Pharmacology and anxiety: inadequacies of current experimental approaches and working models.

Current models concerning the mechanisms of punishment suppression and anxiolytic drug effects fail to account for several treatment-test interactions in pharmacological studies. This applies in the first place to some important "double dissociation" phenomena. For example, in rats benzodiazepines are effective in conflict tests (Geller- and Vogel-type) but not in go-no go avoidance discriminations, while the converse is true in the case of antimuscarinics. Such a situation makes it necessary to postulate a plurality of mechanisms which can serve punishment suppression in various conditions, and can operate at least partly "in parallel" rather than "in series". In addition, different varieties of a particular test can show quite different sensitivities to the same type of agent and/or different profiles in studies using various types of anxiolytics and antagonists. This does not preclude the use of one or the other test as a convenient assay. It appears, however, that we have only limited knowledge on the mechanisms involved in the production of behavioral effects which are assumed to be typical of the anxiolytic profile.