Intermediate or brainless phenotypes for psychiatric research?

For highly heritable brain disorders, such as schizophrenia and autism, investigating genetic effects on the level of neural systems seems an obvious approach. Nevertheless, the usefulness of the intermediate phenotypes ('endo' phenotypes) continues to be debated energetically. We argue that, while not all intermediate phenotypes are created equal, the hypothesis-driven investigation of the translational cascades linking genetic variation to disturbed behavior is a viable and important strategy that should not be supplanted by an exclusive focus on brainless, clinical/categorical phenotypes investigated in very large numbers of participants.