this method is valid, I assume that using "dependent", "compare_within" will be a valid way of comparing conditions?

Thank you in advance for your assistance and for providing such a valuable tool for the research community.

Best regards, 

Malthe Brændholt