Hello,
I am trying to understand these but it gets complicated when the replier explained it with "con-time1 con-time2 pat-time1 pat-time2 con1 con2 pat1 pat2 pat3 " and then the matrix.
Could somebody help me with this?
I want to study a longitudinal difference between lets say 4controls and 2 patients with age and sex as covariates...also I always demean the age but is it necessary?
Thanks,
Rodrigo
Subject:
Tue, 9 Feb 2010 20:45:47 +0000
Content-Type:
Reply
Hi Adnan,
say you've got 2 controls and 3 patients (n=5), you'll have 2x2+n EVs, for instance:
con-time1 con-time2 pat-time1 pat-time2 con1 con2 pat1 pat2 pat3
1 0 0 0 1 0 0 0 0
1 0 0 0 0 1 0 0 0
0 1 0 0 1 0 0 0 0
0 1 0 0 0 1 0 0 0
0 0 1 0 0 0 1 0 0
0 0 1 0 0 0 0 1 0
0 0 1 0 0 0 0 0 1
0 0 0 1 0 0 1 0 0
0 0 0 1 0 0 0 1 0
0 0 0 1 0 0 0 0 1
If you want to know where there is an interaction between groups and time points (so where the patients significantly differ in their changes from the controls):
1 -1 -1 1 0 0 0 0 0
Please also have a look at the recent archives with the subject:
"Randomize design and contrast matrix for Fractional Anisotropy"
Hope this helps,
Gwenaëlle
--- En date de : Mar 9.2.10, Adnan Majid <[log in to unmask]> a écrit :
> De: Adnan Majid <[log in to unmask]>
> Objet: [FSL] Design matrix for longitudinal change in two groups
> À: [log in to unmask]
> Date: Mardi 9 Février 2010, 19h59
> Hi everyone,
> I'm trying to use TBSS to study the
> differences in longitudinal change in two groups (ctrls vs.
> exp) over the period of one year. Essentially, I'd like
> the design matrix and contrasts for the randomise
> step to model a 2-factor ANOVA with repeated measures on
> one factor, though I'm not sure exactly how to set it
> up. There are some good examples on the FEAT
> detail page, but I wasn't sure if any example really
> represented what I'm looking for.
>
> Thanks for any help on how to set up such a
> design matrix and contrast file. Best wishes!
> Adnan
>
|