Tianeptine enhances while paroxetine inhibits serotonin reuptake into neurons; however, both show an antidepressive action. A subgroup of 38 depressed patients from a drug trial comparing the efficacy of tianeptine with that of paroxetine was studied with regard to their effects on sleep regulation, especially in relation to treatment response. We recorded sleep EEGs at day 7 and day 42 after the start of treatment with either compound, which allows measurement of changes due to the antidepressive medication in relation to the duration of treatment. Spectral analysis of the non-REM sleep EEG revealed a strong decline in the higher sigma frequency range (14-16 Hz) in male treatment responders independent of medication, whereas nonresponders did not show marked changes in this frequency range independent of gender. The patients receiving paroxetine showed less REM sleep and more intermittent wakefulness compared to the patients receiving tianeptine. REM density after 1 week of treatment was a predictor of treatment response in the whole sample. Psychopathological features with regard to the score in single items of the HAMD revealed predictive markers for response, some of which were opposite in the gender groups, especially those related to somatic anxiety. Changes in REM density were inversely correlated to the changes in HAMD in the paroxetine, but not the tianeptine, group. Our data suggest the importance of taking gender into account in the study of the biological effects of drugs. The study further points to the importance of the higher sigma frequency range in the sleep EEG of non-REM sleep and REM density as a marker of treatment response.