Red light and the brain — what families can actually lean on

1 The short take

Red and near-infrared light can affect cells in a lab dish, and a few small human studies have measured behavior or muscle scores after light. That is not the same as proof that a helmet grows a child’s brain, treats epilepsy, or replaces usual care.

• Autism: a handful of small trials and chart reviews. The best-known sham comparison is 30 children, industry-funded, with serious methods limits (below). Symptom scores moved. Brain size was not measured.
• ADHD: the controlled human work we indexed is in adults, mostly one session or one week of a research laser. We did not find a solid trial in children.
• Epilepsy: still no human trial showing that this light stops seizures. What exists is rats and mice. Any “it helped seizures” sentence from those papers is an animal result, not a correlation you can apply at home.
• Brain growth / neurogenesis: we did not find a child study that measured new brain tissue or bigger brain volume after red light. “More energy in cells” on a website is a theory, not a scan result.
• Cerebral palsy: several clinic studies aim light at muscle or nerves for stiffness, not at “repairing” the early brain injury. The one placebo-controlled calf-muscle pilot did not beat the device switched off.

2 Autism — the studies, including the fine print

The study families hear about most is Fradkin and colleagues, Frontiers in Neurology, 2024 (DOI 10.3389/fneur.2024.1221193^[2], archive card pbm-20261003-asd-fradkin-rct). Thirty children aged 2–6 (16 active, 14 sham) received pulsed 850 nm light, 40 times a second, twice a week for eight weeks, from an investigational headset called Cognilum. The autism rating (CARS-2) changed more in the active group: difference 7.23 points (95% CI about 2.4 to 12.1; abstract p=0.011). No moderate or severe side effects were reported inside that trial. EEG shifts were mostly borderline.

A separate methods card (pbm-20261003-asd-fradkin-methods) is why we do not treat that 7.23-point gap as settled efficacy:

• CARS-2 really was the registered primary outcome (posted before the first participant), so this is not a case of swapping the endpoint after the fact.
• The sample of 30 was a feasibility size. There was no power calculation.
• Four early dropouts were replaced. One child assigned to sham was then given active light. The analyzed groups were 16 versus 14.
• The test was an unadjusted completer comparison, not intention-to-treat.
• Research assistants who could see the random list turned the device on or off, so operators were not blinded. The paper says caregivers, children, and assessors were blinded.
• The CARS rater is a co-founder of the device company and holds equity. The company funded the study and took part in design, data collection, analysis, the decision to publish, and the manuscript.
• Sham was the device switched off, not a low-power lookalike. The paper’s device description (850 nm, light up to about 6 minutes) does not match what the public registry results page posted (830 nm, up to 15 minutes, and a different side-effect statement).
• Children with a seizure history or on regular medicines were excluded. Sessions were in two New York offices, not at home.

Other autism papers in the archive are weaker still. An open-label series in ages 2–7 (pbm-20261003-19) gave everyone the light, so it cannot separate the device from hope and practice. A home chart review of 21 children (pbm-20261003-23^[4]) also had no sham group. A 2018 laser-versus-placebo paper on irritability (pbm-20261003-25^[3]) reports a large score change, but the abstract does not show wavelength or who was blinded. Adult open-label work (pbm-20261003-24) is not a result for young children. A single-child urine-metal case (pbm-20261003-26) does not show that red light treats autism.

When a score moves in a tiny study, that is an association in that sample. It is not proof the light caused a brain change, and it is not a correlation you can generalize to the next child — especially a child who has seizures or takes daily medicine, because those children were left out of the main trial.

3 ADHD — adults in the lab, not a children’s protocol

A 2026 sham-controlled study (pbm-20261003-27^[5]) gave adults one session of 1064 nm laser over the right prefrontal area. Active light was linked with a small attention-task difference and higher prefrontal oxygen signals than sham. No adverse events were listed. The trial was registered after it had started. It is one session in adults.

A second adult study (pbm-20261003-29) used seven days of the same kind of laser in 48 people. Memory and attention scores rose, but the abstract does not describe a sham group, so practice alone could explain part of it. A rat paper (pbm-20261003-28) is not a child result.

We did not index a randomized trial of transcranial red light in children with ADHD. A narrative review that mentions ADHD alongside autism (pbm-20261003-ndd-narrative) is a map of small studies, not a combined proof.

4 Seizures and “brain growth” — say the correlation carefully

Families sometimes hear that near-infrared light “calms the seizure brain” or “helps the brain grow.” Here is the careful version.

Seizures. Reviews such as pbm-20261003-epi-review discuss animal work and propose a future idea. Rat and mouse papers (for example pbm-20261003-30 through pbm-20261003-35, and the earlier kainic-acid card) report fewer seizure-like events in those animals. That is a real laboratory signal. It is not a human trial, and it is not evidence that a living-room lamp and a child’s seizures move together. One rat study even found that adding light to a high valproate dose made some seizure measures worse (pbm-20261003-32). Do not combine gadgets with antiseizure medicine on the basis of that paper.

Some people have photosensitive epilepsy, where certain flashing visible lights can trigger seizures. Near-infrared research lights are not the same thing as a strobe, and these papers do not prove that photobiomodulation either causes or prevents human seizures. The autism trial above excluded children with a seizure history. A separate brain-stimulation review (pbm-20261003-ndd-nibs-sr) mentions a seizure during magnetic stimulation in ADHD, which is a different tool, and still a reason stimulation studies use strict safety rules when epilepsy risk is high.

Brain growth. None of the child studies we indexed measured neurogenesis or brain volume as an outcome. If a website says light “grows the brain,” that sentence is ahead of the evidence. A change on a behavior checklist is not a picture of new brain tissue. Treat marketing that jumps from a cell-energy theory to a cure as a claim, not a result.

5 Cerebral palsy, rehab rooms, and home lamps

Occupational therapists and physio teams do use lasers in some clinics, especially on tight muscle. What we could verify are research papers, not a professional-society guideline that says occupational therapists should use red light. We did not tag any new card as occupational-therapy guidance, because the papers do not name occupational therapy even when they describe hospital functional training (Japan series pbm-20261003-17 and pbm-20261003-18).

Those series and several Brazil and Egypt muscle studies report less stiffness or better mouth opening for a while. Methods are often weak: no switched-off device, or training at a different hospital used as the comparison. The clearest placebo try on the calf (pbm-20261003-cp-pilot-rct^[6]) improved in both the light group and the device-off group, with no meaningful difference between them. A 2026 scoping review of light aimed at the head for pediatric cerebral-palsy spasticity (pbm-20261003-cp-tpbm-scope^[7]) calls that use unproven.

Home devices. One autism chart review really was at home (pbm-20261003-23), with parent training and no control group. A two-patient PANS/PANDAS note (pbm-20261003-37) mixed a home 1070 nm device with neurofeedback, so the light cannot take the credit. A pediatric clinic blog (pbm-20261003-39^[8]) tells parents that panels and handheld lamps are not proven for autism, ADHD, detox, or “brain function” in children, and that eyes need protection. Shop pages that promise a cure are claims. We did not find a named family’s public diary that we could verify and archive in this pass.

6 Suggestions for parents and caregivers

• Keep the therapies and medicines that a clinician already prescribed. Red light is not a replacement.
• If you are curious, bring the specific device and the specific paper — not a slogan — to the child’s own doctor, neurologist, or therapist. Ask what was measured, how many children, and whether there was a sham.
• Be extra cautious if the child has seizures, a photosensitive epilepsy diagnosis, eye disease, or a photosensitive drug. Do not experiment with pulsing lights or helmets on your own.
• Do not look into therapeutic LEDs or lasers. The cerebral-palsy muscle trial that used light also used eye protection.
• Treat “brain growth,” “cure,” and “detox” wording as marketing until a study measures those things in children and shows them.
• A lower score on an autism or irritability checklist in 10–30 people can be meaningful to those families and still be too fragile to buy a device for. Ask who paid for the study and who scored the child.
• Pain, a burn, a new behavior change you cannot explain, or more seizures after any new device is a reason to stop and call the care team — not a reason to “push through.”