Hi Rick,
Sorry for the multiple replies! But now I see another problem with your suggested method 1 (model the 0s stimulus with GAM or BLOCK(1.5,1), and model the 13s stimulus with 16 seconds of TENTs):
I was thinking about using TENT because I don't have any specific assumption about the shape of the HRFs. If I use GAM or BLOCK(1.5,1), then I feel like I'm ignoring what is happening, say, 1.5s after stimulus presentation. But actually I'm interested in the entire 13s period (i.e., from 0s to 13s). Is there a way to solve this problem?