Hi Ankit,
sorry, missed the fact that you have modelled the mean in your first ev! You therefore do *not* need the -D option in randomise...
Sorry for the confusion!
Gwenaelle
> De: Matthew Webster <[log in to unmask]>
> Objet: Re: [FSL] GLM correlation with a covariate
> À: [log in to unmask]
> Date: Mercredi 25 août 2010, 11h58
> Hello,
> This should be fine (
> note that the Group column is superfluous here though ).
>
> Many Regards
>
> Matthew
>
> > Hello,
> >
> >
> > I know have asked a similar kind of question
> earlier but this question is a little different than
> previously asked. I am trying to run a TBSS analysis wherein
> I am trying to find correlation between a group of subjects
> and some continuous variable and also intend to include age
> in the analysis at a covariate to check if age affects the
> correlation in anyway.
> >
> > I have two group of subjects (I have merged them into
> one for finding the correlation between the data and
> continuous variable and called them group1)
> > EV2 is the continuous variable of interest. I have
> demeaned it and I am using the following design. Age is
> demeaned.
> >
> >
> > Group1 EV1 (data)
> EV2(variable) EV3(age)
> > 1
> 1
>
> x
> a
> > 1
> 1
>
> y
> b
> > 1
> 1
>
> z
> c
> > .
> .
>
> .
> > .
> .
>
> .
> > .
> .
>
> . and so
> on.
> >
> >
> > Contrast
> >
> > A>B
> 0 1
> 0
> > B>A
> 0 -1 0
> >
> > Is this the correct way of proceeding. Kindly advise
> >
> >
> > Thank you
> > Ankit Master
> >
--------------------------------------------------------------------
Gwenaëlle Douaud, PhD
FMRIB Centre, University of Oxford
John Radcliffe Hospital, Headington OX3 9DU Oxford UK
Tel: +44 (0) 1865 222 523 Fax: +44 (0) 1865 222 717
www.fmrib.ox.ac.uk/~douaud
--------------------------------------------------------------------
|