Hi, yes you should demean the covariates (before padding with zeros).
Note that strictly, for any given contrast, you should move any non-
involved covariates into a separate confound matrix and use the -x
option.
Cheers.
On 1 Feb 2007, at 16:40, Bum Seok Jeong wrote:
> Hi,
>
> I'd like to compare FA values with age, gender as covariate (like
> univariate GLM) in TBSS.
> My design matrix is below
> group EV1(control) EV2(patient) EV3(age_control) EV4
> (age_patient) EV5(gender_control) EV6(gender_patient)
> 1 1 0 19 0 1 0
> 1 1 0 26 0 2 0
> 1 1 0 22 0 1 0
> ....
> ....
> 2 0 1 0 23 0 2
> 2 0 1 0 18 0 1
> 2 0 1 0 24 0 1
> ....
>
> My contrast is below
> EV1 EV2 EV3 EV4 EV5 EV6
> control>patient 1 -1 0 0 0 0
> patient>control -1 1 0 0 0 0
>
> My question is 'demean'.
> Do I have to demean all EVs to control age, gender like below?
> EV1 EV2 EV3 EV4 EV5 EV6
> control>patient 1 -1 0 0 0 0
> patient>control -1 1 0 0 0 0
> mean_control 1 0 0 0 0 0
> mean_patient 0 1 0 0 0 0
> mean_control_age 0 0 1 0 0 0
> mean_patient_age 0 0 0 1 0 0
> mean_control_gender 0 0 0 0 1 0
> mean_patient_gender 0 0 0 0 0 1
> and then, type 'randomise -i all_FA_skeletonised -o tbss -m
> mean_FA_skeleton_mask -d design.mat -t design.con -n 5000 -c 3 -V'
>
> OR do I just add -D in radomise command line without input of
> 'mean_control, mean_patient, mean_control_age,
> mean_patient_age....' in my contrast like below?
> 'randomise -i all_FA_skeletonised -o tbss -m mean_FA_skeleton_mask -
> d design.mat -t design.con -n 5000 -c 3 -V -D'
>
> Thanks,
>
> bsjeong
------------------------------------------------------------------------
---
Stephen M. Smith, Professor of Biomedical Engineering
Associate Director, Oxford University FMRIB Centre
FMRIB, JR Hospital, Headington, Oxford OX3 9DU, UK
+44 (0) 1865 222726 (fax 222717)
[log in to unmask] http://www.fmrib.ox.ac.uk/~steve
------------------------------------------------------------------------
---
|