Dear FSL experts, I ran MELODIC + dual regression to estimate the average time series of each component for each subject (stage 1 outputs). I want to use these for further analyses. Now I'm wondering how exactly the individual time series are estimated. Some of my components contain negative as well as positive connections. So if the stage 1 outputs are generated by averaging time series within each component (as I have read in a few papers), won't this lead to faulty results if there are positive and negative connections? I guess that "averaging" might just be wrong terminology here and the negative connections are somehow accounted for. Still, I want to make sure because this might explain some of my results. Thanks a lot for your help! Best, Katharina