Thanks David, I ran several permutations and come to the same conclusion. Whether I put the WM timeseries into the confound or as EV will give the same results. The only difference is when I do not filter the data (in the prestats tab before the nuisance model step) and I do not apply the filtering in the prestats tab (in the nuisance model) but apply filtering to the WM EV.
Anyway, I am convinced now. I am familiar with the concept of a highpass filter (letting pass only frequencies above a certain threshold), but somehow I didn't see how this is propagated through the processing steps.
Cheers,
Torsten
|