← Research library
ResearchNot supported

Five beat frequencies, two ways of delivering them, and no change in the EEG at all

López-Caballero and Escera (2017) · Frontiers in Human Neuroscience · Summary by Sentiosonics research agent
Year
2017
Stimulus
Binaural beats
Design
Within-subject crossover, ten randomized blocks, pink-noise control epochs
Sample
n = 14

Summary

The premise behind every brainwave-entrainment product is that playing a 10 Hz beat raises 10 Hz activity in the brain. This study tested that premise across five frequencies and two delivery methods, with the electrodes on, and found nothing. The title says so without hedging: a failure to enhance EEG power.

Fourteen people sat in a shielded, sound-attenuated chamber watching a subtitled silent film while ten blocks of audio played. Each block ran six minutes: 90 seconds of pink noise, then 180 seconds of beats delivered 10 dB above that noise, then 90 more seconds of pink noise. Five beat frequencies were used, spanning the bands the industry sells, at 4.53 Hz for theta, 8.97 Hz for alpha, 17.93 Hz for beta, 34.49 Hz for gamma and 57.3 Hz for upper gamma. Every frequency was played twice, once as a binaural beat with one tone to each ear, and once as an ordinary acoustic beat where the two tones are mixed before they reach the listener. Block order was randomized for each participant.

For each of the ten blocks the authors compared EEG spectral power in the band being stimulated against the pink-noise baseline that preceded it. No frequency produced a significant increase, with binaural or with acoustic beats. A follow-up analysis of scalp topographies, looking for an effect that might be regional rather than global, also came back negative. Heart rate and skin conductance response did not change for any beat frequency either, which rules out the gentler claim that the beats were doing something arousing even if they were not entraining anything.

The inclusion of acoustic beats is what makes this more than one more null. An acoustic beat physically exists in the air, so the ear has an actual amplitude modulation to follow, and it is the easier case. Not finding an effect there as well means the result is not about the fragility of the binaural illusion. Across the whole design, a sustained tone pair at the volume people use did not measurably pull cortical rhythms toward its beat rate.

We grade this unsupported, by the rubric’s result override: the paper tested the claim and did not find the effect. The score is 3.7, which reflects a well-instrumented, randomized, properly controlled experiment, and the grade reflects what it found rather than how well it was built.

Limitations. Fourteen people, and the authors are the first to say that a sample this size can hide a real but small effect. A null is not proof of absence, and this one cannot rule out effects that need longer exposure than three minutes, a different carrier, a task instead of a film, or responders who differ from the group. Watching a subtitled movie occupies attention, which may suppress a mechanism that depends on attending to the sound. The beat epochs were 10 dB louder than the baseline epochs, so loudness is not held constant. And the measure is scalp EEG power, which later methods work argues is the wrong instrument for deciding whether entrainment in the strong sense happened at all, a criticism that cuts against positive findings more than against this one.

Our analysis

How we graded this study

Each criterion is scored 1–5. Study design and replication count double in the overall score (how we grade). Graded from: full text read.

Study designWithin-subject crossover. Ten blocks, one binaural and one acoustic block for each of five beat frequencies, with block order randomized per participant and the constraint that the same frequency never ran twice in a row. Each beat epoch is bracketed by pink-noise epochs rather than compared against a separate control arm, and there is no washout beyond 45 to 75 seconds between blocks.
Sample sizeFourteen participants, five male, aged 20 to 31. The authors raise the sample size themselves as the most likely explanation for null results and say it may be too small to reach significance.
Control conditionA continuous pink-noise background at 65 dB SPL for 90 seconds before and 90 seconds after each 180-second beat epoch, so the comparison is a sustained sound with no beat in it. Not a full active sham, because the beats were delivered 10 dB above the noise floor, making the beat epochs louder as well as beat-carrying.
BlindingNo blinding, and none is claimed. Every outcome is instrument-recorded rather than reported by the participant, which is why this is scored 2 rather than 1. Participants watched a subtitled silent film throughout, which holds attention roughly constant across conditions.
Outcome measureEEG spectral power within each stimulated band, analysed per epoch, plus scalp topographies of spectral power, heart rate and skin conductance response. All objective and instrument-measured.
ReplicationThe null is the better-replicated side of this question. Goodin and colleagues 2012 found the same absence of an EEG power change against a white-noise control in 31 people, and this paper has been cited about 87 times, largely as the reference null. Partial rather than full, because individual positive studies in narrower conditions have not all been directly re-tested.
Conflicts of interestUniversity group (Brainlab, Barcelona). The authors declare the research was conducted in the absence of any commercial or financial relationships that could be construed as a conflict of interest.
Peer review & venuePeer-reviewed and PubMed-indexed (PMID 29187819).
Overall grade · 3.7 / 5Not supported

Graded by our research agent on Nov 16, 2026.

Get updates from the field

A monthly letter: new research, graded, and the claims worth checking. No hype.

Monthly, one click to unsubscribe. See our privacy policy.