Gidsen 

A color setting describes appearance, not a clinical result

Separate three questions: What does the light look like? What wavelengths reach the skin? What outcome has been tested? A mode label answers only the first unless it comes with technical documentation. Wavelength is expressed in nanometers (nm); a spectrum shows how optical output is distributed across wavelengths.

Purple can be a mixture

Red and blue light together can appear magenta. The two components remain present; mixing them does not turn them into one wavelength halfway between red and blue. Blue and green can similarly produce cyan, and red and green can produce yellow in additive color mixing. [1]

That means a mode called “purple” could describe a red-and-blue combination, whereas a violet-emitting channel is a different optical arrangement. A yellow setting could use a yellow-emitting LED or a mixture. Ask which arrangement the device uses. Do not infer it from a photograph, a screen icon or the mode name.

Seven modes need not mean seven independent light sources. As a simple engineering example, three independently switched channels permit seven nonempty on/off combinations. This explains how a device could offer several colors using fewer channels; it does not identify the construction of any particular mask.

A mixed mode needs its own evidence

Evidence about red light and evidence about blue light do not automatically establish the effect of their combination. You still need to know each channel’s output in the mixed setting and whether it operates simultaneously or in sequence. The clinical question is whether the tested combination produced the claimed outcome, not whether its components have appeared in separate studies.

FDA’s draft PBM submission guidance asks for wavelength, radiant dose, irradiance, output mode and relevant pulsing details. It remains a US draft, not an international treatment protocol. Its technical framework helps show why a color name alone is incomplete. [2]

A color-to-claim table that leaves room for unknowns

Use this table to check a claim, rather than to select a treatment. The clinical entries refer to the sources reviewed here; they are not a complete inventory of all LED research.

Near-infrared also needs a specified wavelength and output. A visible mode indicator is not a measurement of an infrared channel. Treat undocumented components as unknown, rather than filling the gaps with a standard rainbow chart.

Green, yellow and amber have different evidence questions

Green: an early human result with a narrow scope

A 2025 study combined mouse melanoma-cell experiments, reconstructed skin models and a human mask study. The human component enrolled 15 healthy women aged 45–57. One randomly assigned facial side received 505 nm LED exposure; the other did not, and evaluators were blinded. Over eight weeks, the authors reported improvements in pigmentation-related measurements on the exposed side. [4]

This is more informative than a cell experiment alone, but the sample was small, participants continued their usual skincare, and several authors were affiliated with the mask manufacturer. It does not establish treatment of diagnosed melasma, every skin tone or another device’s green setting. Cell-pathway findings and human appearance measurements remain different forms of evidence. [4]

Yellow: promising findings do not create a universal “calming” mode

The 2018 systematic review describes 590 nm yellow LED studies evaluating redness after laser or IPL procedures, with short follow-up. Those studies concern the LED intervention in a specific recovery setting; they do not make LED interchangeable with the preceding laser or IPL procedure. [3]

A 2026 randomized study of 30 volunteers with sensitive skin compared an emollient-only control with 590 nm yellow LED and a 590/630 nm combination. Its publisher abstract reports improved symptom scores in both LED groups and improvements in several measurements with yellow light. Full methods were not verified here, so these findings remain an early, study-specific signal rather than a home-use protocol or a rosacea-treatment claim. [5]

Amber: the endpoint matters as much as the color

A 2025 randomized melasma pilot compared amber PBM with topical tranexamic acid over 12 weeks in 21 women. According to the publisher abstract, neither within-group nor between-group changes in the main clinical severity measures were significant. The trial ended early and was underpowered. Quality-of-life improvement does not establish pigment reduction, and a nonsignificant difference does not by itself prove the treatments equivalent. Full methods were not accessible here. [6]

These examples explain why “yellow soothes” and “amber removes pigmentation” are inadequate summaries. A color label, a biological hypothesis and a measured clinical outcome are separate things.

Ask for a separate evidence file for each mode

The most useful documentation connects the setting you can select to the light delivered and the outcome being claimed. Use the following questions as a verification checklist, not as settings to calculate yourself.

1. Identity: Does the document name the exact model and version?
2. Measured spectrum: Can the manufacturer provide a wavelength-versus-output plot for each setting, including mixed modes? Request peak wavelengths, bandwidth and measurement conditions.
3. Delivered output: What irradiance reaches the treatment surface, at the intended geometry? For a mixture, can output be separated by channel?
4. Mode behavior: Are channels simultaneous, alternating or pulsed? Are reported values peak or time-averaged?
5. Human evidence: What device, participants, comparator, outcome and follow-up were used? Is the paper available beyond its abstract?
6. Use boundaries: What does the current instruction manual say about the intended area, exposure limits, eye protection and suitability?

A specification sheet answers an optical question. A human study answers a clinical question. Neither substitutes for the other. A measured spectrum also does not establish safety on its own.

Read regulatory records mode by mode

FDA record K223544 names Guangdong Newdermo models FM-01, FM-02 and FM-03. Its summary identifies separate red, blue, infrared and mixed modes, with the mixed mode listing multiple wavelengths. The indications are also mode-specific, including an infrared topical-heating purpose. This is a useful document-reading example, not evidence about any other brand or seven-color device. [7]

A record must match the manufacturer, model and claimed use. A 510(k) clearance for named indications cannot be expanded into approval of every color-based claim. Likewise, an infrared heating indication should not be presented as proof of nonthermal PBM. [7]

Common misunderstandings

• “More colors mean broader proven benefits.” Mode count is not a count of successful human trials.
• “The color shown in a photo proves the wavelength.” A photograph is not a spectrum measurement.
• “A purple setting combines all the benefits of red and blue.” Combination performance needs combination evidence.
• “If the color has little research, it must be harmless.” Missing efficacy evidence says nothing definitive about safety.
• “Any study at a similar-looking color applies.” UV, lasers, IPL, photosensitizer-assisted PDT, bright-light lamps and infrared heating address different exposures or purposes; their results cannot simply be assigned to an ordinary LED mode.

Questions worth asking

Can two settings with the same color deliver different exposures?

Yes. The label does not specify spectral width, output, pulsing or geometry. Compare the measured specifications and the clinical context before treating settings as equivalent. [2]

Can I compare devices using total electrical wattage?

Electrical input is not the optical irradiance reaching skin. Request the latter at the documented treatment position, with measurement conditions. The distinction is one reason the FDA draft asks for radiant power and irradiance separately from a simple device label. [2]

Does a cycling mode count as using every color’s tested treatment?

No. Without a documented schedule and channel outputs, you do not know each component’s exposure. Even a fully described cycle needs evidence for the intended outcome; do not lengthen it to recreate separate study routines.

What if only the abstract of a study is available?

An abstract can establish that a study was conducted and describe its reported findings. It may omit details needed to judge blinding, delivered output or applicability. Record that limit, and avoid turning the summary into a device-specific protocol.

The useful question is what each mode can document

Read a multi-color LED device as a set of exposures to verify. A defensible claim needs a documented spectrum, delivered output, relevant human evidence and clear instructions. When any of those are missing, keep the conclusion limited to what is actually known.