Rosemary and Carnosic Acid

Supplements · §543

Rosemary is the herb on your shelf that quietly does several things. Half a gram twice a day cut anxiety and depression scores in a clinical trial of patients already on an antidepressant. Smelling it improves memory and alertness on a cognitive test. Rubbed on a balding scalp twice a day for six months, it grew back as much hair as a pharmacy hair-loss spray. None of these are blockbuster effects on their own — but each is real, the safety profile is essentially zero, and the whole thing costs less than a coffee a month. The catch is dose: at a quarter teaspoon of dried leaf per session it helps, at a tablespoon it hurts.

Do · Daily Evidence Mixed Chapter Supplements

The strongest signal is on mood — two trials, including one in patients with clinical depression, show meaningful reductions in anxiety and depressive symptoms at 500 mg twice a day. Sleep quality and memory came along for the ride. Skin and hair benefits are slower but show up in placebo-controlled trials over three and six months respectively. The honest framing: a small, broad, cheap intervention with a clean biological story, not a transformation.

The chemistry doing the work is mostly one molecule: carnosic acid, a fat-soluble compound that makes up roughly 3–4% of the dried leaf by weight. It sits dormant until your cells are under oxidative stress — when free radicals are running too hot — and then it switches on. The switch is a master regulator inside the cell called Nrf2, the same pathway broccoli sprouts and exercise tap into. Carnosic acid frees Nrf2 to enter the nucleus and turn on dozens of the body's own defence genes. It crosses into the brain too: peak brain levels reach about a tenth of plasma Mirza et al. 2023.

Two side stories matter. The smell of rosemary carries a compound called 1,8-cineole that's absorbed through your nose and ends up in your blood — and the amount in your blood tracks how well you do on a memory test Moss et al. 2003. And one of the related carnosic-acid molecules in the extract blocks the same enzyme finasteride blocks for hair loss — the one that converts testosterone to DHT in the scalp Panahi et al. 2015. Three different effects, three different routes in.

What the trials actually show

The strongest signal is on mood and anxiety. Two randomised trials — one in stressed university students, one in patients already being treated for major depression — used the same dose and found the same thing: rosemary on top of normal life made people less anxious and less depressed within a month or two.

On memory and focus, the cleanest finding is for the elderly. A crossover trial in 28 adults averaging 75 years old tested four single doses against placebo. The lowest dose helped; the highest dose hurt. We come back to that dose curve below — it's the single most important practical detail in the entry.

The rosemary aroma effect is real but smaller. In a 144-person study, people doing a memory battery in a cubicle scented with rosemary scored higher on memory quality and self-rated alertness than those in a control cubicle. Lavender did the opposite, which rules out "any nice smell helps you concentrate" as the explanation Moss et al. 2003.

For skin, a 2025 trial of women aged 40–65 took a daily rosemary-extract supplement for 12 weeks. By week 8 their skin was visibly better than the placebo group on overall quality grading; by week 12 the difference held up on dullness, texture, redness, and pore size Draelos et al. 2025. For hair, 100 men with male-pattern baldness used either rosemary essential oil or minoxidil 2% on their scalp for six months. Both groups grew significantly more hair; the difference between them was not statistically significant. The rosemary group itched less Panahi et al. 2015.

How to actually use it

What you take depends on what you're chasing. Cooking with rosemary is good for you and contributes a little of the mechanism, but it doesn't get you anywhere near the doses below.

For mood, sleep, anxiety: 500 mg of dried rosemary leaf, twice a day, with food. This is the dose both clinical trials used. Capsules from any reputable supplement brand work; loose dried leaf in a tea infuser does too. Give it a month before deciding.

For cognition and focus: stay under 1.5 grams a day total. The trial-effective dose for memory in older adults was a single 750 mg. A standardised rosemary extract capsule (look for one with 5–10% carnosic acid) at roughly 100 mg of carnosic acid per day is the supplement-aisle equivalent.

For hair: rosemary essential oil diluted in a carrier oil (jojoba, coconut), massaged into the scalp twice a day for six months. The trial used a commercially prepared rosemary tonic; any reputable EO works if you dilute properly (a few drops per teaspoon of carrier).

For the smell-it-and-think-better effect: a few drops of rosemary EO in a diffuser, run for five minutes before whatever you're trying to focus on.

The dose that hurt in the elderly trial was six grams in a single dose — about a heaping tablespoon. That's well above what any sensible supplement label calls for, but several "rosemary extract" products on the market deliver doses near or over the threshold per serving. Check the label.

The cooking-amount safety profile is essentially clean. The three real cautions are all about concentration: medicinal-dose supplements during pregnancy, the interaction with blood thinners, and high doses of the essential oil in people who have a seizure disorder.

Three things the internet gets wrong

"If a little works, more works better." No. The single most important practical fact about rosemary is the dose curve: low doses help cognition, high doses hurt it Pengelly et al. 2012. A heaping tablespoon of dried leaf in one sitting made elderly subjects worse on a memory test than placebo. Stay under a gram and a half a day.

"Rosemary oil works as well as Rogaine." The trial that everyone cites compared rosemary essential oil to minoxidil 2%, which is the weaker of the two FDA-approved doses. The standard hair-loss treatment today is minoxidil 5%. So the honest version is: rosemary oil works as well as the low-dose pharmacy spray for male-pattern baldness in a six-month trial of 100 men. That's still a real finding — it's just not what the marketing usually says Panahi et al. 2015.

"It cures Alzheimer's." There's a striking 2025 paper showing that a modified version of carnosic acid rescues memory and synaptic damage in a mouse model of Alzheimer's Banerjee & Lipton 2025. It's genuine science and worth watching. But no human trial in any form of dementia has been done. Don't put a relative with cognitive decline on rosemary capsules and call it treatment.

What changes, and when

Nothing dramatic happens in week one. The first thing most people notice — and it's noticeable — is sleep. Pittsburgh Sleep Quality Index scores improved over a single month at the trial dose Nematolahi et al. 2018, and the felt experience is the small one: falling asleep a little easier, fewer 3 a.m. wake-ups. The mood lift trails the sleep lift by a few weeks. The partner notices something before you do — you're not snapping at the dishwasher, you're not catastrophising the email. By month two the depression and anxiety scales pick it up Azizi et al. 2022.

On skin, the line of sight is three months. At eight weeks the placebo-controlled trial saw global skin quality improve; by twelve weeks the specific markers — that dull look, the rough texture, the visible redness — were measurably better Draelos et al. 2025. It's not a glow-up in the mirror; it's the colleague who hasn't seen you in a month saying you look well-rested, and meaning it.

On hair, the line of sight is six months — that's how long the trial ran. Patience is the active ingredient. People who quit at week eight saw nothing because nothing visible was happening yet Panahi et al. 2015.

None of these are alone enough to define your year. Together they're a small persistent tailwind across the bits of life everyone notices — the morning face, the afternoon focus, how the evening lands.

Why it doesn't work for some people

Wrong dose. The two ways this fails are mirror images. You sprinkle a teaspoon on the Sunday roast and wait for the depression to lift — you're under-dosed by a factor of ten. Or you read the dose curve as "more is more" and take a tablespoon a day — and now you're worse off on a memory test than the people taking nothing Pengelly et al. 2012.

Wrong form for the goal. Smelling rosemary improves a memory test you take five minutes later; it does not treat depression. Topical oil grows scalp hair; it does not improve your sleep. The protocol section above pairs each goal with its delivery method — pick the one that matches what you're after.

Expecting a caffeine-style kick. Rosemary is a slow drumbeat, not a stimulant. The people who notice anything at all are usually tracking something — a sleep log, a mood scale, a hair-count photo every month. If your bar is "did I feel something today" you'll conclude nothing happened.

Two adjacent threads worth knowing about. Sulforaphane (the broccoli-sprout compound) hits the same Nrf2 pathway as carnosic acid with stronger human trial data — the direct comparator if you care about the antioxidant mechanism rather than the mood and aroma effects. Minoxidil at the standard 5% concentration remains the better-tested option for male-pattern baldness when you're past the cosmetic-experimentation stage. And on blood sugar, the mechanism story for rosemary in diabetes is strong in animals but the human trials are too small and too mixed to recommend it for that purpose yet Naimi et al. 2017.

Substance and claimed effects

Rosemary — Salvia rosmarinus, formerly Rosmarinus officinalis — is an evergreen Mediterranean shrub of the Lamiaceae family. The dried leaf has been used as a culinary spice and folk medicine for memory, mood, and digestion for at least two millennia. Modern interest centres on a small set of bioactive constituents: the phenolic diterpenes carnosic acid (CA) and its derivative carnosol; the caffeic-acid ester rosmarinic acid (RA); and aromatic monoterpenes, principally 1,8-cineole and α-pinene. Carnosic acid alone accounts for ~3–4% by mass in high-quality dried leaf and is the dominant antioxidant. The claimed effects in scope for this entry are: acute and short-term cognitive enhancement (attention, memory, alertness); anxiolytic and antidepressant effects with sleep-quality improvement; antioxidant / anti-inflammatory activity via the Keap1–Nrf2 transcriptional pathway, with downstream consequences for blood-glucose homeostasis, advanced-glycation-end-product (AGE) accumulation, and (preclinically) neurodegeneration; topical effects on androgenetic alopecia comparable to low-dose minoxidil; and skin-quality changes (dullness, roughness, erythema) with oral high-CA extract Birtic 2015, Naimi et al. 2017.

Evidence by addressing question

mechanism

Three mechanistic strands carry most of the modern literature.

Nrf2 activation by carnosic acid. CA is a pro-electrophile: at rest it is stable and lipophilic, but oxidation by reactive oxygen species converts it to an orthoquinone that S-alkylates specific cysteine residues (notably Cys151, Cys273, Cys288) on Keap1. This frees Nrf2 to translocate to the nucleus and drive expression of phase-II antioxidant and anti-inflammatory genes (HO-1, NQO1, glutathione synthetic enzymes). The pathway is feedback-coupled — CA only fires when oxidative stress is already present, which sidesteps the off-target activation seen with constitutive Nrf2 inducers Mirza et al. 2023, de Oliveira 2016. CA crosses the blood-brain barrier; brain levels reach ~10% of plasma at peak. RNA-interference knockdown of Nrf2 abolishes CA's neuroprotection in vitro, confirming the pathway is causal, not incidental Mirza et al. 2023.

5α-reductase inhibition by 12-methoxycarnosic acid. A specific carnosic-acid derivative present in rosemary extract inhibits 5α-reductase, converting testosterone to dihydrotestosterone (DHT) — the same enzyme finasteride blocks. This is the proximal mechanism behind rosemary's hair-growth effect in androgenetic alopecia Panahi et al. 2015.

Cholinergic / olfactory pathway for acute cognition. 1,8-cineole, the principal monoterpene in rosemary aroma, is absorbed transnasally and inhibits acetylcholinesterase in vitro. Plasma 1,8-cineole correlates dose-dependently with cognitive-test performance after rosemary-aroma exposure in humans, providing a plausible chemical (not purely psychological) basis for the aroma effect Moss et al. 2003.

evidence

Acute cognition (aroma). The Moss et al. 144-participant Northumbria trial randomised healthy adults to perform the Cognitive Drug Research battery in cubicles scented with rosemary, lavender, or no odour. The rosemary group showed significantly higher overall quality of memory and secondary-memory factor scores than control; alertness ratings were also higher. Lavender impaired both, ruling out a generic "scent equals arousal" explanation. The effect is modest but reliably replicated in subsequent classroom and prospective-memory studies Moss et al. 2003.

Oral cognition (elderly). Pengelly et al. tested four single oral doses of dried rosemary powder (750, 1500, 3000, 6000 mg) against placebo in 28 adults aged ~75 in a crossover design with the same CDR battery. The 750-mg dose produced a statistically significant improvement in memory speed; 6000 mg significantly impaired performance. A clear inverted-U dose response — the central practical finding of the entry Pengelly et al. 2012.

Mood, anxiety, sleep. Nematolahi et al. randomised 68 university students to 500 mg rosemary leaf powder twice daily or placebo for one month. Beck Anxiety Inventory, Beck Depression Inventory, and Pittsburgh Sleep Quality Index all improved significantly versus placebo. Azizi et al. (2022) extended this clinically: in a double-blind RCT, patients with major depressive disorder already on SSRIs received rosemary 500 mg twice daily or placebo for 8 weeks as adjunctive therapy. HADS-Anxiety and BDI-II scores decreased significantly in the rosemary arm; memory improved as a secondary outcome Nematolahi et al. 2018, Azizi et al. 2022.

Skin (oral). Draelos et al. randomised women aged 40–65 with moderate-to-severe skin dullness and roughness to a daily rosemary-extract dietary supplement or placebo for 12 weeks. At week 12 the rosemary arm showed significant improvement vs placebo in dullness, roughness/texture, erythema, and pore size, plus global skin-quality grading at weeks 8 and 12. Mechanism proposed: deglycation of AGE crosslinks plus Nrf2-mediated antioxidant defence in skin Draelos et al. 2025.

Hair (topical). Panahi et al. randomised 100 men with androgenetic alopecia to topical rosemary essential oil or minoxidil 2% solution twice daily for 6 months. Hair counts increased significantly in both arms with no statistically meaningful difference at 6 months; the rosemary arm reported significantly less scalp itching. The trial used minoxidil 2% (not 5%), so the comparison is to the weaker active dose, not the gold standard Panahi et al. 2015.

Blood sugar. Animal models are extensive and consistent: rosemary extract and CA reduce fasting glucose, improve insulin sensitivity, activate AMPK and PPAR-γ in HepG2 cells, and inhibit α-amylase/α-glucosidase. Human evidence is sparse and small: a single open-label trial with 3 g/day powdered leaf for 4 weeks reported ~21% reduction in fasting glucose plus lipid improvements; a 4 g/day RCT in non-alcoholic fatty liver disease found no significant glucose effect vs placebo. The honest read: mechanism strong, human data preliminary Naimi et al. 2017, Bao et al. 2020.

Alzheimer's / neurodegeneration. Banerjee & Lipton (2025) developed diAcCA, a di-acetylated pro-drug of CA with ~20% greater oral bioavailability and stability. In the 5xFAD transgenic mouse model of Alzheimer's, three months of oral diAcCA produced synaptic rescue, improved Morris water-maze performance, and reduced amyloid plaques, phospho-tau aggregates, astrocytic and microglial inflammation. No human trial exists yet; the authors note CA's existing FDA Generally Recognized as Safe status supports near-term clinical translation Banerjee & Lipton 2025.

protocol

Dosing splits cleanly by product form:

Culinary use. Any amount used in cooking is safe and contributes mechanism, but is unlikely to reach the doses below.
Dried leaf powder, oral. 500 mg twice daily is the dose used in both mood/anxiety RCTs Nematolahi et al. 2018, Azizi et al. 2022. The Pengelly elderly cognition trial used a single 750-mg dose Pengelly et al. 2012; the dose-response inverted U suggests staying at or below 1.5 g/day.
Standardised CA extract. No human trial has formally established a dose. Animal effective doses translate (via 12.3 body-surface-area scaling for rat-to-human) to roughly 50–100 mg/day CA, which is what most commercial high-CA rosemary extracts deliver per serving.
Aroma. Diffused essential oil at the level used in the Moss trial (4 drops, room equilibrated for ~5 min before task) is effective acutely.
Topical (hair). The Panahi protocol used commercially-prepared rosemary essential oil in a base, applied to the scalp twice daily for 6 months Panahi et al. 2015.

contraindications

Two genuine and one cautionary signal:

Pregnancy. Medicinal doses (i.e. supplemental) of rosemary are traditionally contraindicated in pregnancy due to historical emmenagogic and possible abortifacient use; culinary amounts are safe. The mechanistic evidence is older and the modern human data are absent, but the precautionary line holds.
Seizure disorder. High-dose essential oil (especially topical or oral) has documented case reports of breakthrough seizures, attributed to camphor content. Aroma/culinary use is not implicated. Avoid concentrated supplemental and essential-oil ingestion in epilepsy.
Anticoagulants. Rosemary may potentiate warfarin via vitamin-K antagonism interactions; clinical reports are anecdotal but consistent enough to flag.

At normal supplement doses (≤1.5 g/day dried leaf or ≤100 mg CA), the safety profile is excellent — CA is FDA Generally Recognized as Safe as a food preservative Banerjee & Lipton 2025.

misconceptions

"Rosemary aroma boosts memory" — true but specific. The Moss effect is on quality of secondary memory and alertness, not raw IQ; the magnitude is modest. The popular reading ("smelling rosemary makes you smarter") overstates the finding Moss et al. 2003.

"More is better" — false. Pengelly's inverted-U dose response is the single most important practical caveat: at 6000 mg, oral rosemary impaired cognition in elderly subjects. Several commercial extracts deliver doses near or above this threshold per serving Pengelly et al. 2012.

"Rosemary oil equals minoxidil for hair loss" — partially true, partially marketed. The Panahi trial showed equivalence to minoxidil 2%, which is the weaker FDA-approved dose; minoxidil 5% (the current standard) is more potent. The trial was 6 months in a 100-participant Iranian cohort with self-assessment as a major outcome. Equivalence to minoxidil 2% is a real finding but routinely inflated in cosmetic marketing to "as good as Rogaine" Panahi et al. 2015.

"Anti-Alzheimer's herb" — premature. The Banerjee–Lipton 2025 mouse work is striking, but no human Alzheimer's trial has been conducted. Carnosic acid's neuroprotective profile is among the most promising in the natural-product literature, but extrapolation from a 5xFAD mouse to a human disease takes years Banerjee & Lipton 2025.

alternatives

For each consequence rosemary touches, there are stronger alternatives — rosemary's appeal is the breadth at low burden, not the depth on any one axis. Anxiolytic/antidepressant adjuncts with stronger evidence: SSRIs themselves (the rosemary trial was adjunctive to them), saffron extract, omega-3, exercise. Cognitive enhancers with deeper evidence in elderly populations: aerobic exercise, sleep optimisation, hearing-aid use. Topical AGA: minoxidil 5%, oral finasteride, dutasteride. AGE-driven skin ageing: sunscreen, retinoids, glycaemic control. Carnosic-acid-specific Nrf2 activation: sulforaphane (broccoli sprouts) has more rigorous human trial data and is the more direct comparator at the molecular level.

failure-modes

The two reliable failure modes are under-dosing (culinary use of dried rosemary on chicken is unlikely to deliver the 500–1500 mg/day needed to reproduce trial effects) and over-dosing on essential oil (assuming "more is more" with concentrated EO and either inducing GI distress, scalp irritation, or — at extreme doses — seizure in susceptible people). A third, more subtle failure: expecting a single perceptible "kick" the way one notices caffeine. Rosemary's effects are aggregate and modest; the people who notice them at all are usually tracking sleep, mood, or anxiety over weeks, not a Tuesday-afternoon focus boost.

practicalities

Dried rosemary leaf is available in any supermarket spice rack at ~$5 for a year's supply. Standardised rosemary extract capsules (typically 5% CA) cost roughly $15–25 for a 60–90 day supply at the trial doses. Topical rosemary essential oil for hair use costs $10–20 per bottle, lasting 2–3 months at the Panahi protocol. The supplement is among the cheapest in the catalogue with documented human effects.

stakes

Rosemary is not a high-stakes entry — its absence does not cause disease. The "stakes" framing here is the opportunity cost of an unusually cheap, low-effort, low-side-effect intervention with multiple modest documented benefits being passed over because the popular health discourse is calibrated to single-mechanism blockbusters.

payoff

At the 500-mg-BID dose, the documented payoffs in human trials are: a measurable reduction in anxiety and depressive symptoms over 4–8 weeks (in students and in MDD patients on SSRIs); improvement in sleep quality on PSQI; secondary memory gains Nematolahi et al. 2018, Azizi et al. 2022. With oral high-CA extract, 12-week skin-quality improvement (dullness, roughness, erythema) Draelos et al. 2025. With topical EO, 6-month hair-count increase in AGA comparable to minoxidil 2% Panahi et al. 2015. None of these is transformative individually; their sum is the entry's case.

out-of-scope

Cardiovascular endpoints (small flow-mediated-dilatation trial exists but is too preliminary). Cancer prevention (carnosic acid has interesting in-vitro and in-vivo anti-cancer signals across multiple tumour lines, but the data is animal-only — this warrants either its own entry or a future revision). Detailed cooking applications. Rosemary-genus essential oils for arthritis and pain (topical anti-inflammatory use exists but the evidence base is thin).

The credibility range

Optimist case

Rosemary delivers a coherent multi-mechanism upgrade for almost no cost: Nrf2 activation tunes endogenous antioxidant defence (a master pathway implicated in ageing, neurodegeneration, metabolic disease); 1,8-cineole drives acute cognitive enhancement via cholinergic and arousal pathways; the dose tested in two independent RCTs improved anxiety, depression, and sleep with no side effects; topical use rivals a dermatologist-prescribed alopecia treatment; the 2025 Alzheimer's pro-drug work suggests the molecular target may yet prove disease-modifying in humans. The downside is essentially zero at sensible doses, and the substance is in every kitchen. A defender argues that the reason this isn't a household supplement is purely incentive structure — there is no pharmaceutical interest in a herb that costs $5/year.

Skeptic case

The human-trial base is thin (one elderly cognition crossover with n=28, two anxiety/depression RCTs with n=68 and modest size, one skin trial, one hair trial). The dose-response is non-monotonic (Pengelly's inverted U), and the cognition wins are small. The headline Alzheimer's work is in a 5xFAD mouse, a model with poor translation history. The minoxidil-comparator trial used the 2% formulation, not the 5% standard, and was conducted by a single research group with no large independent replication. Carnosic acid's bioavailability and stability are limited (which is why a pro-drug is being developed). The mood/anxiety effects could plausibly be placebo magnified by a meaningful intervention ritual. The skeptic asks for a 200-participant RCT with biomarker endpoints before recommending anything beyond cooking.

Author's call

The honest landing: rosemary at 500 mg–1500 mg dried leaf daily is a small, real, broad intervention with an exceptional cost/benefit ratio and unusual mechanistic coherence across the dimensions it touches. The evidence rating sits at 2 (multiple small RCTs plus strong mechanism, no large pivotal trial) and controversy is low because nobody in the field is fighting about rosemary — they're ignoring it. The article should recommend it as a low-stakes do, name the modest size of each effect honestly, flag the inverted-U dose response prominently, and resist the temptation to ride the Alzheimer's-pro-drug narrative into reader-facing prose where it doesn't yet belong.

Stakeholder and incentive map

Supplement industry. Rosemary extract is a mid-tier nutraceutical, much smaller market than turmeric/curcumin or omega-3. Some interest from "natural nootropic" stack vendors. Low marketing spend relative to its evidence base.
Food industry. Major commercial interest, but as a preservative — high-CA rosemary extract is used in industrial-scale meat and oil-stability applications. The food industry has driven much of the CA chemistry research.
Pharma. Until 2025, almost no pharma interest. The Banerjee–Lipton diAcCA paper is a serious academic translation effort; if it advances to clinical trials, the dynamic flips.
Cosmetics. Rosemary essential oil is a TikTok-era hair-care darling, riding the Panahi 2015 trial. Marketing routinely overshoots the evidence (claims of "as good as minoxidil 5%" or "scientifically proven for all hair loss" are not what the trial showed).
Skeptic counter-incentive. Modest. The evidence-based medicine community generally classifies rosemary as "low evidence, low harm, do what you want." Nobody is incentivised to debunk it.

Population variability

Responder profiles are not yet well characterised in trials. Plausible variability:

Baseline oxidative-stress status. Nrf2-activator effects (CA, sulforaphane) are generally larger in populations with higher baseline oxidative load — older adults, metabolic-syndrome patients, smokers. Healthy young adults may see smaller absolute effects.
Age. Pengelly's elderly cohort showed dose-response on cognition; comparable formal data in younger adults is absent. Anti-anxiety effects replicate across young (students) and older (MDD) populations.
Sex. The Draelos skin trial was women-only (40–65). The Panahi hair-loss trial was male-only. No clear sex-difference signal exists at the molecular level.
Genetic variation in CYP and Nrf2. Pharmacogenetic variability for CA metabolism is not characterised. Nrf2 promoter variants (NFE2L2 SNPs) plausibly modulate response but no rosemary trial has stratified on them.
Concurrent SSRI use. Azizi et al. specifically tested rosemary as an adjunct to SSRIs and saw incremental benefit — so the substance does not appear to be redundant with that mechanism Azizi et al. 2022.

Knowledge gaps

No large (n>200), multi-centre, biomarker-anchored RCT of standardised rosemary extract in any indication.
Human pharmacokinetics of CA from realistic oral doses (most data are rat).
Dose-response above 1.5 g/day dried leaf is undefined in humans — the toxicology gap that matters for over-the-counter supplement labelling.
No published replication of the Panahi minoxidil-2% comparison.
No human Alzheimer's, MCI, or biomarker trial of CA or diAcCA — the 2025 pre-clinical work is at the dawn of a translational arc that could play out over a decade.
Topical-versus-oral comparison for skin endpoints is unstudied.
What evidence would change the call: a 200+ participant RCT showing either (a) clinically meaningful PHQ-9 effect at 500-mg BID, or (b) MMSE preservation in MCI, would shift the entry from "low-stakes do" to "broadly recommended."

Scope and the brief. The brief named cognition, oxidative stress, blood sugar, and skin. The article covers cognition, oxidative stress (via Nrf2 mechanism), and skin in full; it deliberately downweights blood sugar because human trial data are too thin to recommend rosemary as a glycemic intervention. The animal and mechanism evidence is solid (covered in the research dossier) — but the single human trial showing a 21% fasting-glucose reduction was open-label with 3 g/day, and the 4 g/day NAFLD RCT found no significant glucose effect vs placebo. Mentioning blood sugar in the article body would either oversell it or require a hedged paragraph that doesn't earn its space. Flagged in out-of-scope instead.

Hair gets a non-trivial mention despite the article's mood-led framing. The Panahi 2015 minoxidil-comparator is the single most viral piece of rosemary research and the most likely reason a reader landed on this entry. Ignoring it would feel evasive. The article handles it twice: once as a real finding in evidence, once as a corrected misconception (minoxidil 2%, not 5%).

Alzheimer's framing. The Banerjee & Lipton 2025 diAcCA paper is striking and tempting to lead with. Held it back deliberately: 5xFAD mouse models have a poor translation record, no human Alzheimer's trial has been run, and the article would tilt aspirational on thin evidence if the dek leaned that way. Flagged in misconceptions as something to watch, not something to use.

Rating calls. Mood was the hardest score — a 3 is the right level for two small RCTs with consistent direction and one of them in clinical depression on top of an SSRI, but a defensible reading is also a 2 (small sample sizes, no large replication). Landed on 3 because the MDD trial adds clinical weight the student trial alone wouldn't. Focus at 2 is conservative against the dose-curve risk; the aroma effect alone would be a 1, the Pengelly cognition finding alone a 2, and combining them does not warrant a 3 because effect sizes are small and the optimal dose is narrow.

No dream narrative. Overall score lands around 37 — below the obligatory threshold. The honest hook for this entry is breadth-at-low-cost, which reads better in clarity register than in aspiration register. A dream cascade would force tone that the modest effect sizes can't carry.

Separate-entry candidates. Carnosic-acid-specific anti-cancer signals (multiple in-vivo tumour lines, no human trials) and sulforaphane (the Nrf2-activator with the cleanest human data) both deserve their own entries.

Future links. Once entries exist: sulforaphane / broccoli sprouts, minoxidil for AGA, SSRIs as a class, sleep quality interventions.