Hi Nic,

I split your message from the rest of the thread
since that was old and you have so many little
topics in this one message. I hope that is okay.

It is not clear how you plan to use the TENT betas
for scaling, so I will ignore that at the moment.

Regarding how much censoring is too much, consider
applying a maximum censor number or fraction limit
in a similar manner to how you would apply it with
BLOCK or other basis functions. With TENTs, it is
certainly more likely that censored fraction of TRs
is higher, since each "response" does not cover so
many TRs. The key is just having enough (relatively
independent) samples for each TENT beta. In your
case, overlap is not a concern since this is a slow
event/block design.

The number of "num TRs per stim" entries is large (if
that is what you mean) because instead of using just 1
regressor for BLOCK, say, you are using 7 or 8 TENTs.
Each regressor (of interest) gets an entry on that line.

I will leave it to others to respond to your goal of
"reduction of cerebrovascular components".

Your question about getting test statistics for the
iresp is not clear to me. Exactly how does this
differ from the F-stat of the breath class?

What do you mean by "iresp bricks = 2.5"? Assuming
TENT(12.5,42.5,7), there should be 7 betas for the 6
intervals of length 5s (from duration 30s). And the
iresp output should still be on the TR grid.

- rick