The symptoms of schizophrenia are treated through the blockade of mesolimbic and mesocortical dopamine activity. However, if dopamine activity in the nigrostriatal region is also blocked, extrapyramidal syndromes (EPS) are induced, and EPS are a major cause of noncompliance. It is therefore important for antipsychotic drugs with selective pharmacological profiles to be developed. Animal models for limbic selectivity and dose-response separation between behavioural and pharmacological parameters, which are analogous to EPS and antipsychotic effects, can now be used to differentiate between novel antipsychotics with few or no EPS and classical antipsychotics. Schizophrenic patients also often experience cognitive dysfunction, and it is important that antipsychotic drugs are developed that do not exacerbate this. Again, animal models can be used to give an indication of whether a compound has any effect on cognition. Both behavioural and electrophysiological rat models indicate that novel antipsychotic drugs, such as sertindole, but not classical antipsychotic drugs, such as haloperidol, demonstrate marked limbic selectivity. Nonhuman primate models of EPS show that the novel antipsychotics have a much greater separation between dose-response curves for antipsychotic effect and the development of EPS. The Morris water maze confirmed that sertindole does not adversely affect spatial learning or memory in rats. This paper reports on the differentiation between classical and novel antipsychotics using animal models. Int Clin Psychopharmacol 13 (suppl 3):S7-S14 (C) 1998 Lippincott-Raven Publishers.