Hi Andy,
No, I believe nothing like that is done. It seems
presumptuous, giving more importance to the first
modulators in some way, and as you say, would
lead to varying results, depending on the order.
In 3dDeconvolve, the regressors of interest are not
orthogonalized either, and so it would make little
sense to do so to the modulators. The regression
is solved "as is".
- rick