THE COMPLEXITY of human behavior has engendered a persistent search for simple reproducible criteria of psychiatric diagnosis. Attempts to describe and utilize somatic variables have been legion. In this respect, interest in the electroencephalogram as a psychiatric investigative technique has fluctuated since its inception in 1929. Two comprehensive symposia in the past year15,23 have highlighted not only the progress in this field, but also the complexities and limitations of behavioral correlation with EEG patterns.
Clinically, most attempts to correlate the EEG with behavior have been based on surveys of populations with known behavior disturbances. The absolute percentages vary, but children with behavior disorders,24,27 adult psychopaths,11,18,20,28 and schizophrenics, particularly catatonics,2,22,25,30 have high rates of EEG abnormality. In most of these studies primary reliance on diagnosis raises serious difficulties. The variability, ambiguity, and lack of specificity of the psychiatric nosology has become more apparent when correlation with specifics such as psychopharmacological agents