Thanks for your reply. I think the regressors I'm using here are fine as far as having breaks between runs, but I'm still unclear about my original question:

I'm trying to understand how to use the -nodata option to validate that the stimuli in my task are not collinear. When I use this option but change the TR length, I get the same results, which says to me that the -nodata option is only deciding if the inverse matrix can be calculated since that is only a function of where the stimuli are relative to each other irrespective of HRF length. I would like to use the -nodata option with an estimate of HRF length so I can see if I will be able to resolve task activation. Obviously that is not possible if several hundred trials are occuring in 5.5 seconds which is less than a canonical HRF. So my question is: is there some way to include an HRF estimate in the -nodata option to see if my stimuli are not collinear?

Thank you again,
Michael