Here is what I would like to do:

I have run a GLM on the entire cortex to determine whether there are significant differences in LGI between patients and controls.
I then would like run a different GLM just with the clusters outputted from the previous GLM. 

It would seem that the best thing to do is to set a mask representing the clusters from the first GLM and then run the second GLM on the mask.

Is this a reasonable thing to do? What steps/commands would I perform in order to do something like this?

Respectfully,