Dear AFNI experts,

I have a question about the purpose of demeaning motion parameters in 3dDeconvolve step.

I have 6 runs (they are 3 different tasks), therefore, I preprocess them using afni_proc.py without regress block. So I need to feed in appropriate motion parameters in writing 3dDeconvolve script for each task type. In the earlier version of afni_proc.py, I found that it used actual motion file (dfile.rall.1D), but the later version seemed to be using demean_motion file (demeaned by each run). Since it's been demeaned at each run, the values would be different from using raw dfile.rall.1D.

So I am wondering whether using demean motion file vs. raw motion file make much difference? Is there a difference in terms of interpreting beta values of regressors of interest (output from 3dDeconvolve)?

Thanks in advance!
SJ