Identifying an abnormal electroencephalographic sleep profile to characterize major depressive disorder.

There is little agreement as to the best definition of a categorically abnormal electroencephalographic (EEG) sleep profile to characterize major depressive disorder. Therefore, a series of classification, replication, and validation analyses were conducted to identify such a profile. The EEG sleep studies of healthy controls (n = 44), depressed inpatients (n = 44), and depressed outpatients (n = 181) were utilized, including subgroups of patients studied both before and after nonpharmacologic treatment with either cognitive behavior therapy (CBT) or interpersonal psychotherapy (IPT). A discriminant index score (based on reduced REM latency, increased REM density, and decreased sleep efficiency) was found to: 1) reliably discriminate between depressed inpatients, depressed outpatients, and controls; 2) show good test-retest reliability; and 3) identify a subset of depressed outpatients who were older, manifested a broader array of EEG sleep disturbances, and were less responsive to CBT or IPT. Posttreatment studies of patients indicated that normal sleep profiles were relatively stable, whereas abnormal profiles tended to normalize. These findings provide an empirically validated method that may improve the applicability, efficiency, and prognostic utility of EEG sleep studies of depressed patients.