Inside the extraordinary nose of a search-and-rescue dog
The blizzard arrived in Alpine Meadows without warning.
Hours earlier, before a half moon rose over this California resort town near Lake Tahoe, an uncle and his teenage nephew had been separated from their hunting party. Authorities called veteran rescue dog handler Shay Cook, who rushed to the mountain site with her dog Rixi, never imagining it would be one of the most challenging searches of her 20-year career.
Cook and Rixi arrived at the trailhead at 10 p.m., after a deep drive up the Sierra Mountains. The forecast had called for light precipitation, and a soft glaze of snow was starting to cover the ground as the search party trekked into the woods.
“By 2:30, it was blizzard conditions, and I was in a drainage, so I had no radio reception. My GPS wasn’t working,” Cook said. “But the snow did not seem to pose a problem for Rixi. She was like a little bulldozer and just led a path through the snow.”
Deconstructing Rixi’s nose
A foot of snow buried any obvious traces of the hikers, but Rixi’s snout persevered. The five-year-old German Shepherd charged through the wet flakes, finding the hunting party’s abandoned campsite, which by that point, was buried under snow. From there, Rixi took off, marching uphill as the blizzard closed around them.
Rixi is what rescuers call a trailing dog, specially trained to pursue human scents over difficult terrain. Her nose, like most expert canines, can catch a scent from a mile away — five times farther than we can. These pooches can then follow this smell for hundreds of miles, which allows them to catch bad guys — or, hopefully, save disoriented hikers.
Yet the basic mechanics of these olfactory talents remain a mystery — a mystery that neuroscientist Lucia Jacobs is trying to solve.
Jacobs works at the University of California Berkeley and has teamed with Cook, her rescue dogs and a bevy of other animals to decipher how smell behaviors function across different species.
In the past, Jacobs lab focused exclusively on the spatial navigation of rodents — voles, kangaroo rats and multiple squirrels. Now, as part of a new nationwide collective called Cracking the Olfactory Code (COC), her research is branching into other animals — rescue dogs, humans, hermit crabs, two species of cockroaches and the leopard slug — to determine how smell governs navigation.
“We’re doing all these weird species because they orient the same way,” Jacobs said. “They show the same behaviors when responding to odor, but they’ve got completely different nervous systems. So we think there’s got to be a common algorithm that they’re all using.”
Speed dating with Nobel Prize winners
The scientists behind Cracking the Olfactory Code are a motley bunch. Jacobs keeps her ark of animals. Engineer John Crimaldi visualizes and videotapes smells as they travel through space. Bard Ermentrout, an applied mathematician who lost his sense of smell after catching a cold, codes the smell reflexes of mice into an electronic brain. The seven-member team unites a smorgasbord of talents — and their mission began in the middle of nowhere.
In May 2015, 30 scientists descended on the Janelia Research Campus, a 689-acre facility, where Virginia outback meets modern design. They had been invited to the campus by the White House Brain Initiative and the National Science Foundation to engage in a weeklong “Ideas Lab” to tackle one of neuroscience’s hardest fields: olfaction.
“Your sense of smell is a huge — one of your five senses, yet we don’t understand how the brain does it,” Jacobs said. Consider, she said, the act of smelling coffee. Scientists know the first step (the coffee scent drifts into the nose and activates smell neurons) and the last step (brain thinks coffee) — but much of what’s in between is a black box.
“We have no idea how coffee gets encoded,” Jacobs said. “And that’s crazy, because in vision we know it all the way back. We can trace a photon from hitting your retina to all the way through all your visual brain cortices.”
The week started with a bout of brainstorming, modeled after speed dating. The scientists were directed to form an outside circle of 15 people and an inside circle of 15, and then to pair off into teams of two, Jacobs recalled. Then the organizers blew a whistle, and each team had five minutes to come up with a collaborative project. The first days included other camp-style activities, such as playing with Play Dough, tossing around Slinkys and sketching the faces of other scientists onto paper bags.
“It was a really unusual way of generating a project. Some people who were there called it Survivor for scientists,” said Nathan Urban, an olfactory neuroscientist at the University of Pittsburgh, who is now on the Cracking the Olfactory Code team. “But the idea was to bring together people from a number of different disciplines — biologists, physicists, mathematicians, statisticians — to think about the problems that are confronting the field.”
The seven-member COC team emerged from the fray as victors, splitting a $15 million pot with two other teams. Their goal, they determined at the time, was to answer three overarching questions: What does a smell look like? How do animals react to smells? And can you program those reflexes into a robot brain?
Circling ain’t just for the dogs
Lucia Jacobs’ lab is focused on the the middle question: How do animals react to smells? When Jacobs observes dogs, leopard slugs, a death’s head cockroach or humans, she see similarities in how they move in reaction to smells. Her lab’s contribution to the COC team is to make quantifiable measurements of these behaviors.
“With the dogs, we can put microphones near their nose to measure their sniff rate. We can put GPS trackers and GoPro cameras on them,” Jacobs said. All this tech is needed, because dogs have evolved a range of intricate tricks to follow smells. Take circling for example. Jacobs said dogs and other smell navigators circle when a scent starts breaking up. Odor isn’t a uniform cloud. Its shape is scraggly and wild, with portions that are stretched like taffy.
Sitting inside an odor plume doesn’t tell you much. You’re surrounded by a jumbled mess, but traveling along the edge of an odor offers context and directions.
“You also see the same pattern in homing pigeons, which rely heavily on olfaction when they’re released from a new place,” Jacobs said. “They circle before they take off, then they seem to pick up something and take off.”
But scientists don’t possess much quantitative data on these behaviors. Jacobs plans to commit the same watchful eye to the other animals, such as hermit crabs and cockroaches. Those creatures smell using two antenna, which allow them to sniff things in stereo, akin to hearing with two ears.
Imagine a world of three-dimensional smells, and you’re probably not far from being inside the olfactory brain of a crab.
Inside the nose of a rescue dog
Jacobs and other scientists suspect dogs smell in stereo too, due to their complex noses. A decade ago at Pennsylvania State University, engineers Brent Craven and Gary Settles uncovered evidence for this pattern by taking MRI scans of canine noses. What they found was a labyrinth. The scans revealed a huge maze of folded passageways — called turbinates — that filter air into different parts of the nose.
“If you spread out the folds, it would be the size of a handkerchief,” Settles said. Human noses contain turbinates too, but in terms of surface area, ours are smaller — on the order of three post-it notes.
Dogs and their wolf ancestors may have evolved long snouts to house these many turbinates, which come in two sets. One group is a humidifier that warms air before it heads into a dog’s lungs.
“A grey wolf has to be able to inhale very cold air in the winter, so there’s an intricate web of blood vessels in those turbinates that exchanges heat,” Settles said.
The other set of turbinates detour airflow to the back of the nose, where this nasal maze separates odor chemicals by their weight and solubility — the ability to be dissolved in water. These filtered smells then come in contact with olfactory neurons — a nose’s scent sensors. Filtering is crucial, as a dog like a bloodhound may pack 300 million smell sensors into its nose, compared to our 5 million.
Scents that are very soluble will be picked up first, by the earliest olfactory neurons, while less soluble scents will go straight through. Limonene, the chemical odor behind the pleasant aroma of lemons, for example, doesn’t readily dissolve and may occasionally be missed by dogs.
Inside the nose, turbinates run the odor show. But the outer layer of the noses carries valuable features too. For instance, computer simulations of airflow around the canine nose have found that each nostril pulls in a separate sample of odor. Sniffing in multiple directions at once allow dogs to keep tabs on gradients of odor in the environment. The feat would allow a rescue dog to distinguish between a fresh and an old scent trail, Craven said.
Slits on the side of the nose allows a canine to sniff in and out at the same time. In the length of time that a human takes three breaths, a dog might take six or seven, Settles said. Other super-sniffers like bears and deer sport these slits too.
“What we’ve found is that most “keen-scents animals” all seem to have a very similar architecture of their nose,” said Craven, who now works at the U.S. Food and Drug Administration, where he applies his computer models of fluid mechanics to life-saving devices like blood stents.
Dogs may smell like dinosaurs
Why the evolutionary fuss? What caused dogs and other stellar smellers to develop these complicated mammalian systems, while humans and other primates landed the short end of the smell stick?
Jacobs has a theory, and it starts with with the olfactory bulb, a relay station just above the nose that sends smell information from the initial smell sensors deeper into the brain.
“If you look at the size of the olfactory bulb, that actually scales with how animals use space… with the size of their home range,” Jacobs said.
Millions of years ago, dinosaurs may have relied on the same pattern. Scientists recently spotted this trend by examining X-ray scans of the skulls of theropod dinosaurs, which includes the Tyrannosaurus rex. The olfactory bulbs of these dinosaurs protrude from a little cavity, making it easy to measure.
“We found that hyper-carnivores, like raptors and tyrannosaurs, had larger-than-average olfactory bulbs,” said Darla Zelenitsky, a dinosaur paleontologist who led the study at the University of Calgary, Canada. “Small raptor dinosaurs had olfactory bulbs similar in size to those of turkey vultures, birds today with large olfactory bulbs and well known for using smell to locate food.”
By contrast, plant-eating or omnivorous theropods, like ornithomimids, which are similar to the modern ostrich, tended to have smaller-than-average olfactory bulbs.
Although this is impossible to test in extinct dinosaurs, it is possible that the size of olfactory bulbs in carnivorous dinosaurs correlated with the size of their home range or individual territory Zelenitsky said, adding that large olfactory bulb sizes in these dinosaurs may have aided foraging and hunting at night.
The modern human, by contrast, relies on smells for short distances.
The early research by Craven and Settles was sponsored by the Navy, and focused on whether the canine nose could be reverse-engineered. Likewise, the Cracking the Olfactory Code team aims to build a robot that might one day assist rescue or bomb-sniffing dogs.
The need for such automation is sometimes dire, as Shay Cook can attest. After finding the first campsite, the night of the Alpine Meadows blizzard, Shay and Rixi carried on as long as they could.
“My foot was sliding between the snow and kept getting stuck in the cracks between boulders,” Cook said. “It was a little too dangerous by 2:30 in the morning, with radio reception and GPS not working.
Cook and Rixi halted their search around 3:30 a.m., without finding the hikers, and after about six hours in the blizzard. Later, Cook learned that other rescue team members had been extracted due to weather and altitude sickness. The uncle was found three days later, suffering from severe hypothermia. His 14-year-old nephew had lasted a day before dying.
That’s one unfortunate aspect of wilderness rescue, Cook said: They don’t always find live people. And though a smell robot or a drone is a distant possibility, it could, she said, aid the process and make all the difference.