Hi Rick,
Thanks much for your reply.
Just to make sure I'm getting you, if I were to lowpass at, say, .25 Hz and I have a TR = 2 sec then, in the context of lowpass filtering during regression, this would entail including only a single regressor that completes a cycle once every four seconds and I wouldn't be taking such a big hit in terms of df. Is that correct? If so, it might be worthwhile as I do often see these high frequency variations in our BOLD data.
All best,
Paul