Substance and claimed effects

Red light therapy — clinically called photobiomodulation (PBM) and historically low-level laser therapy (LLLT) — is the application of red (roughly 630–700 nm) and near-infrared (roughly 800–1100 nm) light at non-thermal doses to skin and underlying tissue. Delivery devices range from clinical lasers and FDA-cleared LED arrays to consumer panels (Joovv, Mito, Red Light Rising), hair-growth helmets (HairMax, Capillus, iRestore), and small targeted wands. The category-defining feature is wavelength inside the so-called optical window where hemoglobin and water absorb least and red/NIR photons penetrate deepest — millimeters for red, up to two centimeters for NIR Avci et al. 2013. Anders, Lanzafame and Arany formalised photobiomodulation therapy as the consensus term in 2015 to cover both laser and LED delivery Anders et al. 2015.

This entry covers the beyond-the-skin consequences specifically named in the brief: joint pain (osteoarthritis, tendinopathy, neck pain, rheumatoid arthritis), wound healing (acute and chronic), inflammation reduction, hair regrowth in androgenetic alopecia, and post-exercise recovery. Cosmetic skin effects (wrinkle reduction, collagen synthesis) overlap and are covered as a beauty_direct consequence. Transcranial PBM for depression and cognition, intranasal and intra-articular laser delivery, and circadian/morning-light exposure (a different wavelength regime and a different mechanism entirely) are explicitly out of scope.

Evidence by addressing question

mechanism

The canonical mechanism is photon absorption by cytochrome c oxidase (Complex IV of the mitochondrial electron transport chain). Tiina Karu's group at the Russian Academy of Sciences identified cytochrome c oxidase as the primary chromophore for red and near-infrared light in the 1990s–2000s; her 2005 paper established that nitric oxide bound to the binuclear copper-heme center is photodissociated by red/NIR photons, lifting NO inhibition of mitochondrial respiration Karu et al. 2005. The immediate downstream effects: increased mitochondrial membrane potential, brief and transient reactive oxygen species (ROS) burst, increased ATP synthesis, and modulation of redox-sensitive transcription factors (NF-κB, AP-1) that govern inflammation and cell proliferation. Hamblin's 2017 review synthesises the anti-inflammatory cascade — at low doses PBM produces a brief ROS signal that paradoxically downregulates chronic inflammation via Nrf2-mediated antioxidant response and reduced pro-inflammatory cytokine expression (TNF-α, IL-1β, IL-6) Hamblin 2017.

Secondary mechanisms documented across cell types: increased nitric oxide bioavailability driving local vasodilation, calcium ion flux changes triggering downstream signaling, stem cell activation (satellite cells in muscle, dermal papilla cells in hair follicles, fibroblasts in skin), and modulation of light-sensitive ion channels (opsins, TRP channels) in non-mitochondrial chromophores — a still-debated layer of the mechanism story Hamblin 2017 Mussttaf et al. 2019.

A defining mechanistic feature is the biphasic dose response: too little energy produces no effect, optimal doses produce the expected biological response, and excessive doses produce inhibition — the Arndt-Schulz curve applied to photons. Huang, Chen, Carroll and Hamblin's 2009 review is the field's standard reference; the inverted-U shape has been replicated in dozens of cell and animal studies and explains why higher-power or longer-duration treatment is not better and may be worse Huang et al. 2009.

evidence

Musculoskeletal pain. The strongest aggregated evidence is for knee osteoarthritis. Stausholm and colleagues' 2019 systematic review and meta-analysis of 22 randomised placebo-controlled trials in knee OA found that LLLT delivered at WALT-recommended doses produced clinically meaningful pain reduction ~15 mm on a 100 mm visual analogue scale versus placebo, with effects persisting up to 12 weeks after the treatment course ended Stausholm et al. 2019. Chow, Johnson, Lopes-Martins and Bjordal's 2009 meta-analysis in The Lancet — 16 RCTs in non-specific neck pain — found LLLT reduced pain immediately after the treatment course by ~20 mm on a 100 mm VAS and effects persisted up to 22 weeks Chow et al. 2009. Bjordal et al.'s earlier 2003 dose-response review across chronic joint disorders established that dose matters: trials delivering doses within the WALT-recommended range showed positive effects, trials outside the range did not — explaining much of the field's apparent heterogeneity Bjordal et al. 2003. Tumilty et al.'s 2010 meta-analysis of LLLT for tendinopathy found significant pain reduction in lateral epicondylitis and Achilles tendinopathy when delivered at recommended doses Tumilty et al. 2010.

For rheumatoid arthritis, the Cochrane review (Brosseau et al. 2005, 5 RCTs) found short-term pain reduction (~1.1 cm on 10 cm VAS) and morning stiffness improvement with LLLT versus placebo, with effects lasting up to 4 weeks; the authors flagged dose heterogeneity as the main caveat Brosseau et al. 2005. Cotler, Chow, Hamblin and Carroll's 2015 narrative review compiles the broader MSK pain evidence — neck pain, low back pain, OA, tendinopathies — and discusses why LLLT is increasingly displacing NSAIDs in some clinical guidelines outside the US Cotler et al. 2015.

Hair growth. Two industry-funded but FDA-cleared RCTs anchor the hair evidence. Lanzafame et al. 2013 randomised 41 men with androgenetic alopecia (Norwood-Hamilton classifications IIa–V) to a 655 nm laser comb device (HairMax) or sham for 16 weeks, with the active group showing a 39% increase in terminal hair density versus the sham group Lanzafame et al. 2013. The parallel 2014 trial in 42 women with female pattern hair loss (Ludwig-Savin I-4 to II-2) found a comparable terminal-hair-density increase versus sham at 17 weeks Lanzafame et al. 2014. Avci, Gupta, Clark, Wikonkal and Hamblin's 2014 review compiles roughly a dozen RCTs of LLLT for androgenetic alopecia, noting the field's consistency on direction-of-effect but variability in absolute magnitude — and that essentially all positive trials enrolled early-to-mid stage hair loss, not advanced bald patterns Avci et al. 2014. The FDA cleared the first laser comb for hair growth in 2007 (510(k) clearance, not full approval — meaning safety and substantial equivalence, not proof of efficacy at the New Drug Application bar).

Skin / collagen. Wunsch and Matuschka 2014 randomised 113 adults to LED treatment (611–650 nm or 570–850 nm, twice weekly for 30 sessions) or untreated control. Active groups showed significant increases in intradermal collagen density measured by ultrasound (collagen density change clearly visible on B-scan) and reductions in periorbital wrinkle depth Wunsch & Matuschka 2014. Jagdeo, Austin, Mamalis et al.'s 2018 systematic review of LED phototherapy in dermatology surveys 31 RCTs across acne, wound healing, photoaging, and psoriasis — finding consistent positive effects on photoaging and acne (red+blue combinations) and emerging but more variable evidence for psoriasis Jagdeo et al. 2018.

Wound healing. Mosca, Ong, Albasha, Bass and Arany's 2019 clinical review compiles the wound-healing evidence base — diabetic ulcers, pressure ulcers, surgical incisions, burns — concluding that PBM is a useful adjunct that accelerates closure times and reduces wound area, with the strongest evidence in diabetic foot ulcers and post-surgical scars Mosca et al. 2019. Avci, Gupta, Sadasivam et al.'s 2013 skin review covers the mechanistic case in dermal fibroblasts and keratinocytes (increased proliferation, increased collagen synthesis, accelerated re-epithelialization) Avci et al. 2013.

Exercise recovery and performance. Leal-Junior, Vanin, Miranda, de Carvalho, Dal Corso and Bjordal's 2015 systematic review with meta-analysis of 39 RCTs found that pre-exercise PBM (applied before activity) reduces post-exercise creatine kinase, lactate, and delayed-onset muscle soreness, and modestly increases time-to-fatigue and peak torque Leal-Junior et al. 2015. Ferraresi, Huang and Hamblin's 2016 review focuses on the muscle-tissue mechanism — increased cytochrome c oxidase activity, increased ATP availability, mitigation of oxidative damage during exercise — and concludes the effect is real but small Ferraresi et al. 2016.

Parameter quality concerns. Zein, Selting and Hamblin's 2018 review enumerates the parameter space that determines whether a given trial detects an effect: wavelength, irradiance (mW/cm²), fluence (J/cm²), beam profile, pulse vs continuous, treatment frequency, tissue depth, and chromophore concentration. They argue that much of the literature's apparent heterogeneity is bad parameter reporting rather than absent effect — many negative trials used doses outside the biphasic optimal window Zein et al. 2018.

protocol

The World Association for Laser Therapy (WALT) maintains dose recommendations by anatomical site and condition. Typical clinical protocols deliver 4–10 J/cm² per treatment point for joint pain, with 4–8 points around the joint, two to three sessions per week for 2–4 weeks. For hair regrowth, the cleared devices deliver 3–5 J/cm² per scalp area, three sessions per week, for at least 16 weeks before assessing response. For consumer panels used for skin or recovery, manufacturers typically recommend 10–20 minute sessions at 20–100 mW/cm² irradiance, 3–5 sessions per week.

The biphasic dose response means session length and distance from the panel both matter. Standing too close to a high-irradiance panel for too long can push into the inhibitory zone; standing too far away delivers a sub-threshold dose. Most consumer panel manufacturers publish irradiance-at-distance specs (e.g. 100 mW/cm² at 6 inches, dropping to 30 mW/cm² at 18 inches) that let the user calculate session duration to hit a target fluence Huang et al. 2009 Zein et al. 2018.

contraindications

Absolute or near-absolute: direct eye exposure at high irradiance (proper goggles are recommended for any near-infrared device above ~50 mW/cm² at close range; bright red LEDs are uncomfortable but not damaging at consumer-panel distances, though chronic NIR exposure to the cornea/lens is a real risk for daily users). Active malignancy at the treatment site — PBM stimulates cell proliferation and angiogenesis; the conservative position is to avoid direct irradiation of known tumors or recent excision sites until cleared by oncology. Photosensitizing medications — methoxsalen (PUVA), some antibiotics (doxycycline, ciprofloxacin), retinoids, St John's Wort — increase skin photosensitivity; while the visible/NIR window has different chromophore physics than UV, caution and patch-testing are standard. Pregnancy — abdominal PBM is avoided as a precaution; no documented harm but no studies establishing safety either.

Relative cautions: thyroid (avoid prolonged anterior-neck irradiation given the thyroid's high vascularity and the absence of safety data), recent corticosteroid injections (potential interaction with the anti-inflammatory pathway), tattoos in the treatment area (some inks absorb at the operating wavelength and can heat). Cotler et al. 2015 Avci et al. 2013.

misconceptions

Three durable misconceptions in the consumer marketing and biohacker discourse:

• "More is better." The biphasic dose response makes this false in a specific, replicated way — sessions twice as long don't double the effect, they overshoot the optimal window Huang et al. 2009.
• "It's the heat that works." The clinical effect is photochemical, not thermal — the same wavelengths delivered as heat (an infrared sauna) work via a different mechanism and on different timescales. PBM panels run cool by design; if a device gets hot, that's wasted power, not therapeutic dose.
• "LED is just a weaker laser." Coherent (laser) and non-coherent (LED) light at the same wavelength and dose produce comparable biological effects in most studies — the field's consensus has moved away from coherence as a load-bearing variable Kim & Calderhead 2011 Anders et al. 2015. LED panels are cheaper, broader-area, and adequate for most consumer use cases; medical-grade lasers are still preferred for point-targeted clinical work where depth and precision matter.

audience

The evidence skews toward populations where the literature has clustered:

• Knee OA patients with moderate disease. Stausholm meta-analysis enrolled mostly Kellgren-Lawrence grade II–III; effect sizes were largest in this band Stausholm et al. 2019.
• Early-stage androgenetic alopecia. Lanzafame trials enrolled Norwood-Hamilton IIa-V (men) and Ludwig-Savin I-4 to II-2 (women) — the early/mid bands where viable miniaturized follicles remain. Advanced baldness (complete follicle loss) does not respond; the therapy rescues struggling follicles, not absent ones Avci et al. 2014.
• Tendinopathy and neck pain — moderate symptom severity. Severe cases benefit less than moderate; very mild cases regress to mean and obscure the effect Chow et al. 2009 Tumilty et al. 2010.
• Photoaged adult skin (Fitzpatrick I–III). Most skin trials enrolled lighter skin types; the relative effect on darker skin (greater melanin absorption competing with target chromophore absorption) is less characterised Wunsch & Matuschka 2014.

For exercise recovery, the effect appears across trained and untrained subjects but is more reliable in the recovery markers (CK, lactate, soreness) than in performance gains Leal-Junior et al. 2015.

alternatives

For knee OA pain: NSAIDs (faster, smaller magnitude per dose, GI/cardiovascular costs), intra-articular corticosteroid injection (large short-term effect but cartilage degradation risk with repeat injection), physical therapy (comparable or larger effect, much higher effort), weight loss (the dominant intervention, much harder), eventual joint replacement. Cotler et al. argue PBM is a useful adjunct that reduces NSAID dependence in some chronic cases Cotler et al. 2015.

For androgenetic alopecia: topical minoxidil (5% for men, 2–5% for women; better-evidenced and cheaper but requires twice-daily application indefinitely), oral finasteride (men only; strongest pharmacological effect, sexual side-effect concerns), oral dutasteride (off-label, stronger than finasteride), hair transplant (definitive but surgical and expensive). PBM is most often combined with minoxidil ± finasteride rather than used alone Avci et al. 2014.

For wound healing: standard wound care (debridement, moist dressings, offloading for diabetic foot ulcers) is the floor; PBM is an adjunct that accelerates closure but does not replace primary care Mosca et al. 2019.

For exercise recovery: sleep, protein adequacy, and tapered training are the dominant interventions. PBM, cold water immersion, massage, and compression are second-tier adjuncts with overlapping evidence quality Ferraresi et al. 2016.

failure-modes

The most common reasons users report "tried it, didn't work":

• Wrong wavelength. Heat lamps (broadband infrared), incandescent red bulbs, and many cheap "red light" devices emit predominantly outside the 630–700/800–1100 nm therapeutic windows. The therapeutic windows are defined by tissue absorption; the visible-red colour is only a proxy Zein et al. 2018.
• Insufficient irradiance or session length. A 5 mW/cm² device used for 2 minutes delivers far below the biphasic optimal — the user is in the sub-threshold zone and sees nothing.
• Excessive session length. The user, reasoning that more must be better, runs 45-minute sessions and pushes into the inhibitory zone — also seeing nothing or worse Huang et al. 2009.
• Inconsistent use. Hair-growth trials used 3 sessions per week for 16+ weeks before assessment. Most consumer attempts give up at 4 weeks.
• Wrong indication. Advanced bald patterns, end-stage joint disease, fresh acute injuries (where inflammation is the wanted signal, not the suppressed one) — wrong targets.

practicalities

Hardware market segments:

• Consumer LED panels — $200–$2,000 range. Joovv, Mito Red Light, PlatinumLED, Red Light Rising are the better-known brands. Larger panels deliver more body-area per session; smaller panels are cheaper and targeted. Irradiance at distance is the spec that matters; total wattage is marketing.
• Hair growth devices — $200–$3,000. FDA-cleared options include HairMax (the original laser comb), Capillus, iRestore, Theradome. Cleared devices have published RCT support; many uncleared alternatives have not been tested.
• Targeted wands/handhelds — $100–$500. Useful for joint and tendon work where panel coverage is wasted.
• Clinical lasers — used by physical therapists, chiropractors, dermatologists; typically billed at $50–$150 per session and a 6–12 session course.

Time cost: a 10–20 minute session, 3–5 days per week, for at least 8 weeks before assessing response (longer for hair). Most users report it slots into morning or evening routine; many do it during meditation or while reading. Eye protection (provided goggles or simply closed eyes) is the friction point for whole-body sessions Cotler et al. 2015.

payoff

Timescale of felt effect, by indication:

• Acute MSK pain — measurable within the first treatment course (4–8 sessions, 2–3 weeks). Stausholm meta-analysis: pain reduction at end-of-treatment, persisting up to 12 weeks Stausholm et al. 2019.
• Chronic MSK pain — slower; 4–8 weeks of consistent sessions before noticeable steady-state reduction.
• Wound healing — accelerated closure visible within days for acute wounds, weeks for chronic ulcers Mosca et al. 2019.
• Hair regrowth — at least 16 weeks before terminal-hair-density changes are measurable. Subjective perception ("more hair") typically lags terminal-hair counts by another 4–8 weeks Lanzafame et al. 2013.
• Skin (wrinkles, tone) — visible at 12–16 weeks of consistent use; collagen biopsy changes visible at 30 sessions Wunsch & Matuschka 2014.
• Exercise recovery — same-day to next-day effect on soreness and CK after pre-exercise application Leal-Junior et al. 2015.

out-of-scope

Transcranial PBM (1064 nm helmets for depression, cognitive performance), intra-articular fiber-optic laser delivery, blue light (different chromophore physics, used for acne and seasonal mood), UV-based therapies (vitamin D, psoriasis PUVA), and circadian light exposure (the morning-sunlight entry's territory — different mechanism, different goal) are deliberately excluded from this entry.

The credibility range

The optimist case

The mechanism is one of the best-characterised in physiology — Karu's cytochrome c oxidase model is replicated across cell types and species and has held since the late 1990s Karu et al. 2005. The clinical evidence is real and growing: a Lancet meta-analysis for neck pain (Chow 2009, 16 RCTs), a Cochrane review for rheumatoid arthritis (Brosseau 2005), a systematic review for knee OA showing 12-week effect persistence (Stausholm 2019, 22 RCTs), FDA clearances for hair growth devices backed by published industry RCTs (Lanzafame 2013, 2014), and a consistent meta-analytic signal for exercise recovery (Leal-Junior 2015, 39 RCTs). The therapy is unusually clean on adverse events: across hundreds of trials and decades of clinical use, serious adverse effects are essentially absent at appropriate doses. The biphasic dose response — initially seen as a confounder — has become a unifying principle explaining why trials outside the optimal window failed, and why trials inside it succeeded. WALT, NICE (in some indications), and clinical practice in Europe and Brazil have moved ahead of US clinical guidelines in adoption.

The skeptic case

The literature is heterogeneous in ways that don't resolve into a single tidy story. Wavelengths, irradiances, fluences, beam areas, treatment frequencies, and session counts vary across trials; meta-analyses must either pool aggressively (losing signal) or restrict to narrow parameter bands (losing power). Many positive trials are industry-funded; the Lanzafame hair trials were sponsored by the device manufacturer. Effect sizes for MSK pain are real (15–20 mm on a 100 mm VAS) but modest — comparable to NSAIDs at peak effect — and require 2–3 sessions per week for weeks to deliver. Consumer panels marketed at non-clinical irradiances often miss the therapeutic window entirely; the gap between what RCTs deliver and what a Joovv panel delivers at 18 inches is large enough that consumer outcomes likely under-deliver versus the published clinical data. The wellness influencer and biohacker subcultures have over-extended the claims — life extension, telomere preservation, "mitochondrial rejuvenation" as a general anti-aging frame — well past what the literature supports. Active malignancy is a genuine concern for chronic whole-body irradiation by enthusiasts. And the parameter complexity is a perpetual escape hatch: any negative trial can be blamed on suboptimal dose, which makes the theory hard to falsify Mussttaf et al. 2019.

The author's call

Real, modest, dose-dependent, with the strongest evidence in three specific indications (knee OA pain, androgenetic alopecia at early/mid stages, post-exercise recovery markers) and a credible but smaller evidence base for tendinopathy, wound healing, skin/collagen, and rheumatoid arthritis. Not a panacea; not a longevity intervention in the catalogue sense; not snake oil either. The right framing for the reader is "useful adjunct in specific conditions, at the right device, dose, and schedule — not a stand-alone cure for any of them." Evidence score is in the middle band (3) — multiple RCTs and meta-analyses, but parameter heterogeneity and industry funding pull the rating below the level of, say, statins for cardiovascular risk. Controversy score also middle (3) — the biohacker–establishment gap is wide; mechanistic skeptics and parameter-optimist defenders argue past each other; and the FDA's 510(k) clearance language is regularly mis-cited by marketers.

Stakeholder and incentive map

• Device manufacturers — Joovv, Mito Red Light, PlatinumLED, HairMax, Capillus, iRestore, Theradome, Thor Photomedicine, Multi Radiance. Consumer panel market has scaled rapidly since ~2018; pricing and irradiance claims compete openly but few brands publish independent third-party irradiance verification.
• Clinical practitioners — primarily physical therapists, chiropractors, sports medicine, some dermatologists. WALT (World Association for Laser Therapy) is the field's main professional body; it publishes the canonical dose tables and runs the international clinical training. Adoption is far higher in Europe, Brazil, and Australia than in the US.
• Researchers — Michael Hamblin (formerly Harvard/Wellman, now Laserowl) is the most-cited synthesiser; the Karu lineage (Russian Academy of Sciences) is the mechanism foundation; Bjordal and Lopes-Martins on dose-response and clinical translation.
• Regulators — FDA has cleared (510(k), not approved) multiple devices for specific indications: hair growth, temporary pain relief, wound treatment. Clearances are at the safety-and-substantial-equivalence bar, not full efficacy. EU CE marking is comparable. This regulatory band is often misrepresented in marketing.
• Wellness / biohacker community — large podcast audience (Huberman, Attia, Rogan-adjacent), strong device sales, advocacy of whole-body daily protocols. Has overshot the science on systemic claims (longevity, mood, "mitochondrial health") while sometimes accurate on local pain and hair claims.
• Skeptics — academic dermatology and rheumatology (especially in the US) have been slower to adopt; reasonable position that parameter heterogeneity, small trials, and industry funding warrant caution; less reasonable position that the mechanism is implausible (it isn't).

Population variability

Responder profile shifts noticeably by indication:

• Hair growth — responders are early/mid-stage androgenetic alopecia with miniaturised but living follicles. Norwood-Hamilton VI–VII (advanced male pattern) and Ludwig-Savin III (advanced female pattern) do not respond meaningfully; the therapy can't regrow follicles that no longer exist. Genetic responder/non-responder bands within early-stage AGA are still being characterised Avci et al. 2014.
• Skin tone — Wunsch's trial and most dermatology trials enrolled Fitzpatrick I-III. Higher melanin in Fitzpatrick V-VI absorbs more incident red light, reducing the dose reaching dermal fibroblasts; the practical implication is longer sessions and higher tolerance, but the literature characterising this directly is thin.
• Joint pain — inflammatory vs degenerative — RA trials (inflammatory) and OA trials (mechanical + secondary inflammation) both show effects, but the mechanism may differ. RA's inflammatory cytokine suppression is a tighter mechanistic match to PBM's anti-inflammatory signal; OA's effect may be more about secondary synovitis than primary cartilage repair Hamblin 2017.
• Exercise context — Leal-Junior's meta-analysis pooled trained and untrained subjects; the recovery-marker effect (CK, lactate, soreness) is more reproducible than the performance-marker effect (peak power, time to exhaustion). Trained athletes squeezing already-optimised recovery may see less marginal benefit than recreational lifters Leal-Junior et al. 2015.
• Diabetic vs non-diabetic wound healing — diabetic foot ulcers show some of the largest PBM effect sizes in the wound literature, plausibly because the baseline impaired microcirculation has more room to be rescued Mosca et al. 2019.
• Age — most adult-age trials replicate across age bands. Pediatric and elderly evidence is thinner; safety appears comparable.

Knowledge gaps

What hasn't been settled:

• Optimal dose for consumer panels. Clinical lasers and clinical LED arrays have published WALT doses; consumer panels at standard-use distances may be sub-optimal, but few head-to-head trials of consumer hardware exist.
• Long-term whole-body exposure safety. Decades of localised clinical use show no significant adverse signal at therapeutic doses. Daily whole-body sessions by enthusiasts are a different exposure pattern with much shorter follow-up; chronic cumulative dose effects on melanocytes, lens (cataract), and thyroid are not well characterised.
• Combination therapy synergy. Minoxidil + LLLT, finasteride + LLLT, NSAIDs + LLLT — combinations are common in practice but rigorously trialled in only a few studies.
• Mechanism beyond cytochrome c oxidase. Opsins, TRP channels, and direct photon effects on water structure are all candidate secondary mechanisms; the relative contribution to the therapeutic effect is unresolved.
• Coherence (laser) vs non-coherence (LED) at deep targets. Field consensus is that LED is adequate for most indications, but a residual question remains for deep-tissue targets where laser focusability may matter.
• Effect persistence after treatment cessation. Stausholm's 12-week persistence and Chow's 22-week persistence in MSK pain are encouraging but raise the question of how long re-treatment intervals can run before effect decays.

Evidence that would change the call: large multi-centre RCTs powered to detect dose-response across standardised consumer-hardware protocols (would tighten the practical guidance); long-term safety cohort following daily whole-body users for 5+ years (would settle the chronic-exposure question); a definitive negative trial in a high-claimed area (e.g. transcranial PBM for depression) would tighten the boundaries of credible claim.

Scope decisions. The brief named joint pain, wound healing, inflammation, hair growth, and recovery — all covered. Inflammation is woven through the mechanism section and the joint-pain subsection rather than carrying its own subsection, because the trial literature treats it as the layer underneath pain reduction rather than as an endpoint in its own right (no consumer-facing inflammation biomarker the reader can act on). Skin/cosmetic effects (collagen, periorbital wrinkles) are kept inside scope because they overlap with the same hardware and dose ranges the brief's indications use — readers buying a panel for joint pain almost always also point it at their face.

Deliberately excluded — separate-entry candidates.

• Transcranial photobiomodulation — the 1064 nm helmet / forehead-pad literature for depression, cognitive performance, traumatic brain injury, and Alzheimer's. Different chromophore depth, different dose tables, different risk-benefit calculus, different evidence quality. Warrants its own entry (transcranial-photobiomodulation).
• Infrared sauna — the heat-based intervention is a different mechanism (heat-shock proteins, cardiovascular load) despite wavelength overlap. Likely its own entry (infrared-sauna).
• Morning sunlight / circadian light exposure — different chromophore (melanopsin), different goal (clock entrainment), different wavelength regime. Belongs in its own entry under the same category (morning-sunlight).
• Blue light for acne and UV phototherapy for psoriasis — distinct categories worth their own entries.

Future cross-links to wire when those entries exist: transcranial-photobiomodulation, morning-sunlight, infrared-sauna, minoxidil, finasteride, knee-osteoarthritis.

Rating difficulties.

• beauty_cumulative (3). Score is conditional on indication: the +39% terminal hair density finding (Lanzafame 2013, 2014) only applies to early/mid-stage androgenetic alopecia responders. For advanced bald patterns the score is effectively 0. I scored against the responder cohort because that is who actually buys the device — the non-responder is unlikely to complete a 16-week course. Considered 2 (small contribution averaged across the unselected population); landed on 3 because for the realistic user it is a meaningful long-term aesthetic improvement.
• health_short_term (3). Strongest meta-analytic evidence in the knee OA / neck pain block; the 12–22 week post-course persistence is unusual enough to push the score above 2. Held at 3 rather than 4 because the magnitude (15–20 mm on a 100 mm VAS) is real but not transformative.
• energy (2). The Leal-Junior meta-analysis (39 trials) covers exercise recovery markers, which is real but narrow; this is not a general-vitality intervention. Scored against the active reader; would be 0–1 for the sedentary one.
• evidence (3) vs controversy (3). Both held in the middle band. Evidence has multiple meta-analyses but parameter heterogeneity is a real signal-quality problem (Zein 2018; Mussttaf 2019) and the Lanzafame trials are manufacturer-sponsored. Controversy reflects the wide gap between WALT-aligned European/Brazilian clinical practice, US academic dermatology/rheumatology skepticism, and biohacker overselling — three credible camps talking past each other.
• longevity (0). The mitochondrial mechanism gets stretched into longevity claims in biohacker discourse; the trial literature does not support a mortality or healthspan endpoint. Honest zero.

Hard calls during the write.

• Whether to gender-restrict audience. Hair-loss patterns differ between sexes and the trials enrolled them separately, but the rest of the indications are not gender-specific and the female-pattern hair trial replicated the male result. Left audience open.
• Whether focus should be non-zero given the transcranial PBM cognition literature. Kept at 0 because transcranial is out of scope for this entry; admitting non-zero focus here would either pull transcranial back in (wrong scope) or claim the brain effect for chest-panel use (unsupported).
• Contraindications field. Only pregnancy made the cut from the closed vocabulary — the warning callout covers the rest (thyroid, photosensitizers, active tumor site, eyes) at the per-paragraph level, which is the right granularity since none of them is a global "this person should not use red light therapy" stop.

Citation note. Heavy reliance on Hamblin-authored reviews (Hamblin 2017, Huang 2009, Avci 2013/2014, Ferraresi 2016) reflects his role as the field's dominant review author rather than a sourcing failure; the underlying primary trials (Stausholm, Chow, Lanzafame, Wunsch, Brosseau Cochrane, Leal-Junior) carry the load-bearing claims.