Your design must be an extension of an unpaired t-test:
http://www.fmrib.ox.ac.uk/fsl/feat5/detail.html#UnpairedTwoGroupDifference
with two additional EVs per covariate, separate EVs demeaned within the group
(separable design).
Because your control group has only subjects with the same gender, I am not
sure how to enter it into your design. You may probably enter gender only for
the patent group.
You should start with the wizard, which will create the basic unpaired t-test
design for you.
Then add your covariates (demeaned within the group):
group con pat sex_pat age_con age_pat
1 1 0 0 a_con1-mean(a_con) 0
1 1 0 0 a_con 2-mean(a_con) 0
1 1 0 0 a_con 3-mean(a_con) 0
1 1 0 0 a_con 4-mean(a_con) 0
2 0 1 1.5 0
a_pat1-mean(a_pat)
2 0 1 -0.5 0
a_pat2-mean(a_pat)
2 0 1 -0.5 0
a_pat3-mean(a_pat)
2 0 1 -0.5 0
a_pat4-mean(a_pat)
con > pat
1 -1 0 0 0
pat > con
-1 1 0 0 0
corr_age_con > corr_age_pat
0 0 0 1 -1
corr_age_pat > corr_age_con
0 0 0 -1 1
Ave,
Auer, Tibor M.D. Ph.D.
Biomedizinische NMR Forschungs GmbH
am Max-Planck Institut für Biophysikalische Chemie
Am Fassberg 11
37077 Göttingen
Germany
Phone/Work: +49-(0)551-201-1725
Phone/Home: +49-(0)551-387-0076
Mobile: +49-(0)176-8012-7921
Mail: [log in to unmask]
> -----Original Message-----
> From: FSL - FMRIB's Software Library [mailto:[log in to unmask]] On
> Behalf Of Miro Drahos
> Sent: Wednesday, August 22, 2012 2:40 AM
> To: [log in to unmask]
> Subject: [FSL] design 2-group comparison with covariates
>
> Hi FSL folks,
> I am running TBSS and could use a little help with stat design. I am
comparing
> two groups (patients, controls) and including additional covariates, to
control
> for effects of sex and age.
>
> I have entered ones for all subjects in the "group" column (exchangebility
> block).
>
> con pat sex age(demeaned)
> 1 0 -1 1.7929
> 1 0 -1 -2.1471
> 1 0 -1 1.6029
> 1 0 -1 -0.3471
> 0 1 1 -1.8971
> 0 1 -1 1.2029
> 0 1 -1 -2.8071
> 0 1 -1 1.6729
>
> with contrasts [1 -1 0 0] and [-1 1 0 0].
>
> 1.) Is this matrix correct to ask about the different FAs between the
groups?
> 2.) How would I include interactions, e.g. age interaction?
>
> Thank you very much for your time and effort.
> Best,
> Miro Drahos
|