Previously, it was shown that subchronic application of the NMDA receptor antagonist ketamine (Ket) induces schizophrenia-related alterations, e.g. decreased non-aggressive behaviour in the social interaction test, which might be a useful animal model in the study of negative symptoms of this disease. In order to further evaluate the predictive validity of this model, the anxioloytic diazepam, the classic neuroleptic haloperidol and the atypical neuroleptics clozapine and risperidone were tested after acute and subchronic treatment. The experiments demonstrated that haloperidol did not normalise the behavioural effects of Ket. After acute administration, diazepam was ineffective in control animals but increased non-aggressive behaviour in Ket-treated animals. Similar effects were found in animals injected with either clozapine or risperidone. Twenty-four hours after discontinuation of subchronic treatment with both substances, there was an increase in the percentage of non-aggressive behaviour in the ketamine group and a decrease in the control animals. This decrease was explained in terms of withdrawal. Different effects in the control groups and the Ket groups were found when the test was performed 1 h after subchronic clozapine or risperidone treatment. The data suggest that atypical antipsychotic drugs (APD) effectively counteract Ket-induced alterations in social behaviour. Regarding false-positive effects by anxiolytic drugs without antipsychotic efficacy, this model may have some predictive validity for identifying anxiolytic effects of novel antipsychotic compounds.