BACKGROUND Numerous studies have been published on postnatal depression (PND) over the last 10 years. A controversy has arisen regarding the specificity of the diagnostic concept. It is based on 2 points: the hormonal environment and the course and recurrence of PND. EEG Sleep studies are also numerous, but this paper presents the first study on EEG sleep profile during PND. METHODS 24 women suffering from major depression according to RDC were placed in 3 groups: 1) Group A (n = 8): PND; 2) Group B (n = 8): depression with a past history of PND; 3) Group C (n = 8): depression without a past history of PND. Women were age-matched and according to the Hamilton 24 item severity score. Group A patients were delivered within less that 6 months, groups B and C within a minimum of 3 years. None were pregnant or alcoholic, and none were physically ill. RESULTS There was no difference between groups B and C. Group A was characterised by a significantly longer stage IV sleep. There was also a strong tendency to a shorter stage I sleep and a better quality of sleep (total sleep time and number of awakenings) during PND. CONCLUSIONS Our study shows that, even if similar to major depression, specific polysomnographic alteration can be found during post-partum depression. This finding is relevant to the hypothesis of PND's diagnostic specificity. The perfect similarity between groups B and C strengthens this evidence. Nevertheless, the significance of SWS alterations are difficult to explain and additional studies are required. CONCLUSIONS The EEG Sleep profile during PND differs from major depression of the same severity. This appears to favour the specificity of the diagnostic concept.