Very few epidemiological surveys have specifically studied relationships between sleep disturbances and psychiatric diseases. In this review, we preferred to use the classification proposed in 1979 by the Association of Sleep Disorders Centers. It includes four main categories: insomnias, excessive sleepiness, troubles of the wake/sleep schedule and parasomnias. Evaluating psychiatric disorders among general populations is easier owing to DSM III and DSM III-R criteria, but there are not equivalent criteria in evaluating sleep disorders. It is almost impossible to realize polysomnographic recordings in large samples, therefore sleep disorders are to be detected by questionnaires. It has been shown that there is a good correlation between self-reports and polysomnographic recordings among clinical and general samples. The prevalence of insomnia, defined as difficulties of initiating and maintaining sleep, is estimated between 9 and 31%. It is higher among women, elderly people, separated and divorced subjects, and low educational levels' groups. It has to be noticed that polysomnographic records of some subjective insomniacs are not different from those of good sleepers, sleep latency excepted. These subjective (and not objective) insomniacs have high scores in anxiety scale, depression scale, or psychologic distress. Insomnia is more frequently noted amongst subjects with psychiatric diagnoses, especially major depressive disorders and anxiety disorders. Depressive disorders are present in 21-40% of insomniacs versus 0-1% of non-insomniacs, and anxiety disorders in 13-24% of insomniacs versus 3-10% of non-insomniacs. In depressive disorders, sleep alterations are frequently noted: they are difficulties of initiating and maintaining sleep, decreasing proportion of slow-wave sleep, decreasing time of REM (rapid eye movement) sleep and REM sleep latency, and increasing density of REM sleep. Of these modifications, the last two ones seem to be specific for depression. The relationships between sleep, aging and depression are more complex than previously noted. For example, differences between depressed and non-depressed subjects depend on the age of the population. The prevalence of hypersomnia is lower than the insomnia's. It varies between 2 and 4%. It is more frequently noted among young people, and never married subjects. Two specific aetiologies must be looked for: sleep apnea syndrome and narcolepsy. These diagnoses are respectively found in 45% and 24% of hypersomniacs examined in American Sleep Centers. Hypersomnias are objectived by the Multiple Sleep Latency Test, which measures the physiologic sleep tendency.(ABSTRACT TRUNCATED AT 400 WORDS)