I get what one would expect if using AM2 instead of AM1, sorry about that confusion. Now back to using RT...

> Gang Wrote:
> >
> > This is why I asked before how you would like to run group analysis regarding each modulation
> > effect. Per RT, the question is whether you want to obtain the interaction between tasks and groups
> > while controlling RT at its average? If so, treating RT as a modulation variable seems better
> > than the alternative (dmBLOCK).

> rick reynolds Wrote:
> > If the durations of the stimulus classes vary from 1 to 4 seconds,
> > it might be a good idea to stick with duration modulation. GAM
> > is not such a good fit for a 4-second stimulus.

The RT is more variable than I thought. It actually goes up to 10 seconds, with the mean and stdev of mean(RT) being (2500ms and 850ms). Using a Gamma function, as is, seems inappropriate. I was thinking perhaps I could use each subject's mean RT as a duration parameter in the GAM function, but that still will do a poor job of fitting for those trials where subjects waited ~10 seconds. Doing dmBLOCK w/ RT as the duration modulator makes sense - would it be non-sensical/invalid to imagine RT acting as BOTH amplitude and duration modulation? It seems weird to me - but if you have any insights, by all means. Otherwise I'll be going back to treating it as a duration modulator only, and keeping sum,diff,delay as amplitude modulators.

Thanks so much for your help. Though I have lots to learn, I understand this much more than before.
Hugo