Low-frequency flicker carries a risk of triggering photosensitive reactions

What it is

Within a flicker-frequency band of roughly 3 to 60 hertz (with the highest-risk peak generally placed around 15 to 20 hertz), strong, large-area alternation of brightness or color carries a risk of triggering a photosensitive epileptic seizure. This is an entirely different order of problem from "flicker is distracting" or "flicker looks glaring" — the latter is an experience-optimization matter, while this is a safety issue that can cause a seizure requiring medical intervention.

Worth separating out: this risk does not require the flicker to be subjectively perceived as "flickering" or "glaring" at all — even if the viewer feels no discomfort whatsoever, the risk is still present as long as frequency, area, and contrast fall in the risk band.

Why it happens

Populations of cortical neurons have a tendency toward abnormal synchronized firing in response to periodic, high-contrast visual input, and this is especially true of visual cortex, a region extremely sensitive to high-contrast temporal change. When flicker frequency falls in a window that readily drives cortical neurons rhythmically, large populations of neurons get driven repeatedly and strongly by the same rhythm — in a minority of susceptible individuals (both those formally diagnosed with photosensitive epilepsy and some who carry the susceptibility undiagnosed), this can trigger abnormal synchronized discharge and precipitate a seizure.

Risk rises with the contrast of the flicker and the visual-field area it covers — which is why an intense, full-screen flicker is far more dangerous than a small flickering icon in a corner, even at the exact same frequency. Rapid alternation involving highly saturated red-to-deep-red is considered higher risk still, likely tied to the strong response of certain cone types to that wavelength range.

Studying it

• Standardized screening: clinical and engineering practice both use automated detection pipelines that score risk from the magnitude of brightness change, frequency, and the visual-field area covered by flicker, analyzing video or dynamic content frame by frame to flag segments that exceed safety thresholds.
• Common independent variables: flicker frequency, the proportion of the visual field the flickering area covers, the magnitude of luminance contrast, whether saturated red alternation is involved.
• Use in content review: broadcast and accessibility standards both build automated flicker-risk screening into the pre-release pipeline, used to catch high-risk material rather than relying on a person's subjective "it looks fine" judgment.
• Methodological caution: these risk thresholds come from conservative estimates in epidemiological and clinical observation; individual susceptibility varies widely. The same content being risk-free for most people does not mean "no flagged issue equals absolutely safe for everyone" — the safety threshold is itself a conservative boundary set to protect susceptible individuals, not "the comfort boundary for most people."

Where it stops holding

• People formally diagnosed with photosensitive epilepsy carry markedly higher risk, but a substantial share of people who carry the susceptibility have never had a seizure and don't know they're at risk — "no user has reported a problem" is not evidence that flickering content is safe.
• Risk correlates strongly with the area flicker covers: small, non-full-screen local flicker carries far lower risk than large-area or full-screen flicker, so the same frequency cannot be judged the same way regardless of coverage.
• This frequency band is a risk window, not proof of zero risk outside it — flicker below or above the band carries markedly lower risk, but edge cases still rest on the judgment of professional screening tools; "avoid the peak frequency band" is not a full substitute for that judgment.

Applying it

• Any interface effect involving brightness or color alternation (loading animations, alert flashes, transition effects) should be run through a standardized frame-by-frame flicker-risk screening pipeline before release, rather than relying on a designer's or engineer's subjective "it looks fine."
• Prefer avoiding full-screen or large-area high-contrast flicker; if urgency genuinely needs to be conveyed, use a non-flickering alternative — a color shift, a static emphasis color, haptic feedback — to communicate the same urgency.
• Be especially cautious with fast alternation involving highly saturated reds — even outside the general safety frequency band, very high-saturation red alternation is worth a separate review pass.
• How to check: run an automated flicker-risk screening pass on the final cut or final interactive effect, recording whether flicker frequency, coverage area, and contrast magnitude all fall within safe thresholds — any one metric failing calls for a redesign, not compensating adjustments to the other two to "pass."

Related

• Same group: A1.08.1 Flicker above a critical rate is perceived as steady · A1.08.2 Critical frequency shifts with luminance and retinal location
• Nearby: A1.02.4 Peripheral flicker forcibly captures attention · J4.10 Flicker and photosensitivity
• Search terms: photosensitive epilepsy · general flash threshold · red flash · flicker risk