Thanks Gwenaelle for getting back to me so quickly! Sorry for the basic question, but how exactly do I do that? Do I just add a -D at the end of the randomise command or do I need to do something different to specify the covariate?

Date: Mon, 9 Aug 2010 15:21:24 +0000
From: [log in to unmask]
Subject: [FSL] Re : [FSL] Design and Contrast Matrices
To: [log in to unmask]

Hi Sarah,

you need to demean your IQ covariate... It should then be fine.

Cheers,
Gwenaelle

--- En date de : Lun 9.8.10, segpg _ <[log in to unmask]> a écrit :

De: segpg _ <[log in to unmask]>
Objet: [FSL] Design and Contrast Matrices
À: [log in to unmask]
Date: Lundi 9 août 2010, 15h10

Hi Everyone,

Thank you for all your help so far. I'm getting some unexpected results out of randomise in tbss. Would someone mind checking if my design and contrast matrices (from the Glm gui) are correct. I have two groups and IQ as a covariate.

My setup looks like the following, except with many more subjects.
Design:
group EV1 EV2 EV3
1     1    0    96.1
1     1    0    108.1
1     1    0    115.5
1     1    0    101.6
2     0    1    59.4
2     0    1    51.8
2     0    1    60.7
2     0    1    47.1

Contrast:
group A>B     1.0   -1.0   0
group B>A    -1.0   1.0   0
group A mean    1   0   0
group B mean    0   1   0

Thanks very much!

Sarah

--------------------------------------------------------------------

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

--------------------------------------------------------------------