Let’s say you used FIVE to extract data from a cluster, and now you are wondering what the stats are for the mean value within the cluster.  You should have VOI in your workspace.  Within VOI you will find VOI.F which is what you get from running CreateDesign, and you will also find VOI.ConSpec which has the Contrast specifications from the I structure.

You also have the mean data in VOI.data, and now you can test the effect on the mean values by doing this:

F = estimateGLM(VOI.data,VOI.F,VOI.ConSpec);

an ANOVA table will be spit out at the command line, and various stats will be stored in F.

estimateGLM is just a small version of GLM flex designed for use when you aren’t trying to do things at the map level.  The one caveat with estimateGLM is that it does not utilize any variance/covariance spm_reml corrections.  So if your model contained a correction this won’t give you perfectly comparable results, but it should match up perfectly with results from R or SPSS.