Dear Jennifer and Lara,
> We conducted a PET study on patients on and off drug before and after
> surgery. In addition to the PET scans we have behavioral data. We
> would like to conduct an analysis of covariance using the scores from
> these behavioral tests and the PET scans. In SPM95, we selected
> multisubjects with replications and covariates, but how do we fill in
> the field for covariates of interest or noninterest? Thank you.
Simply enter a vector [list] of scan-specific covariates in the window
(usually by cutting and pasting from a text file). Note that you have
to have a covariate or parameter for each scan. If you only have a
parameter for each session or subject then the effect of this parameter
is already modeled implicitly by the session or block effects. In this
instance you can, however, model, and test for, interactions between
the condition-specific effects and the covariate.
I hope this helps - Karl
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|