How Smell Works

Of the five senses, smell is the oldest. It predates vision and hearing by hundreds of millions of years in evolutionary history. The olfactory system is not a refinement of earlier sensing apparatus — it's the original sensing apparatus, the one vertebrate nervous systems were built around before any of the others developed. This history is visible in the anatomy. Smell is wired into the brain in a way that's fundamentally different from every other sense, and that difference explains nearly everything unusual about how smell actually works — including things that seem unrelated to anatomy, like why a specific smell can unlock a thirty-year-old memory in under a second while seeing a photograph of the same event retrieves only an intellectual reconstruction.

The peripheral apparatus

The detection of smell begins in the olfactory epithelium — a small patch of specialized tissue, roughly five square centimetres, in the roof of the nasal cavity. It's covered in a thin mucus layer and packed with approximately six million olfactory receptor neurons (ORNs), each one a specialized sensory cell with dendrites that project into the mucus, topped with between eight and twenty cilia. The cilia are where the actual detection happens.

Odorant molecules — airborne, volatile, usually organic — dissolve into the mucus layer and bind to receptor proteins embedded in the cilia membranes. These proteins are G-protein-coupled receptors (GPCRs), the same general family that mediates many other sensory and hormonal signals. Binding triggers a classical GPCR cascade: the receptor activates a G-protein (called Golf), which activates adenylyl cyclase, which produces cyclic AMP, which opens ion channels in the membrane, allowing calcium ions in, depolarizing the neuron, and generating an action potential that travels up the axon toward the brain.

The ORNs are unusual in another way: they're exposed neurons. Unlike the neurons in your eyes or ears, which are shielded behind layers of non-neural tissue, olfactory receptor neurons are directly in contact with the environment. They wear out, die, and are replaced — continuously, throughout life — with a turnover cycle of roughly 30 to 60 days. This makes olfaction one of the most neuroplastic systems in the body; the peripheral sensor array is constantly rebuilding itself. It also makes it vulnerable: anything that damages the epithelium can disrupt smell, which is why viral infections (including COVID-19) can cause anosmia.

The combinatorial code

The human genome encodes roughly 800 olfactory receptor gene sequences. Approximately 400 of these are functional — the rest are pseudogenes, inactive remnants of a receptor repertoire that was likely much larger in our more olfaction-dependent ancestors. Mice have around 1,000 functional OR genes. Dogs have around 1,000 too, which is one source of their extraordinary sensitivity.

Four hundred receptors sounds like a limited system for an animal that can distinguish between tens of thousands of distinct odors. The resolution comes from the combinatorial code.

Each odorant molecule doesn't bind to a single receptor. It binds to multiple receptors simultaneously, with different affinities — weakly to some, more strongly to others. Each receptor responds to multiple odorants, again with different affinities. The odor percept isn't encoded in which receptor fires; it's encoded in the pattern of activation across the whole receptor array. Think of it as a chord rather than a note: each olfactory "chord" is a specific pattern of which receptors are active and at what intensity, and this pattern is the molecular identity of the smell. Four hundred receptors generating combinatorial patterns can theoretically encode an enormous space of distinct percepts — estimates range into the hundreds of billions of discriminable odors, though the practical ceiling under normal conditions is much lower.

Importantly, a single receptor gene variant can change the perception of specific odors significantly. Research published in PNAS found that functional variation in a single OR altered perception for 13% of the 68 odors tested — more redundancy was expected. This means individual receptor variants aren't simply backup copies of each other. Some do unique work, and losing them changes what you smell.

The olfactory bulb

The axons of olfactory receptor neurons travel directly through the cribriform plate — a perforated section of the skull at the base of the frontal lobe — and synapse in the olfactory bulb, a small structure sitting above the nasal cavity on the underside of the brain. This is the first processing station.

The olfactory bulb is organized into structures called glomeruli: spherical clusters of synapses where all ORNs expressing the same receptor type converge. Each glomerulus receives input from thousands of neurons expressing one specific receptor, regardless of where those neurons are physically located in the epithelium. The epithelium is distributed; the bulb is organized. The glomerular map converts the distributed receptor array into a structured spatial pattern of activity — a map, loosely speaking, of the chemical space of the odor.

Inhibitory interneurons in the olfactory bulb perform lateral inhibition, sharpening the pattern by suppressing neighboring activity. The output — from mitral and tufted cells — is a refined signal representing the pattern of receptor activation, ready for higher processing.

The unique neural pathway

Here is where the architecture becomes strange.

Every other sense — vision, hearing, touch, taste — routes its signals through the thalamus before reaching the cortex. The thalamus is the brain's central relay station: it receives sensory information, processes it, and distributes it to the appropriate cortical areas for conscious perception. This relay adds a step, but it also adds something important: a degree of cortical filtering and top-down modulation before the signal reaches the emotional and memory centers.

Smell bypasses the thalamus entirely.

Smell is the only sense that bypasses the thalamic relay and has primary access to regions of the brain typically found to be active during emotional processing (the amygdala), long-term memory formation (the hippocampus), and higher-order cognitive reasoning and evaluation (the orbitofrontal cortex).

The olfactory bulb projects directly to the piriform cortex (the primary olfactory cortex, which handles odor quality perception), and simultaneously to the amygdala and to the entorhinal cortex, which is the gateway into hippocampal memory systems. These projections are direct anatomical connections, not routes through the thalamic relay. A smell reaches the emotional and memory brain before it reaches the rational brain.

This is not a metaphor. It's a measurable anatomical fact with measurable functional consequences. Olfactory information can travel directly to the limbic system via the olfactory nerve without requiring any prior stop at the thalamus, unlike other sensory systems. The signal bypasses the cortical filtering step that every other sense undergoes. This is why an emotional response to a smell can precede conscious identification — you feel something before you know what you're smelling.

The Proust effect

The consequences of this architecture are most visible in the relationship between smell and memory, which is qualitatively different from the relationship between memory and any other sense.

Marcel Proust described it precisely in the opening of Swann's Way: the smell of a madeleine cake dipped in tea floods the narrator with childhood memories that no amount of deliberate recollection had produced. The memories are involuntary, vivid, emotionally loaded, and seemingly disproportionate to the stimulus. Neuroscientists studying this phenomenon — now called the Proust effect — have confirmed that it's real and specific to olfaction. Odor-evoked autobiographical memories tend to be rated as more emotional, more vivid, older, and less frequently retrieved by other means than memories evoked by any other sensory modality.

The mechanism follows directly from the anatomy. The amygdala, which receives direct olfactory bulb input, is the brain structure responsible for emotional tagging of memories — for determining which experiences get encoded with emotional intensity and therefore retrieved more readily later. The hippocampus, equally directly connected to the olfactory pathway, handles episodic memory formation. When a smell is first encountered in an emotionally significant context, the amygdala-olfactory bulb connection ensures it gets encoded with strong emotional weight. When that smell is encountered again later, the same anatomical pathway fires, reaching the emotional and episodic memory systems directly, before rational processing has even begun.

There's an additional property worth noting: olfactory memories seem to be protected from interference in a specific way. Verbal recollection of a memory tends to subtly alter it each time it's retrieved. Olfactory-triggered memories, possibly because they bypass the verbal cortical systems, appear to retrieve with unusual fidelity. You're not reconstructing the memory from conceptual traces. The smell is activating a direct associative pathway laid down during the original experience.

The consequence for perfumery is not trivial. When someone says a fragrance makes them feel something — relaxed, confident, nostalgic, safe, aroused — this is not poetic language. It's a description of direct limbic system activation, mediated by the most emotionally direct sensory pathway in the brain.

The genetic nose: individual variation at the receptor level

Every human nose is genetically unique in ways that go far beyond simple sensitivity differences.

The olfactory receptor gene family is one of the most variable in the human genome. Each individual has a unique set of genetic variations that lead to variation in olfactory perception, and on average, two individuals have functional differences at over 30% of their odorant receptor alleles. This is not like having slightly different eye colors — it's like having different instruments in your sensory orchestra. Two people smelling the same compound may be activating substantially different receptor arrays and therefore encoding substantially different combinatorial patterns. They may not, in any deep sense, be smelling the same thing.

Specific anosmias make this concrete. A well-studied example: androstenone, a steroid compound present in human sweat and in truffles, is imperceptible to approximately 30–40% of people. Of those who can smell it, some find it pleasant and vaguely musky; others find it sharp and urine-adjacent. These differences correlate with specific variants of the olfactory receptor OR7D4. The same molecule, three qualitatively different experiences — not one, which explains why consensus about androstenone-containing materials (truffle, certain musks, aged cheese, armpit) varies so dramatically between individuals.

Similar specific anosmias exist for isovaleric acid (the sweaty-cheesy compound), for the green-grassy molecule cis-3-hexen-1-ol, and for dozens of others. Many people cannot smell the acetone of diabetic ketoacidosis — a clinically relevant fact, as discussed in the disease article in this series. Some people are broadly hyposmic (reduced sensitivity across many odorants) while having perfectly normal anatomy; the explanation lies in their specific complement of functional OR genes.

The practical implication is that olfactory sensitivity is not a single trait. It's a receptor-by-receptor profile that varies across the entire odor space. You might be exquisitely sensitive to one class of compounds and functionally anosmic to another. This is why no two people navigate the same fragrance identically.

Training and expertise: the cortical dimension

The question of whether expert perfumers, sommeliers, and tea masters smell differently in a sensory sense — whether their peripheral apparatus is more sensitive — is interesting and the answer is largely no. The receptor complement of a trained perfumer is not substantially different from a naive person's. What training changes is the brain's processing of what the receptors send.

Several things happen with olfactory training. The most important is the development of vocabulary: once you have a word for a specific quality, you can encode and retrieve it precisely. Naive subjects trying to identify olfactory stimuli without vocabulary perform significantly worse than subjects who have been given labels — even arbitrary ones — for those stimuli. Language is an organizational scaffold for perception. Perfumers who have been trained with a structured vocabulary (this is citrus, this is the specific quality of "woody-earthy," this is the distinction between cumin and coriander seed) can parse a complex mixture more accurately not because they receive richer input but because they have the encoding structure to process it.

Beyond vocabulary, training changes the allocation of cognitive resources to olfactory processing. Olfaction in naive people is processed largely at a preattentive level — smells register, trigger responses, and are rarely subjected to deliberate analytical attention. Trained evaluators learn to allocate focused attention to the olfactory channel, which increases cortical involvement and enables more analytical decomposition of complex mixtures. Orbitofrontal cortex activity — the area handling higher-level olfactory cognition — increases with expertise.

Memory expansion is the third factor. Expert olfactory practitioners have accumulated enormous libraries of olfactory reference memories. When they encounter an unfamiliar composition, they're cross-referencing it against a catalogue of known materials and combinations that a novice lacks entirely. This is not about sensitivity; it's about pattern recognition built from stored experience.

None of this means the peripheral nose is irrelevant to expertise. Joy Milne's "super smeller" ability may have a genuine receptor-sensitivity component. But for the kind of fine discrimination that professional perfumers practice — distinguishing between two similar rose absolutes, identifying a off-note in a composition, tracking a saffron signature across ten different fragrances — the critical resource is cortical, not peripheral.

Cultural conditioning: what smells pleasant is learned

The most persistent misunderstanding about olfaction is that pleasantness and unpleasantness are intrinsic properties of molecules. They are almost entirely not. What smells good is what has been associated with good contexts, and what smells bad is what has been associated with bad ones.

The clearest evidence comes from cross-cultural studies. Indole — the molecule we've discussed extensively in this series in the context of L'Interdit and white florals — smells animalic and dark to noses accustomed to encountering it in body odor contexts, and floral and attractive to noses whose primary context for indole is flowers. Neither response is "correct." Both are learned associations.

Durian fruit, which contains high concentrations of sulfur compounds, is deeply beloved in Southeast Asia and is described as putrid sewage by most first-time Western encounters. The difference is not the odor. It's the cultural context of the first exposure. Butyric acid smells like rancid butter orthonasal (perceived from outside) and contributes to the savory depth of aged cheese retronasal (perceived through the back of the mouth during eating). The molecule is the same. The context determines the valence.

This is why "acquired tastes" work: with repeated exposure in positive contexts, the brain re-encodes an initially aversive stimulus as neutral or attractive. Whisky, coffee, wine, strong cheese — all of these are initially aversive to most people, and all become pleasurable through repeated positive contextual exposure.

For perfumers and connoisseurs, this is liberation, not a constraint. It means the space of what's possible to enjoy is bounded only by exposure and association, not by some fixed set of intrinsically pleasant molecules. The "difficult" materials — civet, castoreum, high-concentration indole, cumin at the edge of its animalic threshold — become interesting, even beautiful, once the context of first encounter has been established through knowledge and deliberate exposure.

The retronasal dimension

A final structural property of olfaction that's frequently overlooked: what we call "taste" is mostly smell.

The tongue's taste buds detect five qualities: sweet, sour, salty, bitter, and umami. Everything else you experience as "the taste of" something — coffee, strawberry, cinnamon, wine — is olfaction. Specifically retronasal olfaction: odorant molecules from food in the mouth travel backward through the nasopharynx and reach the olfactory epithelium from behind. The olfactory bulb processes this retronasal signal differently from orthonasal smell (the same molecules, coming from different directions, are processed partly as separate percepts), but the receptor layer is the same.

This is why food loses almost all of its interesting flavour when you have a nasal blockage from a cold: the five basic tastes remain fully operational, but the olfactory component — 80–90% of the perceived flavour complexity — is gone. Coffee becomes bitter liquid. Wine becomes sour. Strawberries become sweet. The taste buds give you the skeletal outline; olfaction provides almost everything else.

The implication for understanding perfumery is that everything we've discussed about how olfactory receptor neurons work, how the combinatorial code functions, how the limbic system processes smell — all of it applies equally to what people experience as flavor. A chef and a perfumer are operating on the same biological substrate. The aesthetic domains are different; the sensory apparatus is largely shared.

Smell is old. It's wired deep. It talks directly to the parts of the brain that handle memory and fear and desire without stopping to ask the rational cortex for permission. It's genetically unique in every person, culturally shaped in every culture, and trainable in ways that expand what's perceptible without fundamentally changing what's detected. It's the sense that most powerfully defies language — we consistently describe smells by analogy, by reference to objects, by the emotions they produce, because we never developed a dedicated vocabulary for olfactory qualities the way we did for color. The limbic system that processes it doesn't speak in words.

This is the right sense to build an art form around if you want to bypass cognition and reach something more immediate. Perfumery understood this before neuroscience confirmed it.