In a sense, psychosis , loosely defined as “loss of contact with reality,” is a state of being that happens to all of us each night when we fall asleep. The association between psychosis and sleep is even further strengthened when one considers that psychosis implies not simply a lack of connection with external reality, but a bizarre transformation of consciousness where the mind sees what is not there, emotional content is intensified, the flow of information or time is disjointed, and the sense of one’s self is confused. This is what usually happens when we dream. A link between psychosis and sleep has been hypothesized for a long time. In De Somniis, Aristotle stated, “the faculty by which, in waking hours, we are subject to illusion when affected by disease, is identical with that which produces illusory effects in sleep” [1]. More directly, the renowned German philosopher Arthur Schopenhauer, whose writings would significantly influence Sigmund Freud, remarked “a dream is a short-lasting psychosis and psychosis is a long-lasting dream” [2]. The bond between psychosis and sleep seems to have gained even more momentum after the turn of the century as the inquiry into dreams became a central theme in psychiatry. It is noteworthy that almost all of the early papers reporting on sleep in schizophrenia begin with a quote or a reference relating dreams and psychosis . If not from the earlier mentioned philosophers, then from Jung, “Let the dreamer walk about and act like one awakened, and we have the clinical picture of dementia praecox” [3] or noted British neurologist Hughlings Jackson, who quipped, “Find out about dreams and you will find out about insanity.” [4]. If so many of the great thinkers recognized it, how could it not be that sleep would tell us all we needed to know about psychosis and schizophrenia?

At about the same time dream research was making an impact in psychiatry, Kraepelin [5] and Bleuler [6] developed the clinical criteria for schizophrenia. Prior to their work, it is likely that most signs of the disease fell under the general rubric of “madness” [7]. Schizophrenia is a complex thought disorder that typically includes several symptoms from the following categories: a psychotic dimension including hallucinations and delusions; a disorganized dimension that includes disorganized speech, bizarre behavior, and inappropriate affect; and a negative dimension including lack of volition and blunted affect [8]. Cognitive impairments, like working memory deficits and attention difficulties, often accompany the disease. Schizophrenia is widely regarded as the most devastating of psychiatric illnesses based on the early age of onset and the socially debilitating nature of the disease for many sufferers. Although treatment has certainly improved outcomes for many sufferers, schizophrenia still places an immense burden on immediate family and society in general and represents a still largely unsolved puzzle for mental health providers. Despite a century of intense research, the origin and pathophysiology of schizophrenia remain elusive.

Psychiatrists were interested in studying the relationship between sleep and schizophrenia even before the discovery of rapid eye movement (REM) sleep and its association with dreaming . Although Kraepelin himself noted, in his seminal work, that “during the whole development of the disease, the sleep of the patients is frequently disturbed even when they are lying quiet,” [5] early sleep research focused almost exclusively on the content of dreams in schizophrenic and psychotic patients. The first published work in a scholarly journal studying hallucinations and dreams can be traced back to Trapp in 1937, which included 15 patients diagnosed with dementia praecox, a term initially introduced by Kreepelin to describe schizophrenia [9]. Trapp reported that the content of waking hallucinations often influenced dream content, and that when hallucinations subsided, dream content no longer related to previous hallucinations. Later research, however, found no obvious difference in the amount or content of dreams in schizophrenics relative to normal subjects [10, 11].

After its discovery in 1953, REM sleep quickly became a primary target for the investigation of sleep in schizophrenia [12]. Beginning with Dement’s classic work in 1955, researchers looked at how REM sleep might be altered in schizophrenia [13]. Dement was the first to use the electroencephalogram (EEG) and electrooculogram (EOG) to examine alterations in REM sleep in schizophrenic patients. The basic assumption was that if REM sleep represented an electrophysiological correlate of dreaming, and dreaming and psychoses were to share a common neural mechanism, then the study of REM in schizophrenic patients might reveal important insights into this disorder. It was generally postulated that this common mechanism might be overactive in schizophrenics, causing them to have abnormal amounts of REMs during sleep. However, early studies found few consistent differences in the amount of REM sleep between schizophrenics and normal subjects [13, 14], even during periods of the illness when subjects were acutely hallucinating [15, 16].

Influenced heavily by Freud and the psychoanalytic tradition prevalent at the time, some researchers interpreted these early negative results in an alternative light. Instead of schizophrenic patients having too much “dreaming” sleep, perhaps the suppression of dreaming sleep or some modification of dreaming control may lead to waking psychosis in schizophrenics [17]. In other words, was schizophrenia a case of dreaming sleep “bleeding” into wakefulness to cause waking psychoses? Subsequent research would initially diminish support for this hypothesis, as investigators were unable to find consistent evidence during waking of electrophysiological signs of REM sleep in schizophrenics [18]. Nor did it appear that REM sleep-specific deprivation experiments had a unique effect on schizophrenics [19], despite early reports of modified REM rebound [20]. Researchers tended to interpret findings of decreased REM latency, the time it takes to reach the first REM period, as reflecting increased dreaming pressure in schizophrenia [21, 22] but this finding was also not consistently replicated [23]. Attempts to quantify REM density, a measure of the frequency of REMs during REM sleep [23], or other REM phasic events (i.e., middle ear contractions or periorbital integrated potentials, [24] also failed to yield abnormalities specific to schizophrenia.

Although the preponderance of early research efforts examining schizophrenia focused on REM sleep for theoretical reasons, investigators also reported noticeable deficits in the amount of nonrapid eye movement (NREM) sleep in schizophrenic patients. Several studies suggested that there was a deficit specifically in the amount of deep NREM or slow-wave sleep (SWS) in schizophrenics relative to healthy controls [14, 25, 26]. However, it was clear even at the time that these results lacked specificity for schizophrenia [27]. Still, since it was well known that sleep, especially during periods of acute psychosis , was often severely affected in schizophrenics, including periods of profound insomnia or a complete reversal of the sleep–wake periods, research into the role of sleep in the pathophysiology of the disorder persisted. The late 1970s to early 1990s saw a vast improvement in recording techniques, more consistency in sleep and psychiatric classification, and improved data analysis opportunities as a result of access to computing resources. While the number of studies examining sleep architecture in schizophrenia remained high, subsequent meta-analysis of this work suggested that the most consistent sleep disturbances in schizophrenia were congruent with signs of insomnia (i.e. increased sleep latency and waking after sleep onset), revealing relatively little in terms of diagnostic specificity for schizophrenia [28–30].

Still, scientific interest in the relationship between sleep and schizophrenia has continued. The last two decades have seen a flourishing of sleep research in schizophrenia. This has been propelled by a better understanding of the neural circuitry and the functional significance underlying the two main rhythms that typify NREM sleep-slow waves and spindles. Animal and human studies have demonstrated how sleep slow waves are generated cortically by neuronal intrinsic conductances and are synchronized through cortico–cortical connections [31]. Spindles, instead, originate in a structure called the thalamic reticular nucleus (TRN). They are synchronized and sustained by the TRN connections with the cortex and with the thalamus, long recognized as the sensory gating mechanism for the cortex [31]. Therefore, since the late 1980s, researchers have less often turned to sleep macrostructural elements, such as amounts or timing of a particular sleep stage, in order to explain the pathophysiology of schizophrenia, differentiate schizophrenics from normal controls, and/or characterize various schizophrenic subtypes. Instead, current investigators are using sleep as a unique opportunity to examine the spontaneous activity of brain circuits of schizophrenics, as reflected by sleep-specific EEG rhythms. Exploring those rhythms, particularly during NREM sleep, allows researchers to probe neuronal integrity during a time when other concomitants of the disorder (like hallucinations, diminished attention, and decreased motivation) are relatively diminished.

Feinberg proposed his “synaptic pruning” theory of schizophrenia in 1982, in part based on his observation of sleep slow waves during normal human development . He noted that there were two periods of dramatic change in slow waves, a rapid increase in the first year of life and a precipitous fall during adolescence [32]. As a corollary, cortical synaptic density (the number of neuronal connections within a given area of tissue) appeared to follow a similar pattern, although based on relatively scant postmortem evidence at the time [33]. Feinberg argued that errors in synaptic pruning made during this critical period of neuronal synaptic reorganization could be responsible for schizophrenia [32]. Feinberg and others have focused on the observation that slow waves are homeostatically regulated during sleep. Specifically, larger and more numerous slow waves occur at the beginning of the night proportional to the time spent awake, whereas less frequent and smaller slow waves characterize the end of the night after sleep pressure has been diminished, a finding consistent with the idea that slow waves reflect synaptic plasticity in the cortex [34]. Evidence supporting the idea that changes in synaptic strength are mirrored by changes in slow wave activity (SWA), a measure of EEG power between 0.5 and 4 Hz typically used to capture the amplitude of slow waves, continues to accumulate [35]. While it is not known whether slow waves are actively participating in, or merely reflecting, neuronal change, the observation that the dramatic decline of slow waves during adolescence coincides with the typical age of symptom onset in schizophrenia deserves attention. Intriguingly, longitudinal studies of grey matter volumes in schizophrenia have shown accelerated grey matter losses during adolescence, especially in early onset schizophrenia [36–38], as would be predicted by the Feinberg model. Moreover, recent work has demonstrated that grey matter matures in a regionally and temporally complex pattern which is mirrored by the developmental changes in SWA [39–41]. Still, as was the case with amounts of SWS, the data regarding SWA abnormalities in post-adolescent schizophrenics are not consistent, with some, but not all, showing decreased SWA relative to controls, as would be predicted by the theory [42–45]. Moreover, decreased SWA is not specific for schizophrenia [46]. However, it is clear that inasmuch as they can be used as a proxy for cortical development, when combined with the ease of home monitoring of sleep EEG, as has been used by Feinberg et al. to track adolescent development [47] , slow waves may yet provide valuable clues to identify plasticity deficits in schizophrenia, particularly during development. It remains to be seen whether the developmental trajectory of SWA is altered in schizophrenia.