Dear Boris,
> Let me re-phrase my previous post from yesterday regarding high-pass filter settings. I know that there has been a lot written about HPF in the archives. However, I couldn't figure out if the the rule of thumb (2*the longest time between onset of consecutive epochs of the same type) applies also to sparse-sampling experiments, i.e. would the constant delay (7.2s) in TR (10s) affect the rationale for choosing the HPF?
The delay that is set in sparse imaging paradigms shouldn't affect the
decision about a high pass filter—the important thing is the sampling
rate, which is the TR.
> In our specific case, we concatenate 6 same-type conditions for our sparse sampling experiment: a a a a a a b b b b b b a a a a a a b b b b b b c c c c c c b b b b b b etc. So we have alternating 60s blocks (6 x a TR of 10s) randomly distributed throughout the experiment. a) is a motor task, b) is a perception condition, and c) is a silent baseline. The biggest minimum interval would be 180s (6 x a, 6 x b, 6 x c) x 2. Though, would 460s for HPF be reasonable?
Based on your design I think that a 360 sec cutoff would make sense in
terms of not removing task-related signal changes. However, I believe
that's long enough that you'll also probably be keeping some of the
noise that normally would be removed through high-pass filtering, as a
result of the long blocks in your design.
Hope this helps! Good luck.
Best regards,
Jonathan
--
Dr. Jonathan Peelle
Department of Neurology
University of Pennsylvania
3 West Gates
3400 Spruce Street
Philadelphia, PA 19104
USA
http://jonathanpeelle.net/
|