Binaural Beats Need Headphones to Exist: The Delivery Problem Behind Two Decades of Mixed Results

Somewhere right now a person is playing a binaural beats track through a phone speaker on a desk, waiting to feel calmer. Nothing is going to happen, and not because the idea is nonsense. It is because a binaural beat is not on the recording. It is manufactured inside the listener's head from two separate tones, one delivered to each ear, and a speaker sitting on a desk mixes those tones into the air before either ear receives them in isolation. The effect is destroyed in transit.

That failure mode is unusually common, almost never mentioned in the guides, and it goes a long way towards explaining why a field with a plausible mechanism keeps producing such uneven results.

What a binaural beat is, and where it is actually produced

Brainwave entrainment is the use of a rhythmic stimulus, whether sound, light or both, to nudge the brain's electrical rhythms towards a chosen frequency. The underlying phenomenon, the frequency-following response, has been documented for decades: present a steady rhythm and cortical activity tends to align with it. Entrainment is the deliberate application of that tendency, most often aiming at alpha rhythms for calm, theta for deep relaxation, delta for sleep or gamma around 40 Hz for alertness.

A binaural beat is one delivery route among several, and the most fragile. Play 200 Hz into the left ear and 210 Hz into the right, and neither ear receives a 10 Hz sound, because 10 Hz is far below the threshold of human hearing in the first place. The perceived pulsing is constructed in the brainstem, in the superior olivary complex, where input from the two ears is first compared. The beat exists only as a computed difference. This is elegant, and it is also the entire problem: the illusion has a hard hardware requirement, and that requirement is routinely unmet.

Do binaural beats work, and how strong is the evidence

Binaural beats do something measurable, but less than the marketing implies and less consistently than a scan of the abstracts suggests. A 2019 meta-analysis published in Psychological Research pooled the experimental literature on cognition, anxiety and pain perception and found effects that were real but modest, with wide variation between studies. Later systematic reviews have reached similar conclusions: a small-to-moderate signal, considerable heterogeneity, and a body of work built mostly on small samples and short exposures.

Anxiety is the outcome with the most consistent support. Several controlled studies report reduced state anxiety after binaural beat exposure, including in pre-operative and pre-procedural settings where the anxiety being measured is acute and situational. The effect sizes are moderate and the studies are small, so the honest summary is that binaural beats look like a reasonable low-risk adjunct for situational anxiety, and are not a demonstrated intervention for a diagnosed anxiety disorder. Anyone reading a stronger claim than that is reading marketing.

Brainwave entrainment as a whole has a better record than binaural beats considered alone, largely because it includes stimulus types that survive real-world delivery. A 2025 University of Milan review in Brain Sciences examined more than fifty years of audio-visual entrainment research and documented positive results across depression, insomnia, anxiety and attention difficulties. Audio-visual entrainment combines rhythmic sound with pulsed light, which produces a considerably stronger stimulus than a pair of tones through earbuds, and the clinical literature reflects that difference.

The wider picture for consumer digital tools is more settled still. A 2024 meta-analysis spanning 28 systematic reviews and 118,970 participants found significant improvement from digital therapy for insomnia, depression and anxiety. That is a statement about digital tools as a category, not about tones specifically, and conflating the two is one of the more common sleights of hand in this market.

The delivery failure the studies rarely control for

Laboratory studies use calibrated stereo headphones. Real listeners frequently do not, and the gap between those two conditions is where most of the disappointment lives. A binaural beat is silently destroyed by every one of the following, none of which produces any warning that the effect has been lost:

Any speaker. Phone, laptop, smart speaker, car stereo. The two tones combine acoustically before reaching the ears, and no interaural difference remains for the brainstem to compute.
A single earbud. One ear receives both channels summed, the other receives nothing. This is a mono signal wearing a stereo file's clothing.
Bone conduction and open-ear designs. Vibration crosses the skull, so channel separation collapses even though two transducers are present.
Poor channel separation generally. Cheap wireless earbuds, some audio processing modes and any device with mono downmixing enabled will flatten a stereo track without announcing it.
Ordinary asymmetric hearing loss. If one ear is significantly less sensitive at the carrier frequency, the two inputs no longer arrive at comparable strength.

None of these produces an error message. The track plays, it sounds like something is happening, and the mechanism the whole thing depends on has been removed. When a person reports that binaural beats did nothing for them, checking the delivery route is a more useful first response than concluding the evidence is fake.

Isochronic tones, and what layering the two changes

The practical difference between binaural beats and isochronic tones is the delivery route, and it favours isochronic tones in almost every everyday situation. An isochronic tone is a single tone switched on and off at the target rhythm. The pulse is physically present in the audio, so it survives a phone speaker, a single earbud, bone conduction and mono playback intact. There is no illusion to break because nothing has to be computed from a difference between the ears. The trade-off is that isochronic tones are more obtrusive to listen to, a clicking pulse rather than a soft wavering, which is one reason the smoother binaural version dominates the market despite being the more failure-prone of the two.

Some tools address this by running both at once rather than choosing. 6th Mind, a free audio-visual entrainment app developed by a psychiatrist and psychologist team at a practice office in Sofia, layers isochronic tones as the base signal, adds binaural beats on top for anyone wearing stereo headphones, and can add stroboscopic pulses from the phone's camera flashlight as a third pathway. The architecture means the session degrades gracefully rather than silently: without headphones the binaural layer is lost and the isochronic layer still carries the rhythm. The frequency sequences themselves, along with the ramp timings between steps, are derived from data on more than 800 documented clinical entrainment sessions rather than chosen for how they sound.

That last point matters more than it appears to. Most consumer tracks hold a single fixed frequency for the duration. Clinical protocols typically walk through a sequence, with gradual transitions between steps, on the reasoning that a brain already in an agitated beta state does not follow a stimulus dropped abruptly into theta. Whether a fixed tone or a sequence performs better in healthy adults has not been settled by controlled comparison, and anyone claiming otherwise is ahead of the evidence.

Are binaural beats safe, what the side effects are, and when professional care matters

Binaural beats are safe for the large majority of listeners, and the reported side effects are mild and uncommon: headache, mild dizziness, ear fatigue or irritability, most often at high volume or during long sessions. The most concrete risk is the least exotic one, which is hearing damage from listening loudly through headphones for extended periods. Keeping the volume low is genuinely the main safety instruction, and the audio should be comfortably quiet rather than immersive.

Brainwave entrainment is safe under the same general conditions with one significant exception, and it belongs to the light component rather than the sound. Stroboscopic light can provoke seizures in people with photosensitive epilepsy, a condition affecting a small minority of people with epilepsy but a serious matter for those it does affect. Anyone with epilepsy, any history of seizures, or a family history of either should avoid flashing-light sessions entirely or use audio-only versions, and any responsible tool shows that warning before first use and allows the light to be switched off mid-session. People recovering from concussion, and those on medication that lowers the seizure threshold, should take the same precaution.

The limitations worth stating plainly are broader than the safety profile. Entrainment audio is a complementary tool that supports conventional care, not a replacement for therapy or medication, and it has not been shown to resolve a diagnosed condition on its own. Sessions do nothing for the causes of most persistent anxiety, low mood or insomnia, which sit in circumstances, health and history rather than in cortical rhythms. Anxiety that interferes with work or relationships, insomnia lasting more than a few weeks, low mood accompanied by loss of interest or hopelessness, and any thoughts of self-harm are all reasons to speak to a doctor or a mental health professional. A 2024 randomised trial of stepped care found that starting with a digital tool and escalating to a therapist when needed outperformed direct referral, which is a good description of where audio of this kind belongs: at the start of a path that has a clinician further along it, not instead of one.