In　neuroimaging　studies　of　human　cognitive　abilities,　in　vivo　MRI-derived　measurements　of　human　cerebral　cortex　thickness　are　of　particular　interest,　but　those　data　typically　use　univariate　analyses　that　do　not　explicitly　test　the　interrelationship　among　brain　regions.　Among　the　existing　spatial　covariance　models,　Scaled　Subprofile　Model　(SSM)　has　been　highly　successful,　particularly　in　capturing　sources　of　between　and　within　group　variance,　identifying　group　differences　in　regional　network.　Current　article　describes　feasibility　of　applying　spatial　covariance　models　on　cortical　thickness　map,　followed　by　an　application　to　discriminate　normal　control　group　and　Alzheimer　disease　group.　The　results　showed　that　SSM　can　not　only　capture　patterns　of　difference　between　groups　but　also　summarize　the　degree　to　which　individual　subjects　express　anatomical　topography.　©　2012　IEEE.