Effective brain connectivity

Discussions
1

Hi everyone!
I would like to ask your opinion about measuring brain effective connectivity using EEG:

It's been said that Granger's causality is an effective connectivity measure and there are some papers using it in that way, however, since Granger's Causality is not actually a true causality objective measure, I was wondering if I can use directed eeg connectivity measures (Phase Transfer entrophy, Imaginary Coh,etc) as effective connectivity?

What do you think?

Thanks in advances!
:D!

1 Like
2

@hossein27en @Sylvain @Raymundo.Cassani @Marc.Lalancette?

3

Not sure there exists a "true causality objective measure" in electrophysiology yet. Hence, maybe consider testing possible effectivity connectivity effects in your data with both Granger and PTE? Note that imaginary coherence is not a directed measure.

4

Thanks for the reply!

well, you are right, maybe I said it wrong,

I was referring to whether EEG connectivity directed measures can be considered effective connectivity estimators,

I know if I use squared Coh in some research, then that research will inform about functional connectivity, but I wonder whether using directed measures can be reported as effective connectivity.

:grinning:

Thanks in advance!

ps:

You are right, I meant the original version on Imaginary Coherence (Nolte, 2004)

5

I found two papers that may help:

https://arxiv.org/ftp/arxiv/papers/2108/2108.13611.pdf

https://www.researchgate.net/publication/330704230_Towards_Detecting_Connectivity_in_EEG_A_Comparative_Study_of_Parameters_of_Effective_Connectivity_Measures_on_Simulated_Data

What do you think?

6

I think you can use PTE or Granger to test for possible asymmetric interactions between brain regions. I would refer to those as possible manifestations of directed connectivity.

1 Like
8

speaking of Granger causality, what about this implementation?

Let's discuss this in the GitHub issue, this thread was about a different topic.

9

Indeed none of the mentioned measures are about "true" causality (or all of them are). As @Sylvain said, you can call them directed or asymmetric, or conditioned-on-the-past measures. Effective connectivity was introduced by Aertsen et al. a long time ago, then popularized in the context of dynamic causal modelling, i.e. generative models of brain activity with some form of fitting/validation.

1 Like
10

thanks!