Good morning

I was hoping to ask a couple of questions about using RSFgen, waver and 3ddeconvolve with the -nodata option in order to choose optimal stimulus sequences.

I am planning on running an experiment using a fast event related paradigm with 5 experimental conditions and 1 baseline condition. (the number of trials for all of the 5 conditions is equivalent to the number of baseline trials).
I am hoping to have 4 separate runs, containing an equal number of trials for each of my conditions.

My questions are as follows:

should i create 1D files for each of my runs individually or for my experiment as a whole? and what would be differences between the two options?

In running waver the AFNI how-to file suggests using a -dt 1. However, 1 would not be the value i would use running waver on my actual experimental 1d files as i am going to use a 3s TR. what is the impact of different -dt values to figure out optimal stimulus sequences and does it have to match the subsequent usage (meaning the actual analysis)?

i am sorry if the questions seem to basic but i cannot figure it ou from the help files alone

thanks
giulia