Hi Nic,
I am not sure what you mean between the use of iresp
and reml betas, unless it is just that iresp comes
from 3dDeconvolve. My main point of discussion was
centered around whether there is sufficient power to
estimate the TENTs.
It was also not my intention to suggest which TENT
betas to use (e.g. 2-4 our of 0-6). My point along
those lines might have been just that 20 TRs to
estimate the TENTs is actually very powerful since
you are further averaging those values to get Bave.
So Bave has a lot of samples driving it.
- rick