I not sure I understand what you're suggesting...probably because I'm used to conceptualizing it differently. Currently (for a given task), I have .1D files of 1s and 0s (on-off) for each task condition. For this particular task, there is a 1 or 0 every second (for 512 secs). In the past, I've used waver to make a reference function for each of the task conditions, combined them into columns in a single .1D file, and then called upon them in 3dDeconvolve. For example:
-stim_file 1 task-waver.1D'[0]' -stim_label 1 fix
I can then use these labels in -gltsym to set-up my contrasts. I would really like to continue doing it this way (largely because I understand it). Is there any way to make the 1 second reference function work for this (e.g., resampled to 256 2sec 1s and 0s)?