← Research library
ResearchMixed evidence

Tested on two kinds of memory at once, 10 Hz beats helped one and hurt the other

Rakhshan and colleagues (2022) · BioMed Research International · Summary by Sentiosonics research agent
Year
2022
Stimulus
Alpha
Design
Within-subject five-arm randomized trial with pure-tone and silence controls
Sample
n = 31

Summary

Most working memory studies of binaural beats test one kind of memory and report one answer. This one tested visuospatial and auditory-verbal working memory in the same task at the same moment, and the two came apart: the 10 Hz condition improved the visual side and made the verbal side worse.

Thirty-one healthy right-handed adults each completed five 8-minute blocks in a randomized, concealed order: 10 Hz, 16 Hz and 40 Hz binaural beats, a 240 Hz pure tone, and silence. During every block they played a dual 2-back task, which asks you to track a position on a grid and a spoken letter at once and report when either repeats what came two steps earlier. Because both streams run together, one session yields a visuospatial score and an auditory-verbal score from the same eight minutes. Alongside working memory capacity the authors took signal detection measures and response-time variability, and they recorded the volume each participant chose rather than fixing it.

In the visuospatial modality, the 10 Hz alpha condition shortened response times, reduced the trial-to-trial variability of those response times, and slowed the decline in capacity, sensitivity and hit rate that otherwise set in across a block. In the auditory-verbal modality, the same condition reduced the hit rate and sensitivity, along with the false alarm rate. The authors’ own conclusion is that eight minutes of alpha entrainment “may have a few, slight enhancing effects within the visuospatial modality, but not in both modalities combined”. The 16 Hz and 40 Hz conditions did not produce the enhancing pattern.

We grade this mixed, and the reason is the result rather than the method. A benefit in one modality paired with a decrement in another, from a single intervention in a single session, is not a finding that supports “binaural beats improve working memory”, and it is not a null either. The most defensible reading is that a 10 Hz beat competes with verbal material for the same auditory channel while leaving visual material alone, which would predict exactly this split. The paper does not test that explanation, so it stays a hypothesis.

The other result worth taking from it has nothing to do with beats. Short-term training across the five blocks reliably increased hit rate, working memory and sensitivity and reduced false alarms and response time. Practice on the task moved the numbers more dependably than any sound did, which is a problem for every study of this kind that runs its conditions back to back.

Limitations. Thirty-one people, and a design that puts all five conditions in one session with 3-minute rests and no washout, so the practice effect the paper itself documents is entangled with the condition order. The study is explicitly unblinded. The outcomes run into the dozens across two modalities and five arms, which is a large number of tests for 31 participants, and the paper reports many of them as significant at p ≤ 0.05 without a correction that covers the full set, so individual findings here are fragile even where the overall pattern is interesting. Single session only, so nothing is known about repeated use. One session also means no measure of whether the visuospatial gain lasts past the eight minutes in which it appeared.

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 designFive-arm within-subject trial with the order of conditions randomized and concealed. Crossover with 3-minute rests and no washout period, and the paper has to model a substantial practice effect across blocks, which is the weakness of running all five arms in one session.
Sample size31 adults, each completing all five conditions, for 155 sessions in total. The sample size was set from earlier binaural beat studies rather than from a power calculation for these outcomes.
Control conditionA 240 Hz pure tone matched to the carrier frequency, plus a silence arm. Sound volume was also measured rather than fixed, which most studies in this area skip.
BlindingThe paper states plainly that it was not blinded. The researcher played the audio and the participants heard it. The authors note participants could not tell the similar sound conditions apart, silence excepted, but decline to call the study blinded. Outcomes are computer-scored, which is why this is not scored lower.
Outcome measureA dual 2-back task in Matlab giving working memory capacity plus signal detection measures (hit rate, false alarm rate, sensitivity, response bias) and response-time variability, in both modalities simultaneously.
ReplicationSingle study. No replication found of a design that assesses both modalities at once, which is the feature the authors claim as new.
Conflicts of interestDoctoral thesis work at an academic institute, no conflict of interest declared, nothing commercial riding on the result.
Peer review & venuePeer-reviewed and PubMed-indexed (PMID 35655482), but BioMed Research International was delisted from Web of Science during Clarivate's review of Hindawi titles after paper-mill activity in the special-issues programme, and reindexed in 2024. The paper itself shows no sign of that problem; the venue still warrants less weight than a journal with an unbroken record.
Overall grade · 3.6 / 5Mixed evidence

Graded by our research agent on Nov 2, 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.