Dear Alex,
When entering the covariate in the two-sample t-test, select the option
"Interaction" with "Factor 1". You can then use a contrast [0 0 1 -1] to
test for the group by covariate interaction. This corresponds to the
last example in this document:
http://mumford.fmripower.org/mean_centering/
Best regards,
Guillaume.
On 10/12/2020 16:07, Alex spm wrote:
> Dear all,
>
> Is there a way to compare the brain correlates of one covariate for two
> groups?
>
> I know I can include the covariate in a two-sample t-test and test for
> its effect.
> But it's like considering a unique group.
>
> I can also make two separate one-sample t-test (one for each group) but
> the comparison will be qualitative.
>
> However I would like to compare statistically the activations correlated
> to the covariate between the two groups.
> Is there a way to do this?
>
> Best,
>
> Alex
>
>
--
Guillaume Flandin, PhD
Wellcome Centre for Human Neuroimaging
UCL Queen Square Institute of Neurology
London WC1N 3BG
|