Dear Thomas,
> > Use 'mean and exponential decay'. This is an equivalent formulation in
> > terms of within-epoch adaptation' suggesting the earlier responses
> > supervene over later responses. Furthermore these two basis functions
> > allow more easily interpretable contrasts.
> >
> > With best wishes - Karl
>
> Dear Karl,
>
> If I got this right, this model allows to remove the effects of
> within-epoch adaption. In my case I am interested in the effects of
> adaption, so in the SPM96 model I would have looked at the late
> responses. How can this be done using the 'mean and exp. decay' model
Simply define a contrast testing for the decay term with a '1' for
the appropriate [second] basis function associated with the trial[s]
of interest and '0' elsewhere.
With best wishes - Karl
|