ok thanks , 

I was afraid this would be the case!!

I did chopped the timeseries ans selected specific blocks (frames); but now I have less timepoints and the data is not continuous anymore, so it's harder to do causality analysis, say like Granger or some other time-varying FC