Professor, thank you very much. I have read the afni16_soup_to_nuts.pdf. It is a very good materials! As the direction, I think what I did before, bringing the 6 motion parameters as actual regressors via -stim_file into 3dDeconvolve, is right. But the regressors of interests should be given via -stim_times( change from 3dDeconvolve 2004). This point, I am wrong. Thank you very much again.