Hi all,

I have previously posted here and received help with various aspects of my analysis. One of the key aspects of my analysis was to use AFNI to remove all regressors of no interest (including baseline/trend) so as to have a relatively "clean" signal to input to my non-AFNI analyses.
Given that I had not calculated PSC pre-3dDeconvolve, and that I want to use 3dSynthesize to remove the baseline and motion aspects of the signal, which option for calculating PSC is most theoretically sound?:

1. Subtract motion, take that result and divide it by the baseline? or
2. Subtract both motion and baseline, take the mean of the outcome, and divide by that?

I'm sure this must be relatively simple to answer, but I can't quite figure out what the key bit is. To clarify, the later analyses will be done on a per-voxel basis, but since I expect results in clusters, I will have to plot time courses per cluster.

Thanks,
Mahen