Diagnosis and subtyping of depressive disorders by quantitative electroencephalography: I. Discriminant analysis of selected variables in untreated depressives.

Of 88 inpatients admitted for assessment of affective illness, 70 met RDC criteria for major depressive disorder. All patients had a quantitative electroencephalogram (QEEG) after ten days drug-free. Several QEEG variables were examined by multivariate stepwise discriminant analysis of data from patients and controls. Measures of interhemispheric coherence, beta activity, and slow wave excess were potent discriminators. Depressives were discriminated from normals with an accuracy of eighty-six percent. Primary depressives were discriminated from alcoholics, primary dementias and normals with an overall accuracy of 76% (25% expected by chance). Primary and secondary major depressives were correctly classified with 77% accuracy. Unipolar depressives were discriminated from bipolar depressives with 88% accuracy. Differential diagnosis and treatment implications are discussed.