Dear Rik,
Yes, just use a parametric modulation with a rating value per trial.
Then form a t-contrast on the parametric modulation regressor and enter
the contrast images (one per subject) to a second level one-sample t-test.
Best regards,
Guillaume.
On 07/08/17 14:51, Rik Sijben wrote:
> Dear SPM'ers
>
> I have a data set from a study in which participants were presented stimuli of 5 conditions. After each stimulus they provided a rating on a 100 point scale. I'm trying to find out whether activation of certain regions during stimulus perception correlates with the rating they provided on the scale.
>
> How would I proceed to properly model this? Running a multiple regression at the first level requires a single value per scan rather than per trial while adding covariates at the second level doesn't allow me to differentiate between trials. Would a parametric modulation be valid for this? I'm more used to seeing it for ordinal differences within a condition rather than a 100 point scale.
>
> Thank you for any help you can offer on this.
>
> R. Sijben
>
--
Guillaume Flandin, PhD
Wellcome Trust Centre for Neuroimaging
University College London
12 Queen Square
London WC1N 3BG
|