Hi Folks,
In my search of past message board postings, have found some related discussions, but no direct answer to my question, so here goes:
Our proposed design includes stimulus presentation at a rate of 0.5 seconds and our TR is 2 seconds. It seems that there are two places where we can deal with this mismatch of sampling rate. One would be to use the -tstim option in waver to specify the exact time that each stimulus occurs. The other would be to use the -stim_nprt option in 3dDeconvolve. What are the pros and cons of these two options?
One con that I see for the -tstim option is that we are trying to use the batch file in HowTo Number 3 to iterate through thousands of potential randomizations. So hand-translating each stim file generated by RSFgen into timing that can be used in the -tstim option seems prohibitive. (Unless someone has written a program to do this, or if waver could be altered to output .1D files sampled at a different rate than the input). On the other hand, one of the previous posts seemed to suggest that there was a loss of statistical power when the -stim_nprt option is used with 3dDeconvolve.
Any advice or clarification would be appreciated!
Lisa