Psychopharmacological Prediction by Artificial Neural Networks

Abstract

Background: Artificial neural network (ANN) can offer specific advantages with respect to classical statistical techniques. This article is designed to acquaint psychiatrists with concepts and paradigms related to ANN. Method: Literature dealing with pharmacological prediction of depression and schizophrenia was reviewed. Results: In most studies, ANN was found to have similar or better predictive performance than logistic regression. Models combining clinical and genetic data had a higher predictive accuracy than those using clinical data alone. Family of ANN, when appropriately selected and used, permits the maximization of what can be derived from available data and from complex, dynamic, and multidimensional psychopharmacological data. Conclusion: Future prospective studies can use the ANN models in real-life, and diverse clinical settings are critical in determining whether this type of system will have important clinical impact on patient outcome.