> Doesn't defining a TENT function actually look pretty similar to the Cohen
> gamma variate curve in terms of time to peak? Isn't it just another approach
> that still constrains the (assumption of the) impulse response?
No, there is no constraint at all on the response shape when TENT basis functions are adopted, similar to the old approach of deconvolution with -stim_minlag and -stim_maxlag in 3dDeconvolve. You might have confused it with another option, Jim? See "3dDeconvolve -help" for more details.
Gang