Psychiatric animal models are a challenge by their nature. Whether a drug is blocking tumor growth in a rodent is easy enough to measure, although still hard to translate. But how does one figure out what a mouse is thinking? Actually, one doesn’t. There is “no way in heck I’m going to claim that I can model a thought disorder in rodents, so forget about that,” Bita Moghaddam told her audience at the opening keynote of the European College of Neuropsychopharmacology (ECNP) annual conference this weekend. But other aspects of mental disorders, she argued, can be usefully modeled.