The aim of the study was to evaluate REM sleep parameters, especially the temporal characteristics of rapid eye movement activity, in depressed patients, and to compare three different methods for scoring of REM density. The sleep of 15 nonmedicated depressed patients and 13 healthy controls was recorded during two consecutive nights. Sleep recordings were scored by raters blinded to the diagnosis. In comparison to healthy controls depressed patients showed an increased REM density and increased REM activity. Both groups differed also regarding the pattern of REM density changes between REM sleep periods (REMPs). Whereas in healthy controls REM density in the first REMP was significantly lower than in the successive REMPs, no such difference was found in depressed patients. On visual inspection we failed to find any significant differences in the time course of REM activity within the first REMP in depressed patients. All applied methods for scoring of REM density distinguished depressed patients from healthy controls with comparable accuracy.