Neurogastronomy: How the Brain Creates Flavor and Why It Matters

The number of tastes is infinite, since every soluble body has a special flavor which does not wholly resemble any other…. Up to the present time there is not a single circumstance in which a given taste has been analyzed with stern exactitude…. Men who will come after us will know much more than we of this subject, and it cannot be disputed that it is chemistry which will reveal the causes or the basic elements of taste.

There are indeed men and women who have come after, who definitely do know about the chemistry of taste in terms of the molecular composition of the foods we consume. It will be useful to review some of the main types of these molecules in order to relate them to the brain mechanisms that create their flavor.

As mentioned in the introduction, studying food odors and flavors is the job description of several kinds of scientists with distinct but overlapping expertise and aims. Psychologists and psychophysicists have a natural interest in testing human perception. Organic chemists who work in the flavor and fragrance industries routinely test the odors of the chemicals they are synthesizing in order to develop new flavors and fragrances. Food scientists try to understand how a particular food gives rise to a particular flavor, simulating or heightening it with artificial odor molecules. They classify odors on an empirical basis related to particular types of behaviors or to specific types of food flavors or perfumes. To do this, they use modern instrumentation to pull apart the molecular composition of a food to relate it to the flavor perceptions of that food.

To illustrate the complexity of the smell of food, coffee is a good example. It has been calculated that the taste and aroma of coffee are due to some 600 ingredients. How do we relate a particular ingredient to a particular aroma? The main instrument available at present is the gas chromatograph/mass spectrometer (GCMS), widely used to study the composition of different materials. A gas containing vaporized molecules of the substance under study is inserted through a long tube in the GCMS, so that the different molecules are deposited on a collecting surface in accordance with their molecular weight. The different molecules give off different wavelengths of light when illuminated. This enables them to be identified by the mass spectrometer, which can measure a substance down to a concentration of around one part in a million. The readout consists of peaks at different sites (allowing identification of the substance) and with different peaks (showing the amount of a given compound).

The GCMS is a valuable tool for analyzing the molecular composition of a substance. However, the odor world is too clever. There is often a poor correlation between the heights of the peaks in the MS readout and the strength of the odor from that compound. Often the main odor qualities of a foodstuff are due to molecules which are scarcely detectable in the MS.

A good analogy for why this should be can be found in gemstones. It is usually the case that a precious stone, such as a ruby or sapphire, owes its striking color not to its main mineral constituents but to a trace element in the stone. So it is with odors: a given foodstuff may owe its most characteristic odor to a trace amount of a type of odorous molecule. The experiments of the psychophysicists tell us that we can detect an odorous compound even though it is not detectable in the GCMS. Thus we humans, despite the bad publicity about our smell abilities, have a sense of smell that is better than the most powerful molecule-detecting devices that our brains have been able to devise.

To understand how our brains construct perceptions of smells, we need to start with an understanding of the molecules themselves. Peter Atkins of Oxford University is one of the best popularizers of the molecular nature of things. His book Molecules gives a good sense of the chemical structures that make up our material world, many of which are able to give rise to smell perceptions. We summarize here some of this information to give a sense of the principles of how these different odor molecules are related, and how they are able to carry the information about their sources in the perceptions to which they give rise. This is as fundamental for molecular gastronomy as it is for neurogastronomy.

Humans have descended from a line of primates that ate primarily fruit, so it will be appropriate to begin by considering fruit odors.

Ethylene (ethene) (C₂H₄) plays a central role in the ripening of fruit. It is a volatile molecule, meaning that it can be released as a gas into the air. Its release is the signal that the peak of ripening is occurring. This may reflect an action of the gas in activating the metabolic processes involved in ripening, including softening of the cellulose of the cell membranes. As Atkins relates, fruit is often shipped unripe and then exposed to ethylene gas to bring it to ripening when sold. (Judging from the lack of natural flavor in most of the produce at the supermarket, this process is not very effective in practice.)

Apart from this role in food, ethylene is also produced during refining (known as cracking) of petroleum. It can form long chains, producing polyethylene, which is a good insulator because of the inert nature of the bonds. As such, it is the prototype of many other types of plastics. Atkins notes many similar examples of how a given molecule can be used in totally different ways.

Food scientists are carrying out detailed analyses of the dozens or even hundreds of compounds that may be present in a given food. They want to know the way each compound contributes to the flavor as a volatile substance released from ingested food in the mouth to stimulate the nose by the retronasal route, and at different stages during ingestion, chewing, and swallowing.

An example of this kind of study is an analysis of the aroma of a banana carried out by D. Mayr and his colleagues at the Nestlé Research Center in Lausanne, Switzerland, in 2003. This is described in the article as a “breath by breath analysis of volatiles released in the mouth during eating of ripe and unripe banana using Proton Transfer Reaction-Mass Spectrometry.” In other words, they wished to show the production of smell molecules in successive breaths during chewing of unripe as opposed to ripe bananas. The aim was to gain insight into what signals the consumer is getting to detect and eventually decide whether a fruit is ripe.

The retronasal smell of the ingested fruit morsel comes with the first outward breath, carrying the odor molecules from the morsel through the back of the mouth and up to the olfactory sensory cells in the nose. The smell consists of dozens of kinds of volatile molecules released from the banana cells. Chief among these are the simple alcohols, such as the two-carbon ethyl alcohol familiar in alcoholic drinks. (Some suggest that the craving of our primate ancestors for eating ripe bananas to excess disposed humans to alcoholic drinks and susceptibility to alcoholism.)

These simple alcohols give rise to a sensation that we call sweet, but this is only in analogy with the sweet sensation arising from the action of sugar in the taste pathway. We assume that the real meaning of the term sweet has to do with the taste of sugar and, by analogy, to the smell of the alcohol. But is this a result of our use of language, rather than an intrinsic property of either the sugar or the alcohol? Does a mute primate distinguish the two sensations? Did early prelanguage humans distinguish them? As language developed, how did the terms describing these two sensations emerge? These considerations indicate that relations between smell and language are a rich field, which we explore further in chapter 24.

The study of how flavor is produced from eating a food can be said to have begun with Brillat-Savarin’s section “Analysis of the Sensation of Tasting”:

He who eats a peach… is first of all agreeably struck by the perfume which it exhales; he puts a piece of it into his mouth, and enjoys a sensation of tart freshness which invites him to continue; but it is not until the instant of swallowing, when the mouthful passes under his nasal channel, that the full aroma is revealed to him…. Finally, it is not until it has been swallowed that the man, considering what he has just experienced, will say to himself, “Now there is something really delicious!”

In another section, “The Supremacy of Man,” he analyzes the process even further, with special attention to the movements of the lips, mouth, tongue, and throat that appear to him to be unique to humans:

[M]an’s apparatus of the sense of taste has been brought to a state of rare perfection; and, to convince ourselves thoroughly, let us watch it at work.

As soon as an edible body has been put into the mouth, it is seized upon, gases, moisture, and all, without possibility of retreat.

Lips stop whatever might try to escape; the teeth bite and break it; saliva drenches it: the tongue mashes and churns it; a breathlike sucking pushes it toward the gullet; the tongue lifts up to make it slide and slip; the sense of smell appreciates it as it passes the nasal channel, and it is pulled down into the stomach… without… a single atom or drop or particle having been missed by the powers of appreciation of the taste sense.

It is, then, because of this perfection that the real enjoyment of eating is a special prerogative of man.

This pleasure is even contagious, and we transmit it quickly enough to animals which we have tamed.

In our day, food scientists pursue the detailed study of the eating process with sophisticated instruments to analyze both sensory qualities and motor activity.

The banana is probably the most analyzed of all foods by these methods, which seems appropriate. It is an icon of primate preference for fruit, and more bananas are consumed annually worldwide than any other fruit. It contains more than 300 volatile compounds. A typical study focused on two of the main types of these compounds, esters and carbonyls, as registered by proton-transfer-reaction mass spectrometry (PTR-MS). An artificial mouth that performed artificial chews was used to test for the release of volatiles from unripe and ripe bananas. It found that unripe fruit released relatively few ester and carbonyl volatiles, even with the highest chewing rates. By contrast, ripe fruit released many volatiles, and these carried the main attractive qualities of a ripe fruit, such as being fruity, apple, candy, floral, caramel, and yes, definitely “banana-like.”

As a general rule, “fruitiness” is due primarily to ester molecules—that is, molecules with a double-bond oxygen on an internal carbon—that are produced when, through enzymatic action in the ripening plant, an acid is combined with an alcohol. In the natural state, as chewing of the fruit proceeds, there are added retronasal smells from the products of the actions of the enzymes in the saliva. These fill out the sensory profile of the fruit and provide clues to slight variations in degree of ripeness and varieties of a given fruit that are stored in memory in the brain and that guide future feeding preferences.

The volatiles are described in terms of the main notes (fruity, candy [sweet], banana), and secondary notes (cheesy, green, apple, pineapple, floral, chocolate, caramel, mushroom). Is this telling us there is a smell “image” for ripe banana that overlaps to some extent with the smell “images” for these other food objects? These specific molecules are the “signature” of a ripe banana. As we shall see, the brain percept of these molecules is a pattern in the brain that monkeys, and humans, like or crave in a naturally ripened delicious fruit.

According to Harold McGee, “plants in general are biochemical virtuosos,” giving off multiple kinds of aromas. McGee is a studious person with an engaging personality who could well pass for a professor of English literature, which is what he started out to be until sidetracked by a passion for the chemistry of food. Now he is widely recognized, along with others such as Shirley Corriher, as a pioneer in understanding the science behind cooking and the food flavors it produces. We met at a workshop where I gave a talk on smell images and the brain flavor system and he was one of the attendees, learning still more about food and the brain.

Through his books McGee has given insights into the molecular basis of foods and their flavors that are of practical benefit for everyone from food novices to professional chefs. Here are some observations from his book On Food and Cooking: The Science and Lore of the Kitchen on the molecules in plants that give them their virtuoso smells and flavors.

• Green. This aroma is faint until the plant tissue is torn apart, or cut, as by a knife, or chewed. These actions damage the cell membranes, causing an oxidizing enzyme called lipoxygenase (lipid = fat, oxygenase = to break down by combining with oxygen) to act on the fatty molecules that make up the cell membrane to break them down into small, volatile fatty acids, which are further dismantled by other enzymes in the cell contents. The aroma of green, therefore, is not primarily an intrinsic smell—that is, one given off by the natural leaf itself in any great amounts—but rather one produced secondarily on preparing or eating the plant.

• Terpene. These are five-carbon molecules that can take many forms. They are common constituents of plants as well as of fruits, herbs, and spices. They are referred to as ethereal because of the etherlike “lightness” of sensation they produce. The distinctive smell of pine trees in the forest is due to terpenes in the tree resins. They are highly volatile and therefore act quickly when raw vegetables are cut or chewed, and they are quickly lost in cooking. They are also highly reactive with each other and with other molecules.

• Phenols. These are six-member carbon-ring molecules with a variety of side groups hanging off them. Different phenolic compounds are responsible for the main “notes” of different herbs and spices.

• Sulfur. Sulfur-containing molecules often are produced by a plant for defensive purposes. They have an aroma with an “edge” that gives a “pungent” quality to the smell, often because of direct stimulation of touch fibers in the nasal membranes.

Among the plant virtuosos, the herbs and spices reign supreme when it comes to smell and flavor. Since recorded history began, they have been delivering intense aromas to human cuisines. One can hypothesize that they may have been doing this in prehistoric times as well, perhaps even enticing the early humans into migrating out of Africa, as the plants must have been flourishing during those times.

Herbs and spices play dominant roles in the defining flavor elements of many traditional cuisines. Whereas most plants contribute nourishment as well as flavor, herbs and spices are added to a cuisine primarily for their flavor contribution. Strangely, they are almost completely ignored in public policy discussions of how to construct healthy diets.

Each plant actually consists of many volatiles that contribute to its characteristic odor. As with other food sources, the characteristic aroma is due to a few key molecule types. Because of their strong stimulating abilities, these molecules are often used in studying the brain responses to odors.

Finally, as Atkins notes, 2-tert-butyl-4-methoxyphenol (BHA) (C₁₁H₁₆O₂) is an antioxidant by virtue of the fact that it opposes the ability of oxygen to degrade molecules. These actions have been recognized only in recent years. Many spices, such as sage, cloves, rosemary, and thyme, are rich sources of compounds resembling BHA. Already in the Roman Empire, spices were used for their preservative and antibacterial properties as well as for flavoring.

The great leap forward in the development of human cuisines, and the associated emergence of human culture and language, was the discovery of the use of controlled fire to cook foods, as described by Richard Wrangham in his book Catching Fire: How Cooking Made Us Human. The greatest effect was achieved by cooking meat.

There does seem to be something about cooked meat that humans instinctively like. It is mostly due to the smell, both the aroma breathed in through the nose (orthonasal smell) and retronasal smell from ingested food in the mouth associated with flavor. The most attractive volatile molecules from cooked meat are produced by the “Maillard reaction.” The special flavors produced by cooking meat are described by McGee: “Raw meat is tasty rather than flavorful. It provides salts, savory amino acids, and a slight acidity to the tongue, but offers little in the way of aroma.” This means that for the dog and other carnivores, not only does the long nasopharynx limit the aromas coming from the mouth, but the raw meat gives off little flavor to begin with.

When humans invented cooked meat, they not only added flavor to the meat, but they had the short nasopharynx to enjoy it. McGee continues:

Cooking intensifies the taste of meat and creates its aroma…. [Up to] the boiling point… its… flavor is largely determined by the breakdown products of proteins and fats. However, roasted, broiled and fried meats develop a crust that is much more intensely flavored, because the meat surface dries out and gets hot enough to trigger the Maillard or browning reaction. [These aromas] have a generic “roasted” character, but some are grassy, floral, oniony or spicy, and earthy. Several hundred aromatic compounds have been found in roasted meat.

Thus, in contrast to raw meat, the properties produced during cooking are varied and dramatic, and at the heart of much of human flavor.

Heat causes the muscle cells of meat to break down, releasing the molecules that smell meaty and also breaking down other protein and fat molecules to form new molecules—particularly esters, ketones, and aldehydes—that give the meat new smells such as fruity, floral, nutty, and grassy. Fat is in fact responsible for most of the characteristic flavors of different meats. These smells are produced at relatively moderate cooking temperatures. One can imagine that these new smells were a delightful discovery by early humans, enough by themselves to motivate the development of cooking meat as a desirable part of the human diet.

At high temperatures that produce a crust on the meat, a level is reached at which new reactions take place, especially between sugar molecules from the breakdown of carbohydrates and amino acid molecules from the breakdown of proteins. The reactions that form these new molecules are called Maillard reactions, after Louis-Camille Maillard, the French scientist who first described them almost a hundred years ago. These reactions cause the dark brown color of the surface of seared meat. The molecules produced are highly volatile and have even more intense smells of the meaty, floral, and fruity varieties.

With increasing heat there comes a trade-off between the volatile smell-producing molecules that are characteristic of a given meat (or vegetable, for that matter) and the new molecules produced by the Maillard reaction. (According to McGee, the latter are more intense but less individualized.) Expert cooks explore these trade-offs with their different methods of lower heat preparation (such as boiling) and highest heat (frying).

The aroma of cooked meat can be sensed before cooking by inhalation through the nose; however, these are smells only from volatized molecules at the surface of the meat. In contrast, the full amplification and diversification of aromas from all the molecules within the meat produced by cooking can be sensed only when released from within the mouth, and when the other sensory pathways involved in flavor are activated in parallel, especially the pathways for taste and touch.

The texture of cooked meat is due to the balance between the proteins of its muscle cells, the collagen that makes up its connective tissue, and the water contained in or around both. This balance alters during heating. McGee describes three stages: rare, characterized by early juiciness when the muscle fibers begin to coagulate; medium-rare, when the collagen fibers become denatured and shrink, squeezing out water and making the meat drier and chewier; and well-done, when the collagen softens into gelatin while the muscle fibers become looser. The balance between these factors obviously depends on the type of meat, the amount of heat, and the duration of heating.

Dairy products, such as butter and cheese, arise from the milk of domesticated cows and goats, so it is assumed that they became a part of human cuisine only during historic times, in the past 10,000 years. The flavors of butter and cheese, largely because of their aromas, have made them an integral part of the diets of many human cultures. Here we follow Atkins in noting a couple of molecules responsible for that attraction.

Butanedione (C₄H₆O₂) has a cheesy, buttery, acrid smell. It is a ketone (also called diacetyl) because it contains a carbonyl group C = O; this type gives rise to a wide range of smells and flavors. Butanedione has a cheeselike smell, which is responsible for the main flavor of butter. It also contributes to sweat and armpit odor. Psychophysical testing has shown that these types of odors can give rise to either the cheeselike or armpitlike perception, depending on the context.

Butanedione is added to margarine to give it its buttery flavor. Linoleic acid (C₁₈H₃₂O₂) is the main fatty acid in many vegetable oils, such as cottonseed, corn, soybean, and rapeseed oil. It is also used in margarine, shortening, and salad and cooking oils. It has little smell by itself. It is hydrogenated (by bubbling hydrogen through the linoleic oil) to prevent the fat from being oxidized; that is why margarine does not get rancid. However, the hydrogenation leaves the margarine white, so carotene molecules are added to restore the yellow color. Butanedione is added to give it a butter odor. This is a good example of how an understanding of molecules can give insight into the world we live in.

I remember when growing up in Iowa during World War II the dairy farmers insisted that margarine did not taste like real butter, so it had to be sold in its white state in a plastic bag, with the color in a separate capsule, which we had to break to knead in the color. This nonsense ended when, at a conference of the dairy industry, the butter served at the closing banquet was later revealed to have been margarine—and no one had complained. This shows that the appearance of a food influences its flavor, as we shall see in chapter 15.

In addition to these generic smell molecules, cheese contains molecules reflecting the grass that the cows consume. These can be shown to have specific effects on the aromas and flavors of the cheese. A particularly careful study was carried out in 2004 by Stephania Carpino, Guilelmo Licitra, Terry Acree, and their colleagues in Sicily. For her thesis work, Carpino stationed herself in the pastures and charted the specific types of grass that the cows ate through the day and through the season. She correlated this with the GCMS peaks and showed that the molecular composition varied with the type of grass and its maturity in different seasons. She then carried out perceptual testing to show that subjects—and presumably consumers—could distinguish the different cheeses produced.

Licitra is head of a dairy cooperative in Sicily that is using modern dairy methods to a practical end: to produce traditional cheeses with enhanced flavor qualities. Similar attempts are being made in many countries—including, fortunately, more and more in the United States—to counter the bland produce available in mass-market supermarkets with foods that carry the full sensory qualities of locally produced items.

In sum, each food has its characteristic molecular composition, modified by how it is prepared. By themselves, foods have no flavor. They are the raw materials out of which the brain creates flavor.