What the research supports,
and what it does not

Focused attention enhances processing of the attended sound as early as ~50 ms after onset in primary auditory cortex, with robust modulation at ~100 ms in specialised non-primary areas. The effect appears within seconds of shifting attentional focus.

This is the single best-evidenced thing Verbenix is asking a user to do.1

Six 30-minute sessions over three weeks cut localisation error substantially: azimuth 28° → 10°, elevation 15° → 3°, distance 574 mm → 152 mm. Post-training fMRI showed increased recruitment of superior frontal, superior/middle temporal and parietal regions — higher-order attention and spatial areas — with reduced reliance on basic sensory processing.

Caveat that matters: n = 6. Treat the direction as credible and the magnitude as unproven.2

Congruency training between what you see and what you hear enhanced dorsolateral prefrontal activity and improved electrophysiological correlates of auditory selective spatial attention.

This is the direct justification for Verbenix showing the room visually while the sound moves — the picture is not decoration, it is the congruency signal.3

A meta-analysis across 22 studies and 35 effect sizes found a medium, significant overall effect on memory, attention, anxiety and analgesia. But a 2023 study of ~920 adults doing abstract reasoning, concentration and working-memory tasks at home found binaural beats worsened performance versus baseline — regardless of the frequency assigned or what participants were told to expect.

Anxiety is the more defensible endpoint: a 2025 systematic review of 14 perioperative trials (1,047 participants) found reduced anxiety versus blank audio. Verbenix does not use binaural beats.4,5,6,7

The GENUS line of work is real and interesting — an hour a day of 40 Hz audiovisual stimulation for two years produced no adverse effects and slowed MMSE decline by roughly 0.8–1.2 points a year versus matched controls in late-onset patients. But the open-label extension had five participants, and early-onset patients showed no clear benefit.

It is a clinical intervention under investigation for Alzheimer’s disease. It is not evidence that a healthy person meditating gets anything from a 40 Hz tone, and Verbenix does not imply otherwise.8,9

The "ancient Solfeggio" story is a 20th-century invention; the historical solfège system names scale relationships, not hertz values. A 528 Hz sound wave is metres long and cannot interact with a nanometre-scale molecule — there is no physical mechanism to argue about.

The most rigorous test of tuning itself is a 43-patient randomised crossover comparing a body monochord tuned to 432 Hz versus 443 Hz. Both improved emotional well-being, fatigue, anxiety and stress; cardiovascular differences between the two tunings were mostly not significant. The authors themselves cannot say whether anything came from 432 Hz specifically.10

Photosensitive epilepsy — the well-known flashing-light trigger, affecting about 3% of people with epilepsy — is a visual phenomenon. The auditory analogue, musicogenic epilepsy, is a rare reflex epilepsy estimated at roughly 1 in 10 million, and is triggered by the emotional and memory content of specific music via temporal-lobe pathways, not by tone frequency.

No controlled binaural-beat study has induced a seizure. Neurologists still advise people with epilepsy to check with their clinician, which costs us one line of copy.11,12

The WHO–ITU safe-listening standard sets an adult allowance of 80 dB(A) for 40 hours per week (75 dB for children). The exchange is brutal: at 90 dB the weekly allowance collapses to four hours. A 45-minute daily session is about 5 hours a week — safe at 80 dB, not safe at 90.

Broadband noise makes this worse in a specific way: it has no melodic contour, so people calibrate it by "presence" rather than loudness and tend to push it higher than they would music.13,14

Verbenix ships with an output ceiling, a session-time read-out, a weekly listening total kept on your device, and a one-tap stop. Loudness is a safety control, not a feature — so it gets a calibrated scale, not a 0–100 slider with the top end unmarked.