Gang,

Sorry this is getting so involved, but I am still confused.

I don't understand your statement: "since your irf's have a peak of one"?

I have an original signal time series "normalized" as percent signal change, and I have a regressor that is based on the output of the waver program. Waver has been given an input that looks something like this:

0 0 0 0 0 0 0 0 1 1 1 1 1 0 0 0 0 0 0 0 0 1 1 1 1 1 0 0 0 0 0 0 0 0 1 1 1 1 1 0 0 0 0 0 0 0 0

(i.e. there are three "on" periods and each "on" period consisted of stimulus presented continuously over a 5 TR duration)

I want to use one "lag" (using the output of waver as the only irf-regressor of my 15-sec stimulus) in 3dDeconvolve. Will this single beta coefficient (calculated for a single (waver-produced) regressor that peaks at 1.604) for the stimulus be meaningful in terms of % signal change? Or does the peak of the output function from waver need to be made to be =1?

I have just run a little test, where I have made the output of waver that I use for my regressor to peak at 1, and another output I have made to peak at 2. When I put these two regressors through 3dDeconvolve separately, I get, for an example voxel, a beta coefficient for the waver with peak "1" = 8.0817, and the beta coefficient for the waver with peak "2" = 4.014718 (the baseline coefficients for both of these 3dDeconvolve output regressions is the same). So, the fact that the peak value of the regressor I use impacts the magnitude of the beta, brings me back to my question.

Should I make the output function of waver peak at 1, or do I use the output that waver gives me (where peak is set =1 but actually results in a higher peak because my stimulus design is not an event-related design, and so the gamma waves overlap)?

I know that in terms of statistics, if all the data are treated similarly, this will not impact whether I find a significant difference between two stimulus conditions. But, if I want to report a value for % signal change for each of those conditions, I'd like to be sure that I have estimated it correctly.

Again, I apologize for my ignorance, but it's just not so clear to me.

Thanks for your help - Liz