40 Hz light stimulation: failed replications in mice (Soula 2023)

Two independent labs failed to replicate the core mouse findings: 40 Hz flicker did not entrain native gamma oscillations and gave no significant amyloid change. A human pilot (n = 15) was a feasibility study in which standard cognitive scales were flat.

A plain-language summary from binaural's reading notes, edited by Jude Mc. Last updated 10 November 2026. Not medical advice.

The question: Can independent labs reproduce the finding that 40 Hz light flicker entrains gamma activity and reduces amyloid in Alzheimer's-model mice?

What it found

• In Soula 2023, 40 Hz flicker did not entrain native gamma in visual or entorhinal cortex or hippocampus, and gave no significant amyloid or microglial change.
• The mice showed behavioural aversion and stress in response to the flicker.
• In Yang and Lai 2023, 5 weeks of flicker at 24, 40 and 80 Hz gave no amyloid reduction (all p > 0.69).
• A human pilot (Chan 2022, n = 15) was a feasibility study not powered for efficacy: there were some MRI and atrophy signals, but standard cognitive scales (MMSE, ADAS-Cog) were flat.
• Lee 2021 found that 40 Hz may not be the optimal frequency for human entrainment; 34–38 Hz drove stronger gamma.

Limits

• The human pilot was small and not powered for efficacy.
• Positive Phase 2 results reported by a company developing a device are post-hoc and not fully peer-reviewed.
• Whether an ongoing pivotal trial (HOPE) changes the picture was unknown as of 2026.

Where this sits

More in the library on light, flicker and 40 hz. For the wider picture, read 40 Hz, gamma and the brain.

This is our plain-language reading of the paper. It is not medical advice, and one study rarely settles a question.