Dear Clara,
If you set "orthogonalise modulations" to "yes" (the default), SPM will
perform serial orthogonalisation within condition over parametric
modulations. In particular, all parametric modulations will be
orthogonalised to the unmodulated regressor (i.e. the first one), which
corresponds to mean-centre them.
If you set "orthogonalise modulations" to "no", the parametric modulated
regressors are created using the unmodified user-specified values.
Best regards,
Guillaume.
On 25/05/2021 22:14, Clara Sanches wrote:
> Hi,
>
> Ok, thank you for your answer!
>
> So, just to be sure, in SPM12, the regressors are not automatically
> mean-centered regardless of whether I set the spm_orth argument to 0 or 1?
>
> Best regards,
> Clara
>
> Guillaume Flandin <[log in to unmask] <mailto:[log in to unmask]>>
> escreveu no dia segunda, 24/05/2021 à(s) 04:20:
>
> Dear Clara,
>
> Orthogonalising with respect to the unmodulated regressor corresponds to
> mean-centring, so, in your case, you could enter mean-centred parametric
> modulators with "orthogonalise modulations" set to "no".
>
> Best regards,
> Guillaume.
>
>
> On 22/05/2021 00:34, Clara Sanches wrote:
> > Hello,
> >
> > I am reviving this discussion on how to set parametric regressors
> in SPM12 since I am also trying to find how to orthogonalize two
> parametric regressors only with respect to the unmodulated regressor
> without orthogonalizing them with respect to each other (SPM8
> orthogonalizes each regressor with respect to the ones preceding it,
> I’m not sure if this is the case for SPM12 but from what I’ve been
> reading it is, please correct me if I'm wrong).
> >
> > So, from the answers here, there is no way of doing this in SPM12?
> I would have to set orthogonalization to ‘no’ and enter my
> regressors modified? Or define two separate models with the order of
> the parametric regressors changed.
> >
> > Since this thread is from 2019 I was wondering if someone has a
> found way (maybe through code changing) of orthogonalizing
> parametric regressors with respect to the unmodulated regressor only?
> >
> > Any suggestion is much appreciated!
> >
> > Thank you,
> > Clara
> >
>
> --
> Guillaume Flandin, PhD
> Wellcome Centre for Human Neuroimaging
> UCL Queen Square Institute of Neurology
> London WC1N 3BG
>
--
Guillaume Flandin, PhD
Wellcome Centre for Human Neuroimaging
UCL Queen Square Institute of Neurology
London WC1N 3BG
|