The research library

These are the 53 papers we read while building binaural: randomized trials, meta-analyses, reviews and experiments on binaural beats, pulsed tones, flickering light, noise and slow breathing. Each has a one-page summary of what was done, what was found and how far to trust it. Null and negative results are kept in.

By Jude Mc, founder of binaural. Last updated 10 November 2026. Not medical advice.

Each summary is labelled by how much weight the study can carry. One lamp per study:

Attention, memory and thinking

Summary: Binaural beats for focus and studying

• A month of daily 6 Hz binaural beats and attention (Chockboondee 2024)Randomized trial · Mixed evidenceCompared with a silent control group, four weeks of daily listening raised auditory P300 amplitude and shortened P300 latency, and reaction times were faster. The control was silence rather than a sham sound, so expectancy effects are not ruled out.
• Alpha binaural beats and memory recall: a pilot study (Mujib 2021)Pilot study · PreliminaryMemory score rose significantly only in the 10 Hz alpha group (p = 0.0000235, Cohen's d = 1.15). The study had no control group, so practice effects cannot be separated from any effect of the beats.
• Binaural beats and creativity (Reedijk 2013)Experiment · PreliminaryThere was no overall effect. Divergent thinking improved only in people with a low spontaneous eye-blink rate (a proxy for low striatal dopamine), while those with a high rate were unaffected or even impaired; convergent thinking was not affected.
• Binaural beats, brain entrainment and sustained attention (Melnichuk 2025)Experiment · Mixed evidenceEEG showed entrainment at the target frequency (16 or 40 Hz), but binaural beats did not reduce the decline in performance across the task. One combination, gamma beats on a 340 Hz carrier, improved overall accuracy (p = 0.002).
• Binaural beats, memory and attention: meta-analysis (Basu 2022)Meta-analysis · Mixed evidencePooled across 15 studies, binaural beats showed a near-moderate benefit for memory and attention (Hedges' g = 0.40, 95% CI 0.14 to 0.67). Heterogeneity was high (I² about 77%) and study-level results conflicted.
• Binaural beats, vigilance and mood: a double-blind study (Lane 1998)Experiment · Mixed evidenceBeta beats gave more correct target detections (mean 153.5 vs 147.6 of 180, p<0.05 one-tailed) and fewer false alarms (6.6 vs 8.7, p<0.02) than theta/delta beats. The effects were small in absolute terms.
• Gamma binaural beats and attentional focus (Colzato 2015)Experiment · PreliminaryPeople who heard 40 Hz beats showed a smaller global-precedence effect than controls (36 ms vs 57 ms, p<0.05), meaning attention was more locally focused. Suppression of irrelevant information was not affected.
• Gamma binaural beats and the attentional blink (Reedijk 2015)Experiment · Preliminary40 Hz gamma beats eliminated the attentional blink, but only in people with a low spontaneous eye-blink rate (a proxy for low striatal dopamine). Those with a high rate showed no benefit, and alpha beats had no effect in either group.
• Gamma binaural beats, working memory and OCD symptoms (Hodhodi 2024)Experiment · PreliminarySymptom severity was significantly reduced (P<0.05) at post-test and at one-month follow-up. Working-memory improvement was in the expected direction but not statistically significant (P>0.05).
• Listening fatigue after conversations in virtual noisy rooms (Hládek 2025)Experiment · PreliminaryAfter effortful conversation in noise, response times got faster and accuracy did not systematically change, yet self-reported effort and fatigue rose. The speed-ups were partly training artifacts.

Anxiety and stress

Summary: Binaural beats for anxiety

• Binaural beats around surgery: anxiety and pain meta-analysis (Xiong 2025)Meta-analysis · Mixed evidencePooled effects were large and significant for perioperative anxiety (SMD = −1.38, 95% CI −1.89 to −0.87) and postoperative pain (SMD = −0.61, 95% CI −1.03 to −0.19). Heterogeneity was very high (I² = 91.6% and 75.8%), so the pooled estimates are fragile.
• Binaural beats for dental anxiety and pain: a meta-analysis (Shukla 2025)Meta-analysis · Mixed evidenceBinaural beats significantly reduced dental anxiety versus control (VAS mean difference −1.50, 95% CI −2.34 to −0.65). The effect on pain was inconclusive against a no-intervention control, and the certainty of the evidence was rated low.
• Binaural beats for everyday stress: a systematic review (Platt 2024)Systematic review · PreliminaryThe evidence was equivocal. About a third of the 12 randomized trials found no significant reduction in stress or anxiety, and for theta beats 4 of 7 studies favoured binaural beats while 3 found no effect.
• Binaural beats vs music in young adults: systematic review (Elnazer 2026)Systematic review · Mixed evidenceAcross 10 trials the review reports consistent, moderate improvements in sleep quality, mood and stress regulation, with small-to-moderate effects (Cohen's d about 0.3–0.6). Binaural beats gave more consistent benefits (d about 0.34–0.59) than music therapy alone.
• Daily binaural beat listening for 60 days: a pilot study (Wahbeh 2007)Pilot study · PreliminaryFour of 33 measures changed significantly: trait anxiety decreased (p=0.004), quality of life increased (p=0.03), and IGF-1 (p=0.01) and urinary dopamine (p=0.02) decreased. With 8 participants and no control group, the findings cannot be generalized.
• Hypnosis for distress around medical procedures (Tefikow 2013)Meta-analysis · Mixed evidenceA meta-analysis of 34 randomized trials (n = 2,597) found moderate effects on distress (Hedges' g = 0.53) and pain (g = 0.44). For general, everyday stress the evidence is markedly weaker and equivocal.
• Isochronic tones versus binaural beats at 10 Hz: a trial (Kanzler 2023)Randomized trial · Mixed evidenceSymptoms fell across the active arms: binaural beats showed significant drops in anxiety (p = 0.028), stress (p = 0.012) and depression (p = 0.035), and isochronic tones significantly improved stress (p = 0.027) and sleep quality (p = 0.041). The groups were too small to show that one method differs from the other.

Sleep

Summary: Binaural beats for sleep

• Binaural-beat music and sleep in older adults (Lin 2024)Randomized trial · Mixed evidenceAfter 14 days, the binaural-beat music group showed significant improvement in sleep quality (Pittsburgh Sleep Quality Index) and a significant reduction in sympathetic nervous activity versus control, plus improvement in depression.
• Pre-sleep audio-visual alpha stimulation for pain and sleep (Halpin 2023)Pilot study · PreliminarySelf-reported sleep improved (total sleep time +29 min, p = .0033), but objective actigraphy showed no significant change. The study had no control group.

Pain

• Binaural beats and chronic pain: a double-blind trial (Gkolias 2020)Randomized trial · Strong designIn this small trial, pain ratings fell from 5.6 to 3.4 after 30 minutes of binaural beats (p < .001) versus 5.2 to 4.8 with sham (p = .78). Analgesic use over the following week was also significantly lower with binaural beats.

What happens in the brain

Summary: What are binaural beats?

• Binaural beats and alpha coherence between the hemispheres (Solcà 2015)Experiment · Mixed evidenceBinaural beats increased coherence between the auditory cortices compared with monaural beats and rest, but at alpha frequencies, not at the beat frequency itself. There was no behavioural aftereffect on a dichotic digit task.
• Do binaural beats entrain the brain? A systematic review (Ingendoh 2023)Systematic review · PreliminaryThe evidence is inconclusive. Of 14 EEG studies, only 5 were in line with the entrainment hypothesis, 8 reported contradictory results and 1 was mixed.
• EEG responses to binaural beats at four frequencies (Gao 2014)Experiment · PreliminaryNo. The dominant EEG frequency did not shift toward the beat frequency in any condition; instead there were transient, band-specific changes in power and connectivity.
• EEG-guided binaural beats for relaxation and cognition (Kahathuduwa 2025)Randomized trial · Mixed evidenceThe EEG-guided audio lowered dominant EEG frequency below 8 Hz in 100% of participants (median 7.4 min), while sham reached that state in essentially nobody. Cognitive effects were modest: one significant result (novelty-encoding reaction time, p = 0.039) among several non-significant trends.
• Short binaural beats, EEG and vigilance: a null result (Goodin 2012)Experiment · No support foundNo. Vigilance reaction times did not differ between beta beats, theta beats and white noise (all p > .27), and EEG spectral power showed no entrainment (beta p = .227, theta p = .625).

Light, flicker and 40 Hz

Summary: 40 Hz, gamma and the brain

• 10 Hz flicker gets in the way of attention (Gulbinaite 2017)Experiment · Strong designYes, for the worse: 10 Hz flicker interfered with the task when attention demands were high, and the effect was larger the closer 10 Hz was to the person's own alpha rate.
• 40 Hz light stimulation: failed replications in mice (Soula 2023)Preclinical study · Mixed evidenceTwo independent labs failed to replicate the core mouse findings: 40 Hz flicker did not entrain native gamma oscillations and gave no significant amyloid change. A human pilot (n = 15) was a feasibility study in which standard cognitive scales were flat.
• 40 Hz sensory stimulation in Alzheimer's-model mice (Iaccarino 2016)Preclinical study · Mixed evidenceIn 5XFAD mice, 40 Hz stimulation reduced amyloid-β, shifted microglia toward a clearance state and improved memory in mouse tasks. These are preclinical findings in mice.
• Audio-visual entrainment in psychological disorders (Pino 2022)Randomized trial · PreliminaryResults were mixed: depression scores decreased (p = .022) and Performance IQ rose (p = .016), but state and trait anxiety significantly increased in the audio-visual entrainment group. The trial had only 15 participants.
• Audio-visual entrainment: a review of the evidence (Rahmani 2025)Review · PreliminaryFindings are heterogeneous and include null or limited effects. The review concludes that it is premature to say the technique is effective and that rigorous, standardized, adequately controlled studies are needed.
• Blue and red light in the afternoon dip (Sahin 2013)Experiment · Mixed evidenceEEG markers of sleepiness (alpha, alpha-theta and theta power) were lower after light exposure, and the authors conclude that melatonin suppression is not needed for light to have an alerting effect.
• Colour therapy: a review that concedes little evidence (Azeemi 2005)Review · No support foundNo controlled-trial evidence is offered. Even this sympathetic review admits a "severe lack of scientific work" on chromotherapy.
• Flicker made of colour changes instead of on and off (Henney 2024)Experiment · Mixed evidenceYes, and the size of the response depends on the pair of colours: amber-red was strongest, and pairs that included blue or red were consistently stronger than those without.
• Flicker with a matching binaural beat (Ávila 2026)Experiment · Mixed evidenceAt 8 and 20 Hz, yes: the rhythmic response to the flicker was larger with the matching beat. At 40 Hz there was no clear rhythmic gain once background activity was separated out.
• Flickering light and geometric hallucinations (Amaya 2023)Experiment · Mixed evidenceRhythmic light flicker reliably induced geometric hallucinations and phosphenes in sober participants, peaking around 10 Hz. An arrhythmic control with matched luminance strongly reduced the effect.
• Ganzfeld: imagery from a uniform visual field (Wackermann 2008)Experiment · Mixed evidenceDeprived of contrast and edges, most observers spontaneously reported imagery ranging from elementary percepts to vivid, dream-like scenes. The EEG resembled an activated wakeful state, not sleep onset.
• How much light gets through a closed eyelid (Bierman 2011)Measurement study · Strong designVery little: on average the eyelid cuts light between 450 and 650 nm by a factor of about a hundred, and it blocks short wavelengths (blue) far more than earlier figures suggested.
• Narrow-band green light during migraine (Lipton 2023)Open-label observational study · PreliminaryMany said so: headache was rated improved in 55% of 3,232 attacks. But this was an uncontrolled study of people who had bought the lamp, and only 5% of buyers supplied complete data.
• Red light through closed eyelids and waking up (Figueiro 2019)Crossover experiment · Mixed evidencePartly. Performance on some tasks was better after red light through closed eyelids than after dim light, at light levels that do not suppress melatonin.
• Sharp-edged or smooth flicker (Panitz 2023)Two-laboratory experiment · Strong designYes at slower rates. Square-wave flicker gave larger responses than sine-wave at 6 and 8.57 Hz in both laboratories; the difference was weaker at 15 Hz.
• Two eyes flickering together or in turn (Richard 2026)Experiment · Mixed evidenceChanging the phase between the eyes changed the response, including a smaller response at the flicker rate itself for on-off flicker shown out of step.
• Visual cortex responses to 1–100 Hz flicker (Herrmann 2001)Experiment · Strong designFlicker drove a steady-state response in visual cortex that tracked the stimulus frequency up to about 90 Hz, with resonance peaks near 10, 20, 40 and 80 Hz. No cognitive, mood or clinical outcome was measured.
• What people see at each flicker rate (Allefeld 2011)Experiment · Mixed evidencePartly. For a given person a percept tends to come back at the same rate, and whole percepts are tied to rate more tightly than single features are. But rate alone does not determine what is seen.
• Which colour, brightness and rate of flicker drive the brain hardest (Lee 2021)Experiment · Mixed evidenceRed light entrained gamma most effectively, then white. Brighter was stronger, but the brightest setting caused more side effects. Flicker at 34–38 Hz was stronger and spread further than 40–50 Hz.
• White, red or green: which flicker the brain follows best (Duart 2020)Experiment · Mixed evidenceYes. The middle rate (12 Hz) gave the clearest response, then 5 Hz, then 30 Hz. At 12 Hz white did as well as red.

Breathing and heart rate variability

Summary: Resonance breathing

• Heart rate variability biofeedback methods: a review (Lalanza 2023)Systematic review · BackgroundAcross 143 studies, three breathing protocols are in use and there is no methodological consensus. Almost two-thirds of studies under-report the protocol.
• Heart rate variability biofeedback: a meta-analysis (Lehrer 2020)Meta-analysis · Mixed evidenceA meta-analysis of 58 randomized controlled studies found a significant small-to-moderate effect favouring heart rate variability biofeedback, not different from other effective treatments. Effects were larger against inactive than active controls, though significant for both.

Noise

Summary: White, pink and brown noise: what's the difference?

• Noise as a sleep aid: a systematic review (Riedy 2021)Systematic review · Mixed evidenceLater summaries of this 38-study review state that, by GRADE criteria, the quality of evidence for broadband noise as a sleep aid was low.

Safety

• Flashing lights and seizure risk: expert consensus (Fisher 2005)Consensus statement · Strong designFlashes at about 3–65 Hz are hazardous, with 15–25 Hz the most provocative. Accessibility standards built on this consensus allow no more than 3 general flashes and no more than 3 red flashes in any 1-second window.
• How safe is strobe light outside the lab (Schwartzman 2026)Safety survey and guidance · BackgroundThe main risk is a seizure in a susceptible person. The labs reported 20 non-serious events and no seizures; the two providers reported 14 notable events across roughly 4.2 million sessions, but reporting was passive, so no rate can be calculated.

Background

• Auditory beat stimulation, cognition and mood: a review (Chaieb 2015)Review · Mixed evidenceThe review of 30 articles finds that results are frequently contradictory, largely because of heterogeneous stimulation parameters, and calls for standardization.
• Binaural beats and the brain: a narrative review (Mirmohamadi 2024)Review · Mixed evidenceThe review concludes that binaural beats are a plausible non-invasive adjunct, but that effects depend heavily on stimulation parameters (duration, frequency, carrier) and personality traits, and that more rigorous research is needed.
• Cymatics: the physics of sound patterns, not a therapyBackground · No support foundCymatic patterns are textbook acoustic physics: standing waves, resonance and nodal geometry. No peer-reviewed study tests a cymatic image as a causal biological agent.