Lydia, for each seed region under a condition, you'd have two regressors: the seed time series (step 2) and the interaction term (step 4). Therefore with four conditions at a seed region, yes you'd have totally five regressors since the seed time series is shared for all four conditions.
History of AFNI updates
