Dear Andre,
The simplest way to do this is to create a “difference” image for each subject, by splitting the 4D output image ( e.g. GM_mod_merg_s3 ), subtracting each subjects B image from their A image, and then recombining into a 4D image with fslmerge. You can then run a model consisting of a column ones and the covariate of interest, with positive and negative contrasts on the covariate to test for differences in both directions as well as an f-test on both contrasts.
Hope this helps,
Kind Regards
Matthew
--------------------------------
Dr Matthew Webster
FMRIB Centre
John Radcliffe Hospital
University of Oxford
> On 30 Oct 2019, at 13:39, André Schmidt <[log in to unmask]> wrote:
>
> Dear FSL experts,
>
> I am interested to test whether temporal gray matter changes (from baseline (A) to 10 year follow-up (B)) are related to a clinical variable (drug dose). I have a within subjects design with 22 patients and successfully conducted all fslvbm steps. While it is clear to me how to compare gray matter between the two time points using a paried t test, I don't know how to define the contrasts for randomise to test whether A>B or B>A is correlating with drug dose.
>
> Please find attached my design file, which I used for the paired t test + the demeaned covariate of interest (EV24). I would be very grateful if you can advise how to define further contrasts to test whether:
>
> 1) A>B is positively related to EV24
> 2) A>B is negatively related to EV24
> 3) B>A is positively related to EV24
> 4) B>A is negatively related to EV24
>
> Many thanks for your help,
> André
>
>
>
> To unsubscribe from the FSL list, click the following link:
> https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=FSL&A=1
> <design.fsf>
########################################################################
To unsubscribe from the FSL list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=FSL&A=1
|