What Dr. Axe (and Almost Everyone Else) Doesn't Understand About Red Light Therapy Masks

By Scott Chaverri, founder and CEO of Mito Red Light | Updated October 10, 2026

The short answer: A red light therapy mask should be judged by how much light reaches your whole face, and how evenly. On the common 66-LED silicone masks, the LEDs sit right against the skin and cover only about 2% of the face. A new 2026 peer-reviewed study of commercial LED masks confirms what that geometry predicts: on a skin-contact mask, light varied up to 40-fold from one part of the face to another. That creates two problems a reading taken over an LED hides:

1. The face-wide average is far lower than advertised. Counted across the whole face, these masks deliver roughly a tenth of what MitoGLOW does.
2. The light comes in massive peaks and valleys. Intense spots over each LED, near-darkness in between. That's a poor way to deliver light therapy, and the research on light dosing explains why.

I've followed Dr. Josh Axe's work for years, and I respect what he's done to educate people about health and wellness. His article "The Best Red Light Therapy Masks of 2026" gets one big thing right: masks should be scored on measured light, not marketing claims. But measuring a mask properly is harder than it looks, and the method behind that ranking (like most mask reviews and spec sheets) misses both problems. Measured output is 25% of every mask's score, so it shapes the whole ranking.

Key takeaways

• New 2026 research: a peer-reviewed study published in May 2026 in Photochemical & Photobiological Sciences measured light varying about 40-fold across the face on one skin-contact mask and about 21-fold on another, and concluded that "different regions of the face may receive substantially different doses during the same session."
• On a typical 66-LED silicone mask, the LED packages cover about 2–2.5% of the face.
• An independent lab measured 63.7 mW/cm² over the LEDs but about 1.46 mW/cm² across the whole surface of one silicone mask, a 44-fold gap.
• Counted across the whole face, the 66-LED contact masks Dr. Axe scored at 26 mW/cm² deliver about 0.5–0.65 mW/cm², roughly a tenth of MitoGLOW's measured 5.7 mW/cm².
• Light therapy research defines dose per square centimeter of tissue, with too little doing nothing and too much reducing the benefit. Peaks and valleys put most of the face outside that window.
• MitoGLOW's hover design, developed more than two years ago, holds its 266 LED packages off the face so their light overlaps: zero gaps.

Problem 1: The face-wide average is far lower

Peak irradiance vs average irradiance

Irradiance is the amount of light energy reaching a surface, measured in milliwatts per square centimeter (mW/cm²). Dose, or fluence, is irradiance multiplied by time, measured in joules per square centimeter (J/cm²).

There are two very different ways to report a mask's irradiance:

• Peak irradiance: the reading with the sensor directly over an LED. It tells you how bright one LED is.
• Face-wide average irradiance: the light averaged across the entire treated area, including the spaces between LEDs. It tells you what your face actually receives.

Look at the illustration at the top of this article: a skin-contact mask with one LED over one cheek, measuring 100 mW/cm² directly on the LED. Taken to its logical extreme, a ranking built on on-LED readings would put it at the top: 100 mW/cm² beats every mask on the list. Averaged across a face of about 300 cm², that single 0.1 cm² LED delivers about 0.03 mW/cm², almost no light at all. It would be a terrible mask. Peak irradiance describes the LED. Average irradiance across the face, and how evenly it's delivered, describes the treatment.

Light therapy researchers have warned about this for years. A widely cited review of light measurement notes that a single irradiance figure "assumes uniform irradiance over the output area, which can be far from accurate," especially given "the non-unform irradiance distribution of LEDs" (Hadis et al., Lasers in Medical Science, 2016). The standard guide to reporting light therapy dose requires the beam area "at the skin," which "is not necessarily the same as the aperture size" (Jenkins & Carroll, Photomedicine and Laser Surgery, 2011). And light therapy pioneer Michael Hamblin's group cautions that quoting very high mW/cm² from a tiny spot "can give the impression that much higher doses of light were given than actually were delivered" (Huang et al., Dose-Response, 2011).

The LEDs cover about 2% of your face

We measured a typical 66-LED flexible silicone mask, the layout used by several popular brands.

	Value
LED package size	about 1/8 × 1/8 inch (3.2 × 3.2 mm), about 0.10 cm²
Total LED area (66 packages)	about 6.7 cm²
Treated area inside the mask	about 265–340 cm²
Share of the face directly under an LED	about 2–2.5%

Even if every package were a full 5 × 5 mm, the LEDs would cover only about 5–6% of the face. The light-emitting chips inside each package are smaller still.

An independent lab found the same pattern

Light Lab International tested the Maysama Aura, a 155-LED silicone mask, and found 63.7 mW/cm² directly over the diodes, peaks covering "only about 2.2% of the mask's total surface area," and about 1.46 mW/cm² averaged across the whole illuminated surface (Light Therapy Insiders). That's a 44-fold gap on the same mask. The simple area calculation predicts it: 63.7 × 2.2% ≈ 1.4 mW/cm². A 66-LED mask has fewer than half the Maysama's LEDs, so its LEDs sit about 50% farther apart and its gaps are larger still.

The same review shows why the word "average" needs a definition. The reviewer's own average of readings from about 15 spots on that mask was about 5.5 mW/cm², nearly four times the lab's whole-surface figure. An average of readings taken over or near LEDs is still a peak-type number.

What the ranking looks like with face-wide averages

Mask	Figure Dr. Axe scored	Face-wide average (Mito method)
KALA (66 LEDs, skin contact)	26 mW/cm²	about 0.5–0.65 mW/cm²
HigherDOSE (66 LEDs, skin contact)	26 mW/cm²	about 0.5–0.65 mW/cm²
MitoGLOW (266 LED packages, hover design)	5.7 mW/cm²	5.7 mW/cm²

Dr. Axe's article says KALA's reading was taken "directly over the red and near-infrared LEDs, so the average across your face is likely lower" (Dr. Axe), yet that reading still counts at full value in the score. HigherDOSE, on the same 66-LED layout, measured exactly the same 15.6 J/cm² in 10 minutes, and its review describes "surprisingly consistent power readings" across the surface (Light Therapy Insiders). That's hard to reconcile with LEDs covering about 2% of the mask unless the readings were taken over LEDs.

Counted across the whole face, the 66-LED contact masks deliver roughly a tenth of what MitoGLOW does.

The 26 mW/cm² figure itself is also hard to rely on, because readings taken in contact with an LED are very sensitive to technique. We explain why in the appendix.

Problem 2: Massive peaks and valleys

A low average is only half the problem. The light that does arrive isn't spread evenly. It's concentrated in intense spots over each LED, with near-darkness in between, exactly the pattern the 2026 study measured.

Why contact masks create peaks and valleys

On these masks the LEDs sit right against the skin, so each LED's light has almost no distance to spread before it lands. Light scatters within skin over only a few millimeters, while the LEDs are 20–25 mm apart. Optics research on LED arrays shows that even light requires LED spacing matched to the distance between the LEDs and the surface (Moreno et al., Applied Optics, 2006). Skin-contact masks don't meet that condition. A 2025 review of LED therapy devices in Light: Science & Applications puts it plainly: in wearable devices, "insufficient illumination persists in the gaps between LEDs" (Wang et al., 2025).

We're not the only ones critical of skin-contact masks

In June 2024, Carter Gottlieb, an MD/PhD candidate who posts evidence-based skincare content as @evibacarter, put a spectroradiometer on popular LED masks and showed the same pattern on camera: bright spots directly over each LED, and readings falling toward zero in the gaps between them (Eviba Carter on TikTok).

We agree with him that silicone skin-contact masks have real design problems: most of the face sits between LEDs, the light arrives in hot spots, and the fit changes from face to face. That's exactly why we designed MitoGLOW the way we did.

Where we disagree is his conclusion that a mask needs panel-level irradiance to help skin. Board-certified dermatologist Dr. Andrea Suarez (Dr. Dray), who has worked with Omnilux, posted a reply two days later making the opposite case: "High fluences are not necessary to achieve a photobiomodulatory outcome" (Dr. Dray on TikTok). The clinical evidence backs her on this point. Light therapy follows a biphasic dose response, so more irradiance isn't automatically better (Huang et al., 2011). In a 2025 randomized, double-blind, sham-controlled trial, a home LED mask with a maximum of 10 mW/cm² of red and 10 mW/cm² of near-infrared light, used for 9 minutes five times a week, improved crow's feet by at least one grade in 86% of users as scored by independent raters, compared with 17% of users of a sham mask (Park et al., Medicine, 2025). Manufacturer-run studies of flexible silicone masks have also reported improvements in wrinkles, texture and elasticity (Omnilux).

So silicone contact masks can help skin, in spite of their design. The point isn't that they do nothing. It's that a mask that delivers its light evenly across the whole face, at a sensible dose, does the job better. That's what MitoGLOW was built to do.

Measured on real masks: the 2026 study

In May 2026, researchers from Universidad San Pablo CEU in Madrid and the University of Barcelona published the most detailed independent measurement of consumer LED face masks we've seen: "Spectral emission variability in commercial LED masks for cosmetic use" (Baeza-Moyano, Bernardez & Sola, Photochemical & Photobiological Sciences, 2026). The authors bought the masks with university research funds and declared no competing interests.

How they measured

• Four commercial masks, chosen for wide online availability, prices of about €13–200, common LED colors and different designs, including different distances between the LEDs and the face.
• A laboratory-grade instrument: a double-monochromator spectroradiometer, calibrated for wavelength and irradiance at the Institute of Optics of the Spanish National Research Council (CSIC) against Spain's national standard for luminous intensity.
• A fixed sensor, so readings weren't affected by hand placement.
• Two kinds of design: two masks (#2 and #4) are worn in direct contact with the skin; the other two (#1 and #3) hold the face about 8 cm and 2.5 cm from the LEDs.

What they found on the skin-contact masks

LED spacing was irregular. On both contact masks, the gaps between LEDs ranged "from about 0.5 cm to 1.5-3.0 cm, depending on the region," with irregular spacing in areas such as the forehead and chin.

Light varied enormously across the face. The researchers took readings from the blue LEDs at several facial regions on both contact masks:

Skin-contact mask	Lowest reading	Highest reading	Spread
Mask #2	1.089 W/m² (0.11 mW/cm²)	43.76 W/m² (4.38 mW/cm²)	about 40-fold
Mask #4	1.41 W/m² (0.14 mW/cm²)	29.42 W/m² (2.94 mW/cm²)	about 21-fold

In the authors' words, the results were "highlighting large spatial differences in the energy received," and "different regions of the face may receive substantially different doses during the same session." On Mask #4, readings at the same spot also varied more between repeated measurements, which the authors say "indicates increased variability in the energy delivered at each measurement point between sessions."

The real spread is likely even larger. The sensor's light-collecting diffuser was 1.5 cm wide, so each reading already averaged light over an area much bigger than an LED. The authors note that "owing to the diffuser size and the LED spacing, the irradiance does not correspond solely to the selected LED." A pinpoint sensor at the skin would show higher peaks and darker gaps.

Fit makes it worse. Contact masks are meant to sit on the skin, but "depending on the user's facial anatomy, they might not fit uniformly."

What they found on the masks held away from the face

The researchers didn't need to map the two masks that hold the face 2.5–8.5 cm from the LEDs, "due to the greater distance between LEDs and face (2.5–8.5 cm), which means that radiation reaches it from more angles." They also found that in those masks "the distribution of LEDs is more regular." That's the principle behind MitoGLOW's hover design: when light reaches the skin from many LEDs at once, the peaks and valleys smooth out.

Why the authors say it matters

• Dose is everything: "therapeutic outcomes depend critically on the radiometric dose: insufficient irradiance or exposure time may result in subtherapeutic stimulation, whereas excessive radiant exposure may trigger adverse or counterproductive responses."
• Mask specs can't be trusted at face value: "None of the devices include spectral irradiance data," and one manual listed a "light blue" wavelength of 160–195 nm, which is ultraviolet-C, not visible light.
• Masks that look alike aren't alike: with similar session times, users were exposed to irradiances "that differ by more than an order of magnitude," and "devices labeled with the same nominal color cannot be assumed to produce equivalent effects." Some "may not perform the intended function, while others may emit more energy than necessary."
• Measure the light properly: the authors conclude that their findings "underscore the importance of characterizing both the spectral and radiometric properties of facial LED masks to ensure consistent and effective photobiomodulation outcomes."

The study measured spatial differences using the masks' blue LEDs, which the authors chose for their biological relevance. The geometry that creates those differences (LEDs against the skin, centimeters apart) applies to every color on a contact mask.

The science is catching up

We designed MitoGLOW's hover shell more than two years ago for exactly this reason. This 2026 study is the first peer-reviewed measurement we know of to document it on commercial masks: skin-contact masks deliver light unevenly across the face, and masks that hold the LEDs away from the face let light arrive "from more angles."

Why uneven light is a poor way to deliver light therapy

Light therapy has a dose window, defined per square centimeter of tissue. Too little light has no effect, and too much can reduce or cancel the benefit. Hamblin's group summarizes it this way: "insufficient power density or too short a time will have no effect... too much power density and/or time may have inhibitory effects," with an optimal balance in between (Huang et al., 2011). In human skin fibroblasts, 5 J/cm² of red light stimulated healing activity while 10 and 16 J/cm² reduced cell viability and proliferation (Hawkins & Abrahamse, Lasers in Surgery and Medicine, 2006). A mask with peaks and valleys puts small patches of skin at the top of the curve and most of the face near zero. Neither is the window.

The same dose delivered at different intensities gives different results. In a wound-healing study using 670 nm LEDs, the same 5 J/cm² produced different outcomes depending on how irradiance and time were combined: "Varying irradiance and exposure time to achieve a specified energy density affects phototherapy outcomes" (Lanzafame et al., Lasers in Surgery and Medicine, 2007). So hot spots plus dark gaps are not equivalent to the same average delivered evenly.

Even beams produce even biological responses. When researchers compared a uniform "flat-top" beam with a standard beam that's brighter in the center, cells at the edges of the uneven beam showed significantly smaller increases in energy (ATP) production, while "ATP synthesis was increased homogenously in the flat-top handpiece, both in the centre and the edges of the beam" (Amaroli et al., International Journal of Molecular Sciences, 2021). A similar comparison found the uniform beam improved bone-cell maturation (Hanna et al., Frontiers in Endocrinology, 2019).

Small spots penetrate less. In skin modeling, penetration rises noticeably as the beam widens from 1 to 5 mm (Ash et al., Lasers in Medical Science, 2017). Tiny spots of light from contact LEDs don't reach as deep as broad, even illumination.

Other light-based medicine already demands even light

Fields that treat light as a medical dose require uniformity:

• Infant phototherapy: the international standard for neonatal jaundice lights (IEC 60601-2-50) defines the effective treatment area as the zone where the ratio of minimum to maximum irradiance is more than 40%. Researchers also found that moving LED lights closer than recommended raised the peak but lowered the median irradiance and the uniformity (Ismail & Horn, South African Journal of Child Health, 2020), the same trade-off a contact mask makes. The 40-fold and 21-fold spreads in the 2026 mask study are nowhere near that standard.
• Photodynamic therapy: clinicians are told that "a homogeneous and reproducible fluence rate delivery... plays a determinant role in preventing under- or overtreatment" (Mordon et al., Translational Biophotonics, 2020). When one commercial dermatology LED array was mapped, "the dose fell to 38% of that received at the central area at a distance of only 2 cm" (Moseley, Photochemical & Photobiological Sciences, 2005).
• Research light sources: laboratory LED arrays built for light therapy research are tuned to vary as little as 2.3% across the illumination field (Kercher et al., Journal of Biomedical Optics, 2020).

Why MitoGLOW was built differently

We designed the MitoGLOW LED mask more than two years ago, and spent over a year developing it, to solve both problems. Instead of pressing LEDs against the skin, its rigid hover shell holds them off the face, so each LED's light widens and overlaps its neighbors' before it reaches the skin: the light reaches the face "from more angles," as the 2026 study put it. With 266 LED packages (1,064 individual chips), four times as many as a 66-LED contact mask, there are zero gaps: even light across the face instead of peaks and valleys.

That's also why we state a conservative whole-face dose of 3 J/cm² per 10-minute session instead of a peak figure. When Light Therapy Insiders tested MitoGLOW, it measured 3.4 J/cm², more than we claim (Light Therapy Insiders). As far as we can tell, it was the only mask in Dr. Axe's ranking that delivered more than its maker states.

Publishing honest, independently measured numbers isn't new for us. Since 2019, we've had our red light therapy panels tested by LightLab International, an ISO/IEC 17025-accredited photometric lab, and published the results: average irradiance across the treatment area, minimum and maximum readings, uniformity, iso-irradiance maps and full spectra, not just a peak figure. You can see every report on our independent lab test data page. That page covers our panels; the MitoGLOW figures in this article come from Light Therapy Insiders' testing.

MitoGLOW also includes:

• four wavelengths (465, 590, 630 and 830 nm) and five modes;
• built-in chin and neck coverage;
• full blackout eye protection;
• a rigid, hygienic shell that's easy to clean.

MitoGLOW is FDA-cleared for the treatment of full-face wrinkles.

How to compare red light therapy masks fairly

Ask two questions of any mask spec or review: what's the face-wide average, and how even is the light? Three measurement methods work, as long as every mask is measured the same way:

1. Peak × coverage: for skin-contact masks, multiply the on-LED reading by the share of the face the LEDs cover. It's a quick, realistic estimate of the average.
2. A full grid: take readings at evenly spaced points across the whole treated area, at the skin, in the same mode, and count every point, including the near-zero gaps between LEDs. The grid also shows evenness: the lowest reading as a share of the average.
3. Total output ÷ treated area: measure the mask's total light output in an integrating sphere and divide by the area it treats. It doesn't depend on where a sensor is placed.

A trustworthy review should also publish where the sensor was placed, how far it was from the LEDs, the sensor's size and calibration, and the mode tested, as the light therapy reporting guidelines recommend (Jenkins & Carroll, 2011; Hadis et al., 2016). The 2026 study's authors make the same call: characterizing a mask's light properly is what ensures "consistent and effective photobiomodulation outcomes." Be wary of any mask that quotes a single big mW/cm² number without saying where it was measured. For more on choosing a mask, see what to look for in a red light therapy mask, how LED face masks work and what the research supports, and red light therapy mask vs panel.

Frequently asked questions

What did the 2026 LED mask study find?

Researchers from Universidad San Pablo CEU and the University of Barcelona measured four commercial LED face masks with a calibrated laboratory spectroradiometer (Photochemical & Photobiological Sciences, May 2026). On the two skin-contact masks, light varied about 40-fold and 21-fold across different parts of the face, and the authors concluded that "different regions of the face may receive substantially different doses during the same session." They also found that none of the masks came with spectral irradiance data, and that irradiance differed by more than tenfold between masks with similar session times.

What is a good irradiance for a red light therapy mask?

Look at the face-wide average and how evenly it's delivered, not the peak. A mask with a modest but even face-wide average delivers more useful light than a mask with a high peak reading and dark gaps between LEDs. Dose matters too: MitoGLOW delivers a conservative 3 J/cm² across the face in a 10-minute session.

Do LED face masks have dead spots?

Skin-contact masks do. When LEDs sit right against the skin, each one lights only a small spot, and the skin between LEDs receives very little light. On a typical 66-LED silicone mask, the LEDs cover about 2–2.5% of the face, and a 2026 study measured up to 40-fold differences in light across the face of a skin-contact mask. Hover designs like MitoGLOW hold the LEDs off the face so their light overlaps.

Do silicone LED face masks work?

They can help. Red light therapy doesn't require very high irradiance: in a 2025 randomized, sham-controlled trial, a home LED mask with a maximum of 10 mW/cm² of red and near-infrared light improved crow's feet in 86% of users, compared with 17% with a sham. But skin-contact silicone masks deliver their light unevenly, with bright spots over each LED and little light between them, so much of the face gets far less than the intended dose. A mask that delivers light evenly across the whole face, like MitoGLOW's hover design, does the job better.

Why is uneven light a problem in red light therapy?

Light therapy works within a dose window: too little light does nothing, and too much can reduce the benefit. Uneven light puts small patches of skin at very high intensity and most of the face near zero, so little of the face gets the intended dose. Research comparing uniform and uneven beams found more consistent biological responses with uniform light.

Why do some mask reviews report very high mW/cm² numbers?

Usually because the reading was taken with the sensor directly over an LED. That measures the brightest spot on the mask. One lab measured 63.7 mW/cm² over the LEDs of a silicone mask but about 1.46 mW/cm² across its whole surface.

Why are readings taken directly on a mask's LEDs unreliable?

Within a few millimeters of a small LED, the reading changes sharply with tiny shifts in the meter's height, alignment, angle and sensor size, and pressing a meter into a soft silicone mask changes the geometry. A single on-LED number with no distance or position recorded can't be reproduced, and it measures the brightest spot rather than what the face receives. When calibrated laboratory instruments measure silicone contact masks, they find around 1 mW/cm² or less.

How should red light therapy masks be measured?

The same way for every mask, with the face-wide average as the headline number and evenness reported alongside it: an evenly spaced grid across the treated area (including the gaps between LEDs), total output divided by treated area, or, for contact masks, the on-LED reading multiplied by LED coverage.

Is MitoGLOW FDA-cleared?

Yes. MitoGLOW is FDA-cleared for the treatment of full-face wrinkles.

We've shared this with Dr. Axe

On October 10, 2026, we emailed Dr. Axe and his team to explain these measurement problems, with our calculations in a short summary: Why a Reading Over One LED Isn't a Whole-Face Dose (PDF). We'll update this article if and when we hear back from Dr. Axe and his team.

The bottom line

Dr. Axe is right that red light therapy masks should be scored on measured light. But a reading over one LED on a skin-contact mask measures about 2% of your face, and it hides both of the problems that matter: a far lower face-wide average, and light delivered in peaks and valleys instead of evenly. A 2026 peer-reviewed study has now measured those peaks and valleys on commercial masks. Count the whole face, and measure how evenly it's lit, and the ranking changes.

We've offered to send MitoGLOW to Dr. Axe and to Light Therapy Insiders for testing under any fair, published method. See how MitoGLOW's hover design works, read more about red light therapy for the face and neck, and how often to use a red light therapy mask. For more of the science behind red light therapy, explore our Research Evidence Hub.

Appendix: Why readings taken on a contact mask are so error-prone

This appendix adds to the main argument but isn't required for it. Even if every on-LED reading were perfectly accurate, the two problems above would still stand.

Dr. Axe's output scores come from Light Therapy Insiders' measurements, and the 26 mW/cm² credited to KALA and HigherDOSE is the number most worth examining. Light Therapy Insiders uses a handheld spectrometer, a good tool for this work (we use the same model ourselves). The problem isn't the meter. It's that measuring light from an LED pressed against the skin is one of the hardest measurements in light therapy, and small differences in technique produce very different numbers.

What the reviews disclose

For KALA, the review says only: "I measured 26 mW/cm² directly over the red and near-infrared LEDs" (Light Therapy Insiders). It doesn't say how far the sensor was from the LED, whether it touched the mask, how many readings were taken, or which reading was reported. For HigherDOSE, the review says "I moved the spectrometer across the mask surface" and gives a dose of 15.6 J/cm², with no positions, distances or individual readings (Light Therapy Insiders).

Light Therapy Insiders acknowledges the limitation itself. In September 2026, it wrote that a spectrometer "measures the light reaching its sensor at one particular point, distance, and moment... It does not automatically tell me what is happening across the panel's entire treatment area" (Light Therapy Insiders). It added an integrating sphere so that "upcoming reviews, particularly those involving masks," would include more comprehensive testing. The KALA (July 2025) and HigherDOSE (November 2025) figures were measured before that.

Seven ways a contact reading goes wrong

1. Height. Within a few millimeters of an LED, light intensity changes steeply with distance. In a simple optics model of a bare LED, lifting a small sensor just 1 mm, from 1.5 mm to 2.5 mm above the LED, cuts the reading by about 64%. On a soft silicone mask, "1.5 mm" and "2.5 mm" are the difference between resting the meter on the mask and pressing it in.

2. Alignment. In the same model, sliding a small sensor 1 mm sideways off the center of the LED cuts the reading by about half, and 2 mm sideways cuts it by almost 90%. Mask LEDs are only about 3 mm wide, so "directly over the LED" covers a wide range of possible readings.

3. Sensor size. A meter reports light per square centimeter averaged over its own window, not over the skin. When the window is wider than the LED, the number depends on the window. In the same model, an LED 1.5 mm away reads about twice as high through a 10 mm window as through a 15 mm one, and a pinpoint sensor would read 12 to 26 times higher than either. Same LED, same distance, three very different numbers.

4. Neighboring LEDs. A wide sensor window also picks up light from nearby LEDs. The 2026 study's authors noted that, "owing to the diffuser size and the LED spacing, the irradiance does not correspond solely to the selected LED" (Baeza-Moyano et al., 2026). Where LEDs are irregularly spaced, as they found on contact masks, the same reading method gives different answers in different places.

5. Steep angles. Up close, much of an LED's light hits the sensor at steep angles. A meter's diffuser is designed to weight light correctly by angle, and that correction is hardest to get right at the extremes. Readings taken almost touching the source are where any instrument is least accurate.

6. The mask itself. Pressing a sensor into a flexible silicone mask can tilt the LED, change the distance and flatten the mask in ways that don't happen on a face. KALA's own reviewer found that on a real face, "some areas, like the cheeks, sit flush, while other parts, like the jaw and the sides of the face, float above the skin" (Light Therapy Insiders). The 2026 study found contact masks "might not fit uniformly" either.

7. Mode and timing. Many masks pulse or cycle through colors, and LED output can shift as the LEDs warm up. A single snapshot can land on a peak, a dip or a transition, so when in the session a reading is taken also matters.

We see all of this in our own testing. Even with a well-calibrated spectroradiometer, a reading taken directly against a mask's LEDs swings wildly when the meter moves by a fraction of a millimeter. That's why light therapy measurement guidance stresses measuring "at, or through, relevant targets at specific distances and geometry" and recording how it was done (Hadis et al., 2016; Jenkins & Carroll, 2011). A single on-LED number with no distance, position or sensor size recorded isn't a measurement anyone can reproduce, and it can't be compared with an average taken on a different mask.

Solar meters make it worse

Many brands' advertised mW/cm² figures come from handheld solar meters. Light Therapy Insiders warns that these meters "are designed to measure sunlight, not LEDs, and they often read two to three times higher than the actual value" (Light Therapy Insiders), and that "in many cases, solar meters produce higher readings than more appropriate scientific equipment" (Light Therapy Insiders). Dr. Axe's ranking checks each mask's measured output against those advertised figures, so the claims being checked may themselves be inflated. It's why we publish accredited-lab averages for our panels rather than solar-meter peaks; see our independent test data.

Same mask, handheld readings vs a laboratory

One mask was measured both by Light Therapy Insiders and by an independent laboratory: the Maysama Aura. The handheld readings ranged "from roughly 3 to 7 mW/cm² across different parts" of the mask, averaging about 5.5 mW/cm². Light Lab International's average across the whole illuminated surface was about 1.46 mW/cm² (Light Therapy Insiders). Even the lowest handheld reading was twice the laboratory's average, and the handheld average was nearly four times higher.

Every laboratory measurement is far lower

Source	Mask type	Measured
Light Therapy Insiders, handheld spectrometer (scored by Dr. Axe)	KALA and HigherDOSE, 66-LED silicone contact masks	26 mW/cm², "directly over the red and near-infrared LEDs"
Light Lab International	Maysama Aura, 155-LED silicone contact mask	about 1.46 mW/cm² averaged across the surface (63.7 directly over the diodes)
Baeza-Moyano et al. 2026, calibrated laboratory spectroradiometer	Two skin-contact masks, red LEDs, sensor centered on an LED	0.23 and 0.93 mW/cm²
Baeza-Moyano et al. 2026	All four masks, red LEDs	0.017 to 0.93 mW/cm²; red dose per session 0.01 to 0.84 J/cm²

The masks differ, and the 2026 study measured visible light only (380–680 nm), so it doesn't include near-infrared. But the pattern is consistent: when laboratories measure silicone contact masks, they find around 1 mW/cm² or less, not 26. Even with the 2026 study's 1.5 cm sensor centered directly on an LED, the brighter contact mask's red light measured 0.93 mW/cm², about 28 times lower than the figure Dr. Axe scored.

References

1. Baeza-Moyano D, Bernardez AM, Sola Y. Spectral emission variability in commercial LED masks for cosmetic use. Photochem Photobiol Sci. 2026;25:1179–1188. https://link.springer.com/article/10.1007/s43630-026-00908-x
2. Axe J. The Best Red Light Therapy Masks of 2026, Scored on Measured Light and Real FDA Records. DrAxe.com, Sept. 26, 2026. https://draxe.com/healthy-home/best-red-light-therapy-masks/
3. Fergus A. Maysama Aura Review: My Testing Revealed a Big Problem (including Light Lab International results). Light Therapy Insiders, Aug. 13, 2026. https://www.lighttherapyinsiders.com/maysama-aura-review-my-testing-revealed-a-big-problem/
4. Fergus A. HigherDose Mask Review. Light Therapy Insiders. https://www.lighttherapyinsiders.com/higherdose-mask-review/
5. Fergus A. Mito Red MitoGLOW Mask Review. Light Therapy Insiders. https://www.lighttherapyinsiders.com/mito-red-mitoglow-mask-review/
6. Hadis MA, Zainal SA, Holder MJ, et al. The dark art of light measurement: accurate radiometry for low-level light therapy. Lasers Med Sci. 2016;31(4):789–809. https://pmc.ncbi.nlm.nih.gov/articles/PMC4851696/
7. Jenkins PA, Carroll JD. How to report low-level laser therapy (LLLT)/photomedicine dose and beam parameters in clinical and laboratory studies. Photomed Laser Surg. 2011;29(12):785–787. https://pubmed.ncbi.nlm.nih.gov/22107486/
8. Huang YY, Sharma SK, Carroll J, Hamblin MR. Biphasic dose response in low level light therapy: an update. Dose Response. 2011;9(4):602–618. https://pmc.ncbi.nlm.nih.gov/articles/PMC3315174/
9. Moreno I, Avendaño-Alejo M, Tzonchev RI. Designing light-emitting diode arrays for uniform near-field irradiance. Appl Opt. 2006;45(10):2265–2272. https://pubmed.ncbi.nlm.nih.gov/16607994/
10. Wang C, Yu Q, Li M, et al. Challenges and opportunities in next-generation LED therapeutic devices. Light Sci Appl. 2025;14:319. https://pmc.ncbi.nlm.nih.gov/articles/PMC12434144/
11. Hawkins DH, Abrahamse H. The role of laser fluence in cell viability, proliferation, and membrane integrity of wounded human skin fibroblasts following helium-neon laser irradiation. Lasers Surg Med. 2006;38(1):74–83. https://onlinelibrary.wiley.com/doi/abs/10.1002/lsm.20271
12. Lanzafame RJ, Stadler I, Kurtz AF, et al. Reciprocity of exposure time and irradiance on energy density during photoradiation on wound healing in a murine pressure ulcer model. Lasers Surg Med. 2007;39(6):534–542. https://pubmed.ncbi.nlm.nih.gov/17659591/
13. Amaroli A, Arany P, Pasquale C, et al. Improving consistency of photobiomodulation therapy: a novel flat-top beam hand-piece versus standard Gaussian probes on mitochondrial activity. Int J Mol Sci. 2021;22(15):7788. https://pmc.ncbi.nlm.nih.gov/articles/PMC8346075/
14. Hanna R, Agas D, Benedicenti S, et al. A comparative study between the effectiveness of 980 nm photobiomodulation delivered by hand-piece with Gaussian vs. flat-top profiles on osteoblasts maturation. Front Endocrinol. 2019;10:92. https://pubmed.ncbi.nlm.nih.gov/30842754/
15. Ash C, Dubec M, Donne K, Bashford T. Effect of wavelength and beam width on penetration in light-tissue interaction using computational methods. Lasers Med Sci. 2017;32(8):1909–1918. https://pubmed.ncbi.nlm.nih.gov/28900751/
16. Ismail MT, Horn AR. The effect of device position and use of transparent covers on the irradiance distribution of LED phototherapy devices. S Afr J Child Health. 2020;14(2). https://scielo.org.za/scielo.php?script=sci_arttext&pid=S1999-76712020000200006
17. Mordon S, Thécua E, Ziane L, et al. Light emitting fabrics for photodynamic therapy: technology, experimental and clinical applications. Transl Biophotonics. 2020. https://hal.science/hal-02866507/document
18. Moseley H. Light distribution and calibration of commercial PDT LED arrays. Photochem Photobiol Sci. 2005;4(11):911–914. https://pubmed.ncbi.nlm.nih.gov/16252048/
19. Kercher EM, Zhang K, Waguespack M, et al. High-power light-emitting diode array design and assembly for practical photodynamic therapy research. J Biomed Opt. 2020;25(6):063811. https://pmc.ncbi.nlm.nih.gov/articles/PMC7156854/
20. Gottlieb C (@evibacarter). "Sorry red light mask fans" [video]. TikTok, June 4, 2024. https://www.tiktok.com/@evibacarter/video/7376731623749995818
21. Park SH, Park SO, Jung JA. Clinical study to evaluate the efficacy and safety of home-used LED and IRED mask for crow's feet: a multi-center, randomized, double-blind, sham-controlled study. Medicine (Baltimore). 2025;104(7):e41596. https://pmc.ncbi.nlm.nih.gov/articles/PMC11835066/