Some research suggests that “olfactory training” — practicing to improve your olfaction, or sense of smell — can actually improve your cognitive abilities. So in the interest of self-improvement, I open my pack of what look like markers and sniff at them one by one. I get whiffs of citrus, lavender, driftwood. From a marker enigmatically labelled “hairy,” I register a greasy smell. From another, ominously labelled “barnyard/fecal”, comes the distinct smell of ass.
More appealing is a little bottle that, when opened, emits a scent of plum so vivid it brings to mind the deeper red closer to the pit, the exact feeling of cutting a piece, the interplay of taut skin and juicy flesh as you bite and chew. It brings to mind not a Platonic plum, but a specific plum I must have enjoyed on some long lost summer day. And yet the particles I inhale to recapture that sense memory were never part of a real fruit. They are a re-creation of the molecular structure of scent particles emitted by a real plum in a different country from where I sit, transported to summer on a January day, courtesy of Osmo Labs, which sent me the samples.
Yes, friends. Digital smell has arrived.
The elusive sense
Of all the senses, olfaction — smell — must be the most primal and evocative, the most magically elusive. With just around 350 olfactory receptors, we can each smell a vast array of different individual molecules, called odorants (one model suggests there could be as many as 40 billion such smellable molecules!).
Research suggests that there are multiple benefits to increasing the diversity of our olfactory environment, which for humans is limited compared to our visual or even auditory or tactile worlds. The Monell Center, a Philadelphia-based independent research center devoted to smell and taste, quotes one of their neuroscientists, Joel Mainland, on olfaction: “Before COVID, [people] didn’t take it seriously. Now, more people are realizing how important their sense of smell is. It’s not a throwaway sense!”
By no means. Disorders of the olfactory system, and growing evidence of how serious they can be, inspired the use of such olfactory training beyond perfumiers or scientists specializing in smell. My sister-in-law, who like thousands of others lost her sense of smell after an early COVID-19 infection, has gradually regained much of it through a disciplined program of smelling things, such as citrus or coffee.
"That’s the only part of the brain that leaves the skull and touches the air."
But many questions about smell remain unresolved. How do we distinguish so many odors? What is the relationship between a molecule’s structure and the scent it evokes? And most importantly, as technology moves virtually everything from analog to digital experiences, why can we not yet send Stinkygrams? Why is the Cloud not scented? Why does 4D cinema suck so bad? As yet, nobody knows — but we’re getting closer to the answer, perhaps making digital smell a future mainstream wing of the digital experience.
The good news: smell exists
Smell is quite different from sight, where the brain relies on information relayed from a limited range of wavelengths of light that is registered by just three types of cones in our eyes, allowing us to perceive a palette of about a million colors. In olfaction, those under-400 olfactory receptors combine, in ways that have been mysterious to researchers, to produce the staggering variety of odors we can perceive.
“You and I have about 350 olfactory receptors but our sets are different from each other,” Mainland, of the Monell Center, told Salon in a video interview from Philly, which he says has become a de facto capital of olfactory research in the U.S. He found, however, that when you set up panels of 15 to 20 people to sample different scents, the average from one group of that size to another really smooths out the variation between individuals. The panels don’t tend to differ from each other significantly, which is allowing scientists to begin developing databases of odors.
Another way in which it’s different from sight: color (probably) doesn’t actually exist out there in the world, but smell does. Or to state it differently, we don’t stick bits of color into our eyeballs to perceive colors, but with smell, we must literally insert odorants into our lucky (or unfortunate) noses in order for us to perceive them. They later succumb to recycling as nasal mucus that finds its way into our digestive systems. And still another difference is that olfaction is the only sense that involves the brain directly reaching out for information from the world: the olfactory bulb in the nose is literally composed of brain tissue, the olfactory epithelium.
Sigmund Freud thought smell related to animalistic behavior and was relevant in humans mostly in terms of behavioral pathology.
“That’s the only part of the brain that leaves the skull and touches the air. Your brain is physically moving through the skull. There’s nerves that go through porous holes and become our sense of smell, and so molecules physically are touching your brain,” Alex Wiltschko, a neuroscientist who worked at Google Brain, using machine learning to create a map of smells, told Salon in a video interview. Wiltschko is now CEO of the company he founded out of that work, Osmo Labs, dedicated to, as its tagline puts it, “giving computers a sense of smell.” Nobel Prizewinner Geoffrey Hinton, the “godfather of AI,” recently joined the board of Osmo.
And yet the sense of smell has been routinely undervalued. Sigmund Freud thought it related to animalistic behavior and was relevant in humans mostly in terms of behavioral pathology. And a 2011 poll of 700 young people around the world by the marketing company McCann Worldgroup found 53% of 16 to 22 year olds and 48% of 23 to 30 year olds polled were ready to sacrifice their sense of smell if they’d otherwise have to give up an item of technology such as their phone or laptop.
“We think of this as a sort of a layer of information that we’re not currently great at interpreting,” Mainland explained.
Most species, including most mammals, really rely on olfaction to make sense of the world. But as humans, we tend to feel that our sense of smell doesn’t give us very precise or clear information. If we could find reliable ways to extract the information inherent in scent, there are so many ways this could be useful. This is where digitization of scent comes in. As recently as 2016, biophysicist Luca Turin described to a journalist from the Guardian two technical obstacles to digitizing scent: the need for an actual chemical in the air on the receiving end, and the lack of a scent equivalent of primary colors you could mix to create more complex smells. That publication also described the commercial failure of a scent messaging device, inventor and Harvard professor David Edwards’ “oPhone” — nobody was terribly interested. Perhaps we’re still culturally traumatized by Smell-O-Vision.
“If you were early in vision, and sort of asking, “why would you want to digitize vision” There are lots of inventions that came about that you never would have conceived of, right? So I think there’ll be lots of applications for this,” Mainland said.
A fresh-cut plum
Which brings us to the plum. A few months ago, Wiltschko’s Osmo Labs hit a major milestone: they digitized the smell of a freshly-cut plum. Osmo sent me dozens of odor training samples, including three newly-synthesized, never-before-smelled molecules, the fragrances created with them, and that sample of the world’s first digitized — in the company’s marketing language, “first teleported” — scent (Edwards similarly called the scent of champagne and macarons he “sent” from Paris to New York in 2014 “the world’s first transatlantic smell message.”)
The way Osmo’s digitization process works is that the researchers trap and concentrate a physical scent emitted from a source (a stinky sneaker, say, or a rose, or a freshly-cut plum) in the form of volatile molecules. The researchers then analyse the scent sample to figure out the exact type of molecules involved, and the proportions of each type of molecule found in that air sample. They send that information, which is essentially the chemical formula for that little pocket of air, to a printer equipped with all the molecules or similar-smelling ones, and recreate the exact chemistry that was found in the original scent. Only the printing process could take place miles, or continents, away from the source of the original smell.
This seems like real digitization, unlike the oPhone’s vials of scent cartridges tagged to release the indicated smell. Still, the company makes bold promises indeed regarding the impacts this service could have.
“I think it will be impossible to catalog all the ways in which digitizing our sense of smell will positively impact our lives,” Wiltschko told Salon. “So what we’ve managed to do is demonstrate that with no human intervention, we can digitize a smell, and now what we can do is do that routinely and cheaply for every smell.”
"It will be impossible to catalog all the ways in which digitizing our sense of smell will positively impact our lives."
There’s no particular limit to the kinds of smell that can be digitized, though Wiltschko says Osmo is at work on a database with the molecular compositions of everyday smells, such as everything you might find at the grocery store, for example. He doesn’t intend to stop at retail, though.
“For instance, I have a newborn coming in January. It’s my first. And I know that babies smell great,” Wiltschko said. “And I want to save the smell of babies. And it’s the same problem as saving the smell of a plum.”
Wiltschko noted that George Eastman, of Kodak fame, introduced the quick and easy preservation of memories dependent on sight back in the 19th century, just as he hopes to do with olfaction here in the 21st. “I really think that we should be able to save the olfactory memories of our family,” the father-to-be told Salon.
Other company initiatives include using chemical sensors and AI training on large data sets to authenticate products — designer sneakers, say, or luxury skin cream or wristwatches — by detecting every volatile molecule released by it; an AI-powered fragrance designer that can generate scents based on your description of a memory or other imagery conveyed in words; and a platform to develop new aromas to repel disease-carrying insects like mosquitoes.
But all of these digitization projects, and others not yet dreamt of, depend on research. We still have a long way to go in understanding how odorants work their magic on our noses — or, rather, how our olfactory systems work their magic on those insignificant-seeming particles.
“We have instruments that basically tell you exactly how loud something is, exactly what color something is, based on physical principles, and we don’t have that in olfaction, right?” Mainland explained. “So there’s no way that I can take an instrument and point it at something and have it tell me that that smells fruity, right? We have to sort of interpret that in a complicated way. And really the only source of truth we have on that is asking people what something smells like. So it’s very different from the other senses.”
Our few hundred olfactory receptors (far fewer than in animals like mice, dogs or elephants) are broadly tuned, with any one receptor responding to many different molecules, and one molecule often activating most or almost all of the receptors, with the relative ratios of their activating apparently providing the olfactory code, a kind of signal that our brain interprets. How exactly the brain pulls this off is still a mystery.
The dataset on which machine learning models are trained are key to using AI to figure it out.
“Part of the challenge here is scaling this up to get enough molecules to people to smell and get reasonable reports of those [so] that we can create a large data set and improve the models on those things.” Mainland explained.
Smellorism terrorism
But what are the downsides of future apps like Stinkygram or Nosebook? It’s not here yet, but Wiltschko imagines that we will be able to send scents via email or other online application one day.
With any new technology, of course, it’s wise to consider whether (or how) it will eventually oppress us. For a start, scent is powerful, which raises the everpresent spectre of military applications. Never mind a letter bomb, what about a stinkbomb that is also a letter? The recipient who unwisely opens an attachment might, for example, unleash thiacetone, which could render them unconscious with a sulfurous odor akin to a demonic leek — as well, possibly, as knocking out people within nearly a half mile radius.
Wiltschko isn’t too concerned. “Any new technology, of course, could be used for multiple purposes. What I love about scent in this context is that it truly is just scent, right? […] You may like it, or you may not like a smell. You may remember it, or you may not remember it. But scent is our brain’s way of sampling the chemical slice of reality, and those molecules come into our nose, and then they’re discarded, and they don’t [leave] a trace. And so I think that there’s something really beautiful and almost innocuous about what it is that we do, which is just to basically create memories through chemistry.”
COVID robbed our sense of smell
There’s more though. The most exciting applications — way better scent-enhanced entertainment aside, of course — may be biomedical.
Through changes in how you smell, “dogs and bees and people and all kinds of animals can definitely detect changes in your health,” Wiltschko said, but noted that it’s still unknown what’s actually being detected. “However, we know that whatever is in your blood and in your body does get to the outside. So your breath is a filtrate of blood. So is your sweat. Your earwax is the metabolic sink of all the molecules in your blood. So everything’s kind of leaking out of you at all times.”
Wiltschko hopes to collect this information using their platform and correlate it to health outcomes. He’s not the only one. Research on olfaction was already an area of booming interest by 2020. The arrival of a brand-new virus that had as a signature feature the tendency to steal one’s ability to smell further stimulated interest, and urgency. Things we already knew about the importance of the sense of smell to our cognitive processes and the association of its loss with neurodegenerative conditions suddenly acquired the extra importance of many tens of thousands of potential sufferers of dementia or other cognitive disorders in the coming years. There’s also been an influx of members to charities like the U.K.-based Fifth Sense or AbScent, which are devoted to people who have lost their senses of smell or taste, often a post-viral complication even before COVID existed.
We need your help to stay independent
And yes, a systematic review published last spring in the journal Neuropsychology Review found that it’s not just those who have lost their sense of smell who benefit cognitively from scent training. They found that through olfactory training, people with smell loss and people with normal olfaction alike can achieve increases in the size or volume of the olfactory bulb and hippocampus. And they can improve the functional connectivity of the different structures related to olfaction. In their review of 18 studies that met their criteria, the researchers even found emerging evidence that olfactory training is associated with improved global cognition, with an especially marked impact on verbal learning and memory as well as verbal fluency.
A fine New Year’s goal! So I continue to practice.
And, now and then, I take out Plum 1.0. It’s not a sophisticated smell with the complexity of a carefully constructed fragrance. Rather, it really does smell like a real-life plum, in a way quite different from the approximations of fruit in the average cherry-scented lip balm or even the most complex and luxurious citrus-inspired perfume — even though, unlike some of those, it contains no essential oils literally extracted from the fruit.
This ability to “print” the exact molecular structure of the desired odorant does seem like a significant advance in the reproducibility of scent. Will that lead us to dramatic improvements in our detection of disease… or merely a revival of scented markers and smelly stickers, sports merchandise that smells like unwashed star athletes, or spam attachments redolent of farts? Only time will tell — but the future smells bright.
Shares