Taste and Smell: Sensing the Chemical World

taste and smell

taste and smell reliefs in Erfurt, Germany (credit: Andreas Praefcke via Wikimedia Commons)

Taste and smell are two senses close to my heart.

I mean not literally, I’m not like one of those weird bugs that can taste all over their body (such as with their legs, wings and genitals—perhaps a post for another time?).

No, it’s because taste and smell, the “chemical senses”, are what we study in my current lab. The olfactory system was also the subject of the very first neuroscience research I did as an undergrad. Back then I remember thinking olfaction was kind of lame. I mean, it just didn’t seem like that important of a sense. I bet a bunch of you are also thinking the same thing. Like if someone told you they were going to get rid of one of your senses for the rest of your life and let you choose which one, I bet you’d probably choose smell.

The importance of smell

It turns out that smell is more important to your life than you might think. Especially if you love stuffing your face with food, like me.

cake eating 2

See, I wasn’t kidding about the face-stuffing.

Have you ever heard the saying that most of what you taste is really smell? That’s actually sort of true. I mean it’s patently false, since these are two separate senses. But it’s true that most of our perception of food, which we typically ascribe to taste, is really mediated by our sense of smell. The combined perception of food through both taste and smell (some people also add in texture and temperature) is what we experts call “flavor”. And here we’re talking about your ability to smell food while you’re chewing it (as opposed to when it’s in front of you), which is mediated by the retronasal pathway that allows air to travel from your throat back to your nose.

It’s commonly cited that 70 or 80% of flavor is smell, not taste, though it’s not clear where those numbers come from.2 While the specific percentage may be debatable, the general idea can be easily tested. First, get on the subway and rub up on one of the 15 people who are sneezing or coughing. Then a couple days later, when you get sick and your nose is all stuffed up, try eating something and you’ll notice that it doesn’t “taste” nearly as good as usual, since your sense of smell is compromised.

Mr Jelly Belly

Mr. Jelly Belly endorses this experiment.

What, you don’t like my experiment? Fine, here’s another one. It’s called the jellybean experiment and I’ve tried it on lots of kids, so you can be assured it’s a bit safer. Hold your nose closed and shut your eyes, and then put a jellybean into your mouth without knowing its color. Chew slowly and try to guess what flavor it is. Turns out, without your sense of smell all the jellybean flavors kind of taste the same, basically like a plain lump of sugar (admittedly not far from the truth).

Then open your nose and see if you can now identify the flavor. When I do this part I’m immediately overwhelmed with new sensations, the sense of orange or cherry or grape or whatever flavor that little guy happens to be. The kids I’ve tried this on seem to agree with this assessment (at least the ones that don’t just grab the jellybeans and run).

Anatomy and neural coding

It’s actually not that hard to explain why smell contributes more to flavor than taste. Have you ever heard about how there are just five basic tastes? Again, this is a saying that’s mostly true (two for two!). The five major tastes are sweet, salty, sour, bitter, and umami (also called savory). The reason there are five tastes is because there are five types of taste cells on your tongue, each of which is specialized for detecting one of these tastes.

Now take smell. Can you guess how many kinds of olfactory cells there are in your nose? Seriously, take a guess. Yes, it’s more than five.

In humans, there are over 300 kinds of olfactory neurons, each expressing a different receptor that recognizes specific odor molecules.3 Mice have over 1000, corresponding with their much better sense of smell.4 So even we humans, with our crappy sense of smell, still have over 60 times more types of olfactory neurons than taste neurons, allowing us to smell many more things than we can taste. Taste can only tell us that the jellybean is sweet and perhaps a little sour—that’s it. Smell is what really differentiates all the various fruity flavors, like orange or cherry.

But actually, the numbers don’t tell the whole story. The reason we can smell more things than we can taste also stems from the way these systems function.

The taste system is nice and simple. A particular taste, such as sugar, only activates one type of taste neuron, the sweet-sensing neurons. Those sweet-sensing neurons are not activated by any other tastes. This is referred to as a “labeled line” system, and means that with five types of cells we can taste five types of things.

labeled line and combinatorial

Schematic of labeled line vs. combinatorial coding. Colored circles represent sensory neurons, which respond to individual types of stimuli in a labeled line system, but are activated by diverse stimuli in a combinatorial system. Sensory neuron signals are then sent to and translated by the brain.

In contrast, the olfactory system utilizes “combinatorial coding”. This basically means that each odor can activate multiple types of olfactory neurons, and each neuron can be activated by multiple odors. As long as the brain can read out what combination of neurons has been activated, now we’re able to detect way more odors than the number of cell types we started with. For example, if there are 300 olfactory neuron types and each odor activates 10 of them, we’d be able to perceive 1.4 x 1018 odors, or over one billion billion!5 A recent study using actual human subjects, not just math, indeed estimated that humans can distinguish trillions of smells.6 This is way more than the number of colors or tones that we can perceive. Still think smell is the lamest sense?

Smell vs. taste: what are they good for?

At this point you might be wondering: if our olfactory system could evolve to be so powerful and make such good use of its neurons, why is the taste system so limited? It doesn’t seem to be just some evolutionary fluke, since the chemosensory systems of insects share the same features as ours despite evolving independently.

Well, let’s go back to the basics. Olfaction detects volatile molecules present at some distance from their source, whereas taste detects compounds by direct contact. So maybe olfaction can afford to provide more nuanced information, like telling you which direction has the most fragrant fruit or the potential for predators, and allowing you to weigh different options before making a decision. In contrast, once you’re tasting something you’re already in direct contact with it, so you’ve got to do something simple and immediate: either eat it or spit it out. Perhaps the organization of the taste system, while seeming quite limited compared to olfaction, is actually optimized for making binary choices such as this—eat this if it’s sweet or savory (meaning it’s food), don’t eat it if it’s bitter or too sour (which signals poison).

Of course, this is all speculation. But that’s part of neuroscience: starting with simple observations about how we perceive things, identifying the underlying molecules and neural circuits that mediate these experiences, and finally trying to make sense of why the brain evolved in this way. At the very least, it’s something to think about the next time you eat a jellybean.



1. I hope you guys appreciate that I refrained from making a “how the nose knows” pun in the title, like everyone else who has ever written about smell.

2. It seems plausible that the 80% figure comes from a 1977 paper that I tracked down (citation below), in which subjects rated the perceived odor and taste intensity of an artificial stimulus composed of two compounds, ethyl butyrate (which has a strong odor but little taste) and sodium saccharin (which has a taste but no odor). As stated in the abstract, “when the nostrils were closed, as much as 80% of the ‘taste’ disappeared”. I also uncovered a paper from 1969 that I find much more convincing because real foods were used as stimuli. This study found that subjects were drastically less competent at identifying the food when they couldn’t smell; the average percent decrease was not stated, but looks to be at least 80%.

Murphy C, Cain WS, Bartoshuk LM. Mutual action of taste and olfaction. Sens Processes 1(3):204-211 (1977).

Mozell MM, Smith BP, Smith PE, Sullivan RL Jr, Swender P. Nasal chemoreception in flavor identification. Arch Otolaryngol 90(3):367-373 (1969).

3. Glusman G, Yanai I, Rubin I, Lancet D. The complete human olfactory subgenome. Genome Res 11(5):685-702 (2001).

Note that this study identified over 300 functional olfactory receptors, which I’m interpreting as the number of olfactory neuron types because each neuron expresses only one receptor.

4. Zhang X, Firestein S. The olfactory receptor gene superfamily of the mouse. Nat Neurosci 5(2):124-133 (2002).

Again, this study identified the number of functional olfactory receptors, which corresponds with the number of olfactory neuron types.

5. I calculated this number as 300 choose 10, i.e. how many different combinations of 10 things can made from a set of 300.

6. Bushdid C, Magnasco MO, Vosshall LB, Keller A. Humans can discriminate more than 1 trillion olfactory stimuli. Science 343(6177):1370-1372 (2014).

Note that the authors define an olfactory stimulus as a mixture of any number of single odorants, since natural smells typically contain hundreds of molecular components. This study didn’t actually test subjects with trillions of smells (obviously), but instead estimated the number of discriminable smells based on how well we can distinguish odor mixtures with varying numbers of shared components. Subjects could generally distinguish stimuli that had less than ~50% overlap of their components. Because the number of mixtures that can be made from even a relatively small number of components, such as 100, is astronomical (many trillions of trillions), the number that overlap by less than 50% is still in the trillions.


Leave a Reply

Your email address will not be published.