I am using Freesurfer to analyze a set of brains of subjects with temporal lobe epilepsy. The brains are being run at two different locations. In addition, the brains at one location have been run on two different computers, with two different freesurfer builds. The problem that I am having is that that each time the data are run, they are markedly different. The differences are substantial, more than should be accounted for by using different versions. In some cases, the differences are from 10-30% off. Has anyone else had this problem?