afni.nimh.nih.gov

I think that's right. There shouldn't be a "trend" to de-trend with just two time points, and replicates of these pairs won't make it a trend. Of course, we don't have any experience with this kind of dataset, so it's interesting to hear about your experience. Thanks!