REST-GCA

Submitted by isa_nz on Mon, 09/21/2015 - 05:21

Hi there,
I've got a question regarding the granger causality analysis. I've got an fMRI block design where I get activations in two main language areas. I'm now interested to test whether region A exerts influence on region B or whether region B exerts more influence on region A. Can I use REST-GCA even though it is a block design?

Thanks very much for your help!

Isabelle

Submitted by zangzx on Mon, 09/21/2015 - 22:58

 Hi Isabelle

Rest-gca was designed to perform gca on resting state data in principle, but you can also use it for task data. There are several suggestions before you use it.

Before you wanna perform gca, you might need an additional preprocessing to get rid of the 'blocks'. Here is one of a recently published paper describing the method:
 2014 Jan 1;84:888-900.

Test-retest reliability of fMRI-based graph theoretical properties during working memory, emotion processing, and resting state.

Cao H1, Plichta MM, Schäfer A, Haddad L, Grimm O, Schneider M, Esslinger C, Kirsch P, Meyer-Lindenberg A, Tost H.

But this method can only provide you the general task signal which means you can't perform condition-specific GCA later on. If you do want to have condition-specific effective connectivity, the best choice might be the dynamic causal modelling.