Mind & body, in brief

Smell starts in a patch of tissue in the roof of your nose. Human DNA carries 391 working smell receptor genes, and that is the whole toolkit. [1]

There is no receptor for coffee. There is no receptor for lemon, or rain, or your grandmother's kitchen. Each one responds roughly, to many different molecules. [1]

The standard account is a pattern. A molecule switches on some receptors and not others, and which set lights up is the smell. That is how a few hundred parts cover far more than a few hundred smells. [1]

Each smell nerve cell picks exactly one receptor gene and silences all the others. In mice that is one out of more than 1,000, and a protein called TRIM66 does the silencing. [4]

Every cell that picked the same receptor sends its wire to the same bundle in the smell bulb, the first stop in the brain. A mouse bulb holds a median of 2,851 bundles, about 2.5 for each intact receptor gene. [9]

Camillo Golgi described those bundles in 1875. Nobody had counted them completely until this work; every earlier figure was an estimate. [9]

The receptor genes themselves were found in 1991 by Linda Buck and Richard Axel, who shared the 2004 Nobel prize in medicine for it. [5]

Two papers in the journal Cell in May 2026 redrew where the receptors sit. Each one occupies a set position along the roof of the mouse nose, rather than a broad zone picked at random. [2][3]

The two maps line up. Where a receptor sits in the nose predicts where its wires land in the brain, through one shared set of genes. [2]

The pattern account is under pressure from the other side too. In 2025 chemists found human receptors tuned tightly to single molecules: the cork taint in wine, the pepper note, grapefruit, ambergris. [1]

One human receptor, OR52E8, answers to a single acid in armpit sweat. Natural variants of it respond differently from one person to the next. [54]

About 60% of the human smell receptor gene family is broken by mutations, and not the same genes in each person. [7]

So two people in one room are not running the same equipment. Where you live explains 17 to 20% of the difference in how sensitive people are to odours, measured in 1,046 people across 19 places. [16]

What people find pleasant is more shared than that. In 909 people across 16 regions, region explained under 10% of the variation, and intensity varied more than pleasantness. [17]

The pattern arrives with no name attached. People correctly name under half of common everyday smells, and over 90% of familiar objects they can see. [6]

Smell takes the shortest route to feeling of any sense: only two nerve-cell junctions separate the nose from the brain's emotion circuits. [57]

It is also the sense health systems skip. About 22% of people have some loss of smell and about 5% have none at all, and no country's public health plan covers it. [12][13]

Why it matters. Watch whether the tightly tuned receptors found in 2025 turn out to be the exception or the rule; if many receptors are that specific, the map from molecule to smell gets far easier to build. [1][5]

In a US study of 3,525 people, 66% of the 1,563 previously infected who reported no change in smell scored as having reduced smell on a 40-odour test, 8.2% of them severely. Among those who did report a change, 79.8% did. [19]

Why it matters. How common smell loss is depends on whether you test people or ask them.

A review of 21 studies covering more than 4,000 people found losses lasting from several months to over two years, with distorted and phantom smells appearing during recovery. [20]

Why it matters. The recovery is not a return to normal so much as a slow, strange rebuild.

Between 26.5% and 46% of people who had COVID still had smell loss one year after infection, falling to 8.3% at two years. [21]

Why it matters. Most people do recover, and the tail is long enough to be a public health problem.

The standard treatment, olfactory training, is smelling four different scents for 10 to 20 seconds each, twice a day, for at least three months. [24]

Why it matters. It is the main thing medicine has to offer for a sense that a fifth of people have lost some of.

A meta-analysis of eight randomised trials found training effective, with the largest gain when it was combined with the supplement PEA-luteolin, and higher recovery rates for combinations generally. [23]

Why it matters. The sniffing is the part everyone can do and the smaller part of the effect.

Two 2025 meta-analyses, one of four trials in 198 people and one of seven studies in 789, found injections of the patient's own blood platelets beat placebo at 1, 3 and 12 months. Mayo Clinic reports 87% of its patients clinically improved a year on. [21][22][18]

Why it matters. The first treatment in decades to be added to a hospital protocol for smell loss.

Between 67% and 78% of people with chronic sinus inflammation have impaired smell, and the presence of polyps is among the strongest factors. [25]

Why it matters. The commonest cause of lost smell is not a virus, it is a blocked and inflamed nose.

A shortened five-item smell test was run in 1,472 people over 60 with no Parkinson's diagnosis. Scores fell with age, were worse in men, and tracked with a REM sleep disorder screen. [26]

Why it matters. Smell loss can precede the tremor by years, which is why a cheap test is attractive.

Of 1,044 people screened with a rapid smell test in Iran, 25% of those over 50 had smell loss with no identified cause, against under 10% of younger people. [27]

Why it matters. The screening pool is large, and most of those people will never develop Parkinson's.

In 94 people, only a detailed perceptual profile separated Parkinson's from other smell loss, telling the groups apart with 88% accuracy, rising to 94% when age and sex were matched. Between 75% and 90% of people with Parkinson's have smell loss. [28]

Why it matters. The finding is not that smell fades but that the same lemon stops being enjoyable.

Olfactory loss in Parkinson's reflects widespread degeneration across the movement and rear cortical networks, not damage confined to the nose. [29]

Why it matters. It explains why smell tracks with later thinking problems rather than only with the nose.

A review of olfactory testing in mild cognitive impairment found identification and discrimination the domains most impaired in Alzheimer's, and called for tests adapted to different cultures. [30]

Why it matters. A test built around Western smells fails everywhere else, and most such tests are.

In 5,142 older US adults followed 9.6 years, poor smell carried about twice the risk of coronary heart disease at year two, falling to no measurable difference by year 9.6. [31]

Why it matters. A signal that decays like that usually means something already under way, not a cause.

Smell disorders affect about 22% of people overall, but only about 2% of patients at specialist smell clinics are under 18. Age-specific normal ranges for children's tests barely exist. [32]

Why it matters. A child who cannot smell gas or spoiled food has the same safety problem as an adult.

Fourteen mixed martial arts fighters scored lower than 14 matched controls on the 40-item smell identification test, a statistically significant gap, with no difference in taste. [33]

Why it matters. The nerve fibres running to the brain pass through a thin plate of bone, and a blow shears them.

In 80 women, those between six and ten years past menopause were far less likely to score well on odour identification than women of reproductive age. [34]

Why it matters. One of the few studies looking at smell against a hormonal timeline rather than age alone.

In 228 adults, 58.3% of those over 50 had reduced smell against 19.8% of the younger group. A questionnaire correlated with formal testing, which costs around $30 a go. [35]

Why it matters. The gap between what testing costs and what a clinic will pay is why nobody is tested.

In mice, the inflammatory signal IL-17a builds up in ageing smell tissue and blocks its renewal. Blocking IL-17a restored nerve cell regeneration and reversed the tissue change. [36]

Why it matters. It reframes age-related smell loss as inflammation rather than simple wear.

In mice, a virus destroyed the smell lining, immune cells moved in and stayed, and bone formed where sensory tissue had been. The smell loss was long-term. [37]

Why it matters. A tissue that normally rebuilds itself can be replaced by something that cannot smell.

Glands beneath the smell lining that normally only produce mucus were shown to generate replacement support cells, but only after severe damage. [38]

Why it matters. A reserve nobody knew was there, held back until the usual repair route fails.

Single-cell profiling of olfactory stem cells found a resting group already primed for activation, with the genetic switches open around genes not yet switched on. [39]

Why it matters. It shows the tissue keeps a set of cells ready rather than starting from scratch.

Silencing the immature smell neurons in healthy mice impaired their ability to detect an odour, but not their ability to tell two odours apart. [40]

Why it matters. The half-built cells were assumed to be spare parts; they are a separate input stream.

A review links shifts in the nasal microbiome to chronic sinusitis, allergic rhinitis and post-viral smell loss, while stating the evidence is largely indirect. [41]

Why it matters. A live hypothesis that has not yet shown which way the arrow points.

In 106 people, 77% recognised an odour they had met one to three days earlier, and 55% of those recognitions brought back the episode attached to it. Emotional strength predicted recognition best. [42]

Why it matters. Recognising a smell and remembering what it belongs to are two separate steps.

A review of 24 studies found people worse at remembering places by smell than by sight, by a moderate margin. [43]

Why it matters. The sense that binds hardest to feeling is one of the weakest at telling you where you are.

In 22 people, smelling lemon while hunting for a lemon among other fruit improved speed and accuracy. A mismatched smell hurt the strongest searchers most. [44]

Why it matters. The senses are not searching separately; one primes what the other is looking for.

A personally pleasant smell slowed breathing, deepened it, lowered heart rate and raised heart rate variability. A personally pleasant piece of music did none of that. [45]

Why it matters. Smell is sampled by breathing, so it can reach the calming machinery that sound cannot.

A review of 21 studies in pregnancy found reduced nausea and anxiety. A review of 13 studies in new and expectant mothers found smaller effects on anxiety and sleep, with study-to-study disagreement of 88.7%. [46][47]

Why it matters. Disagreement that high means the average is describing very few of the studies in it.

In 90 adults, a commercial essential oil blend and a sage control both beat no aroma on memory and executive function tasks, measured alongside blood oxygen in the brain. [48]

Why it matters. The control condition was another smell, which is the harder and rarer comparison.

Inhaled lavender oil eased migraine-like pain in mice through a specific brain circuit running from the smell region to the hypothalamus. Blocking one receptor in that circuit abolished the effect. [49]

Why it matters. A mechanism means the effect can be tested rather than only reported.

Forty-three horses exposed to pads carrying human fear odour startled harder, gazed longer at novel objects and touched the person less than with joy or control pads. [50]

Why it matters. The chemical signal crosses species without either side intending to send it.

Across 42 women, pregnancy and the phase of the menstrual cycle changed how attractive they were to yellow fever mosquitoes. Twenty-seven volatile compounds were identified, one of them enough on its own to shift preference. [51]

Why it matters. Who gets bitten is a chemistry question, and it changes over weeks.

Mice trained to fear an odour grew more neurons carrying the matching receptor. So did their own offspring, who had never met the odour. [52]

Why it matters. The receptor mix a nose is built with is not fixed at birth by genes alone.

A cryo-electron microscopy structure of a class II smell receptor showed an unusually large water-repelling pocket for fatty acid molecules and an unconventional way of switching on. [53]

Why it matters. Smell receptors have been almost unphotographable, which is why so few are matched to smells.

Pairs of people who discussed a smell and agreed a joint answer outperformed the better of the two individuals, on both discrimination and identification. [10]

Why it matters. Words fail for smell, and talking still helped, which nobody expected.

An analysis of 29,372 reviews of 36 perfumes found descriptions leaning heavily on naming the source of a smell rather than on qualities of it. [55]

Why it matters. With no vocabulary for smells themselves, every language borrows the names of things.

Phantosmia is smelling something that is not there. Reliable validated tests for these distortions of smell quality do not currently exist. [56]

Why it matters. A symptom that cannot be measured cannot be counted, treated or trialled.

Individual dogs carry 779 to 807 working smell receptor genes, against about 400 in humans and 1,100 in mice. A dog's nose is reported to hold 125 to 300 million receptors, against roughly 5 million in a person. [8][14]

Why it matters. The first number is how many kinds; the second is how many copies. The gap is the size of the array, not a different kind of sense.

Across 27 mammal species living only above about 1,000 metres, the genes that had broken were heavily concentrated in smell receptors: about 23% fewer of them, and smell bulbs about 18% smaller than in lowland relatives. [15]

Why it matters. Thin air carries fewer odour molecules, and unused genes are not maintained.

Genomes from 50 Indigenous people of the Malay Peninsula showed fewer broken smell receptor genes than other populations. One farming group carried a distinct version of OR12D3, a gene also linked to insulin release. [7]

Why it matters. How people get their food appears to have shaped which smells they can still detect.