'Bivariate Granger Causality(spectral)NxN' in Connectivity

The results make me confused. There are two main colors, the red and the blue, between two nodes. I wonder is the blue gradully changing to red representing the cause to result? I can’t tell the causal relationship between them.

In this figure, the changing color represents the direction of the interface: from blue to red.

Secondly, wether any models like multi-variant autoregressive models or not are used in Granger Causality?
Thirdly, when using the Granger Causality to analyse the connectivity between two channels, which time scope is recommeded under the different sampling rate?

@Sylvain @leahy @hossein27en @Guiomar ?