Your independent source for Harvard news since 1898 | SUBSCRIBE

Your independent source for Harvard news since 1898

Features

The Mystery of Smell

COVID-19 shines a spotlight on a once-obscure field of science.

November-December 2021

Sandeep Robert Datta

Sandeep Robert Datta

Photograph by Stu Rosner


Sandeep Robert Datta

Photograph by Stu Rosner

In the early weeks of the pandemic, as scientists and physicians scrambled to find the edges of this new, dangerous disease—how it spread from person to person, how it behaved inside the human body, and how they might be able to stop it—one emerging symptom sent a jolt of recognition through Sandeep Robert Datta: the sudden disappearance of many patients’ ability to smell.

A professor of neurobiology, Datta studies olfaction: what happens between nose and brain as sensory neurons pick up a smell and the signal makes its way to the olfactory cortex, where the information is transformed into something we recognize as coffee, or roses, or dirty socks. But the news that smell loss could be a symptom of COVID-19 jolted him on another level too—as a graduate student in his early 20s, Datta had briefly lost his own sense of smell and taste. For him, it was a side effect of chemotherapy. “It was pretty horrible,” he recalls. Emotionally, he felt disoriented and disconnected, strangely set adrift. Physically, it became almost impossible to eat: “I just found nothing palatable at all.” What finally worked was fried egg and cheese sandwiches—even without taste or smell, the saltiness was perceptible, and the texture made them easier to swallow. For a while, this was the only thing he ate. “I had a cholesterol of 300,” he says, “but it got me through.” And after a couple of months, his smell and taste recovered. “So, I have some sense,” he says, “of what it’s been like for people with this virus.”

As a scientist, though, he also knew something else: there weren’t going to be many answers for those patients’ questions, at least not yet. Despite years of research in labs like his, much about olfaction remains, essentially, a mystery. “So much is still just open science,” says Datta, who last year led a study that uncovered how COVID-19 seems to disrupt the sense of smell. “Right now, there’s a lot of intense interest in smell,” he says, “from physicians and from the many millions of patients who’ve had their sense of smell affected. And it has really highlighted, collectively, how little we know about all aspects of our sense of smell.”

The Bonus Sense

Researchers do have a grasp of the rudiments: that specific odor molecules bind to matching receptor proteins in the nose’s sensory neurons like keys in a lock, and that when each lock is opened, an electrical signal travels to the brain’s olfactory bulb, which in turn relays the message to other parts of the brain, where it is processed further—the piriform cortex, which identifies smell; the thalamus, which acts as a relay station; the orbitofrontal cortex, which is involved in taste. But even this knowledge is somewhat recent. The landmark genetic study identifying hundreds of different olfactory sensors in the nasal neurons was published only in 1991. That breakthrough won the Nobel Prize for its authors, biologists Linda Buck and Richard Axel, and opened the door to a whole new universe of research. Still, 30 years later, much of the olfactory system remains unmapped.

One reason for the persistence of this mystery is sociological, Datta says: humans are visual creatures. Losing the sense of sight is a psychologically devastating change that substantially impairs people’s ability to navigate everyday existence. Fully one-third of the human brain is devoted to processing visual information. No surprise, then, he says, that from the beginning, modern neuroscience focused most intensely on deciphering sight (and, then, hearing). “We have tended to think of olfaction as a kind of bonus sense, an aesthetic sense,” Datta says, “an accessory to these more essential sensory processes.” That comes through in the paucity of language to describe it. Vision and hearing abound with adjectives, but humans’ vocabulary for what something smells like is fuzzy and fragmentary and highly variable. This is a huge hurdle for science: it’s extraordinarily difficult for people to convey their olfactory perceptions in a way that is comprehensible to researchers. 

Scientists have no such command over the levers of smell. “I mean, what is smell, precisely?” Datta asks. “If I take a sniff of my morning coffee, that’s not actually a thing. It’s 800 separate volatile chemicals.”

Then there is the problem of smell itself. Scientists have a robust understanding about the dynamic components of a visual object or a sound—attributes like shape, color, light intensity, pitch, volume, frequency, direction—and they can vary these elements with precision during laboratory experiments to stimulate the brain and study how, for instance, a brighter color or a louder sound is processed. But scientists have no such command over the levers of smell. “I mean, what is smell, precisely?” Datta asks. “If I take a sniff of my morning coffee, that’s not actually a thing.” Or at least, not one single thing: “It’s 800 separate volatile chemicals that are coming off into the headspace above the cup, all of which exist at different concentrations, which my nose detects and my brain synthesizes into a unitary percept of coffee.”

It’s not clear, Datta says, how these chemicals interact with each other, or with the nose; it’s also not clear which chemicals mean the most to the olfactory system, and which it ignores, or under what circumstances it registers any given chemical as pleasant or unpleasant. Sulfur, the rotten-egg smell so noxious to humans it is added to odorless natural gas as a safety warning for gas leaks, is also an essential component of garlic, onions, and certain perfumes; the purified compound MMB, which is what makes cat urine smell intolerable, is also sold, at low concentrations, as a food additive to enhance flavor. “We don’t know the axes that odor chemistry is organized along, or how they matter to the brain,” Datta says. Changing as little as one odor molecule can dramatically alter olfactory perception “in ways that we, as scientists, simply do not understand right now.”

The Olfactory Cocktail Party

And yet, olfactory research holds a tantalizing promise: by unraveling the intricacies of smell, neuroscientists might be able to crack open deeper mysteries of how the brain itself works. That’s partly because smell is thought to be the earliest evolved sense in mammals. The olfactory bulb sits near the bottom of the brain, layers below the more recently developed folds of the neocortex. Its neural circuits are ancient, and—unlike vision and hearing, whose signals must trace a longer path—intimately connected to other primordial brain centers: the hippocampus, where memories are stored, and the amygdala, responsible for processing many aspects of the emotional world, including fear and threats. (The amygdala is thought to play a major role in anxiety and post-traumatic stress disorders.)

Odors have the power to trigger intense memories and emotions, and can profoundly influence mental health—as many COVID patients attest, people who suddenly lose the ability to smell often struggle with depression and emotional wellbeing. Loss of smell is linked to increased mortality risk and considered an early warning signal for neural illnesses like Alzheimer’s, Parkinson’s disease, and schizophrenia. Children with autism have a heightened sensitivity to smell and a different sniff response from neurotypical children. “This kind of intimacy between sense of smell and these parts of the brain that are fundamental to our human experience is super intriguing,” Datta says. “The simplicity of the circuits, and the directness of their connections to places like the hippocampus or the amygdala, offer a potential window into how our brains might sense information and transform it into a memory or an emotion, and then ultimately a behavior.”

That same curiosity is also what attracted Venkatesh Murthy, Erikson life sciences professor of molecular and cellular biology and Finnegan Family Director of the Harvard Center for Brain Science. For Murthy, too, olfaction seemed a useful way in to larger questions about the brain. The reasoning was partly practical. In neuroscience, one common research subject is mice, whose neural circuitry approximates humans’ in important ways; and for mice, smell is dominant, the most profound sense they have for navigating the world. (In mice, it is smell that takes up one-third of the brain.) Researchers can train rodents to respond to specific odors remarkably easily. “We think of animal training as difficult,” Murthy says. “For people who train dogs or horses, for instance, it’s very laborious. You cannot just tell them the rules, because they can’t understand you, and so you have to reinforce the right behavior for each command over and over.” But communicating through smell is remarkably smooth. Essentially, scientists can tell mice the rules. “It’s such an intuitive sense for them,” he says, “that we are able to train animals to do very complicated—or at least, what we think are very complicated—tasks. And they are able to do them beautifully.”

Some of those tasks explore the connection between olfaction and memory: how animals use smells to store memories, and how that process might alter neural activity for those smells. (Murthy has found that it takes mice only a few minutes to make the connection between an odor and a specific corresponding reward, that a whiff of vanilla, for instance, means a sip of water, or a chocolate chip.)

 Other experiments, which Murthy calls “the olfactory cocktail party,” use mice to help decipher the brain’s strategies for sorting through the cacophony of messages arriving from the tens of thousands of neurons in the nose. “Let’s say you’re in a room where somebody is brewing a fresh pot of coffee and they also put on some really nice perfume,” Murthy says. “And let’s say there are also flowering plants in the background, and food on the stove.” How does a person recognize any one smell, when it’s embedded in so much clutter and chaos? Especially given that each odor is not a single chemical, but numerous ones? “This is a long-term question,” Murthy says, but mice offer clues: in a mixture of up to 16 smells, they are consistently able to identify whether a particular smell—say, banana—is present. One hypothesis is that animals somehow train their sense of smell; after all, sommeliers and coffee tasters manage a similar feat. Or the explanation could be more purely biological—in mice, the synapses of the olfactory cortex appear more plastic than in the visual cortex, Murthy says. “So maybe the brain does have this ability to rewire connections and make associations.”

Recently, Murthy has been investigating a new mystery: how animals follow scent trails. It’s a deeper question than it seems. “You might think, how hard can that be?” Murthy says. “But actually, it’s not so easy. I mean, close your eyes and imagine that you smell something. What do you do next? Where’s the next part of the trail? It’s completely unobvious.” To approach this problem Murthy uses an inkjet printer to print rose-scented trails—sometimes straight, sometimes curved—onto long strips of paper, which he fastens on to a contraption that acts as a treadmill for mice. Four separate video cameras capture every move as they sniff their way down the page, and AI software converts the videos into data. “The next step,” Murthy says, “is to use electrodes to start understanding which parts of the brain are involved.”

In the meantime, he’s taken a potentially elucidating sidestep: running the treadmill experiment with carpenter ants and pheromone trails. It’s yielded some intriguing results, he says. On his laptop, he pulls up a brief clip of one of the ants, up close and in black and white. It proceeds slowly, losing its way, correcting the mistake, turning around, proceeding again. Ants do not have noses—their sensory neurons are located in their antennae, and in the video, the antennae sweep from across the trail, sometimes opening wide like windshield wipers, sometimes bending and stretching like an extra pair of legs. “We were shocked to find they were so active with their antennae,” Murthy says. Ants are too small to measure their neural activity, “so we can only watch this behavior, we can’t look inside yet,” he says. Turnng back to the video, he watches the mystery unfold in front of him, while the ant on the screen keeps moving slowly forward, smelling its way back to the trail, not knowing exactly where it’s headed.

COVID-19 and Renewed Urgency

In his own lab, Datta is trying to answer a similar question: how smells help the brain build models of the world. Those models allow animals (and humans) to make predictions about their surroundings and then decisions about what to do—whether to turn to the right or the left, whether to run, or eat, or fight, or mate. One study Datta has been working on probes a phenomenon called adaptation. “You know how, when you first step into someone’s kitchen, you smell all the smells of the cooking food?” he asks. “But then, over time, it dissipates and you stop noticing all the smells.” That’s adaptation, the olfactory system’s way of allowing the brain to focus on what’s new or important, rather than what’s simply there. “So, if something’s burning, you’ll smell it, as opposed to having your senses overwhelmed.” Datta wanted to learn how that process works, specifically in the olfactory neurons in the nose. “Traditionally, people have thought this process occurs in the brain, but we’ve been asking whether it actually happens before information even gets to the brain.” As part of the study, he sequenced cells in mice’s olfactory epithelium (the thin tissue of neurons and surrounding cells lining the upper nasal cavity) to determine what RNA each cell expressed.

That was the project Datta was working on when the pandemic struck and everything abruptly shut down. Marooned from his lab and reading the proliferating accounts of COVID patients losing their sense of smell, he realized that the sequencing data he’d amassed—and similar stockpiles in the hands of other smell researchers he knew around the globe—might point to an explanation. “We were lucky that we had all this stuff on our hard drives,” he says. A few months later, in July 2020, Datta and 24 coauthors published their findings. Early analyses had shown that the virus attaches to its host using the ACE-2 receptor protein. But that protein is not expressed by the olfactory neurons; instead, it’s expressed by cells surrounding the neurons—stem cells, which allow damaged neurons to regrow, and “sustentacular” cells, which provide physical and metabolic support. The researchers theorized it was those surrounding cells the virus was infecting.

This idea might also explain why some patients recovered their smell quickly, and some patients not at all. If the infected cells were so damaged that the neuron also died, it would take months for the neuron to regrow. And in some cases, perhaps the viral destruction in the epithelium, especially to the stem cells, was so complete that the neurons would never be able to regrow, and sense of smell would never return.


Venkatesh Murthy
Photograph by Stu Rosner

Before COVID, Datta says, “It was hard to get many people—some scientists included—to pay attention to smell as a legitimate wedge that one might use to understand the brain. And I think that’s really changed now.” For researchers like him and Murthy, the fresh urgency directed at their field is an unfamiliar feeling, but a galvanizing one. “I’m excited to begin to think more about the underlying problem of smell itself,” Datta says. Murthy foresees a renaissance in olfaction research. Lately he’s been contemplating how breathing and smelling might intertwine neurologically.

Much of the new primacy felt by researchers like Datta and Murthy has to do with the increasingly acute, COVID-driven need for therapies. “Right now, we have no clinically validated treatments for the loss of smell as a result of a virus or trauma,” Datta says. Some studies suggest modest efficacy in the practice of “smell training,” in which patients try to recover their sense by regularly breathing in specific odors, “But for the most part,” Datta says, “I don’t have a drug that I can give you that will fix your broken olfactory system. We don’t even know for sure what level those interventions should be made at. If you lose your sense of smell because your nose has been damaged by the coronavirus, is it enough if I simply fix your nose? Do I also have to fix the brain? We just don’t know.”


Eric Holbrook
Photograph by Stu Rosner

Eric Holbrook began hearing about the COVID-related smell loss a few weeks before it started showing up in his clinic. Director of the rhinology division at Massachusetts Eye and Ear and associate professor at Harvard Medical School, Holbrook had seen bad colds and other viruses occasionally knock out patients’ olfactory systems for long periods, but he’d never seen the sheer number of smell-loss patients that COVID-19 produced: whole families, friend circles, half the floor of a single dorm. And those with long-term loss skewed younger—rather than in their 30s or 40s, COVID patients were often college-aged, or not much beyond, Holbrook recalls. “I saw one kid who was nine.”

The lack of proven, reliable treatments for smell loss can drive patients to desperation, Holbrook says. “It puts physicians in the position of wanting to try everything. And that can be dangerous.”

About 85 percent of COVID patients with mild infections seem to suffer loss of taste and smell—for many, it is the earliest symptom, or the only one. And although most regain their ability to smell within three or so weeks, for as many as 35 percent, the loss lasts longer. Those conversations with patients can be hard, Holbrook says. There is usually not much he can do. Smell training sometimes works: a months-long regimen of inhaling smells twice a day in an effort to reactivate the neurons. But the lack of proven, reliable treatments can drive patients to desperation, Holbrook says. “You wouldn’t believe the number of people who ask me about so-called therapies”—medicines, devices, surgeries abroad—“that have no proof, or even very much scientific rationale. It puts physicians in the position of wanting to try everything. And that can be dangerous. It can also lead to false hope.”

Like Datta and Murthy, Holbrook is surrounded by unknowns. Although he has some ideas, he is not sure what to make of the numerous reports from COVID patients who say they are incapable of detecting only certain odors—“bathroom smells” especially. And although he knows it’s usually a good sign when patients progress from an absent sense of smell to a distorted one—in which coffee might smell like sewage, or food like cigarettes—he cannot yet explain exactly why. Doctors have long recognized it in trauma patients as an indication that the olfactory system is working to heal itself, and the neurons perhaps “mis-wiring” as they grow back, but the mechanism remains murky.

Olfactory sensory neurons are a rare part of the human nervous system capable of regeneration. That’s what first gripped Holbrook, as a medical student sitting in a classroom listening to a lecture on neuroanatomy. After neurons die, they can come back. “That was completely fascinating to me,” he says. It still is. For the past several years, he has been collaborating on possible therapies for smell loss. One, led by researchers at Tufts University, involves stimulating the system’s stem cells. “In a lot of cases, after damage has occurred,” he says, “those stem cells are sitting there, very quiet. And it looks like there are ways we can tell them to start dividing and making neurons again.”

Another project, which Holbrook sometimes calls “a cochlear implant for the nose,” would use electrodes to stimulate the nerves in the olfactory bulb. It’s still a distant dream, but in 2018, he and several collaborators conducted a small experiment—“a proof of concept”— placing electrodes inside the nasal cavities of five patients with intact senses of smell, very close to the olfactory bulb. After the researchers administered an electric current, three patients described experiencing sensations of antiseptic, sour, and fruity aromas. “One said it was ‘onion-like,’ but not an onion,” Holbrook says. The results were encouraging enough to warrant more research. Neurosurgeon Mark Richardson, who founded the Brain Modulation Lab at MGH, is pushing that effort forward.

In the paper Holbrook and his colleagues published after their 2018 experiment, they estimated that five percent of the population suffers from total smell loss. After COVID-19, that number seems almost certain to rise. “The implications of this virus are huge,” Holbrook says. Some of his patients who lost their sense of smell early in the pandemic still have not regained it. “There’s a big, significant number of people who are going to be potentially without smell for the rest of their lives,” he says. “There is so much work that still needs to be done.”  

You Might Also Like:

Illustration of a seesaw with a house on one end, and an apple, representing teaching, on the other.

Illustration by John Tomac

Optimizing Public-School Spending

A smokestack spewing smoke with a wind turbine in its shadow

Illustration by Taylor Callery

China’s Excess Wind Energy

A sample of slime mold grows clockwise in a petri dish

A sample of Physarum polycephalum sends out slimy, finger-like tendrils to sense information about its petri dish environment.

Image courtesy of Nirosha Murugan, Levin lab, Tufts University, and Wyss Institute at Harvard University

Can Slime Molds Think?

You Might Also Like:

Illustration of a seesaw with a house on one end, and an apple, representing teaching, on the other.

Illustration by John Tomac

Optimizing Public-School Spending

A smokestack spewing smoke with a wind turbine in its shadow

Illustration by Taylor Callery

China’s Excess Wind Energy

A sample of slime mold grows clockwise in a petri dish

A sample of Physarum polycephalum sends out slimy, finger-like tendrils to sense information about its petri dish environment.

Image courtesy of Nirosha Murugan, Levin lab, Tufts University, and Wyss Institute at Harvard University

Can Slime Molds Think?