Thank you very much!
Sent from my iPhone
On 16 Nov 2011, at 16:40, "Yingying Wang" <[log in to unmask]> wrote:
>
>
> >>> Xiaozhen Li <[log in to unmask]> 11/16/2011 4:37 AM >>>
> Dear FSL experts,
>
> I am running TBSS to compare 3 groups (26, 19, and 30subjects) and including 2 covariates of age and gender. I generated my matrices with Glm
> But I am not sure if it is right.
>
> My design files look like:
>
> Mat file:
> /NumWaves 5
> /NumPoints 75
> /PPheights 1.000000e+00 1.000000e+00 1.000000e+00 8.000000e+01 2.000000e+00
>
> /Matrix
> 1.000000e+00 0.000000e+00 0.000000e+00 5.500000e+01 2.000000e+00
> 1.000000e+00 0.000000e+00 0.000000e+00 7.200000e+01 2.000000e+00
> ....
> 0.000000e+00 1.000000e+00 0.000000e+00 5.500000e+01 1.000000e+00
> 0.000000e+00 1.000000e+00 0.000000e+00 7.200000e+01 2.000000e+00
> ....
> 0.000000e+00 0.000000e+00 1.000000e+00 6.500000e+01 1.000000e+00
> 0.000000e+00 0.000000e+00 1.000000e+00 6.600000e+01 1.000000e+00
>
> Con file:
> /ContrastName1 "ad_mci"
> /ContrastName2 "mci_ad"
> /ContrastName3 "ad_hc"
> /ContrastName4 "hc_ad"
> /ContrastName5 "mci_hc"
> /ContrastName6 "hc_mci"
> /NumWaves 5
> /NumContrasts 6
> /PPheights 1.327501e+00 1.327501e+00 1.129511e+00 1.129511e+00 1.290120e+00 1.290120e+00
> /RequiredEffect 1.637 1.637 1.187 1.187 1.553 1.553
>
> /Matrix
> 1.000000e+00 -1.000000e+00 0.000000e+00 0.000000e+00 0.000000e+00
> -1.000000e+00 1.000000e+00 0.000000e+00 0.000000e+00 0.000000e+00
> 1.000000e+00 0.000000e+00 -1.000000e+00 0.000000e+00 0.000000e+00
> -1.000000e+00 0.000000e+00 1.000000e+00 0.000000e+00 0.000000e+00
> 0.000000e+00 1.000000e+00 -1.000000e+00 0.000000e+00 0.000000e+00
> 0.000000e+00 -1.000000e+00 1.000000e+00 0.000000e+00 0.000000e+00
>
>
>
> I have read through the numerous posts on this topic, but I still have some questions:
> 1. Is my matrics right?
>
> -- Your covarites needs to be demeaned, which can be done in spread sheet xls for age. Take the mean off. For gender, you can use 1 for females, -1 for boys. If your design is balanced between sexes, it should be demeaned then.
>
> 2.Do I need demean the covariates? If so, how to do it? Is it okay when I do randomise with -D option?
>
> -- See above answer. Do not need to use -D because not all of your column has zero mean. Otherwise, you will get warning message.
>
> Thanks for your help!
>
> Best Regards,
> Xiaozhen
>
|