I am interested in doing a beta series correlation analysis and thus want to use -stim_times_IM to extract a beta estimate for each trial. However, in my event-related design, the event of interest (the delay period in a working memory task) varies in duration across trials. So I've been using -stim_times_AM1 to perform duration modulation for this regressor. My question is, is there a way to combine the stim_times_IM and stim_times_AM1, and thus get a beta weight for each trial of a series of duration modulated events?

Alternatively, it seems like one option would be to use -stim_times_IM and seperate the delay period trials into groups with the same duration (there are only three different durations). In other words, split the single delay period -stim_times_AM1 regressor into three -stim_times_IM regressors, each with their own specific duration built into the Rmodel. Is this the approach I need to pursue?

Thanks!
Ryan