Dear Dorian,
no matter how you set up your design, you have to model all features
that you assume to be present in your data. Even if you are not
interested in some modulation, some event type etc. you must not omit
them from your design. If you do omit them and there is an effect which
can be described by the omitted regressor, it will end up in the
residuals.
If your modulations are not orthogonal to each other, you may want to
orthogonalise them before entering them into SPM. This gives you control
how common variance between two modulations will be assigned.
Hope this helps,
Volkmar
Am Freitag, den 13.02.2009, 12:48 +0100 schrieb Dorian P:
> Dear all,
>
> Two basic questions:
> Will the results of a parametric modulation be different if:
> 1. Analyzed in a GLM dedicated only to that modulation OR
> 2. Analyzed together with different other parametric modulations in
> the same GLM?
>
> And will the result of the 1st order modulation be exactly the same as
> the linear regressor in a 2nd order modulation?
>
> The answers to these questions may save me a lot of time by using the
> GLM I already have.
>
> Thank you.
>
> Dorian
> PhD Student
> Ruhr University Bochum
>
--
Volkmar Glauche
-
Department of Neurology [log in to unmask]
Universitaetsklinikum Freiburg Phone 49(0)761-270-5331
Breisacher Str. 64 Fax 49(0)761-270-5416
79106 Freiburg http://fbi.uniklinik-freiburg.de/
|