Jasmin,

Yes, let it run longer! This is one of my biggest problems with PESTICA, the slowness of this second stage. You could also skip the second stage, it depends on how wide the filter is, but I don't have an objective recommendation for whether you should or not.

For the dummy scans, most scanners are setup so the acquisition is off for the first 3 seconds of scanning, so most users will see data that is already equilibrated. To check yours, open a dataset in AFNI and open the graph on a view, and see if there is a much stronger value for the first volume that gets much weaker for volume 2 and is equilibrated by volume 3 and beyond. If your scanner acquires those first volumes, then you'll have signal that is stronger on volume 1 than anywhere else. If you don't, then the signal will be flatter. When in doubt, ask a physicist or one of the AFNI people to look if the signal is equilibrated. Or please feel free to post an image of the timeseries and I would be happy to look.

Erik