Many patients suffering from the majority of anxiety disorders complain about their sleep by reporting difficulties in initiating and maintaining it. Polysomnographic studies have shown that, in comparison to normal subjects, the sleep of patients with panic disorder is characterized by longer sleep latency, increased time awake and reduced sleep efficiency. Sleep architecture is normal and there are no significant changes in REM sleep measures. Nocturnal panic attacks are non-REM-related events and occur without an obvious trigger in 18-45% of panic disorder patients. Regarding generalized anxiety disorder, the patients complain of 'trouble sleeping' in 60-70%, while polysomnography has shown increased sleep latency and decreased sleep continuity measures. The findings in REM sleep and sleep architecture generally do not show any aberration to exist. In patients with obsessive-compulsive disorder (OCD) and post-traumatic stress disorder (PTSD), results from the sleep laboratory do not seem to support the subjective complaints of poor sleep. The early reports of shortened REM latency in OCD could not be replicated by recent studies. A dysregulation of the REM sleep control system has been reported for patients with PTSD. Finally, no significant differences were found in all sleep parameters between social phobia patients and controls.