Hi,

What you need is a repeated measure 2-way ANOVA design: (with courtesy 
to Jeanette Mumford)
http://mumford.bol.ucla.edu/setting_up_designs_2009.pdf
78. slide

Bests,

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]



2010.11.22. 22:12 keltezéssel, Gelareh Ahmadi írta:
> Dear FSl Expert,
>
> I want to design a GLM to be able to show whether there is a decreased FA between 2 time points (baseline and 12 months) in the whole brain in each patients and also each controls. Then compare the result of two groups to see if decreased FA during one year in patients is significant in compare to controls.
>
> I designed different GLMs but I am not sure if they are correct. I would appreciate if you could give me a hand and let me know which one is the proper one please.
>
> Examples regard 2 patients and 2 controls. Each sunjects had two time points MRI scan:
>
> 1-      Use multi session multi subject method:
> Group            EV1         EV2
>   1                   1              0
> 1                    1              0
> 1                     0              1
>   1                    0              1
>   2                    1              0
>   2                    1              0
>   2                    0             1
>   2                    0              1
>
>
> Contrast   EV1      EV2
> Patinet1     -1          0 ( as I am interested in decreased FA not increased)
> Patinet 2     0           -1
> Control1     -1          0
> Control2      0           -1
>
> Then run randomise in TBSS and get the result like tbss_vox_p_tstat1, tbss_vox_p_tstat2, tbss_vox_p_tstat3, tbss_vox_p_tstat4 which show areas with decreased FA in each patients and controls during one year .Then use the “unpaired t-test” to compare the whole result of two groups.
>
> Hence I should merge the tbss_vox_p_tstat1, tbss_vox_p_tstat2, tbss_vox_p_tstat3, tbss_vox_p_tstat4 to get the 4D images to put it as an input for the randomise and design below GLM to see if decreased FA during these 2 timepoints in patients is significant in compare to controls.
>
> Group            EV1     EV2
>   1                   1              0
>   2                   0             1
>
> Contrast        EV1        EV2
> 1-2                 1               -1  (which shows decreased FA in patients is significantly more than comtrols).
>
> 2-      Using Fslmaths to subtract FA map at 12 months from baseline for each patients and control and threshold it for below zero to get the result of decreased FA for each data and then use these images and run simple unpaired t-test to compare the result between patients and controls.
>
> 3- By using “Pair t-test” find common areas with decreased FA during one year amongst patients and also amongst controls. So there would be two results ctl_vox_p_tstat1.nii.gz and pt_vox_p_tstat1.nii.gz.
> Merge them to get the 4D images as an input for running TBSS randomise. and Use unpaired t-test GLM design to see if decreased FA in patients is significant.
>
> I would highly appreciate if you could help me.
>
> Best Regards,
>
> Gelareh