Psychometric tests have long been used to infer something about the brain. These tests have been used for decades with clear assumptions as to what aspects of cognitive function they measure. Recent advances in functional connectivity analyses have established links between such external measures and the patterns of functional organization within the brain. Such models can be used to better understand the circuits that support scores on the external measure. Alternatively, the brain models can be inverted and used to develop better instruments. The talk will review brain-behavior modeling and then focus on how inverting such models can guide the development of new external test instruments that better reflect specific circuits of interest. A generalized brain-based framework for developing tests that measure either cognitive or clinical scores will be described. Additional benefits such as improved predictive power and reduced test burdens will be discussed.