[Sleep electroencephalography in depression and mental disorders with depressive comorbidity].

Traditional scoring of sleep EEG in depressed patients shows abnormalities in sleep maintenance, sleep architecture, REM sleep, the distribution of slow wave and REM sleep during the night. Computerized analysis that comprises the period-amplitude analysis procedure and spectral analysis discloses changes in delta activity and distribution of delta activity. However, these methods of analysing EEG sleep are not able to distinguish the various concepts of depression: endogenous and non-endogenous depression, unipolar and bipolar depression, psychotic and non-psychotic depression. Polysomnographical data in patients with recurrent depression show alteration during remission suggesting trait-like abnormalities of sleep in depression illness. Shortened REM latency is not specific in depression. This sleep parameter is defined in many different ways explaining the heterogeneousness of study results and the failure of constituting a biological marker. Many sleep parameters are affected by several factors such as age, gender and severity. Several physiopathological hypotheses have been proposed to explain EEG sleep alterations. They refer either to circadian rhythms such as the two process model of Borbély, the phase advance hypothesis and the circadian amplitude hypothesis, or to neurotransmitter abnormalities such as the cholinergic hypothesis. None of them takes sufficient account of all the sleep abnormalities. Sleep abnormalities have also been described in other psychiatric disorders such as mania, panic and obsessional-compulsive disorders, generalized anxiety, phobias, post-traumatic stress disorder, eating disorders, borderline personality, schizophrenia and dementia. None of them have a particular sleep EEG profile which allows to differentiate between them. A concomitant episode of major depression cannot be uncovered by sleep recordings.