Hi, I have a quick question regarding the process of orthogonalization
of parametric regressors in the case where one has included more than
one parametric modulator of a single predicted response. Let:
X = predicted response (post-convolution), assuming fixed amplitude
P1 = first parametric modulator
P2 = secondar parametric modulator
My understanding is that P1 is orthogonalized with respect to X by
simply being demeaned prior to computing the crossproduct of X and P1.
However, I do not understand how P2 is orthogonalized with respect to
P1.
Thank you for any explanation you can offer.
Cheers,
Bob
----------------------------------------------------
Bob Spunt
Doctoral Candidate
Department of Psychology
University of California, Los Angeles
|