Dear Dario,
I don't use the TFCE toolbox so I cannot be specific regarding the error
you get but in your flexible factorial design, you should only include
'main effect' of subject and 'interaction' group x time. Testing for a
group by time interaction in this design will be with an F-test [1 -1 -1 1].
Best regards,
Guillaume.
On 18/01/17 10:08, Dario Zaremba wrote:
> Dear SPM experts,
>
> we have a flexible factorial design on longitudinal CAT12 data in SPM12 and want to run TFCE toolbox on it. We get the following error:
> Error using cg_tfce_estimate>correct_xX (line 1234)
> Weird design was found that cannot be analyzed correctly.
>
> Our design is as follows:
> - subject factor
> - group factor (patients n=41 vs. controls n=14)
> - time factor (baseline vs. 2yr-follow-up)
> - covariate TIV
> - covariate age
>
> We included following main effects & interactions
> - main effect subject
> - main effect group
> - main effect time
> - interaction group x time
>
> The contrast of interest is a F-contrast on the interaction group x time:
> 0 0 0 0 1 -1 -1 1
> 0 0 0 0 -1 1 1 -1
>
> Is it possible to estimate TFCE on this SPM.mat and specific contrast?
>
> Thank you very much in advance,
> Dario
>
--
Guillaume Flandin, PhD
Wellcome Trust Centre for Neuroimaging
University College London
12 Queen Square
London WC1N 3BG
|