Thanks Doug for your reply. I forgot that the number of columns is number of classes by number of covariates +1.
So, I only want you to confirm me that the extra zero that I've seen in the FSwiki example is an erratum (see contrasts 7 and 8 in Four Groups (2 Factors, Two Levels), One Covariate. DODS only. QDEC-able) and that I don't have to add any extra zero per covariate when testing the interaction effect.
Many thanks again.
Jose
----- Mensaje original -----
De: Douglas N Greve <greve@nmr.mgh.harvard.edu>
Fecha: Martes, 18 de Enero de 2011, 16:53
Asunto: Re: [Freesurfer] contrast matrix for 2 discrete factors and 3 continuous variables
A: Jose Luis Cantero Lorente <jlcanlor@upo.es>
CC: freesurfer@nmr.mgh.harvard.edu
> Hi Jose, with 6 classes and 3 continuous variables, I think you
> should have 24 regressors, which means 24 columns in the
> contrast matrix. Yours only has 18. Otherwise the contrast to
> test between B and C is correct (just need to add 6 zeros). For
> the interaction, you would need
>
> 0 0 0.5 -0.5 -0.5 0.5 0 0 ....
>
> doug
>
> Jose Luis Cantero Lorente wrote:
> >
> >Dear Freesurfers,
> >
> >
> >My experimental design includes 2 discrete factors and 3
> continuous variables:
> >
> >
> >Discrete factors:
> >
> >- Diagnosis with three
> levels (A, B, C)
> >
> >- Gender with two levels
> (Males, Females)
> >
> >
> >Continuous variables:
> >
> >- Mean cortical thickness
> in the left hemisphere (CTLH)
> >
> >- Age
> >
> >- Performance in a memory
> test (TASK)
> >
> >
> >The FSGD file would be as follows:
> >
> >
> >GroupDescriptorFile 1
> >
> >Title Thickness-subj
> >
> >Class AMale
> >
> >Class AFemale
> >
> >Class BMale
> >
> >Class BFemale
> >
> >Class CMale
> >
> >Class CFemale
> >
> >variables ZCTLH ZAGE
> >
> >Input control_01 AMale -1.04398 0.24789
> >
> >Input control_02 AMale 1.22853 -0.72343
> >
> >Input control_03 AFemale .13773 -1.37097
> >
> >Input control_04 AFemale .41043 0.24789
> >
> >Input mci_01 BMale -.77128 0.24789
> >
> >Input mci_02 BMale -2.40749 0.73354
> >
> >Input mci_03 BFemale -.86218 -1.0472
> >
> >Input mci_04 BFemale .50133 0.23777
> >
> >Input ad_01 CMale -.75678 0.22789
> >
> >Input ad_02 CMale -2.88749 0.63047
> >
> >Input ad_03 CFemale -.96015 -1.4452
> >
> >Input ad_04 CFemale .80712 -0.37567 …
> >
> >
> >We are assuming different offsets and different slopes (DDOS).
> >
> >
> >The following contrast would test the null hypothesis whether
> there is a difference between B and C (T-test) regressing out
> the effect of the remaining variables.
> >
> >
> >0 0 0.5 0.5 -0.5 -0.5 0 0 0 0 0 0 0 0 0 0 0 0
> >
> >
> >Could you help me to build the contrast for testing
> interactions in the above experimental design? In the FSwiki
> examples, you add one column when you have two factors with two
> levels each, and one continuous variable. I am not sure how many
> columns should I add in our particular case and why.
> >
> >
> >Null hypothesis: is there an interaction between diagnosis (B >
> C) and gender regressing out the two remaining variables?
> Assuming that you add one column per variable that is regressed
> out, the contrast would be as follows:
> >
> >
> >0 0 0.5 0.5 -0.5 -0.5 0 0 0 0 0 0 0 0 0 0 0 0 0 0
> >
> >
> >Is this correct? Any help will be very appreciated.
> >
> >
> >Many thanks in advance.
> >
> >
> >Best regards,
> >
> >Jose
> >
> >
> >---
> >Jose L. Cantero, Ph.D.
> >Laboratory of Functional Neuroscience
> >Department of Physiology, Anatomy and Cell Biology
> >University Pablo de Olavide
> >Ctra. de Utrera, Km.1
> >41013 - Seville
> >- Spain -
> >Phone: +34 954 977433
> >Fax: +34 954 349151
> >Email: jlcanlor@upo.es
> >http://www.upo.es/neuroaging/en/
> >----------------------------------------------------------------
> --------
> >
> >_______________________________________________
> >Freesurfer mailing list
> >Freesurfer@nmr.mgh.harvard.edu
> >https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
>
> --
> Douglas N. Greve, Ph.D.
> MGH-NMR Center
> greve@nmr.mgh.harvard.edu
> Phone Number: 617-724-2358 Fax: 617-726-7422
>
> Bugs: surfer.nmr.mgh.harvard.edu/fswiki/BugReporting
> FileDrop: www.nmr.mgh.harvard.edu/facility/filedrop/index.html
>
>
>
> The information in this e-mail is intended only for the person
> to whom it is
> addressed. If you believe this e-mail was sent to you in error
> and the e-mail
> contains patient information, please contact the Partners
> Compliance HelpLine at
> http://www.partners.org/complianceline . If the e-mail was sent
> to you in error
> but does not contain patient information, please contact the
> sender and properly
> dispose of the e-mail.
>
---
Jose L. Cantero, Ph.D.
Laboratory of Functional Neuroscience
Department of Physiology, Anatomy and Cell Biology
University Pablo de Olavide
Ctra. de Utrera, Km.1
41013 - Seville
- Spain -
Phone: +34 954 977433
Fax: +34 954 349151
Email: jlcanlor@upo.es
http://www.upo.es/neuroaging/es/