Hello all,
Is there a way to specify waver so that the waveform isn't convolved with the reference function in a linear way? For example, I have been doing a block design experiment but I do not think the hrf is additive in a perfectly linear fashion. So, as it is now, when I use waver to convolve the -WAV waveform to a 1s and 0s file, I end up with a graph that only shows one peak for one block period. In my experiment the block lasts for 20 seconds and I am using a TR of 2 seconds. So, I have 10 1's in a row and instead of seeing just one peak, it would be nice to create a model that shows the peaks of BOLD response deconvolved/separated a bit. I realize it might be over-fitting the data if I represent the model with 10 differnt peaks during the block period, but it would be nice to see three or four peaks for one block period, thus representing the data a little bit better. Please let me know if this makes sense and if it is possible. Thanks.

Nicolas