Current neuropsychological, electrophysiological, and other imaging data strongly suggest the existence of a neurobiological basis for obsessive-compulsive disorder (OCD), which was long considered to be exclusively of psychogenic origin. The positive response of some OCD patients to neurosurgery, as well as the efficacy of agents that selectively block serotonin reuptake, lends further support to a biological involvement. However, a survey of the treatment literature reveals that only 45-62% of OCD patients improve with these specific medications. In a pilot study using a quantitative electroencephalographic (QEEG) method known as neurometrics, in which QEEG data from OCD patients were compared statistically with those from an age-appropriate normative population, we previously reported the existence of two subtypes of OCD patients within a clinically homogeneous group of patients who met DSM-III-R criteria for OCD. Following pharmacological treatment, a clear relationship was found between treatment response and neurometric cluster membership. In this study, we have expanded the OCD population, adding patients from a second site, and have replicated the existence of two clusters of patients in an enlarged, statistically more robust population. Cluster 1 was characterized by excess relative power in theta, especially in the frontal and frontotemporal regions; cluster 2 was characterized by increased relative power in alpha. Further, 80.0% of the members of cluster 1 were found to be nonresponders to drug treatment, while 82.4% of the members of cluster 2 were found to be treatment responders.(ABSTRACT TRUNCATED AT 250 WORDS)