Hi, Stef-

I am not quite certain why the EPIs wouldn't be processed together... Going back to the initial question you had asked about this, our thoughts were these:
[afni.nimh.nih.gov]
... and I'm not sure why that isn't appealing? One main thing to avoid is having drifts within a segment dominate the overall correlation after being concatenated together.

Having a lower motion censoring limit seems fine. Indeed, one basically does treat naturalistic scanning as resting state, from a processing point of view. The polynomial order seems to be what about the afni_proc.py default is (though I think that one is based on time duration, and not just number of time points).

Re. last question about resting state: yes, you should be able to process your resting state in the same way. You should be able to put both runs into the same script.

--pt