Gamma Entrainment

Gamma entrainment is the driving of ~40 Hz neural oscillation by rhythmic sensory stimulus — light, sound, or both. The research line is usually called GENUS (Gamma ENtrainment Using Sensory stimulus), after the MIT protocol.

Wikipedia has an article on gamma waves and one on brainwave entrainment; it has none on GENUS, on 40 Hz light therapy, or on invisible spectral flicker. This page exists to cover the applied literature, including the parts of it that did not replicate.

The four delivery methods

Luminance flicker — the strobe. Most studied, least comfortable.
Invisible Spectral Flicker (ISF) — modulate spectral composition at constant perceived luminance. Same 40 Hz SSVEP, materially better comfort scores.
Auditory — 40 Hz clicks or an amplitude-modulated carrier. Eyes closed, no photosensitivity exposure at any frequency. See Auditory Steady State Response.
Combined audiovisual — what the pivotal clinical trials actually run.

The evidence, stated with its failures

The positive line. Phase 2A over three months in mild probable Alzheimer's reported lesser ventricular dilation and hippocampal atrophy, increased default-mode connectivity, and better face–name delayed recall against control. 40 Hz auditory stimuli produce the largest EEG response and raise regional cerebral blood flow in healthy participants.

The pivotal trial has not read out. HOPE (NCT05637801): status ACTIVE_NOT_RECRUITING, enrolment 670, primary completion June 2026 estimated, no results posted. Anything asserting clinical efficacy today is asserting it ahead of the trial.

The flagship preclinical result has a prominent failed replication. Soula M, …, Buzsáki G, Nature Neuroscience 2023;26(4):570–578 — in APP/PS1 and 5xFAD mice, 40 Hz flicker did not engage native gamma oscillations, with no significant change in plaque burden, microglial morphology or Aβ levels, and avoidance behaviour. A page that cites the original and omits this is not reporting the field.

The systematic review agrees. Bolland E et al., npj Aging 2025, 62 studies: animal-model findings on amyloid clearance and neuroinflammation "were not consistently replicated in human studies", and neurophysiological responses did not reliably predict cognitive or biological outcomes.

Tolerability is the better-supported claim. Sato S et al., Healthcare 2025, n=28: 96.4% adherence, six mild-to-moderate device-related adverse events, auditory channel.

The frequency-safety inversion

The intuition that a higher frequency is more dangerous is backwards, and this is the most important practical fact in the field.

Fisher RS et al., Epilepsia 2005;46(9):1426–1441, verbatim: "Frequencies of 15–25 Hz are most provocative, but the range is 1–65 Hz." So 40 Hz sits inside the provocative range, not above it — there is no frequency-based safety argument for 40 Hz visual flicker — but the 13–26 Hz band is the dangerous one. A protocol that silently converts a 40 Hz request into ~20 Hz has moved the subject from the safer end of the range to the peak. See Display Refresh and Flicker Delivery.

Harding G et al., Epilepsia 2005;46(9):1423–1425: "A flash is a potential hazard if it has luminance ≥20 cd/m², occurs at a frequency of ≥3 Hz."

Photosensitivity affects 0.3–3% of the general population and 2–14% of people with epilepsy, with peak incidence between ages 7 and 19.

Duration, which the standards address and the devices do not

ITU-R BT.1702-3 (11/2023), verbatim: "a sequence of flashing images lasting more than five seconds might constitute a risk even when it complies with the guidelines above." A GENUS session is 12 to 60 minutes — 150 to 700 times that. Its Annex 5 names three aggravating factors: a dark room, a bright or large screen, and close viewing.

WCAG 2.3.1 (Level A) permits nothing flashing more than three times per second unless it is below the flash thresholds or occupies no more than 25% of any 10-degree visual field. A full-screen 40 Hz strobe fails both limbs. No configuration rescues it; only a small flashing region conforms, which is a materially weaker stimulus than the protocol calls for. That tension should be stated to users rather than engineered around.

Contraindications

Epilepsy or family history of seizures; migraine with aura; implanted electrical devices (pacemaker, ICD, DBS, VNS); pregnancy; sleep deprivation; alcohol or benzodiazepine withdrawal; medication that lowers seizure threshold; recent head injury or stroke; age under 20. Saturated red is a hazard channel independent of luminance in both WCAG and BT.1702-3 and should never be used.

Eyes closed is not safer. BT.1702-3's normative remedy for the most susceptible is a long-wavelength-red-reflecting filter; closed eyelids transmit predominantly long-wavelength light, so an eyes-closed bright strobe is physically closer to a red strobe at the retina. Presented here as precaution grounded in the red-flash rule and eyelid transmission, not as an established finding.