Animal models are key to our understanding of many human diseases and psychiatric disorders are no exception. Animal models have provided insight into the neurotransmitter systems and brain circuitry underlying psychiatric illnesses that have enabled the screening of potential psychiatric medications for efficacy and guided the search for new pharmacotherapies. However, modeling complex psychiatric disorders in animals presents distinct challenges. Modeling psychiatric disorders in animals is difficult due to the complexity of human thoughts, emotions, and behavior; the heterogeneity of many psychiatric disorders; and the requirement of self-reporting of internal state for diagnosis. According to the DSM-IV, most psychiatric diagnoses are made when a patient displays a certain number of diagnostic criteria. However, patients with the same diagnosis can differ significantly in specific symptoms, perhaps implying differences in the underlying etiology of their disorders and explaining the heterogeneity of response to pharmacotherapy. Many symptoms are identified by the patient's report of internal state (e.g., obsessive obtrusive thoughts or ruminations). This leads to questions of how we can best model disorders in animals that are, by definition, both heterogeneous and dependent on report of internal state. Rather than attempting to model a psychiatric disorder in its entirety, most neuroscientists focus on individual aspects or dimensions of a disorder and use physical manifestations and measurable behaviors when modeling a particular aspect.
Inferences from experiments with animal models have the potential to impact clinical practice; therefore, stringently validating such models is of utmost importance. The strengths and weaknesses of a model can be conceptualized with different types of validity. Three important types of validity are suggested as predictive, face, and construct validity. In the first one, the animal should display reduced anxiety when treated with anxiolytics (predictive validity). In the second type, the behavioral response of an animal model to a threatening stimulus should be comparable to the response known for humans (face validity). Finally, in the third type of validity, the mechanisms underlying anxiety as well as the psychological causes should be identical (construct validity). In the meantime there is no consensus about which type of validity is superior. Some authors argue that predictive validity is most essential but the others suggest that constructive validity is the most crucial. On the other hand, these three requirements are difficult to achieve in any animal model for each anxiety disorder. Moreover, in the related literature, the term"animal model of anxiety" is used for animals that are altered in their anxiety-related behavior, as well as for test assays conceptualized to assess anxiety-related behavior in animals.