Sleep and Anxiety: The Bidirectional Problem

Almost everyone with anxiety has experienced the vicious cycle: anxiety keeps you awake, and being sleep-deprived makes the anxiety worse. This bidirectional relationship is well-documented, but pulling apart cause from consequence is one of the more difficult challenges in anxiety research.

The experimental evidence: sleep loss increases anxiety

The clearest causal evidence comes from sleep deprivation experiments. When healthy people are kept awake for a night and then scanned with fMRI, the results are striking: amygdala reactivity to emotional stimuli increases by roughly 60%, while prefrontal cortex regulation of the amygdala decreases [2]. In other words, sleep loss makes the brain’s threat-detection system more reactive and its control system less effective.

Self-reported anxiety also increases after sleep deprivation, even in people without anxiety disorders [1]. These are not subtle effects. One night of poor sleep is enough to shift emotional processing measurably toward greater anxiety.

These experiments establish a causal direction: sleep loss causes increased anxiety. But they use total sleep deprivation in healthy volunteers, which is more extreme than the typical insomnia experienced by anxious people. Whether partial, chronic sleep disruption produces the same effects is less clear from experimental data.

The longitudinal evidence: insomnia predicts anxiety

Longitudinal studies follow people over time to see whether sleep problems precede the development of anxiety disorders. A large Norwegian study followed over 25,000 people for 11 years and found that chronic insomnia at baseline significantly predicted the development of anxiety disorders, even after controlling for baseline anxiety and depression [3].

This suggests insomnia is not merely a symptom of anxiety but a risk factor for it. However, longitudinal studies have their own limitations — residual confounding, reliance on self-report, and the possibility that subclinical anxiety drives both the insomnia and the later diagnosis. For how to think about this kind of evidence, see correlation vs causation.

The intervention evidence: treating sleep helps anxiety

Perhaps the strongest evidence for the sleep-to-anxiety direction comes from intervention studies. If treating insomnia reduces anxiety, that supports the idea that poor sleep actively contributes to anxiety maintenance.

A meta-analysis of CBT for insomnia (CBT-I) trials found that improving sleep produced secondary reductions in anxiety symptoms [4]. A large randomized trial of digital CBT-I in 3,755 participants — one of the biggest sleep intervention trials to date — found significant reductions in paranoia, hallucinations, depression and anxiety in the treatment group [6]. The anxiety reduction was a secondary outcome, not the primary target, but the effect was clinically meaningful and the trial was well-designed.

These findings don’t prove that sleep problems cause anxiety disorders. But they suggest that poor sleep is part of what maintains anxiety, and fixing sleep can help break the cycle.

Why the relationship is hard to study

Symptom overlap. Sleep disturbance is a diagnostic criterion for generalized anxiety disorder, among others. When insomnia is both a symptom and a proposed cause, disentangling the two is inherently circular.

Measurement. “Sleep quality” is a subjective experience shaped by anxiety itself. People with anxiety tend to perceive their sleep as worse than objective measures suggest [5]. When studies use self-report sleep questionnaires, the anxiety and sleep measures may partly be capturing the same thing — worry and negative appraisal — rather than two independent constructs.

Temporal resolution. Anxiety and sleep interact on timescales from minutes (lying awake worrying) to months (chronic insomnia preceding anxiety onset). Different study designs capture different timescales, and they may tell different stories.

Confounding. Many things that disrupt sleep also increase anxiety: stress, illness, substance use, major life events. Studies try to control for these, but residual confounding is always possible.

Questions to ask about sleep-anxiety claims

1. Does the study establish a temporal sequence (which came first)?
2. Is the evidence experimental, longitudinal or cross-sectional?
3. Was sleep measured objectively (actigraphy, polysomnography) or by self-report?
4. Were baseline anxiety levels controlled for in longitudinal analyses?
5. If an intervention improved both sleep and anxiety, was the sleep improvement a mediator or just a co-occurring change?

What remains uncertain

Whether sleep problems have a different causal relationship to different anxiety disorders (panic vs GAD vs social anxiety) is barely studied. Whether treating insomnia early could prevent anxiety disorder onset is an intriguing but untested idea. What the optimal sequencing of treatment should be — treat insomnia first, anxiety first, or both simultaneously — needs more evidence. The bidirectional relationship is well-established in broad strokes, but the clinical implications are still being worked out.

Sleep and your anxiety. If anxiety is disrupting your sleep, or poor sleep is making your anxiety worse, both are worth discussing with a clinician. Effective treatments exist for both insomnia and anxiety, and addressing one often helps the other. If anxiety is affecting your daily life, talk to a GP or mental health professional. If you are in crisis or thinking about harming yourself, contact emergency services or a crisis line now: in the US, call or text 988; in the UK and Ireland, Samaritans on 116 123.