Hi - you want something equivalent to the paired-t-test case on the web
page. something like:
acq acq acq sub
1-4 2-4 3-4 1
sub1acq1 1 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0
sub1acq1 0 1 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0
sub1acq1 0 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0
sub1acq1 -1 -1 -1 1 0 0 0 0 0 0 0 0 0 0 0 0 0
sub2acq1 1 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0
sub2acq1 0 1 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0
sub2acq1 0 0 1 0 1 0 0 0 0 0 0 0 0 0 0 0 0
sub2acq1 -1 -1 -1 0 1 0 0 0 0 0 0 0 0 0 0 0 0
....
Thanks, Steve.
On Fri, 25 Apr 2003, Conny Schmidt wrote:
> Hi Steve,
>
> thanks for your fast reply. I already have done the simple griup
> analyses with mixed effects.
>
> Do I understand it right, to handle each acquisition strategy applied as
> an own group? So having 4 groups with 14 subjects each? Wouldn't there
> be a conflict in statistics, as those groups wouldn't be independent
> from each other like groups of different subjects?
>
> Than you very much!
> Best regards,
> Conny
>
> Stephen Smith wrote:
>
> >Hi Conny, this sounds like an interesting study.
> >
> >Probably the first thing to check is just to run 4 simple group FEAT (ie
> >mixed effects) analyses for the 4 cases. Running them separately will let
> >you test that all looks sensible and give you a feel for how the different
> >acquisition strategies compare. Note, however, that higher mixed-effects Z
> >doesn't _necessarily_ mean better acquisition, but it often should do.
> >
> >Then, if all is making sense, you can actually test for significant
> >differences between the strategies by using a complete group-level for all
> >data, as you say, with appropriate additional covariates for the different
> >strategies.
> >
> >Thanks, Steve.
> >
> >
> >On Thu, 24 Apr 2003, Conny Schmidt wrote:
> >
> >
> >
> >>Dear FSL community,
> >>
> >>I have a study of 14 subjects and 4 acquisition parameter sets tested.
> >>For each the same stimulation paradigm was used.
> >>What would be the best way in FSL/FEAT to check, which acquisition
> >>strategy turnes out to be the best?
> >>
> >>I think there should be some kind of "repeated measurements statistics"
> >>applied. Furthermore, is there a way to inlude then also additional
> >>covariates?
> >>
> >>Best regards,
> >>Conny
> >>
> >>
> >>
> >>--
> >>------------------------------------------------------------------------
> >>
> >>Conny Schmidt
> >>Institute for Biomedical Engineering
> >>ETH and University Zurich Phone: +41-1-632 71 34
> >>Gloriastr. 35 Fax: +41-1-632 11 93
> >>CH-8092 Zurich, Switzerland Email: [log in to unmask]
> >>
> >>
> >>private:
> >>Oerlikonerstr. 5
> >>CH-8057 Zurich, Switzerland Phone: +41-1-363 13 39
> >>
> >>
> >>
> >
> > Stephen M. Smith MA DPhil CEng MIEE
> > Associate Director, FMRIB and Analysis Research Coordinator
> >
> > Oxford University Centre for Functional MRI of the Brain
> > John Radcliffe Hospital, Headington, Oxford OX3 9DU, UK
> > +44 (0) 1865 222726 (fax 222717)
> >
> > [log in to unmask] http://www.fmrib.ox.ac.uk/~steve
> >
> >
>
>
> --
> ------------------------------------------------------------------------
>
> Conny Schmidt
> Institute for Biomedical Engineering
> ETH and University Zurich Phone: +41-1-632 71 34
> Gloriastr. 35 Fax: +41-1-632 11 93
> CH-8092 Zurich, Switzerland Email: [log in to unmask]
>
>
> private:
> Oerlikonerstr. 5
> CH-8057 Zurich, Switzerland Phone: +41-1-363 13 39
>
Stephen M. Smith MA DPhil CEng MIEE
Associate Director, FMRIB and Analysis Research Coordinator
Oxford University Centre for Functional MRI of the Brain
John Radcliffe Hospital, Headington, Oxford OX3 9DU, UK
+44 (0) 1865 222726 (fax 222717)
[log in to unmask] http://www.fmrib.ox.ac.uk/~steve
|