Sleep Interventions and Psychosis Risk

confidence: high · 6 sources

⚠
Working research — not a recommendation.

This page records what the evidence says a policy does, not what anyone should do about it. It has been synthesized from the cited sources but has not been independently reviewed.

What the evidence says

The strongest evidence on this page is for sleep outcomes, not psychosis outcomes. A meta-analysis of CBT-I in patients with psychotic disorders, covering eight studies, reports improvements in insomnia and sleep quality in both the short and long term. It also reports short-term decreases in psychotic symptoms and increases in mental well-being. A systematic review calls CBT-I the most frequently studied intervention and consistent across delivery formats. It gives no pooled effect size.

For people at clinical high risk (CHR-P) or ultra-high risk, the evidence is smaller. In a randomized feasibility trial of 40 young people at ultra-high risk, SleepWell decreased insomnia severity at three and nine months compared with usual care alone. Preliminary decreases in psychotic experiences and depressive symptoms were also seen. That trial was not powered for psychosis outcomes. A review also reports improved sleep variables, psychiatric symptoms and social functioning in young CHR-P patients after CBT-I.

No source on this page shows that any sleep intervention decreases conversion from CHR-P to psychosis. Omega-3 fatty acids decrease insomnia, but a meta-analysis found no measured effect on conversion. Antipsychotics have no established evidence for preventing conversion. Melatonin has not been studied in CHR-P or early psychosis. Whether better sleep lowers conversion risk is not answered here.

The observational evidence gives a possible mechanism. In a prospective study, sleep disturbance was strongly correlated with heightened CHR symptoms over time. Sleep predicted CHR symptoms, but CHR symptoms did not predict sleep. The two directional effect sizes were similar, so this does not establish causation. Depression accounted for half of the association between sleep and symptoms. This suggests mood is one route linking sleep to symptoms. It is estimated from observational data only.

Other approaches have thinner support. Image Rehearsal Therapy in people with persecutory delusions produced large effect sizes for nightmare severity and insomnia, and a medium effect size for paranoia. The confidence intervals were wide, and the authors called for a larger trial. Light-based and circadian approaches rest on small uncontrolled studies. Pharmacological sleep agents give variable short-term benefits, with concerns about side-effects, tolerance and dependency. Several claims that sleep treatment "may" change long-term outcomes in early psychosis are forward-looking statements, not tested results.

Where the evidence splits

Baseline sleep disturbance did not predict conversion to psychosis in a CHR cohort. The same cohort found sleep disturbance strongly correlated with heightened CHR symptoms over time. The two findings use different outcomes: conversion to psychosis, and a continuous symptom measure. The sources on this page do not explain why the results differ.

Thresholds and dose-response

No threshold established in the current sources.

Open empirical questions

Findings

CBT-I decreases insomnia and short-term psychotic symptoms and increases sleep quality in patients with psychotic disorders

confidence — direction: highconfidence — magnitude: mediumconsensus: settled for sleep outcomes; thin for psychotic symptoms
mechanism
CBT-I targets insomnia directly. Symptom changes may follow from better sleep, but the sources do not test this.
effect
Meta-analysis of eight studies: improvements in insomnia and sleep quality in the short and long term. Short-term psychotic symptoms decrease and mental well-being increases. A systematic review reports consistent effectiveness across delivery formats, with no magnitude given.
conditions / moderators
Holds across delivery formats. Pooled effect sizes are not given in the review abstract. Long-term psychotic symptom outcomes are not reported.

Sleep-focused CBT decreases insomnia severity in people at high risk of psychosis and in psychosis spectrum diagnoses

confidence — direction: high for insomnia; medium for psychotic experiences and depressionconfidence — magnitude: medium for insomnia; low for other outcomesconsensus: thin, leaning toward decreased insomnia
mechanism
Cognitive and behavioral treatment of insomnia, delivered in person (SleepWell) or by smartphone with weekly therapist contact.
effect
SleepWell decreased insomnia severity at three and nine months versus usual care alone (40 participants at ultra-high risk). Preliminary decreases in psychotic experiences and depressive symptoms were also seen. A review reports increases in sleep variables, psychiatric symptoms and social functioning in young CHR-P patients. A single-arm smartphone study of 28 people reported ISI reductions and gains in depression and quality of life.
conditions / moderators
The SleepWell trial was a feasibility trial, not powered for psychosis outcomes. The smartphone study had no control arm and a small sample.

No intervention on this page has a measured effect on conversion from CHR-P to psychosis

confidence — direction: mediumconfidence — magnitude: lowconsensus: thin; the sources report absence of evidence, not measured null effects, except for omega-3 and baseline sleep
mechanism
not established
effect
Omega-3 fatty acids: no measured effect on conversion in meta-analysis, although they decrease insomnia and increase sleep quality. Antipsychotics: no established evidence for preventing conversion. Melatonin: not researched in this population. Baseline sleep disturbance did not predict conversion.
conditions / moderators
The omega-3 conversion data come from a meta-analysis, and the sleep-quality data from a narrative review. Trial size is the stated limit.

Sleep disturbance is associated with higher CHR symptoms over time, and depression carries half of that association

confidence — direction: highconfidence — magnitude: mediumconsensus: thin, one cohort
mechanism
Depression mediates part of the path from sleep disturbance to CHR symptoms.
effect
Strong positive association between sleep disturbance and CHR symptoms. Sleep predicted symptoms but not the reverse, with similar bidirectional effect sizes. Depression accounted for half of the association.
conditions / moderators
Observational data with cross-lagged panel models. Similar bidirectional effect sizes limit causal inference.
  • Eva Velthorst (2025), Carolina Digital Repository (University of North Carolina at Chapel Hill) · doi:10.17615/2jvh-b829 · https://doi.org/10.17615/2jvh-b829 (retrieved 2026-10-03)

Image Rehearsal Therapy decreases nightmare severity, insomnia and paranoia in people with persecutory delusions

confidence — direction: mediumconfidence — magnitude: medium for the trial; low for the case seriesconsensus: thin
mechanism
Targets nightmares, which are linked to insomnia and paranoia. The sources do not test the causal chain.
effect
Large effect sizes for nightmare severity and insomnia at post-treatment and follow-up. Medium effect size for paranoia. A case series of five participants reported descriptive decreases in nightmare distress, vividness and intensity.
conditions / moderators
Confidence intervals were wide. The authors deemed a larger trial necessary. The case series had no control group.

Light and circadian interventions show preliminary increases in sleep, well-being and psychopathology scores

confidence — direction: mediumconfidence — magnitude: lowconsensus: thin
mechanism
Realigning disrupted circadian rhythms through light exposure and structured sleep scheduling.
effect
A 10-week circadian simulation lighting study of 20 inpatients with schizophrenia reported improvements in BPRS total score and MMSE, with no adverse events. L-DART showed preliminary improvements in subjective sleep and well-being. A review states that adjunctive therapies show promise in specific subpopulations, with no pooled estimates.
conditions / moderators
Both studies were uncontrolled. The samples were small, and the L-DART study was a feasibility design. Benefit is limited to "specific subpopulations".

Pharmacological sleep agents show mixed evidence in psychotic disorders

confidence — direction: high that results varyconfidence — magnitude: lowconsensus: thin
mechanism
not established in the sources
effect
Variable short-term benefits, with concerns about side-effects, tolerance and dependency. No effect sizes are given.
conditions / moderators
not established

A four-week sleep intervention in early psychosis shows mixed evidence: no measured effect on sleep duration, increases in REM sleep and well-being

confidence — direction: mediumconfidence — magnitude: lowconsensus: thin
mechanism
not established
effect
Sleep duration rose 26 minutes (p = 0.272), with no measured effect. Fitbit sleep score rose from 80.7% to 81.5% (p = 0.158), with no measured effect. REM sleep increased from 21.9% to 22.5% of total sleep time (p = 0.023). WEMWBS increased from 49.8 to 52.9 (p = 0.037, effect size 0.32). PANAS negative scores decreased. Wake time after sleep onset was positively correlated with the PANAS negative scale (rho = 0.676).
conditions / moderators
No control group. The correlation is cross-sectional and does not show that the intervention caused the change.

Authors state that sleep improvement may decrease psychotic symptoms, but no intervention trial here tests it

confidence — direction: lowconfidence — magnitude: lowconsensus: thin
mechanism
Sleep disturbances are described as modifiable targets. This is stated in discussion, not tested.
effect
The sources say better sleep "may" alleviate delusions and paranoia, and that sleep-targeted interventions "may" influence long-term outcomes. No magnitudes are given.
conditions / moderators
not established