Near Infrared Light Therapy: Benefits, Evidence and Safety
Near infrared light therapy uses invisible light just beyond red, at about 780 to 1,400 nm, and its clearest support in trials is for knee pain, exercise performance and facial wrinkles. It is light that acts in cells, not heat. Below: what near-infrared is, how it differs from a sauna, how it works, the benefits with the best evidence, 850 versus 1064 nm, trial settings, safety, and how to use our mask and panel.
Key takeaways
- ICNIRP, the international body that sets light exposure limits, defines near-infrared (IR-A) as 780 to 1,400 nm, light the eye cannot see. ICNIRP, 2013
- A 2019 review of 22 placebo-controlled trials with 1,063 people found knee osteoarthritis pain 18.7 mm lower on a 100 mm scale at the recommended near-infrared doses. Stausholm, 2019
- In a sham-controlled split-face trial of 76 people, 830 nm near-infrared, 633 nm red or both, twice a week for 4 weeks, reduced wrinkles by up to 36%. Lee, 2007
- In a 12-week trial of 48 men, light applied before strength training gave larger strength gains than the other schedules tested. Vanin, 2016
- Because near-infrared gives no visible glare, the eye loses its natural urge to squint or look away, which is why goggles matter with a panel. ICNIRP, 2013
What is near infrared light therapy?
Near infrared light therapy is the use of low-power light at about 780 to 1,400 nm, just past the red end of the visible spectrum, on the skin. It belongs to photobiomodulation, the same research field as red light therapy, and most devices pair the two.
The American Academy of Dermatology describes red light therapy as using "red or near-infrared light", also called photobiomodulation. AAD, 2024
Joint and muscle trials mostly used 785 to 950 nm, skin trials 830 nm, and newer marker studies 1064 nm.
You cannot see most of it. An 850 nm LED may show a faint dull red glow, while 1060 and 1064 nm show nothing, so brightness is no guide to output. Our infrared vs red light therapy guide explains a safe way to check that the near-infrared is on.
Is near-infrared light therapy the same as an infrared sauna?
No: near-infrared is light that acts inside cells, while an infrared sauna mainly warms you with far infrared, which starts at about 3,000 nm. Far infrared is absorbed at the very surface of the skin and is felt mainly as heat.
In photobiomodulation the effect is thought to come from light absorbed by molecules in cells, without heating the tissue, in timed sessions. A sauna works through heat on the whole body. Our far infrared vs near infrared article covers the bands and the sauna research.
How does near-infrared light work, and how deep does it reach?
The leading idea is that near-infrared is absorbed by cytochrome c oxidase, an enzyme in mitochondria, which raises ATP and sets off brief cell signals. Near-infrared reaches deeper than red light, and red reaches deeper than blue.
A 2017 review lists the follow-on effects: more ATP, "a brief burst of reactive oxygen species" and more nitric oxide. Hamblin, 2017 It calls reduced inflammation one of the most reproducible effects. Much of this mechanism work comes from cell and animal studies.
- Forearm study: in two groups of 10 healthy adults, 800, 850 and 1064 nm lasers and an 810 nm LED all raised oxygenated haemoglobin and oxidised cytochrome c oxidase in the forearm, compared with sham. Pruitt, 2022
- Tissue study: in a laboratory study, more 830 nm near-infrared than 633 nm red came through cheek and hand tissue. Jagdeo, 2012
LEDs belong to the same field as lasers: a review of about 40 comparison papers concluded that "photobiomodulation is not dependent on lasers or coherence". Heiskanen, 2018 For more, see our photobiomodulation therapy guide.
Which near infrared light therapy benefits have the best evidence?
The strongest results come from reviews of placebo-controlled trials on knee osteoarthritis, neck pain and exercise, plus a sham-controlled skin trial in which near-infrared alone reduced wrinkles. A 2025 umbrella review rated knee disability and hair density among the outcomes with moderate certainty.
| Area | Evidence | Light used | Main finding |
|---|---|---|---|
| Knee osteoarthritis | Stausholm, 2019: 22 trials, 1,063 people | 785 to 860 nm or 904 nm laser at recommended doses | Pain 14.2 mm lower than placebo across all trials, 18.7 mm at recommended doses (100 mm scale) |
| Neck pain | Chow, 2009: 16 trials, 820 people | Low-level laser light | Pain 19.86 mm lower than placebo (100 mm scale) |
| Exercise | Vanin, 2018: 39 trials, 861 people | 655 to 950 nm laser, LED or both | Better performance and less fatigue than placebo |
| Facial wrinkles | Lee, 2007: 76 people | 830 nm LED, 633 nm LED or both | Wrinkles down by up to 36%, elasticity up by up to 19% |
| Hair | Son, 2025: umbrella review | Red or near-infrared light | Moderate certainty for higher hair density |
In the neck review, side effects were mild and no different from placebo. Chow, 2009
The 2025 umbrella review covered over 9,000 people. Son, 2025 Its authors call photobiomodulation "a promising non-invasive adjunctive therapy", an add-on to usual care.
For hair, the AAD says light may regrow some hair, though "it won't regrow a full head of hair". AAD, 2024 The hair reviews mix red and near-infrared wavelengths.
Does near-infrared help knee and joint pain more than red light?
The joint trials mostly used near-infrared, so that is where the evidence sits, rather than in a head-to-head win over red. The recommended doses in the 2019 knee review were all near-infrared, at 785 to 860 nm or 904 nm.
The review set its recommended dose at 4 to 8 J per treatment spot at 785 to 860 nm, or 1 to 3 J per spot at 904 nm. Stausholm, 2019
Disability was also lower than with placebo, and no adverse events were reported.
These trials used clinic lasers at points over the knee. A hot, red, swollen joint, a new injury or numbness needs a doctor first, and light sits alongside your doctor's or physiotherapist's plan.
Should you use near-infrared light before or after a workout?
Before, in most trials: in a 12-week study of 48 men, light applied before strength training gave larger strength gains than the other schedules tested. A review of exercise trials found benefits mainly with 655 to 950 nm light.
The trial used probes combining 905 nm laser, 875 nm LED and 640 nm LED light, and reported no side effects. Vanin, 2016
A 2018 review of 39 trials with 861 healthy people found "some effect in favor to phototherapy" for performance and fatigue. Vanin, 2018 These were small trials of mixed quality. Light adds to training; it does not replace it.
Can near-infrared light on its own improve facial wrinkles?
Yes, in one sham-controlled trial: 830 nm near-infrared alone improved wrinkles, as did 633 nm red alone and the two combined. Wrinkles fell by up to 36% and elasticity rose by up to 19% after 4 weeks of twice-weekly sessions.
The 76 volunteers had LED light on one half of the face and were followed for 3 months. Lee, 2007 Skin samples showed a "marked increase in the amount of collagen and elastic fibers", with no adverse effects.
The AAD describes subtle to noticeable improvement in fine lines, wrinkles and roughness, with results that take time. AAD, 2024 Our 660 nm vs 850 nm guide covers when to favour red or near-infrared.
Is 1064 nm better than 850 nm near-infrared?
Neither is better across the board: they appear to act in partly different ways and sit well together. In a small forearm study, 1064 nm laser light kept mitochondrial markers raised for longer, while 800 and 850 nm did not differ from each other.
That study was a single 8-minute session in two groups of 10 adults, measuring markers rather than symptoms. Pruitt, 2022 A cell study found that 810 nm light acted mainly through cytochrome c oxidase, while 980 nm light raised calcium inside cells through ion channels. Wang, 2017
Research at 1064 nm is thinner than at 800 to 850 nm. Our 1064 nm article treats it as an addition to red and 850 nm, rather than a replacement.
Is near-infrared light safe for your skin and eyes?
At the moderate levels used in therapy, near-infrared on the skin is well tolerated, and the knee review and skin trial above reported no adverse effects. The eyes need more care, because invisible light does not make you squint or blink.
A 2016 review of fears that near-infrared ages skin concluded: "It is essentially a question of intensity." Barolet, 2016 In the forearm study, lasers warmed skin by about 4°C or less over 8 minutes. Warm is fine; if skin feels hot or stays red, step back, dim the light or stop.
ICNIRP names infrared LEDs among sources with no strong visual stimulus: "The lack of visible stimulus results in loss of the aversion response." ICNIRP, 2013 The AAD adds: "Don't substitute sunglasses for protective goggles." AAD, 2024
- Panel: goggles on and eyes closed every session; never look into it, even when it looks dim.
- Mask: an eye opening with no diodes over the eyes; keep your eyes closed and never look into it from outside while it is on.
See our red light therapy goggles guide for more on eye protection. Our precautions rule out both devices during pregnancy, with epilepsy, a thyroid condition or thyroid medication, and over broken skin or an open wound. They also rule out photosensitising medication, so anyone taking a medicine should check with their pharmacist or doctor before using the device. The panel manual adds breastfeeding, preparing for pregnancy, body sunburn in the past 2 weeks, seizures triggered by light, and Fitzpatrick skin types V or VI.
What settings did near-infrared trials use?
Trials applied near-infrared, mostly at about 785 to 950 nm, to the skin twice a week in the skin trial, or with each workout in the strength trial, for 4 to 12 weeks. Cleveland Clinic describes a typical course as 1 to 3 sessions a week for weeks or months.
| Study | Light | Dose or time | Sessions per week | Weeks |
|---|---|---|---|---|
| Lee, 2007 (sham) | 830 nm, 633 nm or both, LED | Not given in the summary | 2 | 4, then 3 months of follow-up |
| Stausholm, 2019 (22 trials) | 785 to 860 nm or 904 nm laser | 4 to 8 J per spot (785 to 860 nm); 1 to 3 J (904 nm) | Varied by trial | Varied by trial |
| Vanin, 2018 (39 trials) | 655 to 950 nm laser or LED | 20 to 60 J small muscles; 60 to 300 J large muscles | Varied by trial | Varied by trial |
| Vanin, 2016 (48 men) | 905 nm laser, 875 and 640 nm LED | Before each training session | With each session | 12 |
The typical course comes from Cleveland Clinic. Every study put light on intact skin.
More is not always better: the 2017 review describes a biphasic dose response, where low doses stimulate and high doses can inhibit. Hamblin, 2017 Laser joules per spot do not convert into a home session. To stay near the gentler studied range, dim each wavelength, set a shorter timer, or step back from 6 to 12 inches.
How do you use the Kivorla mask and panel for near-infrared sessions?
Use the mask for the face and the panel for joints, muscles and broad body areas, since both include near-infrared. Start gently, keep your eyes closed, and wear the goggles with the panel every session.
The 4-Wavelength Red Light Therapy Mask carries 72 VCSEL laser diodes at 665 nm, 72 at 850 nm and 20 at 1064 nm, in a cordless, full-face silicone shell with one eye opening.
- Modes: the red mode is 665 nm red with 850 and 1064 nm near-infrared; the blue mode is 460 nm blue with 850 nm near-infrared.
- Controls: a handheld controller with a colour screen, or voice; brightness 50, 70 or 90%; timer 5, 10 or 15 minutes.
- Limits: 10-minute sessions up to 4 times a week at high brightness, or 15-minute sessions up to 3 times a week after a week of three 10-minute sessions; never more than 20 minutes a day, counting both modes together.
- Start: 50% brightness on the 10-minute timer, on clean, dry skin, stepping up only if your skin stays calm.
You will see the red diodes and little or nothing from the near-infrared ones, which is normal. For the jaw, neck or scalp, use the panel.
The 9-Spectrum Red Light Therapy Panel has nine wavelengths: 480, 590, 630, 660, 670, 810, 830, 850 and 1060 nm. Its near-infrared channels are 810, 830, 850 and 1060 nm, and each channel dims from 1 to 100%. Most of its output at 6 inches is near-infrared: at least 1,200 of at least 1,567 W/m².
- Position: hang it on a door or wall with the included door hook, pulley, hanging cables and screws, or stand it upright on the floor.
- Distance: about 6 inches for a joint, 12 inches for a broad area such as the thighs or back.
- Channels: near-infrared on, red on or lower, 480 and 590 nm turned down, continuous light.
- Timer: 10 to 15 minutes per area, set from 1 to 30 minutes; it stops at 5 minutes if no time is set. Leave at least 6 hours between 30-minute sessions.
- Neck and head: side-on or from behind, never aimed straight at the face.
- Warmth: the surface stays at or below 40°C.
The near-infrared channels look almost dark, so never look into the panel to check it is on. Goggles and a controller come in the box.
Read our precautions before ordering. Standard shipping to the US, UK and Germany is free. Under our return and refund policy, returns are accepted within 30 days: an unopened device with its seals intact is refunded less a 10% restocking fee, and an opened device that is undamaged and complete is refunded less 20%.
Kivorla devices are general wellness devices, not medical devices, and are not intended to treat any condition. See a doctor or physiotherapist for joint pain, injuries or skin changes, and keep any treatment they have advised.