Hi Giulia,

You should run 3dDeconvolve -nodata the same way that you plan
to use it when analyzing your data.

If you decide to use waver (which is not necessarily recommended,
perhaps you can give the stimulus timing directly to 3dDeconvolve),
presumably the -dt should match your TR.

I gather these stimuli are TR-locked, is that correct?

----

Are you talking about HowTo #3? If you use RSFgen, then the fact
that you want each run to contain an equal number of trials means
you need to run the program once per run, apply waver with the
proper -numout option (_if_ you decide to use waver), and catenate
those runs together for each stimulus type.

The make_random_timing.py program can do all runs at once, still
making sure that each run has the same number of trials. However
it creates files that are not appropriate for waver, as they are
in the -stim_times format.

So my questions for you:

1. Are your stimuli TR-locked?
2. What basis function do you plan to use in the regression?

- rick