Event-Related Potentials (ERPs) provide non-invasive measurements of the electrical activity on the scalp that are linked to the presentation of stimuli and events. Brain mapping techniques are able to provide evidence for the solution of debatable issues in cognitive science. In this paper, a two-step signal classification approach is proposed, extending the use of the Low Resolution Brain Electrical Tomography (LORETA) inversion technique. The first step concerns the feature extraction module, which is based on the combination of the Multivariate Autoregressive model with the Simulated Annealing technique. The classification module, as the second step of the methodology, is implemented by means of an Artificial Neural Network (ANN) trained with the backpropagation algorithm under “leave-one-out crossvalidation”. The ANN is a multi-layer perceptron, the architecture of which, is selected after a detailed search. The proposed methodology has been applied for the classification of first episode schizophrenic patients and normal controls using as input signals the intracranial current sources obtained by the inversion of ERPs using the LORETA technique. Results by implementing the proposed methodology provide classification rates of up to 93.1%. Finally, the proposed methodology may be used for the design of more robust classifiers based on the head-surface measured potentials as well as on the intracranial source locations, which directly relate to cognitive mechanisms.