Interindividual Sleep Variability Across the Psychosis Spectrum: A Systematic Review and Meta-Analysis

Rosario Aronica; John Torous; Amedeo Minichino; Melinda Mills; Philip McGuire; Dominic Oliver

doi:10.1001/jamanetworkopen.2026.24358

. 2026 Jul 22;9(7):e2624358. doi: 10.1001/jamanetworkopen.2026.24358

Interindividual Sleep Variability Across the Psychosis Spectrum

A Systematic Review and Meta-Analysis

Rosario Aronica ^1,^2,^3,^✉, John Torous ⁴, Amedeo Minichino ^1,^2,⁵, Melinda Mills ^6,⁷, Philip McGuire ^1,^2,⁵, Dominic Oliver ^1,^2,⁵

¹Warneford Hospital, Oxford Health NHS Foundation Trust, Oxford, United Kingdom

²Department of Psychiatry, University of Oxford, Oxford, United Kingdom

³Department of Population Health, University of Oxford, Oxford, United Kingdom

⁴Beth Israel Deaconess Medical Center, Harvard Medical School, Boston, Massachusetts

⁵NIHR Oxford Biomedical Research Centre, Oxford, United Kingdom

⁶Leverhulme Centre for Demographic Science, University of Oxford, Oxford, United Kingdom

⁷Nuffield College, University of Oxford, Oxford, United Kingdom

Accepted for Publication: May 26, 2026.

Published: July 22, 2026. doi:10.1001/jamanetworkopen.2026.24358

^✉

Corresponding Author: Rosario Aronica, MD, Department of Population Health, University of Oxford, Oxford OX3 7LF, UK (rosario.aronica@psych.ox.ax.uk).

Author Contributions: Drs Aronica and Oliver had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Aronica, Torous, Minichino, McGuire, Oliver.

Acquisition, analysis, or interpretation of data: Aronica, Mills, McGuire, Oliver.

Drafting of the manuscript: Aronica, Torous.

Critical review of the manuscript for important intellectual content: Aronica, Minichino, Mills, McGuire, Oliver.

Statistical analysis: Aronica, Oliver.

Administrative, technical, or material support: Aronica, McGuire.

Supervision: Aronica, Minichino, Mills, McGuire, Oliver.

Conflict of Interest Disclosures: Dr Torous reported serving as a web editor for JAMA Psychiatry and as an advisor to Precision Mental Wellness outside the submitted work. In addition, Dr Torous reported receiving research support from Otsuka outside the submitted work. No other disclosures were reported.

Funding/Support: This study was funded by the Wellcome Trust (WT/216005/Z/19/Z). Dr Aronica is supported by the Oxford Health NHS Foundation Trust, the Clarendon Scholarship, the Medical Science Division at the University of Oxford, the Lincoln College at University of Oxford, the Department of Population Health at the University of Oxford, and the Consortium on Analytics for Data-Driven Decision-Making at McGill University. Dr Minichino is supported by Wellcome Trust Early Career Award WT/304693_Z_23. Dr Mills is supported by European Research Council Advanced Grant 835079 and by Leverhulme Trust Large Centre and Leverhulme Centre for Demographic Science Grant RC-2018-003. Drs Minichino, McGuire, and Oliver are supported by the National Institute for Health and Care Research (NIHR) Biomedical Research Centre at Oxford Health NHS Foundation Trust.

Role of the Funder/Sponsor: The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Disclaimer: The views expressed are those of the authors and not necessarily those of the NHS, the NIHR, or the UK Department of Health and Social Care.

Data Sharing Statement: See Supplement 2.

^✉

Corresponding author.

PMCID: PMC13392811 PMID: 42485043

Key Points

Question

Is interindividual sleep variability greater in individuals across the psychosis spectrum than in healthy control participants?

Findings

In this systematic review and meta-analysis of 1358 participants from 18 case-control studies, interindividual variability in time in bed, wake after sleep onset, and sleep efficiency was greater in those at clinical high risk for psychosis and those with schizophrenia spectrum disorders compared with healthy control participants. Total sleep time variability was significantly greater only in participants with schizophrenia spectrum disorders.

Meaning

The distinct digital phenotypes of sleep variability across the psychosis spectrum described in this study support further investigations of sleep as a candidate biomarker for stratified care in psychosis.

This systematic review and meta-analysis uses actigraphy sleep measures to quantify interindividual sleep variability among individuals across the psychosis spectrum compared with healthy control participants.

Abstract

Importance

Sleep disturbances are common across the psychosis spectrum and may precede overt core clinical symptoms. Most studies focus on mean differences and overlook interindividual variability, which can be informative for stratified care.

Objective

To quantify interindividual variability in actigraphy sleep measures across psychosis stages among individuals at clinical high risk for psychosis (CHR-P) and those with schizophrenia spectrum disorders (SSDs) compared with healthy control participants.

Data Sources

The PubMed, Embase, MEDLINE, Cochrane Central Register of Controlled Trials, ClinicalTrials.gov, EU Clinical Trials Register, World Health Organization International Clinical Trials Registry Platform, and OpenGrey databases were searched for studies published from inception to April 29, 2024. An updated search was performed to identify studies published between April 30, 2024, and April 25, 2026.

Study Selection

Case-control studies using wrist actigraphy in participants at CHR-P or with SSDs compared with healthy control participants were selected.

Data Extraction and Synthesis

This systematic review and meta-analysis was preregistered with the Center for Open Science and followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses. Data were independently extracted under PRISMA standards. Two reviewers independently applied the inclusion criteria.

Main Outcomes and Measures

The primary outcome was the natural logarithm of the variability ratio (lnVR) for total sleep time (TST). Secondary outcomes were the lnVR and the natural logarithm of the coefficient of variation ratio (lnCVR) for all actigraphy sleep parameters, including time in bed (TIB), sleep latency, wake after sleep onset (WASO), sleep efficiency, and number of awakenings. Variability metrics (lnVR and lnCVR) were pooled using random-effects models. Heterogeneity was assessed with Q, I², and τ² values and 95% CIs. Univariable mixed-effects meta-regression analysis tested antipsychotic use, with age and sex as moderators of the primary outcome. Each key result was tested against a panel of sensitivity analyses.

Results

Of the 19 studies that met criteria, 18 entered the quantitative synthesis (1 study reported outcomes as medians and was excluded). Eighteen studies (n = 1358), including 202 individuals at CHR-P (mean [SD] age, 20.7 [2.8] years; 104 female [51.4%]), 574 individuals with SSDs (mean [SD] age, 37.7 [10.2] years; 363 female [63.2%]), and 582 healthy control participants (mean [SD] age, 31.5 [8.6] years; 311 male [53.4%]), were meta-analyzed. Participants at CHR-P had greater variability in TIB (lnVR, 0.29 [95% CI, 0.04-0.54] minutes; I² = 0%; P = .02), WASO (0.45 [95% CI, 0.07-0.83] minutes; I² = 83%; P = .02), and sleep efficiency (0.35% [95% CI, 0.06%-0.64%]; I² = 72%; P = .02). Participants with SSDs had greater variability in TST (0.46 [95% CI, 0.28-0.64] minutes; I² = 70%; P < .001), TIB (0.38 [95% CI, 0.19-0.56] minutes; I² = 0%; P < .001), WASO (0.52 [95% CI, 0.15-0.89] minutes; I² = 88%; P = .006), and sleep efficiency (0.27% [95% CI, 0.04%-0.50%]; I² = 71%; P = .02). Moderator effects were nonsignificant.

Conclusions and Relevance

In this systematic review and meta-analysis of 18 case-control studies, distinct digital phenotypes of sleep variability were identified among individuals across the psychosis spectrum compared with healthy control participants. These findings support further investigations of sleep as a candidate biomarker to inform stratified care in psychosis.

Introduction

Sleep disturbances are common across the psychosis spectrum, affecting up to 70% of individuals at clinical high risk for psychosis (CHR-P)^1,2 and 80% of those with schizophrenia spectrum disorders (SSDs).³ In most of these studies, sleep was assessed with self-report questionnaires⁴ or polysomnography,⁵ which limits ecological validity. The advent of digital technologies, such as wrist actigraphy,⁶ enables continuous monitoring of sleep and offers a more scalable and powerful approach to characterize sleep disturbances in individuals with psychiatric conditions.⁷

To date, most wrist actigraphy studies in psychiatric populations have reported average alterations across patient samples.^8,9,10 This approach assumes group-level homogeneity and may obscure clinically meaningful variability. While there are evident sleep disturbances in individuals at CHR-P and those with SSDs compared with control participants,¹⁰ these populations are highly heterogeneous in terms of clinical outcomes and underlying pathophysiology.¹¹ Averaging across participants can therefore mask important interindividual differences within the same diagnostic group. For example, in a cohort comprising equal numbers of individuals with very short total sleep time (TST; approximately 3 hours) and very long TST (approximately 11 hours), the mean (approximately 7 hours) would resemble that of healthy control participants, concealing distributions with potentially distinct prognostic or therapeutic profiles (Figure 1). Failing to account for interindividual variability may therefore lead to null findings and overlook potential subpopulations with distinct sleep profiles.^12,13

Three-panel line graphs of total sleep time distributions for patients and controls. Three panels labeled A, B, and C. All panels use the same axes: horizontal axis labeled Total sleep time, h, ranging from 0 to 15; vertical axis labeled Participants, percent, ranging from 0 to 50 with gridlines at 10 percent intervals. Panel A title: Patients at C H R dash P or with S S D s. One orange smooth density curve with two narrow peaks, one near 3 hours reaching about 40 percent and a second near 11 hours reaching about 40 percent, with values near zero between about 5 and 9 hours. A vertical blue dotted line at approximately 7 hours. Near the bottom, an orange horizontal line with end caps spans broadly from about 2 hours to about 12 hours. Panel B title: Healthy control participants. One dark teal smooth density curve with a single narrow peak centered near 7 hours reaching about 40 percent, tapering toward near zero by about 5 hours and 9 hours. A vertical blue dotted line at approximately 7 hours. Near the bottom, a dark teal horizontal line with end caps spans from about 5 point 5 hours to about 8 point 0 hours. Panel C title: Overlap between patients at C H R dash P or with S S D s and healthy control participants. Two overlaid curves: the orange bimodal curve with peaks near 3 and 11 hours at about 40 percent, and the dark teal unimodal curve centered near 7 hours at about 40 percent. A vertical blue dotted line at approximately 7 hours. Two horizontal capped lines near the bottom: orange spanning about 2 to 12 hours and dark teal spanning about 5 point 5 to 8 hours. — Hypothetical distributions illustrate the study rationale. Participants at CHR-P or with SSDs are depicted with a conceptual bimodal distribution (A), reflecting short (approximately 3 hours) and long (approximately 11 hours) sleepers. In contrast, healthy control participants showed a unimodal distribution centered around approximately 7 hours of sleep (B), consistent with a normal pattern. Dashed vertical lines mark the mean, and horizontal lines with caps indicate the SD. While the mean values overlap (C), the bimodal distribution in the CHR-P or SSD group leads to higher SD.

Two metrics from the meta-analysis of variation¹⁴ can quantify interindividual variability across studies: (1) the natural logarithm of the variability ratio (lnVR), which compares the SD of the clinical group with that of the control group, and (2) the natural logarithm of the coefficient of variation ratio (lnCVR), which additionally adjusts for between-group differences in the mean. Clinically, interindividual variability in sleep captures the extent to which individuals within the same diagnostic group differ from one another in their sleep, whether patients cluster tightly around a common sleep profile or instead span a wide range (eg, from severe insomnia to hypersomnia).

Despite these considerations, sleep variability remains underexplored in psychosis. Quantifying interindividual variability in sleep across the psychosis spectrum relative to healthy control participants could yield valuable insights into these conditions and provide a quantitative basis for future works on candidate subpopulations and stratified care.^14,15

To address this gap, we conducted a systematic review and meta-analysis of case-control studies using wrist actigraphy to examine the interindividual variability in digital sleep phenotypes across psychosis stages (at CHR-P or with SSDs) compared with healthy control participants.

Methods

This systematic review and meta-analysis aggregated deidentified data from previously published case-control studies. Ethical approval and informed consent were obtained by the original investigators of each included study. No additional institutional review board approval and no additional informed consent were required for this secondary meta-analysis. This study was preregistered with the Center for Open Science¹⁶ and followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA).

Search Strategy and Study Selection

This study extends prior work by our group on digital sleep phenotypes in CHR-P and SSDs, and it follows the same inclusion and exclusion criteria, data extraction protocols, and risk-of-bias assessments.¹⁰ The PubMed, Embase, MEDLINE, Cochrane Central Register of Controlled Trials, ClinicalTrials.gov, EU Clinical Trials Register, World Health Organization International Clinical Trials Registry Platform, and OpenGrey databases were searched from database inception to April 29, 2024, with no language restriction. The core search syntax combined actigraphy OR actimetry OR wearable OR accelerometer AND psychosis OR schizophrenia OR schizo* OR clinical high risk OR prodromal OR ultra-high risk; the full search syntax is reproduced in the eMethods in Supplement 1. The search strategy and study selection were last updated in April 2026 to identify records published from April 30, 2024, to April 25, 2026; 185 additional records were identified, none of which met the inclusion criteria (Figure 2).

Flow diagram of record identification, screening, exclusions, and included studies. Two vertical streams of light gray rectangular boxes connected by thin gray arrows. Left stream, top box: 2177 records identified in initial search, with sources listed beneath as 1494 MEDLINE, 356 Embase, 272 PubMed, 21 CENTRAL, 14 ClinicalTrials dot gov, 13 WHO I C T R P, 6 OpenGrey, and 1 E U Clinical Trials Register. A downward arrow leads to a side box reading 891 duplicates removed. A downward arrow continues to a box reading 1286 records screened. A right-pointing arrow from this box leads to a box reading 1204 excluded after title and abstract screening. A downward arrow continues to a box reading 82 records retrieved and assessed for eligibility. A right-pointing arrow from this box leads to an exclusions box reading 63 excluded, with indented reasons: 23 wrong study design, 19 polysomnography only, 18 actigraphy of motor activity only, and 3 single-arm study. A downward arrow continues to a box reading 19 studies included in systematic review. A right-pointing arrow from this box leads to a box reading 1 excluded due to method of outcome reporting. A downward arrow continues to the final left-stream box reading 18 studies included in quantitative synthesis, meta-analysis. Right stream, top box: 185 records identified in updated search, with sources listed beneath as 91 Embase, 49 PubMed, 44 MEDLINE, and 1 WHO I C T R P. A right-pointing arrow leads to a box reading 81 duplicates removed. A downward arrow continues to a box reading 104 records screened, title abstract and full-text combined, and excluded. — CENTRAL indicates Cochrane Central Register of Controlled Trials; ICTRP, International Clinical Trials Registry Platform; WHO, World Health Organization.

Due to the high statistical heterogeneity encountered when pooling the CHR-P and SSD groups together, we decided to separate the analyses for each group. We included case-control studies that used wrist actigraphy to assess sleep in individuals at CHR-P or with SSDs and included a healthy control group. Each included study contributed effect sizes to one diagnostic comparison (CHR-P vs healthy control or SSDs vs healthy control); no study contributed to both comparisons. For studies with repeated assessments or multiple monitoring intervals, we selected the longest continuous actigraphy period to maximize the stability of habitual-sleep estimates, consistent with prior recommendations that recording windows of 7 days or longer improve test-retest reliability of actigraphy estimates.^17,18

Outcomes

Interindividual variability was primarily assessed with the log-transformed variability ratio (lnVR) of TST. Secondary outcomes included the log-transformed coefficient of variation ratio (lnCVR) for TST, as well as the lnVR and the lnCVR for the following additional actigraphy-derived sleep parameters: time in bed (TIB), sleep latency (SL), wake after sleep onset (WASO), sleep efficiency, and number of awakenings. lnVR was selected as the primary metric for its clinical interpretability and its ability to capture absolute variability differences linked to sleep instability; lnCVR provides complementary insight into relative variability.

Risk of Bias

Risk of bias was assessed at the study level for the primary outcome using the ROBINS-E tool,¹⁹ by 2 independent reviewers (R.A., external reviewer), with disagreements resolved by team discussion. Per-study and per-outcome ratings are reported in eTable 1 in Supplement 1.

Handling of Missing Data

Where SDs were not reported, they were calculated from the reported 95% CIs or SEs and the sample size, following the Cochrane Handbook for Systematic Reviews of Interventions.²⁰ No study had imputed values for the variables of interest.

Statistical Analysis

We fitted univariable random-effects meta-analyses with restricted maximum likelihood estimation of the between-study variance τ² for each outcome and each diagnostic group (CHR-P vs healthy control and SSDs vs healthy control). Effect sizes (lnVR and lnCVR) were computed from the reported group means, SDs, and sample sizes; the inverse variance–weighted random-effects pool was the primary summary.¹⁴ Statistical heterogeneity was assessed with the Q statistic, I², and τ² values and CIs. Sensitivity subgroup analyses were also conducted by stratifying each primary or significant outcome by sampling epoch (≤30 seconds vs ≥60 seconds) and risk of bias, with the between-subgroup test computed as the omnibus moderator Q from a meta-regression on the stratifier.

Moderator analyses used univariable mixed-effects meta-regression to explore the influence of antipsychotics (percentage of clinical arm participants taking any antipsychotic), age, and sex on TST—the only 3 variables identified as significant moderators in our previous meta-analysis.¹⁰ Multivariable models were not fitted because the small number of studies per moderator (≤3-13) does not support stable multivariable estimation. All P values are 2-sided and statistical significance was set at .05. All analyses were conducted in R, version 4.5.1 (R Project for Statistical Computing), using the meta and metafor packages,²¹ and were last updated in April 2026.

Results

Of the 19 studies meeting the inclusion criteria, 18 were included in the quantitative synthesis (1 study²² reported outcomes as medians and was excluded).^{23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40} The 18 studies contributed 1358 participants: 202 individuals at CHR-P (mean [SD] age, 20.7 [2.8] years; 104 female [51.4%] and 98 male [48.6%]), 574 with SSDs (mean [SD] age, 37.7 [10.2] years; 363 female [63.2%] and 211 male [36.8%]), and 582 healthy control participants (mean [SD] age, 31.5 [8.6] years; 217 female [46.6%] and 311 male [53.4%]). Of the 18 studies, 5 contributed a comparison of participants at CHR-P and healthy control participants^{23,24,25,26,27} and 13 contributed a comparison of participants with SSDs and healthy control participants^{28,29,30,31,32,33,34,35,36,37,38,39,40}; no study contributed effect sizes to both diagnostic comparisons (each study appeared in only 1 of the 2 pairwise pools). Control participants were matched on age within each primary study (Table and eTable 2 in Supplement 1). Figure 3 illustrates digital sleep phenotypes by diagnostic subgroups of patients at CHR-P and with SSDs compared with healthy control participants.

Table. Sleep Variability and Mean-Level Effect Estimates in Participants at CHR-P and Those With SSDs vs Healthy Control Participants.

Outcome	lnVR (95% CI)	lnCVR (95% CI)	MD (95% CI)^a
Total sleep time, min
CHR-P	0.16 (−0.05 to 0.37)	0.17 (−0.03 to 0.37)	−4.88 (−20.57 to 10.81)
SSD	0.46 (0.28-0.64)	0.22 (0.02-0.42)	106.13 (86.02-124.24)^b
Time in bed, min
CHR-P	0.29 (0.04-0.54)	0.27 (0.02-0.52)	3.49 (−181.81 to 188.79)
SSD	0.38 (0.19-0.56)	0.15 (−0.04 to 0.33)	121.58 (88.16-155.00)
Wake after sleep onset, min
CHR-P	0.45 (0.07-0.83)	0.26 (−0.03 to 0.56)	7.96 (−7.00 to 22.91)
SSD	0.52 (0.15-0.89)	0.16 (0.03-0.29)	17.87 (−8.70 to 44.22)
Sleep latency, min
CHR-P	0.13 (−0.36 to 0.61)	0.04 (−0.45 to 0.52)	1.00 (−3.97 to 5.97)
SSD	0.58 (−0.07 to 1.23)	0.05 (−0.28 to 0.38)	13.05 (2.11-24.00)
Sleep efficiency, %
CHR-P	0.35 (0.06-0.64)	0.38 (0.08-0.67)	−2.04 (−3.55 to 0.53)
SSD	0.27 (0.04-0.50)	0.27 (0.03-0.51)	−0.43 (−1.89 to 1.03)
No. of awakenings
CHR-P	0.27 (−0.02 to 0.56)	0.23 (−0.16 to 0.61)	0.98 (−10.83 to 12.80)
SSD	0.25 (−0.33 to 0.82)	−0.29 (−0.81 to 0.23)	1.70 (−5.31 to 9.51)

Open in a new tab

Abbreviations: CHR-P, clinical high risk for psychosis; lnCVR, natural logarithm of the coefficient of variation ratio; lnVR, natural logarithm of the variability ratio; MD, mean difference; SSD, schizophrenia spectrum disorder.

^{^a}

MD values are reproduced from our previous meta-analysis¹⁰ for comparison; lnVR and lnCVR values are new in the present analysis.

^{^b}

In participants with SSDs, the increase in mean total sleep time was associated with antipsychotic use (R² = 88.14%), age (R² = 38.89%), and sex (R² = 26.29%) in our previous meta-analysis.

Two-panel forest plots of sleep variability ratios and mean differences for C H R-P and S S D. Panel A, title at upper left: Interindividual sleep variability. A table at left with columns Outcome, No. of studies, No. of patients at risk slash No. of healthy control participants, and Variability ratio with 95% C I. Two grouped sections labeled C H R-P and S S D. C H R-P rows: Total sleep time, 5 studies, 202 slash 160, 1.17 (0.95 to 1.45); Time in bed, 2 studies, 87 slash 52, 1.34 (1.04 to 1.72); Sleep latency, 1 study, 19 slash 16, 1.14 (0.70 to 1.84); Wake after sleep onset, 4 studies, 183 slash 144, 1.57 (1.07 to 2.29); Sleep efficiency, 5 studies, 202 slash 160, 1.42 (1.06 to 1.90); No. of awakenings, 2 studies, 104 slash 67, 1.31 (0.98 to 1.75). S S D rows: Total sleep time, 13 studies, 574 slash 422, 1.58 (1.32 to 1.90); Time in bed, 4 studies, 145 slash 104, 1.46 (1.21 to 1.75); Sleep latency, 6 studies, 223 slash 160, 1.79 (0.93 to 3.42); Wake after sleep onset, 5 studies, 360 slash 246, 1.68 (1.16 to 2.44); Sleep efficiency, 9 studies, 387 slash 248, 1.31 (1.04 to 1.65); No. of awakenings, 3 studies, 79 slash 79, 1.28 (0.72 to 2.27). At right, aligned forest plot with a vertical dotted reference line at 1 and headers Favors lower variability on the left and Favors higher variability on the right. Horizontal axis labeled Variability ratio (95% C I), with ticks at zero point five, 1, 2, and 4. Each row has a dark square marker and a horizontal 95% C I line. Panel B, title: Mean-level differences in digital sleep phenotypes. A similar table at left with columns Outcome, No. of studies, No. of patients at risk slash No. of healthy control participants, and M D with 95% C I. C H R-P rows: Total sleep time, min, 5 studies, 202 slash 160, minus 4.88 (minus 20.57 to 10.81); Time in bed, min, 2 studies, 87 slash 52, 3.49 (minus 181.81 to 188.79); Sleep latency, min, 1 study, 19 slash 16, 1.00 (minus 3.97 to 5.97); Wake after sleep onset, min, 4 studies, 183 slash 144, 7.96 (minus 7.00 to 22.91); Sleep efficiency, percent, 5 studies, 202 slash 160, minus 2.04 (minus 3.55 to 0.53); No. of awakenings, 2 studies, 104 slash 67, 0.98 (minus 10.83 to 12.80). S S D rows: Total sleep time, min, 13 studies, 574 slash 422, 106.13 (86.02 to 124.24); Time in bed, min, 4 studies, 145 slash 104, 121.58 (88.16 to 155.00); Sleep latency, min, 6 studies, 223 slash 160, 13.05 (2.11 to 24.00); Wake after sleep onset, min, 5 studies, 360 slash 246, 17.87 (minus 8.70 to 44.22); Sleep efficiency, percent, 9 studies, 387 slash 248, minus 0.43 (minus 1.89 to 1.03); No. of awakenings, 3 studies, 79 slash 79, 1.70 (minus 5.31 to 9.51). At right, a forest — A, Digital phenotypes of sleep variability by diagnostic subgroups of patients at clinical high risk for psychosis (CHR-P) or with schizophrenia spectrum disorders (SSDs), back-transformed from the natural log of variability ratio (lnVR). B, Digital sleep phenotypes for mean difference (MD) by diagnostic subgroup (from a previous meta-analysis by Aronica et al¹⁰). For each diagnostic subgroup, the squares represent the variability ratio or MD values for each sleep metric, with square size scaled to study weight; the error bars represent the 95% CIs.

Total Sleep Time

Individuals at CHR-P did not differ significantly from control participants in TST variability (lnVR, 0.16 [95% CI, –0.05 to 0.37] minutes; I² = 47%; P = .14; and lnCVR, 0.17 [95% CI, –0.03 to 0.37] minutes; I² =41%; P = .10) (eFigures 1 and 2 in Supplement 1). Moderator analyses of treatment with antipsychotic medications were associated with greater variability (β = 0.24 [95% CI, 0.06-0.41]; R² = 100%; P = .008) (eFigure 3 in Supplement 1); given the small number of studies (k = 5), this finding should be interpreted as exploratory. Sex (β = 0.08 [95% CI, –0.15 to 0.30]; R² = 0%; P = .50) (eFigure 4 in Supplement 1) and age (β = 0.11 [95% CI, –0.11 to 0.33]; R² = 4.19%; P = .33) (eFigure 5 in Supplement 1) did not significantly moderate this effect.

In contrast, participants with SSDs showed greater TST variability than control participants (lnVR, 0.46 [95% CI, 0.28-0.64] minutes; I² = 70%; P < .001) (eFigure 6 in Supplement 1). lnCVR values also suggested greater variability (0.22 [95% CI, 0.02-0.42] minutes; I² = 76%; P = .03) (eFigure 7 in Supplement 1). No moderation effects were found for antipsychotic medications (β = –0.83 [95% CI, –2.13 to 0.48]; R² = 0%; P = .21), sex (β = 1.20 [95% CI, –2.48 to 0.09]; R² = 7.4%; P = .07), or age (β = 0.01 [95% CI, –0.01 to 0.03]; R² = 0%; P = . 30) (eFigures 8, 9, and 10 in Supplement 1, respectively). In summary, interindividual TST variability was greater in participants with SSDs but not in those at CHR-P.

Time in Bed

Participants at CHR-P showed increased TIB variability compared with healthy control participants (lnVR = 0.29 [95% CI, 0.04-0.54] minutes; I² = 0%; P = .02; and lnCVR = 0.27 [95% CI, 0.02-0.52] minutes; I² = 0%; P = .04) (eFigures 11 and 12 in Supplement 1). Participants with SSDs also had an increase in lnVR (0.38 [95% CI, 0.19-0.56] minutes; I² = 0%; P < .001) (eFigure 13 in Supplement 1) but not in lnCVR (0.15 [95% CI, –0.04 to 0.33] minutes; I² = 0%; P = .12) (eFigure 14 in Supplement 1). In summary, interindividual TIB variability was greater in both the CHR-P and SSD groups compared with control participants.

SL and Number of Awakenings

No significant group differences were observed for SL. In participants at CHR-P, the lnVR was 0.13 minutes (95% CI, –0.36 to 0.61 minutes; P = .60) (eFigure 15 in Supplement 1) and the lnCVR was 0.04 minutes (95% CI, –0.45 to 0.52 minutes; P = .89) (eFigure 16 in Supplement 1). For participants with SSDs, the lnVR was 0.58 minutes (95% CI, –0.07 to 1.23 minutes; I² = 94%; P = .08) (eFigure 17 in Supplement 1) and the lnCVR was 0.05 minutes (95% CI, –0.28 to 0.38 minutes; I² = 78%; P = .77) (eFigure 18 in Supplement 1).

For number of awakenings, neither group showed differences compared with healthy control participants. In participants at CHR-P, the lnVR was 0.27 (95% CI, –0.02 to 0.56; I² = 39%; P = .07) (eFigure 19 in Supplement 1) and the lnCVR was 0.23 (95% CI, –0.16 to 0.61; I² = 66%; P = .25) (eFigure 20 in Supplement 1). In participants with SSDs, the lnVR was 0.25 (95% CI, –0.33 to 0.82]; I² = 84%; P = .40) (eFigure 21 in Supplement 1) and the lnCVR was –0.29 (95% CI, –0.81 to 0.23; I² =81%; P = .28) (eFigure 22 in Supplement 1). In summary, neither SL nor number of awakenings showed significant interindividual variability differences in the CHR-P group or SSD group.

Wake After Sleep Onset

Both participants at CHR-P and those with SSDs had increased WASO variability. In those at CHR-P, the lnVR was 0.45 minutes (95% CI, 0.07-0.83 minutes; I² = 83%; P = .02) (eFigure 23 in Supplement 1). In participants with SSDs, the lnVR was 0.52 minutes (95% CI, 0.15-0.89 minutes; I² = 88%; P = .006) (eFigure 24 in Supplement 1), and the lnCVR was 0.16 minutes (95% CI, 0.03-0.29 minutes; I² = 12%; P = .02) (eFigure 25 in Supplement 1).

Sleep Efficiency

Individuals at CHR-P showed greater variability in sleep efficiency (lnVR, 0.35% [95% CI, 0.06%-0.64%]; I² = 72%; P = .02; and lnCVR, 0.38% [95% CI, 0.08%-0.67%]; I² = 73%; P = .01) (eFigures 26 and 27 in Supplement 1). Participants with SSDs also showed increased variability (lnVR, 0.27% [95% CI, 0.04%-0.50%]; I² = 71%; P = .02; and lnCVR, 0.27% [95% CI, 0.03%-0.51%]; I² = 73%; P = .03) (eFigures 28 and 29 in Supplement 1). In summary, both the CHR-P group and the SSD group showed greater interindividual variability in WASO and sleep efficiency than control participants.

Sensitivity Analyses

Risk of bias was rated as high for 5 of 18 studies and as having some concerns for the remaining 13 (eTable 1 in Supplement 1); no study was rated low risk. Excluding the 5 studies at overall high risk of bias preserved the direction and statistical significance of all primary and significant outcomes (eTable 3 in Supplement 1). This suggests that residual bias from the 5 studies with high risk of bias is unlikely to explain the interindividual variability signal. Influence diagnostics (eFigures 30 and 31 in Supplement 1; summarized in eTable 4 in Supplement 1) showed that SSD findings (TST, TIB, WASO, and sleep efficiency) were robust to single-study omission, whereas CHR-P findings for WASO and sleep efficiency were sensitive to the omission of 2 studies—reflecting the small number of CHR-P analyses (k = 4-5). Stratification by sampling epoch (≤30 seconds vs ≥60 seconds) yielded estimates that were broadly comparable with the all-studies pools across the 4 primary or significant outcomes; full per-stratum estimates are reported in eTable 5 in Supplement 1. Small-study effects were assessed for the only outcome with at least 10 contributing studies (TST lnVR in participants with SSDs, k = 13), as recommended by Higgins et al.²⁰ Results of the Egger regression test were not significant (intercept = 0.49 [95% CI, −0.06 to 1.05]; P = .89), suggesting no statistically detectable funnel plot asymmetry, although the test was moderately powered. The funnel plot and Egger scatter plot are provided in eFigure 32 in Supplement 1. The across-outcome summary is presented in eTable 6 in Supplement 1. Funnel plots for the 6 lower-k secondary outcomes are provided as visual inspection–only diagnostics in eFigure 33 in Supplement 1.

Discussion

This is, to our knowledge, the first systematic review and meta-analysis evaluating sleep variability using wrist actigraphy studies in individuals at CHR-P and in those with SSDs. The findings suggest that both individuals at CHR-P and those with SSDs exhibit greater interindividual variability in TIB, WASO, and sleep efficiency compared with healthy control participants; only individuals with SSDs showed increased interindividual variability in TST compared with control participants. Whether the SSD-specific TST variability reflects progressive divergence of sleep duration with illness chronicity, or a fixed feature of established psychotic illness, requires longitudinal investigation.

Sleep disturbances are well characterized in individuals with psychotic disorders and may precede overt core clinical symptoms.⁴¹ In individuals at CHR-P, we identified increased interindividual variability in TIB, WASO, and sleep efficiency, which may reflect early disruptions in sleep regulation⁴² in at-risk individuals. Our previous meta-analysis showed no mean-level differences in TIB or WASO between those at CHR-P and healthy control participants, concealing this potentially informative signal. The increased variability in these measures may suggest distinct candidate at-risk subpopulations⁴³ and support the view that disrupted sleep may not be merely secondary to psychotic symptoms but a candidate early biomarker of psychosis risk.⁴⁴ Interestingly, the greater variability did not extend to TST,⁴⁵ despite being increased in TIB. One interpretation is that individuals at CHR-P differ in the amount of time they spend in bed (some prolonging it, others shortening it) without corresponding changes in actual sleep duration. This dissociation between sleep opportunity and sleep obtained may represent an early feature of vulnerability,⁴⁶ with increased variability in TST emerging only at later illness stages, as we observed in participants with SSDs.

In participants with SSDs, the variability profile resembled CHR-P but with the additional, statistically distinct increase in TST variability. In our previous mean-level meta-analysis,¹⁰ longer average TST in individuals with SSDs was associated with antipsychotic treatment. The present variance-level analysis shows that the increase in TST variability is not moderated by antipsychotic use; this is more consistent with candidate subpopulations who have SSDs with divergent baseline sleep disturbance (eg, some with hypersomnia, others with insomnia) on which antipsychotics exert a general tendency to lengthen TST.⁴⁷ This pattern would also be consistent with a progressive disruption of sleep homeostatic mechanisms with illness progression.^13,48

The use of wrist actigraphy in this meta-analysis⁴⁹ underscores its translational potential for longitudinal research. Future studies should prioritize measures that were informative in this meta-analysis (WASO, TIB, sleep efficiency, and particularly TST) and collect continuous actigraphy data for at least 1 week to capture consistent sleep patterns.^17,18 Combining actigraphy with ecological momentary assessments,⁵⁰ clinical questionnaires,⁵¹ and passive smartphone data on sleep^52,53,54 is essential to enrich the definition of sleep phenotypes and enhance early detection strategies.⁵⁵ In this context, leveraging large, multisite datasets such as the AMP-SCZ (Accelerating Medicines Partnership Schizophrenia) program,⁵⁶ will be crucial to validate these markers prospectively and study their potential prognostic and therapeutic relevance in clinical settings.

Randomized clinical trials and emulated target trials⁵⁷ will also clarify whether interventions that stabilize sleep⁵⁸ (eg, cognitive behavior therapy for insomnia,⁵⁹ pharmacologic strategies,⁶⁰ or chronotherapies^61,62) reduce symptom severity, improve treatment response, and lower psychosis risk.^63,64 For example, the SleepWell trial demonstrated that cognitive behavior therapy for sleep problems in individuals at CHR-P not only reduced insomnia severity but also lessened symptoms of depression, anxiety, and ideas of reference and persecution.⁶⁵ Other interesting research avenues include novel mechanistic approaches to re-establish sleep homeostasis, such as mitochondria-targeted interventions.^66,67

Limitations

This study has some limitations. Because lnVR and lnCVR pool variances, they can inflate apparent interindividual variability if measurement noise differs systematically between groups. While residual measurement confounding cannot be excluded, each study used a single device and protocol for both clinical and control arms, so methodologic variance is shared between numerator and denominator within each study. A related consideration is that selection of the longest continuous actigraphy interval per study, while maximizing ecological validity, could introduce minor selection bias if recording duration differs systematically across clinical groups; reassuringly, stratification by sampling epoch (≤30 seconds vs ≥60 seconds) produced estimates broadly consistent with the all-studies pools across the 4 primary or significant outcomes, suggesting these methodologic factors did not materially confound the findings.

The case-control design of the included studies precludes causal inference and prognostic claims; therefore, we cannot determine whether elevated interindividual sleep variability precedes clinical change, and prospective studies—in CHR-P samples followed to first-episode psychosis, and in SSD samples followed to relapse—will be needed to address this. Between-study heterogeneity and sample composition imposes further constraints on generalizability: race and ethnicity were not consistently reported in the primary studies, and the CHR-P literature draws on heterogeneous instruments (Comprehensive Assessment of At-Risk Mental States [CAARMS], Structured Interview for Psychosis-Risk Syndromes [SIPS], and Community Assessment of Psychic Experiences [CAPE]) and inclusion criteria (genetic risk and psychotic-like experiences), limiting harmonization.

Statistical power was a further constraint, particularly for moderator analyses in SSDs: at most, 13 SSD studies contributed to the most-populated moderator, with fewer providing usable covariate data on antipsychotic use (with dose unavailable), age, or sex, so null moderator effects in this group should be interpreted cautiously. Finally, because our analyses operate on aggregated study-level statistics, the findings establish increased interindividual variability rather than the existence of discrete clinical subgroups. Individual-participant data are required to determine whether the observed dispersion reflects identifiable subpopulations or a single broader distribution.

Conclusions

In this systematic review and meta-analysis of 18 case-control studies, we identified distinct digital phenotypes of sleep variability among individuals across the psychosis spectrum compared with healthy control participants. Both individuals at CHR-P and those with SSDs showed greater variability in TIB, WASO, and sleep efficiency, while only those with SSDs showed greater TST variability. These findings support further investigations of sleep as a candidate biomarker to inform stratified care in psychosis⁶⁸ using individual-participant data and prospective studies.

Supplement 1.

eMethods. Full Search Strategy and Database Syntax

eTable 1. ROBINS-E Ratings Per Study and Per Domain (Primary Outcome)

eTable 2. Characteristics of Included Studies

eTable 3. Sensitivity Meta-Analysis Excluding Studies at Overall High Risk of Bias

eTable 4. Leave-One-Out Sensitivity Summary for the Primary or Significant Outcomes (lnVR Only)

eTable 5. Subgroup Analyses Stratified by Sampling Epoch

eTable 6. Small-Study Effects/Publication-Bias Assessment Across the Primary and Significant Outcomes

eFigure 1. Total Sleep Time (TST)—Patients With Clinical High Risk for Psychosis (CHR-P): Natural Logarithm of the Variability Ratio (lnVR) (k = 5; Restricted Maximum Likelihood [REML])

eFigure 2. Total Sleep Time—CHR-P: lnCVR (k = 5; REML)

eFigure 3. Total Sleep Time—CHR-P: Meta-Regression by Antipsychotic Use

eFigure 4. Total Sleep Time—CHR-P: Meta-Regression by Sex

eFigure 5. Total Sleep Time—CHR-P: Meta-Regression by Age

eFigure 6. Total Sleep Time—Patients With Schizophrenia Spectrum Disorders (SSD): lnVR (k = 13; REML)

eFigure 7. Total Sleep Time—SSD: lnCVR (k = 13; REML)

eFigure 8. Total Sleep Time—SSD: Meta-Regression by Antipsychotic Use

eFigure 9. Total Sleep Time—SSD: Meta-Regression by Sex

eFigure 10. Total Sleep Time—SSD: Meta-Regression by Age

eFigure 11. Time in Bed (TIB)—CHR-P: lnVR (k = 2; REML)

eFigure 12. Time in Bed—CHR-P: lnCVR (k = 2; REML)

eFigure 13. Time in Bed—SSD: lnVR (k = 4; REML)

eFigure 14. Time in Bed—SSD: lnCVR (k = 4; REML)

eFigure 15. Sleep Latency (SL)—CHR-P: lnVR (k = 1)

eFigure 16. Sleep Latency—CHR-P: lnCVR (k = 1)

eFigure 17. Sleep Latency—SSD: lnVR (k = 6; REML)

eFigure 18. Sleep Latency—SSD: lnCVR (k = 6; REML)

eFigure 19. Number of Awakenings (AWK)—CHR-P: lnVR (k = 2; REML)

eFigure 20. Number of Awakenings—CHR-P: lnCVR (k = 2; REML)

eFigure 21. Number of Awakenings—SSD: lnVR (k = 3; REML)

eFigure 22. Number of Awakenings—SSD: lnCVR (k = 3; REML)

eFigure 23. Wake After Sleep Onset (WASO)—CHR-P: lnVR (k = 4; REML)

eFigure 24. Wake After Sleep Onset—SSD: lnVR (k = 5; REML)

eFigure 25. Wake After Sleep Onset—SSD: lnCVR (k = 5; REML)

eFigure 26. Sleep Efficiency (SE)—CHR-P: lnVR (k = 5; REML)

eFigure 27. Sleep Efficiency—CHR-P: lnCVR (k = 5; REML)

eFigure 28. Sleep Efficiency—SSD: lnVR (k = 9; REML)

eFigure 29. Sleep Efficiency—SSD: lnCVR (k = 9; REML)

eFigure 30. Leave-One-Out Sensitivity Analyses for Total Sleep Time (TST), Time in Bed (TIB), Wake After Sleep Onset (WASO), and Sleep Efficiency (SE) in Patients With Clinical High Risk for Psychosis (CHR-P) and Schizophrenia Spectrum Disorders (SSD)

eFigure 31. Baujat Plots for Total Sleep Time (TST), Time in Bed (TIB), Wake After Sleep Onset (WASO), and Sleep Efficiency (SE) in Patients With Clinical High Risk for Psychosis (CHR-P) and Schizophrenia Spectrum Disorders (SSD)

eFigure 32. Funnel Plot and Egger Regression for the Primary Outcome (Total Sleep Time [TST] lnVR) in Patients With Schizophrenia Spectrum Disorders (SSD)

eFigure 33. Funnel Plots for Key Secondary Outcomes (Time in Bed [TIB], Wake After Sleep Onset [WASO], and Sleep Efficiency [SE]) in Patients With Clinical High Risk for Psychosis (CHR-P) and Schizophrenia Spectrum Disorders (SSD)

eReferences

jamanetwopen-e2624358-s001.pdf^{(3.1MB, pdf)}

Supplement 2.

Data Sharing Statement

jamanetwopen-e2624358-s002.pdf^{(15.6KB, pdf)}

References

1.Seitz-Holland J, Jacobs GR, Reinen J, et al. Incidence, prevalence, and stability of remission in individuals with clinical high risk for psychosis. JAMA Netw Open. 2025;8(8):e2525644. doi: 10.1001/jamanetworkopen.2025.25644 [DOI] [PMC free article] [PubMed] [Google Scholar]
2.Cohen S, Goldsmith DR, Ning CS, et al. Sleep disturbance, suicidal ideation and psychosis-risk symptoms in individuals at clinical high risk for psychosis. Psychiatry Res. 2024;341:116147. doi: 10.1016/j.psychres.2024.116147 [DOI] [PubMed] [Google Scholar]
3.Azargoonjahromi A, Bahrampour Z. Sleep disturbances are associated with psychosis in schizophrenia spectrum disorders: an umbrella review. Sleep Med. 2025;134:106690. doi: 10.1016/j.sleep.2025.106690 [DOI] [PubMed] [Google Scholar]
4.Fabbri M, Beracci A, Martoni M, Meneo D, Tonetti L, Natale V. Measuring subjective sleep quality: a review. Int J Environ Res Public Health. 2021;18(3):1082. doi: 10.3390/ijerph18031082 [DOI] [PMC free article] [PubMed] [Google Scholar]
5.Frase L, Nissen C, Spiegelhalder K, Feige B. The importance and limitations of polysomnography in insomnia disorder: a critical appraisal. J Sleep Res. 2023;32(6):e14036. doi: 10.1111/jsr.14036 [DOI] [PubMed] [Google Scholar]
6.Ancoli-Israel S, Cole R, Alessi C, Chambers M, Moorcroft W, Pollak CP. The role of actigraphy in the study of sleep and circadian rhythms. Sleep. 2003;26(3):342-392. doi: 10.1093/sleep/26.3.342 [DOI] [PubMed] [Google Scholar]
7.Marino M, Li Y, Rueschman MN, et al. Measuring sleep: accuracy, sensitivity, and specificity of wrist actigraphy compared to polysomnography. Sleep. 2013;36(11):1747-1755. doi: 10.5665/sleep.3142 [DOI] [PMC free article] [PubMed] [Google Scholar]
8.Bagautdinova J, Mayeli A, Wilson JD, et al. Sleep abnormalities in different clinical stages of psychosis: a systematic review and meta-analysis. JAMA Psychiatry. 2023;80(3):202-210. doi: 10.1001/jamapsychiatry.2022.4599 [DOI] [PMC free article] [PubMed] [Google Scholar]
9.De Crescenzo F, Economou A, Sharpley AL, Gormez A, Quested DJ. Actigraphic features of bipolar disorder: a systematic review and meta-analysis. Sleep Med Rev. 2017;33:58-69. doi: 10.1016/j.smrv.2016.05.003 [DOI] [PubMed] [Google Scholar]
10.Aronica R, Ostinelli EG, Austin C, et al. Digital sleep phenotype and wrist actigraphy in individuals at clinical high risk for psychosis and people with schizophrenia spectrum disorders: a systematic review and meta-analysis. BMJ Ment Health. 2025;28(1):e301337. doi: 10.1136/bmjment-2024-301337 [DOI] [PMC free article] [PubMed] [Google Scholar]
11.Coutts F, Koutsouleris N, McGuire P. Psychotic disorders as a framework for precision psychiatry. Nat Rev Neurol. 2023;19(4):221-234. doi: 10.1038/s41582-023-00779-1 [DOI] [PubMed] [Google Scholar]
12.Sasidharan A, Kumar S, Nair AK, et al. Further evidences for sleep instability and impaired spindle-delta dynamics in schizophrenia: a whole-night polysomnography study with neuroloop-gain and sleep-cycle analysis. Sleep Med. 2017;38:1-13. doi: 10.1016/j.sleep.2017.02.009 [DOI] [PubMed] [Google Scholar]
13.Boiko DI, Chopra H, Bilal M, et al. Schizophrenia and disruption of circadian rhythms: an overview of genetic, metabolic and clinical signs. Schizophr Res. 2024;264:58-70. doi: 10.1016/j.schres.2023.12.002 [DOI] [PubMed] [Google Scholar]
14.Nakagawa S, Poulin R, Mengersen K, et al. Meta-analysis of variation: ecological and evolutionary applications and beyond. Methods Ecol Evol. 2015;6(2):143-152. doi: 10.1111/2041-210X.12309 [DOI] [Google Scholar]
15.Brugger SP, Howes OD. Heterogeneity and homogeneity of regional brain structure in schizophrenia: a meta-analysis. JAMA Psychiatry. 2017;74(11):1104-1111. doi: 10.1001/jamapsychiatry.2017.2663 [DOI] [PMC free article] [PubMed] [Google Scholar]
16.Variability in digital phenotypes of sleep in individuals at clinical high risk for psychosis and with schizophrenia spectrum disorders: a meta- analysis of variability ratios using wrist actigraphy data. Center for Open Science . Updated April 27, 2026. Accessed June 18, 2026. https://osf.io/5n8bc/
17.Aili K, Åström-Paulsson S, Stoetzer U, Svartengren M, Hillert L. Reliability of actigraphy and subjective sleep measurements in adults: the design of sleep assessments. J Clin Sleep Med. 2017;13(1):39-47. doi: 10.5664/jcsm.6384 [DOI] [PMC free article] [PubMed] [Google Scholar]
18.Lau T, Ong JL, Ng BKL, et al. Minimum number of nights for reliable estimation of habitual sleep using a consumer sleep tracker. Sleep Adv. 2022;3(1):zpac026. doi: 10.1093/sleepadvances/zpac026 [DOI] [PMC free article] [PubMed] [Google Scholar]
19.Higgins JPT, Morgan RL, Rooney AA, et al. A tool to assess risk of bias in non-randomized follow-up studies of exposure effects (ROBINS-E). Environ Int. 2024;186:108602. doi: 10.1016/j.envint.2024.108602 [DOI] [PMC free article] [PubMed] [Google Scholar]
20.Higgins JPT, Thomas J, Chandler J, et al. , eds. Cochrane Handbook for Systematic Reviews of Interventions, version 6.4. Cochrane Collaboration; 2023. Accessed April 26, 2026. https://www.cochrane.org/authors/handbooks-and-manuals/handbook
21.Balduzzi S, Rücker G, Schwarzer G. How to perform a meta-analysis with R: a practical tutorial. Evid Based Ment Health. 2019;22(4):153-160. doi: 10.1136/ebmental-2019-300117 [DOI] [PMC free article] [PubMed] [Google Scholar]
22.Lauerma H, Niskanen L, Lehtinen I, Holmström R. Abnormal lateralization of motor activity during sleep in schizophrenia. Schizophr Res. 1994;14(1):65-71. doi: 10.1016/0920-9964(94)90010-8 [DOI] [PubMed] [Google Scholar]
23.Ristanovic I, Haase CM, Lunsford-Avery JR, Mittal VA. The relationship between stress responding in family context and stress sensitivity with sleep dysfunction in individuals at clinical high-risk for psychosis. J Psychiatr Res. 2022;149:194-200. doi: 10.1016/j.jpsychires.2022.02.038 [DOI] [PMC free article] [PubMed] [Google Scholar]
24.Nordholm D, Jensen MA, Glenthøj LB, et al. Sleep disturbances and the association with attenuated psychotic symptoms in individuals at ultra high-risk of psychosis. J Psychiatr Res. 2023;158:143-149. doi: 10.1016/j.jpsychires.2022.12.041 [DOI] [PubMed] [Google Scholar]
25.Lunsford-Avery JR, LeBourgeois MK, Gupta T, Mittal VA. Actigraphic-measured sleep disturbance predicts increased positive symptoms in adolescents at ultra high-risk for psychosis: a longitudinal study. Schizophr Res. 2015;164(1-3):15-20. doi: 10.1016/j.schres.2015.03.013 [DOI] [PMC free article] [PubMed] [Google Scholar]
26.Hennig T, Schlier B, Lincoln TM. Sleep and psychotic symptoms: an actigraphy and diary study with young adults with low and elevated psychosis proneness. Schizophr Res. 2020;221:12-19. doi: 10.1016/j.schres.2019.09.012 [DOI] [PubMed] [Google Scholar]
27.Bartsch U, Corbin LJ, Hellmich C, et al. Schizophrenia-associated variation at ZNF804A correlates with altered experience-dependent dynamics of sleep slow waves and spindles in healthy young adults. Sleep. 2021;44(12):zsab191. doi: 10.1093/sleep/zsab191 [DOI] [PMC free article] [PubMed] [Google Scholar]
28.Afonso P, Figueira ML, Paiva T. Sleep-promoting action of the endogenous melatonin in schizophrenia compared to healthy controls. Int J Psychiatry Clin Pract. 2011;15(4):311-315. doi: 10.3109/13651501.2011.605954 [DOI] [PubMed] [Google Scholar]
29.Apiquian R, Fresán A, Muñoz-Delgado J, Kiang M, Ulloa RE, Kapur S. Variations of rest–activity rhythm and sleep–wake in schizophrenic patients versus healthy subjects: an actigraphic comparative study. Biol Rhythm Res. 2008;39(1):69-78. doi: 10.1080/09291010701318253 [DOI] [Google Scholar]
30.Kammerer MK, Mehl S, Ludwig L, Lincoln TM. Sleep and circadian rhythm disruption predict persecutory symptom severity in day-to-day life: a combined actigraphy and experience sampling study. J Abnorm Psychol. 2021;130(1):78-88. doi: 10.1037/abn0000645 [DOI] [PubMed] [Google Scholar]
31.Martin JL, Jeste DV, Ancoli-Israel S. Older schizophrenia patients have more disrupted sleep and circadian rhythms than age-matched comparison subjects. J Psychiatr Res. 2005;39(3):251-259. doi: 10.1016/j.jpsychires.2004.08.011 [DOI] [PubMed] [Google Scholar]
32.Mayeli A, LaGoy AD, Smagula SF, et al. ; DiAPAson Consortium . Shared and distinct abnormalities in sleep-wake patterns and their relationship with the negative symptoms of Schizophrenia Spectrum Disorder patients. Mol Psychiatry. 2023;28(5):2049-2057. doi: 10.1038/s41380-023-02050-x [DOI] [PubMed] [Google Scholar]
33.Robillard R, Hermens DF, Naismith SL, et al. Ambulatory sleep-wake patterns and variability in young people with emerging mental disorders. J Psychiatry Neurosci. 2015;40(1):28-37. doi: 10.1503/jpn.130247 [DOI] [PMC free article] [PubMed] [Google Scholar]
34.Skeldon AC, Dijk DJ, Meyer N, Wulff K. Extracting circadian and sleep parameters from longitudinal data in schizophrenia for the design of pragmatic light interventions. Schizophr Bull. 2022;48(2):447-456. doi: 10.1093/schbul/sbab124 [DOI] [PMC free article] [PubMed] [Google Scholar]
35.Skowerska A, Wichniak A, Skalski M. Sleep and circadian rhythm disturbances in schizophrenia. Article in Polish. Psychiatr Pol. 2010;44(5):621-631. [PubMed] [Google Scholar]
36.Tous-Espelosin M, de Azua SR, Iriarte-Yoller N, et al. Clinical, physical, physiological, and cardiovascular risk patterns of adults with schizophrenia: CORTEX-SP study: characterization of adults with schizophrenia. Psychiatry Res. 2021;295:113580. doi: 10.1016/j.psychres.2020.113580 [DOI] [PubMed] [Google Scholar]
37.Wamsley EJ, Tucker MA, Shinn AK, et al. Reduced sleep spindles and spindle coherence in schizophrenia: mechanisms of impaired memory consolidation? Biol Psychiatry. 2012;71(2):154-161. doi: 10.1016/j.biopsych.2011.08.008 [DOI] [PMC free article] [PubMed] [Google Scholar]
38.Waters F, Sinclair C, Rock D, Jablensky A, Foster RG, Wulff K. Daily variations in sleep-wake patterns and severity of psychopathology: a pilot study in community-dwelling individuals with chronic schizophrenia. Psychiatry Res. 2011;187(1-2):304-306. doi: 10.1016/j.psychres.2011.01.006 [DOI] [PubMed] [Google Scholar]
39.Wichniak A, Skowerska A, Chojnacka-Wójtowicz J, et al. Actigraphic monitoring of activity and rest in schizophrenic patients treated with olanzapine or risperidone. J Psychiatr Res. 2011;45(10):1381-1386. doi: 10.1016/j.jpsychires.2011.05.009 [DOI] [PubMed] [Google Scholar]
40.Wulff K, Dijk DJ, Middleton B, Foster RG, Joyce EM. Sleep and circadian rhythm disruption in schizophrenia. Br J Psychiatry. 2012;200(4):308-316. doi: 10.1192/bjp.bp.111.096321 [DOI] [PMC free article] [PubMed] [Google Scholar]
41.Waite F, Sheaves B, Isham L, Reeve S, Freeman D. Sleep and schizophrenia: from epiphenomenon to treatable causal target. Schizophr Res. 2020;221:44-56. doi: 10.1016/j.schres.2019.11.014 [DOI] [PMC free article] [PubMed] [Google Scholar]
42.Zanini M, Castro J, Coelho FM, et al. Do sleep abnormalities and misaligned sleep/circadian rhythm patterns represent early clinical characteristics for developing psychosis in high risk populations? Neurosci Biobehav Rev. 2013;37(10 Pt 2):2631-2637. doi: 10.1016/j.neubiorev.2013.08.012 [DOI] [PubMed] [Google Scholar]
43.Renaud-Charest O, Paquin V, Daneault JG, et al. Subcategories of the clinical high-risk state for psychosis and their relationship to a full first-episode psychosis sample: an exploratory analysis of longitudinal outcomes. Schizophr Bull. 2025;sbaf155. doi: 10.1093/schbul/sbaf155 [DOI] [PMC free article] [PubMed] [Google Scholar]
44.Baldini V, Pasquino F, De Ronchi D, Plazzi G, Pelizza L, Menchetti M. Sleep disturbances in individuals with first episode psychosis and clinical high-risk states: a systematic review. Schizophr Res. 2025;282:123-132. doi: 10.1016/j.schres.2025.06.008 [DOI] [PubMed] [Google Scholar]
45.Reeve S, Nickless A, Sheaves B, et al. Sleep duration and psychotic experiences in patients at risk of psychosis: a secondary analysis of the EDIE-2 trial. Schizophr Res. 2019;204:326-333. doi: 10.1016/j.schres.2018.08.006 [DOI] [PMC free article] [PubMed] [Google Scholar]
46.Chung S. The DBST Index, the discrepancy between desired time in bed and desired total sleep time: the possible new sleep index predicting severity of insomnia. Sleep Med Res. 2022;13(2):85-93. doi: 10.17241/smr.2022.01368 [DOI] [Google Scholar]
47.Valencia Carlo YE, Saracco-Alvarez RA, Valencia Carlo VA, Vázquez Vega D, Natera Rey G, Escamilla Orozco RI. Adverse effects of antipsychotics on sleep in patients with schizophrenia: systematic review and meta-analysis. Front Psychiatry. 2023;14:1189768. doi: 10.3389/fpsyt.2023.1189768 [DOI] [PMC free article] [PubMed] [Google Scholar]
48.Ashton A, Jagannath A. Disrupted sleep and circadian rhythms in schizophrenia and their interaction with dopamine signaling. Front Neurosci. 2020;14:636. doi: 10.3389/fnins.2020.00636 [DOI] [PMC free article] [PubMed] [Google Scholar]
49.Liguori C, Mombelli S, Fernandes M, et al. The evolving role of quantitative actigraphy in clinical sleep medicine. Sleep Med Rev. 2023;68:101762. doi: 10.1016/j.smrv.2023.101762 [DOI] [PubMed] [Google Scholar]
50.Song J, Jeronimus BF, Fisher AJ. Sleep, event appraisal, and affect: an ecological momentary assessment study. J Affect Disord. 2024;361:376-382. doi: 10.1016/j.jad.2024.06.034 [DOI] [PubMed] [Google Scholar]
51.Kraepelien M, Blom K, Forsell E, et al. A very brief self-report scale for measuring insomnia severity using two items from the Insomnia Severity Index: development and validation in a clinical population. Sleep Med. 2021;81:365-374. doi: 10.1016/j.sleep.2021.03.003 [DOI] [PubMed] [Google Scholar]
52.Burns J, Chen K, Flathers M, et al. Transforming digital phenotyping raw data into actionable biomarkers, quality metrics, and data visualizations using cortex software package: tutorial. J Med Internet Res. 2024;26:e58502. doi: 10.2196/58502 [DOI] [PMC free article] [PubMed] [Google Scholar]
53.Langholm C, Breitinger S, Gray L, et al. Using data processing to understand inconsistency in smartphone behavior among patients with serious mental illness: Results of a digital phenotyping biomarker study. Biomark Neuropsychiatry. 2024;11:100105. doi: 10.1016/j.bionps.2024.100105 [DOI] [Google Scholar]
54.Vaidyam A, Halamka J, Torous J. Enabling research and clinical use of patient-generated health data (the mindLAMP Platform): digital phenotyping study. JMIR Mhealth Uhealth. 2022;10(1):e30557. doi: 10.2196/30557 [DOI] [PMC free article] [PubMed] [Google Scholar]
55.Polillo A, Foussias G, Wang W, et al. Care pathways and initial engagement in early psychosis intervention services among youths and young adults. JAMA Netw Open. 2023;6(9):e2333526. doi: 10.1001/jamanetworkopen.2023.33526 [DOI] [PMC free article] [PubMed] [Google Scholar]
56.Wigman JTW, Ching AE, Chung Y, et al. ; Accelerating Medicines Partnership® Schizophrenia (AMP® SCZ) . Digital health technologies in the Accelerating Medicines Partnership® Schizophrenia program. Schizophrenia (Heidelb). 2025;11(1):83. doi: 10.1038/s41537-025-00599-w [DOI] [PMC free article] [PubMed] [Google Scholar]
57.Hernán MA, Wang W, Leaf DE. Target trial emulation: a framework for causal inference from observational data. JAMA. 2022;328(24):2446-2447. doi: 10.1001/jama.2022.21383 [DOI] [PubMed] [Google Scholar]
58.Waite F, Myers E, Harvey AG, et al. Treating sleep problems in patients with schizophrenia. Behav Cogn Psychother. 2016;44(3):273-287. doi: 10.1017/S1352465815000430 [DOI] [PMC free article] [PubMed] [Google Scholar]
59.Tanskanen TE, Wegelius A, Härkönen T, et al. Cognitive behavioural therapy for insomnia (CBT-I) in schizophrenia and schizoaffective disorder: protocol for a randomised controlled trial. BMJ Open. 2024;14(6):e076129. doi: 10.1136/bmjopen-2023-076129 [DOI] [PMC free article] [PubMed] [Google Scholar]
60.Robertson I, Cheung A, Fan X. Insomnia in patients with schizophrenia: current understanding and treatment options. Prog Neuropsychopharmacol Biol Psychiatry. 2019;92:235-242. doi: 10.1016/j.pnpbp.2019.01.016 [DOI] [PubMed] [Google Scholar]
61.Gica Ş, Selvi̇ Y. Sleep interventions in the treatment of schizophrenia and bipolar disorder. Noro Psikiyatr Ars. 2021;58(suppl 1):S53-S60. doi: 10.29399/npa.27467 [DOI] [PMC free article] [PubMed] [Google Scholar]
62.Tsai YC, Jou JH, Hsu CC, Shih MC, The L, Sharma D. Circadian lighting effect for inpatients with schizophrenia: a prospective cohort study. Brain Behav. 2024;14(8):e70003. doi: 10.1002/brb3.70003 [DOI] [PMC free article] [PubMed] [Google Scholar]
63.Fekih-Romdhane F, Hallit S, Cheour M, Jahrami H. The nature, consequences, mechanisms, and management of sleep disturbances in individuals at-risk for psychosis. Front Psychiatry. 2022;13:1011963. doi: 10.3389/fpsyt.2022.1011963 [DOI] [PMC free article] [PubMed] [Google Scholar]
64.Fusar-Poli P, Davies C, Solmi M, et al. Preventive treatments for psychosis: umbrella review (just the evidence). Front Psychiatry. 2019;10:764. doi: 10.3389/fpsyt.2019.00764 [DOI] [PMC free article] [PubMed] [Google Scholar]
65.Waite F, Černis E, Kabir T, et al. ; SleepWell lived experience advisory group . A targeted psychological treatment for sleep problems in young people at ultra-high risk of psychosis in England (SleepWell): a parallel group, single-blind, randomised controlled feasibility trial. Lancet Psychiatry. 2023;10(9):706-718. doi: 10.1016/S2215-0366(23)00203-1 [DOI] [PubMed] [Google Scholar]
66.Sarnataro R. Neurobiology of mitochondrial dynamics in sleep. J Physiol. 2025;603(22):6747-6762. doi: 10.1113/JP288054 [DOI] [PMC free article] [PubMed] [Google Scholar]
67.Sarnataro R, Velasco CD, Monaco N, Kempf A, Miesenböck G. Mitochondrial origins of the pressure to sleep. Nature. 2025;645(8081):722-728. doi: 10.1038/s41586-025-09261-y [DOI] [PMC free article] [PubMed] [Google Scholar]
68.Mirzakhanian H, Singh F, Cadenhead KS. Biomarkers in psychosis: an approach to early identification and individualized treatment. Biomark Med. 2014;8(1):51-57. doi: 10.2217/bmm.13.134 [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials