Hello, 

I'm trying to understand how the processing for the regressors for 1 column vs 3 column differ. Given 1s TRs, I thought that a one column format with say a 1 every 6th TR would produce the same output as a 3 column format with 1 sec durations of height 1 every 6 seconds (see below for an example). When I view the design matrix w/ no convolution, they look similar, but when I convolve w/ the HRF the 3-column format shows substantially greater amplitude changes in predicted signal. What am I missing?

Thanks,

Nathan

1

0

0

0

0

0

1