Dear Stephan,
> In an fMRI event-related analysis, if one has several covariates for
> each event and would like to include these as user-specified
> regressors, there seems to be no easy way to get SPM99 to do this
> unless the regressors are scan-specific, such as movement parameters.
>
> One can use parametric modulation, but this allows only one covariate
> to be entered. If, for example, one also had a second, nuisance
> covariate that was specific to each trial, it is unclear how this would
> be entered using the current GUI.
In fact you can use any number of parametric (trial-specific)
covariates in the paramteric modulation option. The trick is to
specify the trial type of interest several times. For example if you
have three covariates to be applied to trial type 2
parametric modulation 'other'
name of parameter '????'
expansion 'linear'
which trial[s] 1 to 4 2 2 2
parameters for trial2 [2 3 2 6 3 ....
parameters for trial2 [9 2 3 2 6 ....
parameters for trial2 [4 5 2 3 2 ....
I hope this helps - Karl
|