Print

Print
That makes sense, thanks. .. so just to clarify:

Without orthogonalization, the two PMs will be given equal 'weight' in explaining the variance? Whereas with orthogonalization, the first PM will be given priority, and the second PM will only explain the variance left over?

Also, if I only have one PM, then it doesn't make a difference whether I set orthogonalization on or off? (Note that I do have other regressors though, but it seems that the orthogonalization option is specifically related to the PM dropdown).

Thanks,
Joelle

On Tue, Aug 18, 2015 at 2:25 PM, MCLAREN, Donald <[log in to unmask]> wrote:
This means that the first PM will explain as much variance as possible, then the 2nd PM will explain any remaining variance. Without orthogonalization, the PMs compete to explain the variance in the data.

Best Regards, 
Donald McLaren, PhD


On Tue, Aug 18, 2015 at 1:52 PM, Joelle Zimmermann <[log in to unmask]> wrote:
Hi SPMers,

I'm wondering if to orthogonalise modulations under my Parametric Modulator? It would be great if someone could give me a bit of an explanation of what orthogonalising modulators really means, maybe with an example to make this concept a bit clearer.

Thanks in advance,
Joelle