Sleep hygiene is the most frequently recommended non-pharmacological intervention for insomnia. The advice is ubiquitous — consistent bedtime, no screens, cool room, no caffeine after noon. But the evidence base behind specific recommendations varies from well-established to barely tested.
This is a field where the gap between confident recommendations and actual RCT evidence is larger than most people realize.
The hierarchy: CBT-I first, sleep hygiene second
Before sleep hygiene: the most evidence-backed non-pharmacological intervention for chronic insomnia is Cognitive Behavioral Therapy for Insomnia (CBT-I).
Meta-analyses (Trauer et al., 2015, Annals of Internal Medicine): CBT-I reduces sleep onset latency by 19 minutes, wake-after-sleep-onset by 26 minutes, and total wake time by 55 minutes vs. control. Effect sizes rival or exceed pharmacological interventions, with benefits persisting after treatment ends.
Sleep hygiene alone (without CBT-I) has much weaker evidence as a standalone treatment for clinical insomnia. As a component of comprehensive CBT-I, it contributes; as a solo intervention, it's often insufficient.
For subclinical sleep difficulties (poor sleep quality without clinical insomnia), sleep hygiene interventions have better standalone evidence.
Light and circadian entrainment — strongest evidence
Morning light exposure
The suprachiasmatic nucleus (SCN) in the hypothalamus is the master circadian clock, entrained primarily by the light-dark cycle. Bright light in the morning advances the circadian phase — making it easier to fall asleep earlier at night.
Evidence:
- Bright light therapy (2,500–10,000 lux, 30–60 minutes in the morning) is established first-line treatment for Seasonal Affective Disorder (SAD) and Delayed Sleep Phase Disorder (DSPD)
- RCTs in non-clinical populations show morning light exposure reduces sleep onset latency and increases sleep duration
- Outdoor light (10,000–100,000 lux on sunny days) is dramatically more effective than indoor light (~300 lux)
Practical protocol: 10–30 minutes outdoor light exposure within 1 hour of waking. If outdoor not feasible, a 10,000 lux lightbox for 20–30 minutes.
Blue light in the evening
Melatonin suppression by blue-wavelength light (480nm) is well-established. Evening screen light delays melatonin onset and pushes the circadian clock later.
Evidence:
Chang et al. (2014, PNAS): e-Reader use in the evening compared to printed books delayed melatonin onset by 1.5 hours, reduced melatonin by 55%, and reduced REM sleep. Next-day alertness was reduced.
Blue-blocking glasses RCTs: Multiple small studies show improved sleep outcomes; effect sizes are modest but consistent.
Practical guidance: Dim ambient lighting in the 2 hours before sleep. Blue-blocking glasses or Night Shift/f.lux have moderate evidence. The phone-in-bed behavior pattern (engaging, alerting content) may matter as much as the light itself — hard to separate in studies.
Temperature — well-evidenced
Core body temperature drops 1–2°C at sleep onset as part of the circadian process. Accelerating this drop facilitates sleep onset.
Evidence:
- Optimal sleep room temperature: 15–19°C (60–67°F) — established across multiple lab studies
- Warm bath/shower 1–2 hours before bed: vasodilation increases heat loss from skin → core temperature drops faster → faster sleep onset. Haghayegh et al. (2019, Sleep Medicine Reviews): meta-analysis found warm water immersion (40–42.5°C, 10+ minutes, 1–2 hours before sleep) significantly reduced sleep onset latency and improved sleep quality.
This is one of the more surprising and well-evidenced sleep hygiene practices — the warm bath paradox (hot bath → faster sleep) is counterintuitive but mechanistically sound.
Consistent sleep schedule — moderate evidence
Circadian phase is entrained by regular timing. Irregular sleep schedules (different bedtimes across the week) disrupt entrainment and impair sleep quality.
Evidence:
Phillips et al. and multiple observational studies show sleep timing variability (social jetlag) correlates with worse health outcomes. Intervention RCTs are limited — it's logistically difficult to randomize sleep schedule consistency.
Wirth et al. (2022, Sleep): Sleep timing variability associated with worse sleep quality, mood, and next-day functioning in an ecological study.
The weekend sleep schedule problem: Sleeping in on weekends delays circadian phase, making Monday morning harder. This "social jetlag" (Wittmann et al., 2006) is associated with metabolic syndrome, depression, and impaired performance. Limiting weekend sleep drift to <1 hour is the recommendation.
Caffeine — well-evidenced timing
Caffeine's half-life is 5–7 hours. This means 100mg consumed at 2pm still has ~50mg active at 9pm.
Evidence:
Drake et al. (2013, Journal of Clinical Sleep Medicine): Caffeine consumed 6 hours before bedtime significantly reduced total sleep time by 1 hour vs. placebo.
Most "cut off at noon" advice is correct for most people, but the precise cutoff depends on individual CYP1A2 enzyme activity — fast metabolizers can consume caffeine later with less sleep impact.
Adenosine rebound: Caffeine blocks adenosine receptors without preventing adenosine accumulation. When caffeine wears off, accumulated adenosine hits unblocked receptors — the characteristic afternoon crash. This doesn't improve sleep quality; adenosine accumulated through natural wakefulness does.
Exercise — evidence with timing nuance
Meta-analysis (Kredlow et al., 2015): Exercise significantly improves sleep quality across 66 studies. Acute effects: reduced sleep onset latency; chronic effects: improved slow-wave sleep duration.
Timing debate: Older guidelines recommended avoiding evening exercise. More recent RCTs challenge this:
- Stutz et al. (2019, Sports Medicine meta-analysis): Evening exercise (ending >1 hour before sleep) did not impair sleep and may improve sleep quality
- High-intensity exercise within 1 hour of bedtime does impair sleep onset in some subjects (elevated core temperature, cortisol, arousal)
Current evidence: Exercise timing matters less than previously thought for most people. Vigorous exercise within 1 hour of bedtime is the exception to avoid. Otherwise, exercise at whatever time you can do it consistently.
Alcohol — underappreciated sleep disruptor
Alcohol is used widely as a sleep aid — it accelerates sleep onset. But it disrupts sleep quality significantly:
- Suppresses REM sleep in the first half of the night
- Causes rebound arousal as alcohol is metabolized, fragmenting sleep in the second half
- Increases snoring and worsens sleep apnea
- Reduces total REM time significantly
Ebrahim et al. (2013): Dose-dependent REM suppression with alcohol. Even low doses (0.1g/kg, roughly 1 standard drink) produce measurable REM reduction.
The sleep subjectively feels better with alcohol (faster onset). Objectively, it is worse. This explains why alcohol as a sleep strategy is self-reinforcing and ineffective long-term.
What doesn't have strong evidence
Strict "no screens 1 hour before bed": The evidence for content-based arousal (engaging media, stressful content) may be larger than the blue-light effect. Calm content on a dimmed screen may be less disruptive than commonly portrayed.
Counting sheep: Tested. Doesn't work. Imagery distraction (imagining a relaxing scene) was more effective in one RCT.
Milk before bed: Milk contains tryptophan, a serotonin/melatonin precursor. The dose of tryptophan in a glass of milk is too small to meaningfully raise brain tryptophan levels — pharmacologically insufficient.
Weighted blankets: Some evidence for anxiety reduction; mixed evidence for sleep quality specifically. Better evidence in autism spectrum disorder. Limited high-quality RCTs in general population.
The framework applied
For any sleep hygiene or intervention study:
- Clinical insomnia vs. subclinical sleep difficulty? CBT-I evidence applies to diagnosed insomnia; hygiene evidence more relevant to general population
- What was the control condition? Good sleep hygiene vs. no advice vs. sleep restriction — comparison matters
- Subjective vs. objective sleep measures? Polysomnography (PSG) and actigraphy vs. self-report — sleep quality is poorly self-assessed
- What duration? Acute vs. chronic effects of interventions differ significantly
We automated this at Q-SCI. Any study — paste it, get a quality score.
Bottom line
- CBT-I is the most effective non-pharmacological treatment for chronic insomnia — sleep hygiene alone is insufficient for clinical insomnia
- Morning bright light (outdoor, 10–30 min) is one of the most effective and underused circadian interventions
- Evening light reduction (especially screens) delays melatonin onset — dim environment 2 hours before bed is better-evidenced than blue-blocking glasses alone
- Warm bath 1–2 hours before bed (not immediately before): reduces core temperature, speeds sleep onset — better evidence than most people expect
- Consistent sleep/wake timing including weekends (limit social jetlag to <1 hour)
- Caffeine cutoff ~6 hours before sleep (individual variation based on metabolism)
- Alcohol disrupts sleep architecture despite accelerating onset — not a sleep aid
- Exercise improves sleep at most timing windows; avoid very vigorous exercise within 1 hour of bed
More evidence-based analyses at q-sci.org/blog. Score studies free at q-sci.org.
Top comments (0)