While reports of EEG correlates of psychiatric disorders date back five decades, clinical sensitivity of the EEG to psychiatric disorders has been greatly enhanced with the advent of quantitative methods of analysis (QEEG). Using a QEEG methodology known as neurometrics we have identified distinctive electrophysiological profiles associated with different psychiatric disorders. With this method quantitative features are extracted from 2 minutes of artifact- free eyes closed resting EEG data, log transformed to obtain Gaussianity, age-regressed, and Z-transformed relative to population norms. Using small subsets of neurometric features, multiple stepwise discriminant analyses were used to construct mathematical classifier functions, the values of which are different for members of different a priori defined diagnostic groups. Using this approach, we have demonstrated high discriminant accuracy in independent replications separating many populations of psychiatric patients from normal as well as from each other, including major affective disorder, schizophrenia, dementia, alcoholism, and learning disabilities, as well as high accuracy of discrimination between known subtypes of depression (unipolar vs bipolar). The use of classification accuracy curves (CACs) which allow one to assess the sensitivity and specificity achieved by the discriminant functions is discussed. In addition, using cluster analysis, neurometric subtypes can be identified in several clinically homogenous populations. Preliminary results suggest that baseline membership in some neurometric subtypes may be highly correlated with response to treatment.