Bivariate Granger Causality VS Phase Locking Value

Discussions
1

Hi all,

In the Brainstorm I am calculating connectivities using two different methods.
Directed —> Bivariate Granger Causality (Spectral)
Undirected ----> Phase locking Value (PLV)

The problem I am coming across is that after applying graph theoretical measures on these connectivites I am getting contrary results. Functional connectivites obtained using GC is showing increase average nodal degree in group 1 comparision to group 2 while FC obtained using PLV shows totally opposite results.

I am wondering what would be the reason behind it. Is it possible that using directed or undirected connectivity measure cause such a big difference?

Any comments/suggestions would be helpful!!

Thanks,
Amna

2

Hi Amna,

These two measures are based on different concepts in brain connectivity, so you may not expect similar results since they assess different things. However, in some cases, you can find consistencies among these measures.

Selecting the proper measure is mostly based on your research goal and your hypothesis. While PLV usually gives more straightforward results, Granger is in some cases a bit problematic in terms of interpretation.

Bests,
Hossein