I am working with a mixed block event related design and I am looking for a way to make my block/sustained predictor a boxcar without delay. The model contains cue predictors so I don’t plan on having a ramp up (delay) for the boxcar. Currently I am using block as my basis function for the sustained predictor. If I am not mistaken block is a convolution of a gamma function and a boxcar function and it has a delay. Is there a way to eliminate the delay (ramp up) and is there a pure boxcar function option? I am thinking EXPR might be an option.
Thanks for your time