How Does Red Light Therapy Work? The Science and Research

Red light therapy (also called photobiomodulation) is now sold for home use, so it’s worth asking what the red light therapy science actually shows, and what it doesn’t. This page summarises a handful of original studies and reviews, including their sizes and limits, so you can read the claims you see online with a clearer eye.

An important note before you read on: every finding below comes from specific study devices, light settings and groups of participants. Those findings may not apply to any particular device, or to you.

This is general information, not medical advice, and it isn’t about any particular product. For advice about your own health or skin, please talk with your GP, dermatologist or another health professional.

How does red light therapy work?

The short answer is that researchers are still working it out.

The most widely discussed idea centres on mitochondria, the parts of our cells that produce energy. A 2016 review by de Freitas and Hamblin describes an enzyme in the mitochondria called cytochrome c oxidase as one of the most important light-absorbing molecules involved. Their “leading hypothesis” is that light frees nitric oxide that has bound to this enzyme and slowed it down, allowing energy production (ATP) to increase. It also describes a second idea involving light-sensitive calcium channels.[1]

A 2020 review by Quirk and Whelan found that no reliable demonstration of a light-driven effect on cytochrome c oxidase had been reported. The authors suggest nitric oxide production may be closer to the heart of how light acts on cells.[2] A 2019 review by Sommer went further and concluded that cytochrome c oxidase “is not the primary acceptor” of red and near-infrared light, proposing a different model.[3]

So the fairest summary is that the mitochondrial explanation is a leading theory, and scientists continue to debate it.

Red vs near-infrared light therapy

“Red” and “near-infrared” (NIR) describe different parts of the light spectrum, measured in nanometres (nm). A 2026 systematic review of at-home LED masks used these working ranges: red light at about 620–750 nm, and near-infrared at about 780–2,000 nm. Near-infrared is invisible to the human eye. In that review, 88% of the included studies used red light and 63% used near-infrared, often in combination.[5]

Why these colours? De Freitas and Hamblin note that red and near-infrared light are used in research partly because they travel through tissue more readily than other wavelengths, since the body’s own pigments absorb and scatter less of it in this range.[1] Penetration depth varies with wavelength, device, power and skin, so we haven’t quoted figures.

What the research says on skin and collagen

A controlled trial (2014). Wunsch and Matuschka enrolled 136 volunteers (128 completed). Of these, 113 were randomised to red-only (611–650 nm) or 570–850 nm light twice a week; 23 controls received no light. After 30 sessions, researchers recorded differences favouring the light groups in blinded photo ratings, skin roughness and ultrasound-measured collagen density; the two light types didn’t differ. At six months, 77 people were re-measured: still better than baseline, but by less.[4] Limits: the devices were large-area and full-body gas-discharge lamps, not LEDs. Controls, mainly sponsor employees, weren’t randomised. A light-equipment company fully funded the study and paid the lead investigator. The authors flagged possible follow-up selection bias.[4]

A systematic review of at-home LED masks (2026). Das and colleagues found eight studies with 365 participants in total. The included studies reported statistically significant changes in measures of wrinkles, texture and elasticity, and side effects were reported as infrequent, mild and short-lived. Limits: the authors noted that protocols varied a lot, there were few head-to-head comparisons, diverse skin types were under-represented, and light dose often wasn’t reported. They called for larger, standardised trials.[5]

This red light therapy research on collagen and skin is still limited: the studies are fairly small, methods vary, and findings may not carry over to other devices or people.

What researchers measured on post-exercise muscle soreness

Delayed onset muscle soreness (DOMS) is the term researchers use for the stiffness many people notice a day or two after unfamiliar or hard exercise. A 2025 meta-analysis by Tsou and colleagues identified 14 controlled studies, which used wavelengths from 660 to 950 nm. Only four studies had enough data to be pooled. In those four, participants given light rated their soreness lower than participants given a placebo at 72 and 96 hours after exercise, and researchers measured higher muscle strength in the light groups at 24 and 48 hours.[6]

Limits: these pooled findings came from just four studies, the findings varied considerably between them, and the light settings and application methods differed widely. The authors themselves called for more high-quality studies with larger samples.[6] The findings came from specific research equipment and may not apply to any particular device or person.

For personal advice about exercise or how your body feels afterwards, talk with your GP or a physiotherapist.

Why dose matters

More light isn’t automatically better. De Freitas and Hamblin describe a biphasic dose response in laboratory and clinical research: too little light may produce no measurable effect, while too much may also produce none, or even an inhibitory one. “Dose” here includes energy delivered, intensity, session time and how often sessions are repeated.[1] Quirk and Whelan, by contrast, describe the idea of a single “best dose” as questionable, though not necessarily false.[2]

Follow your device’s instructions on distance, session length and frequency. These studies don’t support using a device for longer, or more often, than its instructions say.

How long did the studies run?

If you’ve searched “red light therapy how long to see results”, the honest answer is that we can only tell you how long the studies ran. These timeframes describe the research, not what any reader will experience.

In the 2014 trial, participants had 30 sessions at two per week.[4]
In the at-home mask review, studies ran for 4 to 12 weeks, with sessions of 3 to 21 minutes, 2 to 7 times a week.[5]
The muscle soreness studies took measurements over the four days after exercise.[6]

How to read red light research

When you see a headline or claim, ask:

How many people? Eight studies with 365 people in total is useful, but it’s not a large evidence base.[5]
How long? Most studies lasted only weeks, so long-term effects aren’t known.[4][5]
Was there a comparison group? Findings mean more when the people receiving light are compared with a placebo, sham or control group.[4][6]
Who funded it, and who ran it? Check the funding and conflict-of-interest statements. For example, the DOMS meta-analysis reported no funding,[6] the mask review reported no conflicts,[5] and a light-equipment company fully funded the 2014 trial.[4]
Lab or people? Mechanism studies on cells describe how light might act. They don’t show what happens with everyday use.[1][3]

FAQ

How does red light therapy work? The leading theory is that light interacts with mitochondria, possibly via an enzyme called cytochrome c oxidase and nitric oxide, but researchers actively debate this.[1][2][3]

What’s the difference between red and near-infrared light? Red light sits at roughly 620–750 nm and is visible. Near-infrared sits at roughly 780–2,000 nm and is invisible. Many devices use both.[5]

What has research measured on red light therapy and collagen? A 2014 trial of 136 people recorded higher ultrasound-measured collagen density in the light groups than in controls after 30 sessions. A 2026 review of eight studies (365 people) reported changes in wrinkle and texture measures. Both have limits and used specific devices.[4][5]

What did studies on post-exercise muscle soreness find? A 2025 meta-analysis pooled four studies. Participants given light rated their soreness lower than those given a placebo at 72–96 hours, but findings varied between studies and may not apply to any particular device or person.[6]

Why does dose matter? Research describes a biphasic response, where too little or too much light may produce no measurable effect, so follow your device’s instructions.[1]

Red light therapy: how long to see results? We can only report study lengths: 4 to 12 weeks (mask review) and 30 sessions (2014 trial). That describes the research, not what you’ll experience.[4][5]

For personal advice about whether red light therapy is right for you, please talk with your GP or another health professional.

Sources

de Freitas LF, Hamblin MR. Proposed mechanisms of photobiomodulation or low-level light therapy. IEEE Journal of Selected Topics in Quantum Electronics. 2016;22(3). https://doi.org/10.1109/JSTQE.2016.2561201
Quirk BJ, Whelan HT. What lies at the heart of photobiomodulation: light, cytochrome c oxidase, and nitric oxide. Review of the evidence. Photobiomodulation, Photomedicine, and Laser Surgery. 2020. https://doi.org/10.1089/photob.2020.4905
Sommer AP. Revisiting the photon/cell interaction mechanism in low-level light therapy. Photobiomodulation, Photomedicine, and Laser Surgery. 2019. https://doi.org/10.1089/photob.2018.4606
Wunsch A, Matuschka K. A controlled trial to determine the efficacy of red and near-infrared light treatment in patient satisfaction, reduction of fine lines, wrinkles, skin roughness, and intradermal collagen density increase. Photomedicine and Laser Surgery. 2014;32(2):93–100. https://doi.org/10.1089/pho.2013.3616
Das RK, Thayer WP, Kubiak CA, et al. Efficacy and safety of at-home light-emitting diode photobiomodulation for facial rejuvenation: a systematic review. Aesthetic Plastic Surgery. 2026. https://link.springer.com/article/10.1007/s00266-026-06333-w
Tsou YA, Chang NJ, Chang WD. Effects of photomodulation therapy for delayed onset muscle soreness: a systematic review and meta-analysis. Journal of Functional Morphology and Kinesiology. 2025;10(3):277. https://doi.org/10.3390/jfmk10030277