In the present work an integrated signal processing expert system is presented, which provides support in the psychiatric field. Multichannel EEG recordings, focusing on Evoked Potentials (EPs) are processed by novel techniques, including Artificial Neural Networks (ANN) and noise resistant inversion algorithms, in order to provide a reliable classification system including five categories of psychopathological entities. This will enhance the diagnostic process as well as the assessment of the therapeutic interventions, having in mind the acute lack of objective numerical markers for psychiatric clinical applications.