Hi Danny,
That looks fine at a glance.
One point I am not sure about is the use of -polort 2.
Presumably there was a polort in the original regression.
But why remove even a linear trend in the betas?
While I might be inclined to just demean betas.nii and
then get the correlations with 3dTcorr1D, the results
should be the same (aside from the polort question).
It seems strange though if the single subject results
are only really strong near the seed to have them so
consistent as to generate significance all over the
brain. That seems difficult to address without looking
closely at the data.
- rick
