Recent years have seen an increasing number of publications in which psychopathology is being studied with artificial neural networks. In this article the results of a review are presented. The aim is to give an impression of the possibilities of artificial neural networks in psychiatric research. It is found that research with artificial neural networks has been carried out on a number of different diseases, but that until now extensive research has been limited to the field of schizophrenic pathology. In this field two interesting theories have been developed with artificial neural network simulations, namely that performance deficits in a number of neuropsychological tests are being caused by disturbances in the processing of contextual information and that psychotic phenomena can be explained with the concept 'parasitic attractor'. A limited number of simulations on other psychiatric diseases are reviewed. The value as neurobiological models of the artificial neural networks reviewed are evaluated using criteria on modeling. It is concluded that networks reviewed meet these requirements only partially.