The authors describe a pilot investigation using ambulatory polysomnography (PSG) to assess rapid eye movement (REM) latency in 11 depressed inpatients before and after a course of electroconvulsive therapy (ECT). Prior to beginning ECT, all subjects had REM latencies of 56 minutes or less (mean 22 min). A course of ECT was associated with clinical improvement in every patient and a statistically significant (p < 0.0002) increase of 67% in mean REM latency (37 min). However, individual patients showed marked variability in REM latency both during and after the course of ECT, and 7 of the 11 responders continued to exhibit shortened REM latencies of 34 min or less. Technical considerations related to PSG studies during ECT are reviewed and the clinical and theoretical implications of our findings are discussed.