Red light therapy and brain health: what does the research actually say?

A different kind of application

Every other blog on this site covers conditions where red or near-infrared light is applied to the skin - and the skin is more or less the target. Brain health is different. The research here is asking whether light can penetrate through the scalp and skull, reach cortical tissue, and produce measurable changes in how the brain functions. That is a harder thing to demonstrate, and the evidence accordingly has a different character.

The field studying this is called transcranial photobiomodulation - tPBM for short - and it has expanded considerably over the past decade. Clinical trials have now been published on memory in healthy older adults, working memory in young adults, mild cognitive impairment (MCI), Alzheimer's disease, traumatic brain injury, and depression. A 2023 systematic review from the Chinese University of Hong Kong, covering 35 human tPBM studies, found that 82.9% reported positive improvements in cognitive function. A 2025 meta-analysis of RCTs found statistically significant effects across global cognition, working memory, attention, and executive function.

Those are encouraging numbers. But the picture is more layered than that headline suggests, and this blog tries to give you the full version - what the research found, what it cannot yet confirm, and what it means for someone considering a red light device at home.

How light reaches the brain - and what it does when it gets there

The first question worth answering is whether light can penetrate that far at all. Bone, dura mater, and cerebrospinal fluid are not transparent. Near-infrared light in the 800-1100 nm range does, however, have the deepest tissue penetration of any visible or near-infrared wavelength - studies have demonstrated that a small percentage of photons at these wavelengths reach the cerebral cortex. The proportion is low and depends on the power density of the device, the thickness of the skull at the application site, and the wavelength used, with wavelengths towards the longer end of the NIR range - around 1064 nm - penetrating furthest due to reduced scattering in biological tissue.

Once photons reach neurons, the mechanism is the same as elsewhere in the body. The primary target is cytochrome c oxidase (CCO) - the terminal enzyme in the mitochondrial electron transport chain. Neurons are among the most metabolically demanding cells in the body, requiring a continuous supply of ATP. When CCO absorbs NIR photons, ATP production increases, nitric oxide is released (which improves local blood flow), and reactive oxygen species levels are modulated. In compromised or hypometabolic neurons - the kind found in ageing brains, areas of inflammation, or regions affected by neurodegenerative processes - this stimulation is thought to be most useful, because it is addressing a genuine energy deficit rather than simply accelerating already-efficient cells.

Beyond the direct mitochondrial effect, tPBM has been shown to increase cerebral blood flow, reduce neuroinflammation by modulating microglial activity, support neurogenesis, and increase levels of brain-derived neurotrophic factor (BDNF) - a protein that promotes the survival and growth of neurons and is closely tied to learning and memory. These are not hypothesised mechanisms: they have been measured in both animal models and human studies.

What the clinical studies show, condition by condition

The tPBM literature covers several distinct populations, and the evidence varies in quality and confidence across them. The table below summarises the main areas, with the sections below going into the detail.

Healthy adults: working memory and cognitive performance

Young adults

The most controlled evidence for cognitive enhancement in healthy people comes from studies on young adults, where confounds from disease or medication are minimal. A 2019 systematic review and meta-analysis by Salehpour et al. - including researchers from Harvard Medical School and Massachusetts General Hospital - identified nine studies on tPBM in healthy adults. Of six full-text publications on young adults (aged 17-35) that could be pooled for meta-analysis, the combined effect size was SMD = 0.761 (95% CI: 0.573-0.949), indicating a medium-to-large beneficial effect on cognitive performance. The authors noted high heterogeneity between studies, which they attributed partly to variation in device parameters and the cognitive tests used.

A particularly clean piece of evidence from this category is a paper published in Science Advances by Wang et al., showing that 1064 nm tPBM applied to the right prefrontal cortex improved visual working memory capacity and produced corresponding changes in cortical electrophysiology (contralateral delay activity). The specificity matters here: applying the same wavelength to the left prefrontal cortex, or using 852 nm instead of 1064 nm, did not produce the same effect. That kind of wavelength and location specificity is the signature of a real physiological effect rather than noise.

Older adults

A 2022 study by Qu et al. (Neurophotonics) enrolled 61 healthy older adults in a repeated tPBM protocol (1064 nm, 12 minutes daily for seven days) and found improvements in working memory performance that persisted at a three-week follow-up assessment - suggesting that the effects are not simply acute and transient. A companion analysis by Hu et al. (IEEE J Biomed Health Inform, 2023) examined the neurophysiological mechanisms in the same cohort, finding corresponding reductions in bilateral cortical haemodynamic activation during the working memory task - an indicator of improved cognitive efficiency, where the brain achieves the same performance with less metabolic effort.

A 2023 systematic review on age-related cognitive impairment (Gao et al., Lasers Med Sci) pooled 11 RCTs and found a statistically significant moderate effect on global cognitive function in older populations (SMD = 0.51, 95% CI [0.162, 0.864], p = 0.004). The 2025 meta-analysis by Zhu et al. (Lasers Med Sci), covering RCTs up to October 2024, confirmed beneficial effects on global cognition (SMD = 0.66, p = 0.003), working memory (SMD = 1.41, p < 0.0001), attention (p = 0.01), and executive function reaction time (p = 0.005). Subgroup analyses showed that cognitively impaired subjects benefited across most domains, while healthy adults showed the clearest gains in attention.

Mild cognitive impairment: the population with the most consistent results

Mild cognitive impairment sits between normal age-related cognitive change and dementia. It is characterised by noticeable memory or thinking problems that do not yet significantly impair daily life, but which carry a meaningful risk of progression to Alzheimer's disease. It is also the population where tPBM research has produced some of its most consistent positive results.

The 2023 systematic review by Lee et al. (Ageing Res Rev) found that all nine studies on participants with subjective memory complaints, MCI, or dementia showed positive outcomes. A 2022 RCT by Papi et al. (Mazandaran University of Medical Sciences) enrolled 42 older women with MCI into real versus sham tPBM, finding significant improvements in both cognitive function (assessed by mini-mental state examination) and attentional performance on the Go/No-Go psychomotor task in the active group.

The most methodologically robust MCI study is a double-blind, placebo-controlled RCT by De Oliveira et al. (J Photochem Photobiol B, 2024, PMID 39423445) conducted with 93 older adults with MCI, including the lead author's group from the University of South Santa Catarina alongside Paolo Cassano from Harvard Medical School. Participants received tPBM or placebo over 60 days, with follow-up at 150 days. The active group showed improved cognition scores as the primary outcome. As a secondary finding, serum BDNF - a direct neuroplasticity biomarker - increased significantly in the tPBM group, providing a measurable biological correlate for the cognitive improvement. The authors note that lead investigator Cassano holds financial relationships with photobiomodulation companies, which the study discloses; readers weighing this result should factor that in.

A 2024 dose-response RCT by Lee and Chan (J Alzheimers Dis, PMID 39177599) enrolled 88 healthy older adults to compare single versus double-dose tPBM in a single session, finding that the higher dose produced greater improvements in cognitive efficiency as measured by functional near-infrared spectroscopy (fNIRS) - confirming that the effect follows a dose-dependent pattern, a hallmark of photobiological activity rather than placebo.

Alzheimer's disease: promising but still early

Alzheimer's disease is where the stakes are highest and where the evidence is most honestly described as emerging rather than established. As of a 2025 review by Blivet et al. (J Prev Alzheimers Dis, 2025), nine clinical studies have been published on tPBM in Alzheimer's disease, MCI due to AD, or dementia. Of these, four are double-blind RCTs versus sham. The picture they collectively paint is positive but imperfect.

Among the double-blind RCTs, one study by Berman et al. (2021) found no statistically significant difference between active tPBM and sham after 28 days of treatment - an important counterpoint that is sometimes omitted in more enthusiastic accounts. Open-label studies and case series have generally found positive results. A case series by Saltmarche et al. treated five patients with combined transcranial and intranasal tPBM for 12 weeks, reporting MMSE improvement of 2.60 points (p < 0.003) and ADAS-cog improvement of -6.73 points (p < 0.023), with a notable observation that cognitive performance declined sharply when treatment was stopped four weeks later.

The 2025 review by Blivet et al. summarised the overall picture clearly: all studies demonstrate safety, results are promising particularly in Alzheimer's disease, but the exploratory design of most studies and their inconsistent methodology means the current evidence does not yet support widespread clinical use. Several larger, better-powered RCTs are currently underway or recently completed, including a 2024 Japanese protocol (Yokoi et al., Front Neurol) enrolling 30 participants over 12 weeks. Those results will considerably clarify the picture.

One area of active mechanistic interest for Alzheimer's is the glymphatic system - the brain's waste clearance pathway, which runs primarily during sleep. tPBM has been proposed to support glymphatic drainage of amyloid-beta, one of the protein aggregates central to Alzheimer's pathology. This remains mechanistically plausible but not yet clinically demonstrated in humans.

Traumatic brain injury: consistent results, limited RCT data

Traumatic brain injury presents a compelling biological case for tPBM. The acute injury creates a zone of metabolically compromised but potentially salvageable neurons - cells that are alive but unable to produce sufficient ATP to function normally. The mechanism by which tPBM restores energy production in compromised cells maps directly onto this situation.

The 2023 systematic review by Lee et al. found that 7 of 8 TBI studies (87.5%) reported positive cognitive outcomes. A 2024 systematic review specifically on tPBM in TBI by Zeng et al. (Front Psychol, PMID 38952831) reviewed studies up to October 2023 and found consistent positive trends in cognitive rehabilitation outcomes, though the authors noted that most TBI studies use case series or uncontrolled designs rather than RCTs with sham controls, limiting confidence in the conclusions.

The foundational clinical work in this area comes from Margaret Naeser's group at the VA Boston Healthcare System and Harvard Medical School, who conducted a series of open-label and case series studies over more than a decade, documenting improvements in attention, memory, and executive function in patients with chronic mild TBI. These studies used LED devices applied to specific scalp locations corresponding to affected cortical networks. While the lack of sham controls is a limitation, the clinical improvements were observed in patients who had been symptomatic for months to years without spontaneous recovery - making a placebo explanation more difficult.

Depression and mood: a secondary but real connection

Depression is included here not because it is primarily a cognitive condition, but because the cognitive symptoms of depression - difficulty concentrating, impaired working memory, slowed thinking - are often what most affects quality of life, and because tPBM research has directly addressed this population.

A systematic review and meta-analysis on PBM for depression (Frontiers in Psychiatry, 2024) identified 11 RCTs. The pooled effect on depression severity was SMD = -0.55 (95% CI [-0.75, -0.35], I² = 46%), a medium effect size indicating a clinically meaningful reduction in depressive symptoms. This is consistent with the existing NovaThera blog on depression, which covers the evidence in more detail. The overlap with brain health is that improving prefrontal cortex function - which is often hypometabolic in depression - appears to address both the mood and cognitive dimensions simultaneously.

What the field has not yet resolved

A few things in the tPBM literature deserve honest attention rather than being glossed over.

Most clinical studies use specialist laser devices, not LED panels

The majority of tPBM research - particularly the more positive results - uses high-power laser devices delivering precisely calibrated doses of NIR light at specific scalp coordinates. Many studies use 1064 nm wavelength lasers at 250 mW/cm² for 12 minutes, or 810 nm devices delivering 1-10 J/cm² to specific cortical nodes. At-home red and NIR LED panels used for whole-body wellness deliver a different dosimetric profile: lower peak irradiance, broader wavelength ranges, and no specific scalp targeting. Whether the benefits demonstrated with clinical devices translate to general LED panel use has not been rigorously established.

This does not mean panels have no effect - multiple studies have used LED arrays rather than lasers, and positive results exist across both device types. But someone using a NovaThera panel near their head is not replicating the exact protocols that produced the strongest results. The effect, if present, is likely more modest.

Protocol standardisation is still missing

Wavelength, irradiance, pulse frequency, session duration, number of sessions, and scalp location all vary substantially across studies. The 2025 meta-analysis by Zhu et al. noted this as a major limitation, and the umbrella review by Kang et al. (Systematic Reviews, 2025) observed that most of the underlying meta-analyses were rated low quality by AMSTAR 2. The field is producing positive results despite this heterogeneity - but the optimal protocol for any specific cognitive goal is still unknown.

Publication bias is a reasonable concern

The 82.9% positive rate across 35 studies is striking. It likely reflects genuine effects, but publication bias - the tendency for studies with positive results to be published more readily than null results - cannot be excluded. The Salehpour 2019 meta-analysis explicitly noted that their trim-and-fill analysis suggested some studies with null results may be missing from the literature. The field would benefit from a pre-registered large-scale RCT with a well-powered sample.

Effect sizes and clinical meaningfulness vary

Statistically significant effects do not automatically mean clinically meaningful ones. The Kang umbrella review noted that only a limited number of the cognitive outcomes it examined showed effect sizes (SMD ≥ 0.5) that were likely to translate into meaningful real-world differences. The working memory effect (SMD 1.41) from the Zhu 2025 meta-analysis is large by any benchmark. The global cognition effect (SMD 0.66) is moderate. Attention and executive function effects are smaller. People should not expect dramatic transformation from light therapy - but the moderate effects observed in multiple well-designed studies are real and worth taking seriously.

What this means for someone using a red light panel at home

The honest picture is that the tPBM literature is encouraging - particularly for working memory, attention, and cognitive efficiency across both healthy ageing and MCI populations. The mechanisms are well-characterised, the direction of effects is consistent, and the safety profile is excellent. What is less certain is how much of that benefit transfers from clinical laser protocols to at-home LED panel use.

A few things are reasonably well supported. NIR wavelengths in the 810-850 nm range, delivered at normal panel distances for 10-20 minutes, produce detectable changes in prefrontal cortex oxygenation as measured by fNIRS. That is a biological effect, not nothing. Whether that translates into the working memory and attention improvements documented in clinical studies is a reasonable hypothesis but not yet directly tested with the specific device types most people use at home.

For people using red light panels as part of a broader wellness routine - for sleep, recovery, skin health, or general inflammation - positioning sessions so that NIR light reaches the head and neck is a low-cost addition that sits within the plausible range of the research. Targeted use for cognitive support specifically is where expectations should be calibrated against the honest evidence: promising, biologically credible, but not yet at the level of established clinical recommendation.