14 Chapter 4.7 — Taste: The Sense That Decides Whether to Swallow
Chemosensation at the threshold of the body
14.1 Arriving on familiar ground
We have spent this unit moving outward — from the body’s own interior, which was the business of Unit II, to the surface of the skin, and then out into the light and sound of the world at a distance. Now, for the last two chapters, we are going to do something that looks at first like a step backward. We are going to return almost all the way to the inside, to a sense that operates only when the world is already in your mouth.
That is the right way to feel about taste, and I want to lean into it rather than apologize for it. Of all the senses in this unit, taste is the one that sits closest to the interoceptive story we started with — so close that you can reasonably argue it is an interoceptive sense wearing exteroceptive clothing. Its job is not to build a model of the world out there. Its job is to make a decision about something that is about to cross the boundary into your body: swallow this, or spit it out. That is the most ancient decision a motile organism makes, and we will see that the brain still treats it that way.
Here is the encouraging part, and the reason I have placed taste where I have. You already know most of the anatomy this chapter needs, because you met it earlier in the unit and in the one before. Taste reports, in the first instance, to the nucleus of the solitary tract — the NTS — which is precisely the brainstem hub we built up in Unit II as the great clearinghouse for visceral and bodily information, the place where the vagus nerve dumps its cargo of signals from the gut, the heart, and the baroreceptors. And taste’s thalamic relay is a small nucleus, VPMpc, sitting right beside the VPM territory that we said carries touch from the face. And its cortex is the insula — the very region we built up in the pain chapter as the brain’s map of how the body is doing, the cortex of “something is wrong.” So taste does not arrive as a stranger. It arrives on anatomy you already own, and it arrives carrying a question the insula already knows how to ask. We will make that convergence the spine of the chapter.
But — and this matters, so let me say it before we go further — taste is not simply “pain, part two,” and I am going to work to keep it from reading that way. The shared destination in the insula is real and it is illuminating, but the front end of taste is genuinely its own thing: its receptors are unrelated to the ones that carry pain, its cranial nerves are different, and its evolutionary logic runs back to a chemical decision older than nervous systems themselves. The insula is the bridge between this chapter and the last one. It is not the frame. In the middle of this chapter, taste has to stand on its own legs, and it does — on a molecular story that is genuinely thrilling and a coding controversy that is genuinely unsolved.
14.2 Taste is a gatekeeper, not a reporter
Start with what the system is for, because in taste, more than anywhere else in this unit, function explains structure.
A distance sense like vision is, in a sense, disinterested: the photons that land on your retina from a distant hillside are not about to do anything to you, and the sense’s job is to report the layout of a world you might or might not choose to enter. Taste has no such luxury and no such neutrality. By the time a chemical is on your tongue, the encounter is already half-committed — the substance is in your mouth — and the only question left is the consequential one: does it go further in, or does it come back out? Taste exists to make that call, and to make it fast, before you have swallowed something that will hurt you.
Notice that this is the Euglena problem again — the approach-or-avoid decision we met at the very start of the unit, when a single cell with a light-sensitive eyespot steered toward what was good for it and away from what was not. Taste is that same primal logic, conserved and elaborated, now running in a vertebrate mouth. And it sorts the chemical world along exactly the axis a gatekeeper would care about. The tastes we find innately pleasant mark things the body wants: sweet flags sugars, which are energy; umami (the savory taste of glutamate, the taste of broth and aged cheese and ripe tomato) flags amino acids, which are protein; salt, at the gentle concentrations the body needs, flags the sodium that the whole interoceptive apparatus of Unit II works so hard to keep in range. The tastes we find innately aversive mark things the body should refuse: bitter is the taste of a vast chemical zoo of plant alkaloids and other toxins, and we are built to recoil from it; sour, the taste of acid, warns of unripe fruit and of food turned by spoilage.
I want you to see how cleanly this maps onto the spine of the whole book. The hedonic sign of a taste — whether it feels good or bad — is not a decoration added later by some higher faculty of appreciation. It is the output of the system. Taste is valence-first: it does not so much tell you what a substance is as tell you whether to take it in, which is to say it delivers a verdict, not a description. We said in the pain chapter that pain is the sense that tells you something is bad and commands a response. Taste is its mealtime cousin: the sense that tells you whether something is good for the body and commands you to swallow or spit. Both are sensing in the service of a homeostatic decision. Both, as we will see, converge on the same patch of cortex. Hold that thought; we will return to it at the end.
The five “basic” tastes — sweet, sour, salty, bitter, umami — are the ones with well-characterized receptors and dedicated transduction machinery, and they are the spine of any honest account. Umami is the most recent admission to the club: long resisted in the West as a mere curiosity of Japanese cuisine (the chemist Kikunae Ikeda identified glutamate as its basis in 1908), it is now firmly established, with a receptor of its own. I mention the history because it is a useful caution. The list of “basic” tastes is not handed down from nature with a number on it; it is a working scientific claim about which chemical detectors the tongue actually has, and it has grown by one within living memory.
And its edge is still ragged. There is a serious, active case for a sixth taste dedicated to fat (more precisely, to the free fatty acids released from fats), with candidate receptors such as CD36 and GPR120 — “oleogustus,” some have proposed to call it. There are arguments for detectors of calcium, and for a distinct taste of starch in rodents, and a perennial debate about whether the cooling “taste” of water is a taste at all. I am not going to adjudicate these here, and you should not expect a tidy verdict, because there isn’t one. The useful lesson is the meta-lesson: “how many tastes are there?” is not a question with a fixed answer waiting in a textbook. It is a frontier, and where the frontier is depends on how strict you are about what counts as a basic taste in the first place.
14.3 The receptors: a molecular story the old lectures could not tell
For most of the twentieth century, a chapter like this could describe what the tongue does but had to stay silent about how, because the actual molecular detectors were unknown. That has changed completely, and recently, and it is one of the genuine triumphs of modern sensory biology. We now know, in real molecular detail, what the receptors for the basic tastes are — and the answer has a beautiful internal logic that turns out to be the key to everything that follows. So let me lay it out properly, because this is where taste earns its independence from the pain chapter.
The first surprise is that the five tastes do not all use the same kind of machinery. They split into two families, and the split is not arbitrary — it tracks the chemistry of what is being detected.
Sweet, umami, and bitter are detected by G-protein-coupled receptors — the GPCRs, that vast superfamily of membrane proteins that also includes the receptors for most hormones and neurotransmitters and, not incidentally, the opsins that catch light in your retina. A GPCR is the right tool when the thing you are detecting is a specific molecule whose shape you need to recognize: a sugar, an amino acid, a particular alkaloid. The sweet receptor is a partnership of two proteins, T1R2 and T1R3, working as a pair; the umami receptor is a closely related pair, T1R1 and T1R3, sharing one subunit with the sweet receptor [@nelson2001sweet; @nelson2002umami; @li2002umami]. Bitter is handled by a different and much larger set, the T2R family — roughly twenty-five functional receptors in humans [@adler2000t2r; @chandrashekar2000t2r]. That number is itself telling: there is one sweet receptor because sugars are, chemically, a fairly uniform target worth saying yes to, but there are dozens of bitter receptors because “poison” is not one chemical but thousands, scattered all across molecular space, and the system has cast a wide net to catch as many of them as possible. Some T2Rs are specialists tuned to a single class of compound; others are promiscuous generalists that respond to many [@meyerhof2010bitter]. The asymmetry — one yes-detector, many no-detectors — is exactly what you would build if false positives (refusing good food) were cheap and false negatives (swallowing poison) were potentially fatal.
Salt and sour, by contrast, are detected by ion channels — not by recognizing a molecule’s shape but by letting the relevant ions pass directly through a pore. This is the right tool when the thing you are detecting is an ion. For salt, at least for the appetitive low-sodium response the body actively seeks, the detector is the epithelial sodium channel, ENaC: sodium ions from the food flow straight through it and depolarize the cell [@chandrashekar2010salt]. Sour is the taste of acidity — of protons — and its receptor was, remarkably, the last of the five to be identified, holding out as a genuine mystery until 2018, when the proton-selective channel OTOP1 was shown to be the long-sought sour receptor [@tu2018otop1; @teng2019otop1]. Protons enter the taste cell through OTOP1 and acidify its interior, which the cell reads as sour. I find it a little wonderful that the receptor for one of the five oldest and most basic sensations in biology was nailed down only a handful of years ago — a reminder that “basic” and “well understood” are not the same word, and that even the foundations of a mature field can have a recent and surprising plank.
Now here is the payoff, the fact that makes this more than a list of protein names. The three GPCR tastes — sweet, umami, bitter — though they use different receptors at the cell surface, all funnel into the same downstream cascade inside the cell: the activated receptor switches on a G-protein (one is charmingly named gustducin, a cousin of the transducin that does the equivalent job in your photoreceptors), which drives the enzyme PLCβ2, which ultimately opens an ion channel called TRPM5, depolarizing the cell and triggering the signal to the nerve [@zhang2003coding]. One shared assembly line, three different intake hoppers feeding it.
This is not a tidy detail for its own sake; it has a striking and testable consequence. If sweet, umami, and bitter all depend on the same internal machinery, then a single genetic lesion to that machinery should knock out all three at once while sparing salt and sour — which run on their own independent channels and never touch this cascade. And that is exactly what happens: knock out TRPM5, or gustducin, and the animal goes blind to sweet, umami, and bitter together, but can still taste salt and sour perfectly well [@zhang2003coding]. The chemistry predicted the dissociation, and the dissociation confirmed the chemistry. That is the kind of result that tells you the molecular picture is not a just-so story but a load-bearing account of how the system actually works.
Two asides that students reliably enjoy, because they break the parochial assumption that taste is a thing the tongue does to food in the mouth.
First, these taste receptors are not confined to the tongue at all. The very same T1R and T2R proteins turn up throughout the gut, in the pancreas, the airways, even in sperm — “extra-oral” taste receptors, sensing the chemical environment of the body’s interior and helping regulate everything from insulin release to immune responses in the airway [@depoortere2014extraoral]. The detectors evolved to assess what is about to enter the body did not stay at the entrance; the body uses the same molecular tools to monitor what has already gotten in. This is the interoception/exteroception seam blurring at the molecular level, in the most literal possible way: the identical protein serves as an outward-facing taste receptor on the tongue and an inward-facing chemical sensor in the gut.
Second, you can hijack this machinery, and confectioners and pranksters have. The leaf of Gymnema sylvestre — the Hindi name translates roughly to “sugar destroyer” — contains a compound that plugs the sweet receptor; chew it, then eat sugar, and the sugar is reduced to a gritty nothing, all texture and no sweetness, because you have pharmacologically deleted one taste while leaving the rest intact. The “miracle berry” does the opposite, making sour foods taste sweet. These are not magic; they are receptor pharmacology you can perform on your own tongue, and they are vivid proof that the five tastes really are separable channels rather than a single blended sense.
14.4 From tongue to brain: the wiring, briefly
Now the pathway, which I will keep brief precisely because so much of it is anatomy you already hold from earlier chapters. The point of walking it is less the list of structures than two features that matter for what comes after: where taste enters the brain, and where it ends up.
Taste information leaves the mouth by three cranial nerves, divided by territory. The facial nerve (VII) carries taste from the front two-thirds of the tongue; the glossopharyngeal nerve (IX) carries it from the back third; and the vagus nerve (X) picks up the scattered taste buds further back, on the epiglottis and the upper throat. (You need not memorize this, but notice that taste, unlike the other senses in this unit, does not have a single dedicated nerve — it hitches a ride on three nerves that are mostly doing other jobs, which is itself a hint about its piecemeal evolutionary history.)
All three nerves deliver their taste signals to the same first stop: the nucleus of the solitary tract (NTS) in the medulla. And here is the first feature worth pausing on, because it is the single strongest anatomical bridge in this whole unit back to the one before it. The NTS is not a taste nucleus that happens to share a neighborhood with visceral processing. It is the great visceral integration hub — the structure we built up in Unit II as the destination for the vagal traffic from the gut and the heart and the baroreceptors, the brainstem’s running readout of the body’s internal state. Taste pours directly into it. So from the very first synapse, the brain is not treating the taste of your food as an abstract sensory quality to be analyzed; it is treating it as one more stream of visceral information, fused at the brainstem with signals about how full your stomach is and what your blood is doing. Taste enters the brain through the interoceptive door. This is not a quirk; it is the whole thesis of the chapter written into the anatomy, and I think it is genuinely underappreciated.
From the NTS, the path is one you can now predict, because it follows the canonical plan from the unit overview. Taste ascends (in humans and other primates, largely without the immediate brainstem detour to a pontine taste area that rodents have) to the thalamus — specifically to VPMpc, the parvocellular tip of the ventral posterior medial nucleus, sitting right next to the VPM region that handles touch from the face. And from the thalamus it projects to primary taste cortex in the insula (and the overlying frontal operculum), the cortex we met in the pain chapter as the seat of interoceptive awareness. So the canonical receptor → relay → thalamus → primary cortex skeleton holds for taste — it is, in that structural sense, a “well-behaved” sense that fits the unit’s shared plan, which is exactly why I have used it as the worked example. But notice what is distinctive: its thalamic relay sits beside the face’s, and its cortex is the body’s own affective map. Even while obeying the plan, taste keeps revealing its kinship with the interoceptive, valuative side of the brain. We will see in the next chapter that olfaction does not obey the plan at all.
14.5 Two things about taste that are genuinely unresolved
Here is where I have to slow down and be careful, because we are entering territory where the textbook-confident version and the actual-science version come apart, and the whole ethic of this book is to give you the second one. There are two questions about taste that sound as though they ought to have settled answers and do not. Learning to tell a genuinely open question from a merely difficult one is a real scientific skill, and taste offers two clean specimens.
14.5.1 How is taste quality encoded — labeled lines or population patterns?
We have seen that the receptors are exquisitely specific: a given taste receptor cell on the tongue is largely dedicated to a single quality — a sweet cell, a bitter cell, a sour cell. This is the strongest evidence for what is called the labeled-line model of taste coding: the idea that each quality travels to the brain on its own dedicated channel, a “sweet line” and a “bitter line” kept separate all the way up, so that which line is active is the identity of the taste [@chandrashekar2006receptors]. On this view, taste is like a set of parallel wires, each labeled with its meaning, and the brain reads the taste by reading which wire is hot. The Zuker and Ryba laboratories have assembled a powerful case for this picture at the periphery, and for sweet, bitter, and umami the cell-level evidence really is strikingly clean.
But the story does not stay clean as you move inward. When physiologists record from taste-responsive neurons in the brain — in the NTS, and onward — they find that most of these central neurons are not narrowly tuned at all. A typical neuron responds to several different tastes, more vigorously to one than to others, but by no means exclusively [@erickson2008labeled; @smith2000coding]. This is the evidence for the rival across-fiber pattern (or population, or combinatorial) model: the idea that no single neuron’s activity identifies a taste, and that quality is instead encoded in the pattern of activity across a whole population of broadly-tuned cells — the way a chord, not any single note, defines a harmony. On this view, the brain reads a taste by reading the whole ensemble at once.
So which is it? The honest answer is that this has been argued for half a century and is not resolved, and — this is the part I most want you to take away — the disagreement is not merely empirical but partly conceptual. Some thoughtful people have argued that the labeled-line-versus-population framing is itself something of a red herring, because neither “basic taste” nor “neuron type” has ever been given a definition crisp enough to make the two models cleanly distinguishable; push hard enough and they start to blur into each other [@frank2008redherring]. And there is a genuinely different third idea in play: that part of the code is temporal — that the same neuron can signal different tastes through the time course of its firing, not just its rate, so that information we were looking for in which neurons fire is partly hidden in when they fire [@dilorenzo2009temporal]. Multiple independent laboratories hold these positions and genuinely disagree, which is what makes this a live question rather than a dead one. My own read, for whatever it is worth, is that the periphery is closer to labeled-line than the skeptics once allowed, the cortex is closer to population-coded than the labeled-line camp would like, and the temporal dimension is real and underexplored — but you should hold that as one instructor’s synthesis, not a verdict, because the field has not reached one.
14.5.2 Is there a map of taste in the cortex?
The second open question follows naturally from the first, and it is, if anything, more pointed — because it asks whether taste obeys what the unit overview called the most beautiful organizing principle of sensory cortex: the map. Vision has retinotopy, touch has the somatotopic homunculus, hearing has tonotopy. In each, the cortical sheet preserves the relationships of the sensory world as spatial relationships on the brain. Does taste have the equivalent — a gustotopic map, with sweet over here and bitter over there, laid out in orderly patches across the insula?
For a while it looked as though the answer might be a clean yes. A widely-noticed study reported exactly this: using two-photon imaging in mouse cortex, the investigators described spatially segregated “hotspots,” a little region of cells responding to sweet, a separate region for bitter, and so on — a gustotopic map to sit alongside retinotopy and tonotopy [@chen2011gustotopic]. It was an elegant result and it made the textbooks quickly.
The trouble is that subsequent work has had a hard time confirming it, and the weight of more recent evidence has shifted against the strong version of the claim. When other groups imaged taste cortex in awake, behaving mice — rather than anesthetized ones — they found taste-responsive neurons that were sparse, broadly tuned, and spatially scattered, with no sign of the clean hotspot segregation [@chen2021distributed; @fletcher2017overlapping]. And high-resolution functional imaging in humans has likewise mostly failed to find a spatial map, reporting instead that taste quality is carried by distributed, overlapping patterns of activity across the insula rather than by dedicated taste-specific territories [@chikazoe2019distinct]. One review surveying this human work put the conclusion memorably: there is, it argued, no “Cartesian restaurant” in the brain — no little dining map where each taste has its own table [@avery2021cartesian]. The current center of gravity in the field is that taste cortex is organized as a population code in space — the spatial echo of the across-fiber coding debate above — rather than as a clean topographic map.
I am deliberately not closing this with false confidence, because the question connects to something genuinely deep that we do not understand: why some sensory systems get crisp cortical maps and others apparently do not. Vision and touch and hearing all map a continuous physical dimension — position, or frequency — onto a continuous sheet, and perhaps a map is simply what you get when the thing being represented is itself continuous and the receptors are laid out in order across a surface (the retina, the skin, the basilar membrane). Taste qualities are not obviously continuous in that way — sweet does not shade into bitter the way red shades into orange — and the receptor cells are scattered across the tongue rather than ordered along it, so perhaps there is little for a spatial map of taste identity to be of.
But there is a deeper possibility worth raising, because it reframes the whole question and because it points where this book is going. Perhaps we keep failing to find a clean map because we keep looking for a map of the wrong variable. When we ask whether taste cortex is mapped, we sort it by identity — is there a sweet patch, a bitter patch? — and we mostly find scatter. But a sensory system can be organized by a variable other than the identity of the stimulus, and for the chemical senses the natural candidate is valence: not what the substance is, but whether it is good or bad for the body. And here there is a striking result that suggests the brain does exactly this — that it pulls identity and valence apart and handles them in different places. Tracing the outputs of the sweet and bitter cortical fields, the Zuker laboratory found that they project to topographically distinct, cleanly segregated regions of the amygdala — appetitive (sweet) and aversive (bitter) signals running as “separate lines” to separate targets [@wang2018valence]. And these valence-labeled projections are not merely correlated with behavior; they drive it. Optogenetically activating the sweet-to-amygdala projection makes a mouse treat neutral water as a reward; activating the bitter projection makes it recoil — and the manipulation can even reverse the hedonic value of a real tastant, turning sugar aversive or quinine attractive. Silence the amygdala and something revealing happens: the animal can still identify tastes — it discriminates sweet from bitter as well as ever — but it loses the valence, the behavioral pull of good and bad [@wang2018valence].
So the lesson reorganizes itself. It is not quite that taste “has no map.” It is that the brain appears to separate two different things a taste sense could represent — its identity and its value — and to organize them differently and in different structures: identity carried by a distributed, scattered population code in the cortex, but valence assigned downstream in the amygdala through a clean, segregated, almost labeled-line architecture. The map we kept failing to find in the cortex may have been the wrong map to look for; the orderly organization is there, but it is an organization by value, located past the cortex, in exactly the emotional machinery we will study in the unit to come. That is a hypothesis about the chemical senses in general — we will meet its olfactory version in the next chapter — and I offer it in that spirit. But notice how it turns a dead end into the book’s main road: the senses that exist most transparently to deliver a verdict about the body may be organized, in the brain, not by the chemistry of the stimulus but by the verdict itself. The map of taste, if there is one, may be a map of good and bad — which is to say a map of what to do.
The two debates in this section (labeled-line vs. across-fiber; gustotopy vs. distributed coding) are, in my judgment, the rare cases that genuinely clear the §4A bar — multiple independent programs (Zuker/Ryba; Erickson; Di Lorenzo/Katz; Fontanini; the human-imaging groups) actively disagreeing, not a single-lab effect dressed up as controversy — so I have presented them as live. But two calibration points for your eye:
The gustotopy subsection is written to convey that the recent weight has shifted against the strong Chen/Zuker 2011 hotspot map (awake-behaving rodent imaging plus human high-field fMRI mostly not replicating it). I think that directional read is fair and well-supported, but it is a read — if you would rather present it as fully balanced (“two camps, jury out”) I can soften the tilt. I left the tilt in because pretending the 2011 map is still even-money would itself be a calibration failure in the other direction.
The temporal-coding thread (Di Lorenzo) is real and from a genuine independent program, but it is the least-settled of the three positions and is doing only a light supporting role here; if you want the section leaner, it is the first thing to cut without damaging the argument. The chapter’s spine survives without it.
14.6 When taste breaks: the insula and the loss of the world’s flavor
We tend to take taste for granted until it is gone, and the clinical picture of taste loss is worth a paragraph both because it is humanly serious and because it confirms, from the other direction, that the insula really is where the gustatory world is built.
Damage to the insula — a stroke, most often — can produce profound disturbances of taste. I have described before the patient who, after an insular stroke, reported that food had stopped tasting like food and had come to taste, in his words, like dirt — every meal a chore performed on something inedible. It is easy to underrate how devastating this is. We are built to be guided toward food by its pleasures, and a person for whom eating has become joyless, or actively disgusting, can slide into dangerous weight loss without anyone immediately recognizing why. The lesson for a clinician is to take complaints of lost or distorted taste seriously rather than dismissing them as trivial; the lesson for us is the confirmation that the insula is not merely correlated with taste but necessary for the normal experience of it — knock it out, and the gustatory world does not just dim, it can curdle.
This also previews the chapter’s closing move. Notice that “food tastes like dirt” is not only a sensory complaint but an affective one — the problem is not that the patient cannot detect the chemicals but that the detection has lost its hedonic meaning, its goodness. That coupling of sensation and valuation, broken apart by an insular lesion, is exactly what the insula is for.
14.7 The insula as a recurring character: where the body’s verdicts converge
I want to end by pulling on the thread that has run beneath this whole chapter, because it reaches backward to the pain chapter and forward to the unit that follows, and it is, I think, one of the genuinely unifying ideas in this part of the book.
We have now met the insula twice. In the pain chapter it was the cortex of “something is wrong” — the place where the bare sensation of tissue damage acquires its awful affective weight, the felt sense that this is bad and must stop. In this chapter it is primary taste cortex — the place where a chemical on the tongue acquires its verdict of good or bad, swallow or spit. And in Unit II, standing behind both, the insula was already emerging as the cortical map of interoception, the brain’s readout of the body’s own internal condition. These are not three coincidentally adjacent functions that happen to share an address. They are the same function. Pain, taste, and visceral sensation all converge on the insula because they are all answering a single question, the most basic question a body can ask: is this good or bad for me, and what should I do about it?
This is the resolution of a worry I flagged at the start of the unit, and it is worth making explicit. One might have thought the chemical senses fit the evolutionary story of this book but sat awkwardly with its control-theory spine — because that spine was built, in the distance senses, on the reactive-versus-predictive gradient, the idea that senses earn their keep by buying lead time, and taste buys no lead time at all. That worry is half-right and instructively so. Taste does sit awkwardly on the distance axis. But it sits at the dead center of the other axis — the interoceptive, valuative one — and on that axis it is not the worst fit in the unit but the best. Taste is the sense most transparently in the business of homeostatic gatekeeping: chemosensation at the threshold of the body, delivering a verdict about whether to admit a substance, reporting through the visceral hub of the brainstem to the affective cortex of the body. The control loop did not vanish when we reached the chemical senses. It came back into its purest and most ancient form: eat, or do not eat. And in the next chapter we will find that taste does not run this assay alone — that smell asks the very same question, is this safe to take into the body?, but at a distance and ahead of the decision, sniffing the rot before the food ever reaches the mouth. The bitter recoil and the spit of the last unpalatable thing, and the wrinkle of the nose at something gone off, are two ranges of one ancient instrument.
And this sets up where we are going. The insula will not disappear after this chapter; it is becoming one of the book’s recurring characters, alongside the NTS and the control loop itself. We introduced it at pain, we have deepened it here at taste, and we will meet it again in Unit V, where the brain stops merely sensing whether things are good or bad and starts acting on those verdicts — choosing, valuing, deciding. The insula is where sensing hands off to valuation, and valuation is the doorway to action. Taste, the sense that decides whether to swallow, is a small and ancient instance of the largest problem the brain has: turning information about the world into a verdict about what to do.
But first, one more sense — and it is the strangest one in the unit, the sense that breaks nearly every rule we have established. Taste told us a chemical is good or bad. Its partner sense reaches out into the air to tell us, at a distance and before we have committed to anything, what is out there — and it does so with an architecture that ignores the thalamus, refuses to build a tidy map, and wires itself straight into the machinery of memory and emotion. Onward to olfaction, the sense of smell, and the fitting close to a unit that began with a single cell following a chemical gradient through the sea.
Reasonably settled:
- Taste is a small-dimensional gatekeeping sense organized around ingestion: innately pleasant tastes (sweet/sugar, umami/protein, salt/sodium) mark things to take in; innately aversive tastes (bitter/toxins, sour/acid) mark things to refuse.
- The peripheral receptors are known and recently nailed down: GPCRs for sweet (T1R2+T1R3), umami (T1R1+T1R3), and bitter (~25 T2Rs), sharing a common downstream cascade (gustducin → PLCβ2 → TRPM5); ion channels for salt (ENaC) and sour (OTOP1, identified only in 2018).
- The central pathway: cranial nerves VII/IX/X → NTS (the visceral hub, shared with Unit II interoception) → thalamus (VPMpc, beside the face’s VPM) → insula/operculum as primary taste cortex.
- The “tongue map” of regional taste zones is false; all qualities are detectable across the taste-bearing surface, with only slight regional differences in sensitivity.
- The insula is necessary for normal taste experience: insular lesions can abolish or distort it (“food tastes like dirt”), with real clinical danger.
Genuinely unsettled, and presented as such:
- How taste quality is centrally encoded — labeled lines vs. across-fiber population patterns vs. a temporal code. Specific peripheral cells are narrowly tuned; central neurons mostly are not; multiple independent groups disagree, and part of the dispute is conceptual.
- Whether taste cortex contains a spatial (“gustotopic”) map. An influential 2011 report said yes; awake-rodent imaging and human high-field fMRI have largely not confirmed it, favoring distributed/overlapping coding. The recent weight is against a clean cortical map of taste identity — but the brain appears to separate identity from valence, organizing valence cleanly downstream in the amygdala (sweet/bitter cortex → segregated amygdala targets that drive and can reverse appetitive/aversive behavior). The reframe — that the chemical senses may be organized by value rather than by stimulus identity — is a live and promising hypothesis, not a settled result.
- How many basic tastes there are — the case for fat (“oleogustus”), and possibly others, is serious and open; the number depends partly on what one counts as “basic.”
And, as ever, there is a great deal here we are sure of. The molecular logic of the receptors — different detectors for different chemistries, a shared cascade for the molecule-sensing three, the asymmetry of one sweet receptor against dozens of bitter ones — is among the most satisfying and well-established stories in all of sensory biology. You can rely on it.