Schizophrenia is a major psychiatric disorder that has a massive, long-lasting negative impact on the patients as well as society. While positive symptoms (i.e., delusions and hallucinations), negative symptoms (i.e., anhedonia, social withdrawal), and cognitive impairments are traditionally considered the most prominent features of this disorder, the role of sleep and sleep disturbances has gained increasing prominence in clinical practice. Indeed, the vast majority of patients with schizophrenia report sleep abnormalities, which tend to precede illness onset and can predict an acute exacerbation of psychotic symptoms. Furthermore, schizophrenia patients often have a comorbid sleep disorder, including insomnia, obstructive sleep apnea, restless leg syndrome, or periodic limb movement disorder. Despite accumulating data, the links between sleep disorders and schizophrenia have not been thoroughly examined, in part because they are difficult to disentangle, as numerous factors contribute to their comorbidity, including medication status. Additionally, sleep disorders are often not the primary focus of clinicians treating this population, despite studies suggesting that comorbid sleep disorders carry their own unique risks, including worsening of psychotic symptoms and poorer quality of life. There is also limited information about effective management strategies for schizophrenia patients affected by significant sleep disturbances and/or sleep disorders. To begin addressing these issues, the present review will systematically examine the literature on sleep disorders and schizophrenia, focusing on studies related to 1) links between distinct sleep disorders and schizophrenia; 2) risks unique to patients with a comorbid sleep disorder; and 3) and management challenges and strategies.

Importance: Abnormal sleep is frequent in psychosis; however, sleep abnormalities in different stages (ie, clinical high risk for psychosis [CHR-P], early psychosis [EP], and chronic psychosis [CP]) have not been characterized.Objective: To identify sleep abnormalities across psychosis stages.Data Sources: Web of Science and PubMed were searched between inception and June 15, 2022. Studies written in English were included.Study Selection: Sleep disturbance prevalence studies and case-control studies reporting sleep quality, sleep architecture, or sleep electroencephalography oscillations in CHR-P, EP, or CP.Data Extraction and Synthesis: This systematic review and meta-analysis followed Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) reporting guideline. Stage-specific and pooled random-effects meta-analyses were conducted, along with the assessment of heterogeneity, study quality, and meta-regressions (clinical stage, sex, age, medication status, and psychotic symptoms).Main Outcomes and Measures: Sleep disturbance prevalence, self-reported sleep quality, sleep architecture (total sleep time, sleep latency, sleep efficiency, nonrapid eye movement, rapid eye movement stages, and number of arousals), and sleep electroencephalography oscillations (spindle density, amplitude, and duration, and slow wave density).Results: Fifty-nine studies with up to 6710 patients (n = 5135 for prevalence) and 977 controls were included. Sleep disturbance prevalence in pooled cases was 50% (95% CI, 40%-61%) and it was similar in each psychosis stage. Sleep quality was worse in pooled cases vs controls (standardized mean difference [SMD], 1.00 [95% CI, 0.70-1.30]). Sleep architecture alterations included higher sleep onset latency (SMD [95% CI]: pooled cases, 0.96 [0.62-1.30]; EP, 0.72 [0.52-0.92]; CP, 1.36 [0.66-2.05]), higher wake after sleep onset (SMD [95% CI]: pooled cases, 0.5 [0.29-0.71]; EP, 0.62 [0.34-0.89]; CP, 0.51 [0.09-0.93]), higher number of arousals (SMD [95% CI]: pooled cases, 0.45 [0.07-0.83]; CP, 0.81 [0.30-1.32]), higher stage 1 sleep (SMD [95% CI]: pooled cases, 0.23 [0.06-0.40]; EP, 0.34 [0.15-0.53]), lower sleep efficiency (SMD [95% CI]: pooled cases, −0.75 [−0.98 to −0.52]; EP, −0.90 [−1.20 to −0.60]; CP, −0.73 [−1.14 to −0.33]), and lower rapid eye movement density (SMD [95% CI]: pooled cases, 0.37 [0.14-0.60]; CP, 0.4 [0.19-0.77]). Spindle parameter deficits included density (SMD [95% CI]: pooled cases, −1.06 [−1.50 to −0.63]; EP, −0.80 [−1.22 to −0.39]; CP, −1.39 [−2.05 to −0.74]; amplitude: pooled cases, −1.08 [−1.33 to −0.82]; EP, −0.86 [−1.24 to −0.47]; CP, −1.25 [−1.58 to −0.91]; and duration: pooled cases: −1.2 [−1.69 to −0.73]; EP, −0.71 [−1.08 to −0.34]; CP, −1.74 [−2.10 to −1.38]). Individuals with CP had more frequent arousals vs CHR-P (z = 2.24, P = .02) and reduced spindle duration vs EP (z = −3.91, P < .001).Conclusions and Relevance: In this systematic review and meta-analysis, sleep disturbances were found to be prevalent throughout the course of psychosis, and different psychosis stages showed both shared and distinct abnormalities in sleep quality, architecture, and spindles. These findings suggest that sleep should become a core clinical target and research domain from at-risk to early and chronic stages of psychosis.

Resting-state electroencephalography (EEG) microstates reflect sub-second, quasi-stable states of brain activity. Several studies have reported alterations of microstate features in patients with schizophrenia (SZ). Based on these findings, it has been suggested that microstates may represent neurophysiological biomarkers for the classification of SZ. To explore this possibility, machine learning approaches can be employed. Bayesian optimization is a machine learning approach that selects the best-fitted machine learning model with tuned hyperparameters from existing models to improve the classification. In this proof-of-concept preliminary study based on secondary analysis, 20 microstate features were extracted from 14 SZ patients and 14 healthy controls’ EEG signals. These parameters were then ranked as predictors based on their importance, and an optimized machine learning approach was applied to evaluate the performance of the classification. SZ patients had altered microstate features compared to healthy controls. Furthermore, Bayesian optimization outperformed conventional multivariate analyses and showed the highest accuracy (90.93%), AUC (0.90), sensitivity (91.37%), and specificity (90.48%), with reliable results using just six microstate predictors. Altogether, in this proof-of-concept study, we showed that machine learning with Bayesian optimization can be utilized to characterize EEG microstate alterations and contribute to the classification of SZ patients.