Beatline
The research
What binaural beats actually do, and what they probably do not
17 August 2026
Play a 200 Hz tone into your left ear and a 203 Hz tone into your right. You will hear a third thing: a slow pulse, three times a second, that is not present in either ear.
The two faint traces above are the tones themselves, one per ear. The solid line is what happens when they are summed, and the dashed shape riding over it is the envelope: the slow swell that your ears never actually receive.
This is also why headphones are not a recommendation. On a speaker the two channels mix in the air before they reach either ear, and what arrives is an ordinary acoustic beat, the same wobble you hear when two guitar strings are slightly out of tune. Related phenomenon, different mechanism, and not the one the research is about.
The claim that goes further
The popular version does not stop at perception. It says the brain, having constructed a 3 Hz beat, will then synchronise its own electrical activity to 3 Hz, and that because 3 Hz falls in the delta range associated with deep sleep, you will become sleepy.
This is called frequency-following, or entrainment. It is the mechanism behind almost every binaural beats app on the store, including this one.
What the evidence looks like
Ingendoh, Posny and Heine 2023 reviewed fourteen studies of EEG entrainment in PLOS ONE in 2023. Five supported the hypothesis. Eight contradicted it. One was mixed.
Not "found weak support" - contradicted. A beat at a given rate did not reliably drive cortical activity to that rate. If you have read that binaural beats "tune your brainwaves to the frequency you choose", that is the largest review of the question, and it does not support the claim.
So the mechanism is in trouble. The outcomes are a different story.
The outcomes hold up better than the mechanism
Garcia-Argibay, Santed and Reales 2019 pooled 22 studies and 35 effect sizes in Psychological Research in 2019. The result was g = 0.45, a medium-sized effect on anxiety, attention and memory, and consistent across those outcomes.
Baseanu 2024, reviewing twelve studies on anxiety and depression in Applied Sciences in 2024, found binaural beats did better than silence and better than noise-cancelling headphones. That second comparison matters more than the first, because it is the one that starts to rule out "any quiet is good quiet".
That is a genuinely reasonable body of evidence. It is also where most apps stop reading, so here is the rest.
Two things that should make you cautious
The trials disagree with each other enormously. Xiong 2025 reviewed fourteen trials with 1,047 participants in Complementary Therapies in Medicine in 2025 and found a large effect on anxiety, SMD −1.38. They also reported I² = 91.6%.
I² measures how much of the variation between studies is real disagreement rather than chance. Above 75% is conventionally called considerable. At 91.6% the studies are barely measuring the same thing, and an average across them tells you much less than its size suggests. We quote that number rather than the effect size alone, because quoting −1.38 on its own would be the more flattering and less honest choice.
And at least one large study found binaural beats made a task harder. Not neutral. Worse.
Three rates have a study behind that exact rate
Most of the frequency claims you will encounter are extrapolated from band names. Three in this app are not:
- 3 Hz. Jirakittayakorn and Wongsawat 2018, Frontiers in Human Neuroscience: more delta activity and longer stage N3 sleep. A small experiment, and not an insomnia trial.
- 6 Hz. Jirakittayakorn and Wongsawat 2017, Frontiers in Neuroscience: more theta, including the frontal-midline theta associated with focused attention. An EEG resemblance, which is not the same as the state itself.
- 15 Hz. Beauchene 2017, Journal of Neural Engineering: better verbal working memory accuracy, with changed cortical connectivity. Small, and cognitive findings elsewhere are inconsistent.
Each is one experiment. Three experiments do not make a dosage chart, and the other rates in this app or any other are chosen because they sit inside a named band, not because that specific rate was tested.
Because "the mechanism is contested" and "the experience is worthless" are different statements, and only the first one is supported.
The outcome studies are real. The entrainment story attached to them is probably wrong, or at least not reproducible, and those two facts can sit together. A steady tone is a good thing to put between yourself and a noisy street, and the absence of lyrics, structure or surprise is most of why ambient sound helps people concentrate. None of that requires entrainment to be true.
Every study named here is on the research page with its sample size, its effect size and what undercuts it, and the app shows the same grades next to each sound.
The practical version
If you want to try binaural beats, the honest advice is short.
Use headphones, or the effect does not exist. Pick by what you are doing rather than by a number, because outside those three tested rates the band names are a map of a contested theory, not a dosage. Give it twenty minutes: the sleep and attention experiments ran over sessions of that order, and nothing in the literature suggests anything happens in two.
And if a plain recording of rain works better for you, use that instead. The masking research is the strongest thing in this app: Ding 2025 pooled twelve randomised trials with 1,301 participants in Sleep Medicine and found sleep quality improved. Nobody has to believe anything contested for that one to work.