The Hidden Hazard in Your Home Lighting

What Is Light Flicker, Really?

Light flicker, technically called Temporal Light Modulation (TLM), is the rapid fluctuation of light intensity over time. Despite what the name suggests, you don't always see it—but your brain is processing it regardless.

The Four Faces of Flicker

🌀 3. Stroboscopic Effects (80 Hz - 2-10 kHz)

At these higher frequencies, flicker creates perceptual distortions: moving objects appear as a series of still images (stroboscopic effect), or you see multiple "phantom" images during rapid eye movements. Most can perceive these effects at up to 2 kHz, but there are some reports of people seeing them even above 10 kHz

The Three Numbers That Matter

To understand whether a light is likely to be comfortable or harmful, three engineering parameters matter most:

1. Frequency (Hz) – How Fast It Flickers

Frequency is how many times per second the light cycles on and off. Most AC-powered lights flicker at 100–120 Hz because the light output follows the rectified power waveform.

Rule of thumb: Higher is better, but even into the kHz range you can still get phantom arrays if other parameters are bad.

2. Modulation Depth (%) – How Deep the Flicker Goes

Modulation depth (often reported as Percent Flicker) measures how much the brightness varies in each cycle.

• 0% modulation = perfectly constant light (ideal).
• 100% modulation = light goes fully on → fully off (worst case).

Aim for: ≤5–10% modulation whenever possible. Above ~30% is a red flag, especially below 1 kHz.

3. Duty Cycle & Flicker Index – How the Bright and Dark Parts Are Arranged

Two waveforms can have the same frequency and the same modulation depth, but feel very different to the brain depending on how the “on” and “off” parts are arranged.

• Duty cycle (for PWM) is the fraction of time the light is ON in each cycle.
• Flicker Index measures how “top-heavy” the waveform is—how much area is above the average line vs below it over one cycle.

For an ideal PWM signal (0% when off, 100% when on):

• At 50% duty cycle, the light is on half the time and off half the time → Flicker Index is about 0.50.
• At 10% duty cycle, it’s mostly dark with short bright flashes → Flicker Index ~ 0.90 (extremely harsh).
• At 90% duty cycle, it’s mostly bright with short off gaps → Flicker Index ~ 0.10 (less harsh).

The Engineering Behind the Problem

Here's the paradox: LEDs are the most efficient lighting technology ever invented, yet many LED products on the market create worse flicker than technologies from the 1950s. Why?

LEDs Respond Instantly to Power Changes

Unlike older technologies, LEDs are semiconductor devices that respond to current changes in nanoseconds. Any ripple or modulation in the power supply is immediately translated into brightness flicker. There’s no “thermal inertia” to smooth it out like there was with incandescent filaments.

Cheap Drivers = Bad Flicker

LEDs require direct current (DC) to produce steady light, but your wall outlet provides alternating current (AC). The driver circuit converts AC to DC, and this is where the problem originates.

PWM Dimming Makes It Worse

The most common dimming method is Pulse-Width Modulation (PWM). Instead of reducing current gradually (analog dimming), PWM rapidly switches the LED on and off at full power—often hundreds or thousands of times per second.

While efficient and cheap, PWM inherently creates 100% modulation depth with a harsh square-wave pattern. This is the single worst flicker profile for human neurology.^[3]

The Neurological Pathway from Light to Pain – How Flicker Affects the Brain and Body

Flicker is not just a visual annoyance. It’s a neurological stimulus that can affect the brain, eyes, and even the autonomic nervous system.

The Hyperexcitable Migraine Brain

People who suffer from migraines (~12-15% of the population) have brains in a state of cortical hyperexcitability.^[4] Their visual cortex neurons have a lower threshold for firing and respond excessively to sensory stimuli.

Neuroimaging studies (EEG, fMRI, MEG) consistently show that when migraineurs are exposed to visual stimuli like flicker, their brains exhibit abnormally large and widespread cortical activation compared to controls.^[4] This hyperactivity isn't limited to the primary visual cortex—it spreads to higher-order visual areas and even adjacent brain regions.^[6]

The Vision-Pain Connection

There's a direct anatomical link between your visual system and pain pathways. Research has identified a subcortical pathway where signals from the retina converge in the thalamus with neurons from the trigeminovascular system—the very network responsible for migraine pain.^[7]

In healthy people, the visual cortex normally sends inhibitory signals to dampen pain perception. But in migraineurs, this regulatory system is dysfunctional. When the visual cortex becomes massively over-activated by flicker, instead of inhibiting pain, the signal "leaks" across this compromised gateway and activates pain networks.^[6][8]

Why Flicker Is Particularly "Unnatural"

The human visual system evolved to efficiently process natural environments. Natural scenes have specific statistical properties—their temporal frequency content follows a 1/f power law (lower frequencies dominate).^[3]

Studies show that visual discomfort increases when light patterns deviate from these natural statistics. Flicker—especially the sharp, repetitive square waves from PWM—violates the brain's predictive models of the visual world. This creates a larger "prediction error" that requires excessive neural resources to process.^[3]

For the general population, this inefficient processing leads to gradual eyestrain and fatigue. For the hyperexcitable migraine brain, it's enough to trigger a full attack.

A Spectrum of Vulnerability

Flicker doesn't affect everyone equally. There's a clear spectrum from mild annoyance to debilitating neurological events.

Photosensitive Epilepsy: The Extreme End

Approximately 1 in 4,000 people have photosensitive epilepsy (PSE), where flickering lights can trigger seizures. The danger zone is 15-20 Hz, with risk extending from 3-70 Hz.^[11]

The parallels with migraine are striking: both involve hyperexcitable neuronal networks in the visual cortex, both respond to similar frequency ranges, and headaches are a common comorbidity in PSE.^[11]

Comparing Lighting Technologies: A Historical Perspective

Understanding how different light sources produce flicker helps explain why some are well-tolerated and others are problematic.

Why Incandescents Are Well-Tolerated

Traditional incandescent bulbs flicker at 100/120 Hz like everything else on AC power, but they're remarkably comfortable. The key is thermal persistence: the tungsten filament doesn't cool down completely during brief AC zero-crossings. This acts as a natural buffer, resulting in low modulation (~10-20%) with a smooth waveform.^[12]

The Fluorescent Success Story (That Was Abandoned)

Old magnetic-ballast fluorescents were notorious for causing headaches—they had harsh 100 Hz flicker with ~35% modulation.^[1] But modern electronic-ballast fluorescents solved this by operating at >20 kHz, where the phosphor persistence smooths the output to near-constant light.

This was a successful technological solution to the flicker problem. Unfortunately, in the rush to LED adoption, many manufacturers abandoned this hard-won knowledge and reverted to cheap, flickering designs.

Standards: What IEEE 1789 Actually Says

The most commonly cited standard for flicker safety is IEEE 1789-2015. It defines two main regions:

• No Observable Effect Level (NOEL): Very conservative. If you stay below this line, flicker-related health risks are considered negligible for almost everyone.
• Low-Risk Level: Less conservative. Above this line, the risk of adverse effects increases and becomes harder to justify, especially for sensitive populations.
• The problem: Less conservative. Above this line, the risk of adverse effects increases and becomes harder to justify, especially for sensitive populations.

The Math Behind the Recommendations

IEEE 1789 provides frequency-dependent limits on modulation depth:

• <90 Hz: Max modulation = Frequency × 0.025
• 90-1250 Hz: Max modulation = Frequency × 0.08
• >1250 Hz: No restriction (risk negligible)

Practical Limits for Common Frequencies

The Phantom Array Effect: High-Frequency Flicker You See Only When You Move

Most flicker guides focus on low-frequency modulation that you see even when your eyes are still. But LEDs can create a second, very different problem at much higher frequencies—even at 1–20 kHz. This is the Phantom Array Effect (PAE).^[13]

What Is the Phantom Array Effect?

When you move your eyes quickly across a light source (a saccade), your visual system takes a rapid series of snapshots as the light turns on and off. If the light is being pulsed by PWM or a choppy driver, each snapshot lands at a different point in space. Your brain stitches them together and you see a:

✨ A row of repeated “ghost” images

Instead of a single clean light, you see a dotted trail or comb of repeated images across your field of view. That dotted trail is the phantom array.

Key properties:

• You don’t see it when your eyes are still – only during eye movements or when the light moves.
• It’s strongest for rectangular/PWM waveforms with high modulation.
• It remains visible for some people at frequencies well above 10–15 kHz, even though all other artifacts might be gone.
• Subjective reports link strong PAE to headaches, nausea, and visual discomfort in sensitive users. ^[14]

PAVM – Phantom Array Visibility Measure

Researchers at Pacific Northwest National Laboratory introduced a metric called PAVM (Phantom Array Visibility Measure) to quantify how visible the phantom array is under defined conditions.^[13] In short:

• They start from the time-domain waveform of the light.
• Decompose it into frequency components (Fourier spectrum).
• Weight each component by a human sensitivity curve for phantom arrays across frequency.
• Combine components using a Minkowski summation (exponent ≈ 2.1) to reflect perceptual additivity.

The scale is set so that:

• PAVM = 1 → about 50% of observers see the phantom array at threshold.
• PAVM > 1 → phantom arrays are clearly visible for most people.
• PAVM < 1 → phantom arrays are typically weak or absent for most observers.

PAH – A Practical Phantom Array Heuristic (From Real-World Metrics)

In an ideal world we’d always have the full waveform and compute PAVM directly. In practice, many meters and datasheets only give you:

• Frequency (Hz)
• Percent Flicker (%)
• Flicker Index
• SVM (Stroboscopic Visibility Measure)

To still rank real products by phantom-array risk, I came up with something I call the Phantom Array Heuristic (PAH)—a single-number metric you can compute from those four values. Conceptually, PAH is built to behave similarly to PAVM, using a weighted Minkowski combination of four components:^[13][14]

In code, this looks roughly like:

// Pseudocode (not run on this page)
Sf = frequency_weight(freq_hz)          // peaked around 600–800 Hz
Sm = clamp(percent_flicker / 100, 0, 1)
Si = clamp(flicker_index / 0.6, 0, 1)   // assume FI ≈ 0.6 ~ very harsh PWM
Ss = clamp(SVM / 2.0, 0, 1)             // SVM ≈ 1–2 is clearly visible

p  = 2.1
PAH = ((wf*Math.pow(Sf, p) +
        wm*Math.pow(Sm, p) +
        wi*Math.pow(Si, p) +
        ws*Math.pow(Ss, p)) )**(1/p)

You then interpret PAH on a scale similar to PAVM:

How to Test Your Own Lights

You can get a surprisingly good sense of flicker risk with just a smartphone and a little technique. Dedicated flicker meters give the most accurate picture, but even DIY methods can catch the worst offenders.

What About "Flicker Detection" Apps?

I tested several just to see if they had some type of tech that would let them work better than just taking a slow motion video. They didn't seem to be able to detect any flicker with a very high frequency, but they were at least capable of measuring modulation depth.

Here is one that seemed to work decently well.

Note: If you use one of these apps, make sure to put your phone on a tripod and aim the camera at a still, white piece of paper in front of the light. If you hold it in your hand then it can easily confuse hand motion with real flicker.

What About Your Computer Screen?

Most modern LED monitors flicker when dimmed using PWM backlight control. If you're sensitive to flicker:

• Use monitors that advertise "flicker-free" technology (usually DC dimming)
• Keep brightness at 100% and adjust with software (reduces PWM duty cycle issues)
• Use external color/brightness software like f.lux to avoid native PWM dimming

How to Shop for Flicker-Safe Lighting

Most bulb packaging doesn’t list flicker metrics, but you can still stack the odds in your favor.

1. Look for Clues in the Specs

• Mentions of “low flicker,” “flicker-free,” “flicker-safe” from reputable brands.
• Technical notes like “constant current driver”, “low ripple DC”, or “no PWM dimming”.

2. Prefer Better-Engineered Product Categories

• Integrated LED fixtures (from serious lighting brands) often have better drivers than ultra-cheap screw-in bulbs.
• Commercial/architectural lines tend to be engineered to meet stricter office/standards specs.

3. Avoid the Highest-Risk Categories

• Ultra-cheap “no name” LED bulbs and strips.
• USB-powered LED gadgets with unknown drivers.
• Any product that shows obvious banding in slow-mo video at multiple dimming levels.

WARNING:Most dedicated "flicker free" lighting carries a significant cost mark-up. See below for popular, low-cost brands that I've tested to have safe amounts of flicker even if they don't advertise themselves as such.

Specific Recommendations for Sensitive Individuals

What About Older Technologies?

• Incandescent/Halogen: Still available in some markets. Naturally low flicker (~10-20% at 120 Hz) due to thermal inertia. Good fallback option if you can't find quality LEDs.
• High-Frequency Electronic Fluorescents: If you already have these (post-2000 installations), they're often better than cheap LEDs. The flicker is essentially negligible (<1%).

Environment-Specific Strategies

Why This Matters: The Public Health Blindspot

The flicker problem represents a significant market failure. Here's why:

1. The Technology Exists

High-quality, flicker-free LED drivers using multi-stage conversion and constant current reduction (CCR) dimming are mature, proven technologies. They're used extensively in medical devices, studio lighting, and industrial applications.

2. The Science Is Clear

The link between flicker and health effects is extensively documented. IEEE 1789 provides clear, actionable standards. This isn't controversial or uncertain science.

3. The Problem Is Economic, Not Technical

Manufacturers prioritize minimizing component cost over health outcomes because:

• There are no binding regulations requiring flicker disclosure
• Consumer awareness is minimal
• The health costs (lost productivity, medical treatment) are externalized—not paid by the manufacturer

4. The Impact Is Widespread

This isn't a niche issue affecting a tiny population:

• General population: Cumulative effects on productivity, mood, wellbeing
• 12-15% with migraine: A potent, avoidable trigger for a debilitating neurological condition^[9]
• Children in schools: Developing brains exposed 6-8 hours/day to suboptimal lighting
• Office workers: Baseline doubling of headache incidence is a massive productivity drain^[1]

Summary: Your Action Plan

References