Does Red Light Therapy Really Work?

Every so often a treatment modality arrives that sounds too strange to be real. Shining infrared light through the skull to treat depression has that quality. It sounds like something sold out of the back of a wellness magazine, and for years mainstream neurology treated it that way. Then the mechanistic research piled up, the clinical trials started producing effect sizes, and transcranial photobiomodulation (tPBM) quietly became one of the most active frontiers in non-invasive brain stimulation. This article summarizes the current state of the science, including the wavelength debates, the depression trials out of Massachusetts General Hospital, the Alzheimer’s research coming out of Durham University, and the commercial devices now reaching clinicians, including the Neuradiant 1070 helmet made by the neurotechnology company Neuronic.

A note on why a psychotherapy practice cares about any of this. The emerging model of depression and anxiety is not a simple chemical imbalance. It is a picture of neuroinflammation, oxidative stress, and mitochondrial dysfunction in the frontal and limbic networks that regulate mood. That is the same territory trauma therapy works in from the psychological side. Understanding the biology helps clinicians and clients alike make sense of why some interventions work, why some fail, and why the brain’s energy supply matters as much as its stories.

The Core Mechanism: Light, Mitochondria, and Cellular Energy

The brain is roughly two percent of body mass and consumes about twenty percent of the body’s resting energy. Nearly all of that energy comes from ATP produced by mitochondria. Neural tissue is therefore exquisitely vulnerable to anything that degrades mitochondrial function, and mitochondrial deficits show up again and again in depression, anxiety, traumatic brain injury, and neurodegenerative disease.

Photobiomodulation targets that vulnerability directly. The consensus mechanism, developed over two decades of laboratory work, centers on cytochrome c oxidase, also called Complex IV of the mitochondrial electron transport chain. This enzyme contains copper and heme centers that absorb red and near-infrared light in the 600 to 900 nanometer range. When photons reach the enzyme, electron transfer accelerates, the proton gradient across the inner mitochondrial membrane strengthens, and ATP synthesis increases. Michael Hamblin’s foundational review, Shining Light on the Head: Photobiomodulation for Brain Disorders, remains a useful orientation to this literature, and a broader mechanistic overview is available through the National Library of Medicine’s review of photobiomodulation mechanisms and clinical applications.

A second mechanism involves nitric oxide. Under conditions of cellular stress and inflammation, nitric oxide binds to cytochrome c oxidase and blocks oxygen from its binding site, throttling energy production. Near-infrared light photodissociates that inhibitory nitric oxide. The freed molecule then acts as a vasodilator, increasing local cerebral blood flow while the enzyme returns to work. Near-infrared spectroscopy studies in living human prefrontal cortex have confirmed that as light dose accumulates, both cytochrome c oxidase activity and oxygenated hemoglobin rise together in a tight linear relationship.

Downstream of these immediate effects, transient calcium signaling, cyclic AMP, and mild reactive oxygen species act as second messengers that reach the nucleus and alter the expression of more than a hundred genes. The net shift is toward an anti-apoptotic, anti-inflammatory, pro-plasticity state, including upregulation of brain-derived neurotrophic factor, the growth factor most consistently implicated in recovery from depression.

The Wavelength Debate: 810 nm Versus 1070 nm

The field has split into two engineering camps, and the disagreement is genuinely interesting physics rather than pure marketing.

Longer wavelengths scatter less. Light around 1064 to 1070 nanometers passes through cranial bone more coherently than 810 nanometer light, which scatters heavily on contact. The counterargument is water. The brain is seventy to eighty percent water, and water absorbs light above roughly 950 nanometers far more aggressively than it absorbs 810 nanometer light. Critics of the longer wavelength argue that cerebrospinal fluid and brain parenchyma act as a water trap that converts 1070 nanometer photons to trivial heat before they reach cortical depth.

A 2025 cadaver study by Tittelmeier and colleagues, published in the journal Brain Stimulation, measured transmission directly through intact human head tissue with skin, bone, and dura present. The 810 nanometer helmet transmitted 0.71 percent of emitted light to the inner skull surface. The 1070 nanometer helmet transmitted 0.45 percent. That is roughly 2.2 times more light delivered by the shorter wavelength. The same study raised a harder question for the entire category: over 99 percent of light from every tested device was absorbed or scattered before reaching the inside of the skull, and the authors questioned whether low-power LED helmets of any wavelength deliver enough photons to activate cytochrome c oxidase at cortical depth. The 810 nanometer camp, led commercially by Vielight, has made this dosimetry argument central to its case, and their comparison of the two wavelengths is worth reading as an articulate statement of one side of the debate, with the understanding that it is a manufacturer’s document.

The 1070 nanometer camp answers with a different kind of evidence, which brings us to the Alzheimer’s research.

Depression: The ELATED Trials and the Threshold of Inefficacy

The clinical case for tPBM in major depressive disorder has been built largely by Dr. Paolo Cassano and colleagues at Massachusetts General Hospital through the ELATED series of trials. The ELATED-2 pilot trial, a randomized, double-blind, sham-controlled study published in 2018, delivered 823 nanometer light to the dorsolateral prefrontal cortex twice weekly for eight weeks and found medium to large antidepressant effect sizes on the Hamilton Depression Rating Scale.

The more instructive finding came later. A larger low-dose cohort, treated at parameters that delivered about 2.3 kilojoules per twenty-minute session, showed no separation from sham at all. Rather than sinking the modality, this null result established what researchers now call a threshold of inefficacy. Like transcranial magnetic stimulation and like medication, light therapy has a dose floor. Below a certain irradiance and total energy, nothing happens. Cassano’s group has published a systematic review of transcranial and systemic photobiomodulation for depression that lays out the efficacy, tolerability, and mechanism data, and the ongoing ELATED-3 trial is now testing substantially higher irradiances with PET imaging and EEG outcome measures. A related dosing study is registered at ClinicalTrials.gov.

Safety data across these trials has been consistently reassuring. Even daily sessions over twelve weeks did not increase adverse events, which suggests considerable headroom for more aggressive dosing protocols. Preclinical work supports the human findings: in chronic mild stress rodent models, both red and infrared light reversed anhedonia on the sucrose preference test, apparently by stabilizing mitochondrial respiration and interrupting the cycle of oxidative stress and neuroinflammation that the stress protocol induces.

Alzheimer’s Disease: Microglia, Amyloid, and the Case for 1070 nm

The most striking mechanistic work on the longer wavelength comes from Alzheimer’s models. A 2021 study published in Nature’s Light: Science and Applications demonstrated that 1070 nanometer light pulsed at 10 Hz produced highly specific changes in microglia, the brain’s resident immune cells, in Alzheimer’s model mice. Treated microglia shifted from a resting, branched morphology into an activated state, migrated to amyloid-beta plaques, engulfed them, and reduced overall plaque burden. The treatment also promoted angiogenesis and improved performance on spatial memory tests. Notably, the effect was frequency-dependent: 10 Hz pulsing produced the microglial changes while 40 Hz did not, which underscores how much the pulse parameters matter and how far the field is from settled protocols. Follow-up work in Advanced Science showed that the same light parameters shift microglia from an inflammatory to an anti-inflammatory phenotype whose exosomes reduce neuroinflammation and protect dendritic spines. The full text of the original microglia study is freely available through PubMed Central.

On the human side, the 1068 to 1072 nanometer band has been championed by Professor Paul Chazot at Durham University, whose in vitro work showed that 1068 nanometer light protected neuronal cells from amyloid-beta mediated death, apparently by stimulating chaperone proteins that maintain protein folding and support clearance of misfolded aggregates. A pilot trial led by Chazot and Dr. Gordon Dougal, covered by Genetic Engineering and Biotechnology News, found significant improvements in memory, motor function, and processing speed in healthy older adults using a self-administered 1068 nanometer helmet twice daily. Chazot is careful to say the field remains young and needs large randomized trials before anyone should treat this as established dementia care. His program has since advanced toward Phase 2b clinical work in the United States.

Neuronic and the Neuradiant 1070

This is where the research meets the market. Neuronic is a neurotechnology company whose flagship device, the Neuradiant 1070, is a transcranial helmet delivering 1070 nanometer near-infrared light through 256 LEDs arranged in four independently controllable quadrants. The quadrant architecture is the interesting part for clinicians. A provider can target the prefrontal cortex, or one hemisphere, rather than flooding the whole scalp, and pulse frequencies are adjustable from continuous wave up to 9999 Hz. Because neural networks oscillate at characteristic frequencies, alpha around 8 to 12 Hz and gamma around 40 Hz, pulsed light functions as a candidate entrainment mechanism as well as a metabolic one.

Neuronic pairs the device with quantitative EEG. A clinician maps a client’s brainwave patterns with qEEG, identifies regions of dysregulation, and programs the four quadrants with region-specific frequencies. For a practice like ours that already uses qEEG brain mapping to guide neurofeedback and therapy planning, this is a natural extension of the same logic: measure the individual brain first, then intervene specifically rather than generically.

The company has also invested unusually heavily in research for a consumer device maker. Its published research pipeline includes a King’s College London study on Long COVID brain fog that found improvements in chronic fatigue and cognitive symptoms, a UT Dallas protocol for treatment-resistant anxiety and depression, cognitive studies at Santa Clara University, an ADHD and virtual reality project at Bryn Athyn College, and the EVANTHEA randomized controlled trial in mild cognitive impairment and early-stage dementia. Professor Chazot, who first proposed clinical use of 1070 nanometer technology for neurodegenerative disease, joined Neuronic’s scientific advisory board in 2023. The device lineage runs back through the Quietmind Foundation’s early trials with Baylor Research Institute, some of the first human clinical work pairing transcranial light with neuropsychological and neuroimaging assessment. A Neuronic survey of the more than two hundred providers using the device in practice, published on the company’s own blog, reports the most common clinical populations are dementia, traumatic brain injury, ADHD, anxiety, and depression, with the honest caveat that the data are self-reported.

The Dosimetry Problem Nobody Should Skip

Here is the part a responsible summary cannot leave out. The single most important predictor of whether tPBM works is whether enough light actually reaches the cortex, and independent testing has raised real questions about whether current 1070 nanometer helmets clear that bar. Third-party measurements compiled in an independent 2026 device guide found the Neuradiant’s measured scalp irradiance in the range of 7 to 9 milliwatts per square centimeter against declared specifications of 20 to 40, while premium 810 nanometer devices measured at or above their stated output. Combine that with the Brain Stimulation cadaver data showing 1070 nanometer light transmits less efficiently through the head than 810 nanometer light, and you have a legitimate open question: are the impressive microglial and amyloid effects documented in cell cultures and mouse skulls actually reproducible through five to seven millimeters of human bone at consumer-device power levels?

To Neuronic’s credit, the company has responded by publishing its own measurement methodology and calling for transparency standards around total power, irradiance, and peak wavelength across the industry. That is the right direction. The lesson of the ELATED trials applies to every device on the market: below the dose threshold, the mechanism is irrelevant. Anyone considering a device purchase, whether clinician or consumer, should ask for independently measured irradiance figures, not just the number printed on the box.

What This Means for Therapy Clients

Photobiomodulation is not a replacement for psychotherapy, and the honest reading of the evidence is that it is a promising adjunct with an unsettled dosing science, not a finished treatment. But the framework behind it matters for how we think about healing. Depression and trauma are not only stories and cognitions. They live in inflamed tissue, in exhausted cellular energy systems, in brains that are metabolically depleted by chronic stress. Interventions that restore the brain’s energy supply, whether light, exercise, sleep, nutrition, or reduced inflammatory load, create the physiological conditions under which the psychological work of therapy can actually take hold. That is the same reason we use qEEG brain mapping, neurofeedback, and body-based trauma therapies alongside depth psychotherapy: the mind heals faster when the brain has the fuel to change.

If you are curious whether brain-based approaches make sense as part of your treatment, that is a conversation worth having with a clinician who can look at your whole picture rather than a single device.