Illuminating olfaction to explore brain function

Ula Chrobak was getting ready for her day last January when she realized she'd lost her sense of smell. Her nose couldn't pick up the perfumed scent of the lotion that she was using. She had already tested positive for COVID-19 three days prior, so she wasn't surprised.

Still, Chrobak worried how long the condition might last; some COVID-19 patients lose their sense of smell for months. To her relief, a week later, while driving, she passed a dead skunk on the side of the road and realized that she could discern its pungent aroma. "I was like, ‘Oh my God, I can smell,'" she says.

Scientists don't know exactly how the SARS CoV-2 virus interferes with the olfactory system. In part that's because smell is the least understood of our senses. Researchers know the nose is crammed with millions of olfactory neurons. But they're still discovering how the brain processes smells. For instance, how does the brain discern between the scent of sauerkraut versus chocolate chip cookies? And why does the scent of something from the past-such as our grandmother's perfume-trigger memories?

Photonics plays a major role in this neuroscience research. Investigators shine lasers at odors embedded with fluorescent molecules to visualize smells. They also bounce light off miniature mirrors to precisely excite light-sensitive proteins inside olfactory bulb neurons, allowing them to view the neuron-firing sequence. Finally, to see deeper parts of the brain in live mammals, they employ high-powered microscopes that use long-wavelength, lower energy light that can penetrate far down into tissue without damaging it and, once again, turn on fluorescent molecules inside neurons.

This technology can be found in the labs of researchers who are part of the US National Science Foundation-funded Odor2Action network, as well as the US National Institutes of Health-funded Osmonauts. The former aims to understand how animals use information from odors to guide behavior, while the latter wants to characterize how odor information is sequentially transformed by neural circuits to generate meaningful perception.

Both multidisciplinary projects work in accord with a bigger principle: that understanding the olfactory system can lead to greater insight into the workings of the brain.

Using olfaction to look at brain function

"We're using olfaction as a mechanism or vehicle to try and narrow down our approach of looking at brain function," says John Crimaldi, professor of civil, environmental, and architectural engineering at the University of Colorado, Boulder, who runs a lab that looks at interactions between fluid physics and ecology. "It's not as random as it sounds," he adds. "Even in the primordial soup, bacteria were able to sense chemical gradients around them and that's sort of the precursor to olfaction."

Indeed, because the human brain evolved in parallel with our sense of smell, Crimaldi and others say, the olfactory system may be the best window into how our complicated gray matter works. In addition to understanding why loss of smell is a symptom of some viral illnesses like COVID-19, such research could eventually be a steppingstone to a better understanding of neurological problems related to the sense of smell such as Alzheimer's and Parkinson's disease. Loss of smell is an early indicator of such illnesses; scientists believe studying this might reveal the molecular mechanisms involved in the beginning phases of the diseases.

"Ultimately, we believe this fundamental research will benefit human health," says Justus Verhagen, a fellow at the John B. Pierce Laboratory and an associate professor of neuroscience at Yale University School of Medicine. "That is a big thing."

Nevertheless, making the leap from understanding olfaction to understanding neurodegenerative diseases won't be easy. One of the difficulties of studying smell is that it's hard to control the amount of stimulus that subjects are exposed to. Researchers often use olfactometers, instruments that can deliver a more precise concentration of an odor to a subject's nose for better control. But olfactometers can't quite mimic the natural world. Odors ride along on air or water currents and disperse in different directions over time. They're also invisible.

Crimaldi has found a way around this problem. Inside his ecological fluid dynamics lab, he introduces surrogate odors—chemicals that behave exactly like odors but contain a fluorophore—into a wind tunnel and then shines a laser into it. The light excites the surrogate odor's fluorophore, causing it to re-emit light at a different color wavelength than the laser. Cameras outfitted with filters that only pick up the flourophore's wavelength capture images of the odor plume as it moves through the tunnel.

From there, Crimaldi can quantify precise measurements of a plume and use those measurements to recreate it digitally. He and his graduate students can then model how a virtual mouse might move through such a plume to locate a reward. "It's a virtual mouse in a virtual odor plume," he says. "But the odor plume is very close to an exact representation of what exists in nature."

Crimaldi and Verhagen are now working on using those odor plume measurements in a more realistic setting. They attach tiny olfactometers to real mice and place the rodents into a virtual reality backdrop, so the rodent sees a movie that makes it feel like it's moving. Data sets of odor plumes gathered in Crimaldi's lab then drive the scent released in the olfactometer so that the animal perceives the odor as if it were wafting through the air. This virtual reality setting allows the scientists to determine at what point into the plume and at what level of concentration the animal perceives it. The mice get a reward when they smell the odor and perform a task.

"So, both visually, tactically, in terms of motion, and then from an olfactory perspective, the animal is getting information that makes it feel like it's in an aroma," Crimaldi says.

While he and Verhagen's virtual reality work focuses on creating the most realistic odor stimulus possible, other investigators are looking more at the internal workings of the olfactory bulb itself, a part of the brain that sits just behind the nose and conveys odor information from the nose to the brain. Scientists already know that scents attached to air-bound molecules trigger receptor cell neurons that for animals and humans are located in the lining of the nose. Those cells then send electrical signals to bundles of nerve endings inside the olfactory bulb called glomeruli, which then transmit odor information first to the piriform cortex and then to the olfactory cortex, the brain's main processing site for olfactory information.

Until recently, scientists had not been able to track how a particular smell might play across these clusters of nerve endings. They knew that the timing and order of glomeruli activation was different for each smell but could not pinpoint what those exact patterns might be.

Then, in 2020, Dmitry Rinberg, professor of neuroscience and physiology at New York University, figured out a way using optogenetics. Rinberg took mice that had been genetically engineered to have special proteins in their olfactory bulb neurons that would illuminate when the researchers shined a light on them. Using a digital micromirror device, he activated a pattern in the mice's glomeruli that was known to evoke an odor—though not one found in the natural world—and rewarded the mice when they perceived this odor and pushed a lever.

Once the mice were trained to recognize the signal and receive a reward, Rinberg altered the timing and order of the activated glomeruli to determine how that might affect their ability to perceive the odor and act on it.

"Optogenetics gave us an opportunity to really manipulate this neuronal activity and see which features are more or less important," Rinberg says. "And we have discovered many interesting things. For example, odor perception is based on a temporal sequence of receptor activation, like a melody is a sequence of different notes. And we found that for odor identification the first note is more important than the second."

Rinberg says he hopes this discovery will help researchers figure out the minimum number and type of receptors needed by the olfactory bulb to identify a particular smell. He also dreams of being able to figure out how to recreate a real-world smell in an animal's olfactory bulb and watch how it responds. "We are not able yet to reproduce any odor found in nature, like a rose," he says, "because the smell of a rose has a lot of complicated features. So, this is still a challenge for us."

Indeed, Rinberg can't say exactly what odor his mice smelled. Instead, he had to infer that they were smelling an odor by experimenting with turning on patterns of glomeruli until he hit one that triggered the mice to get their reward as quickly as they would have when smelling a natural odor. "The logic here is that if these were very artificial smelling signals, it would take them a much longer time to learn the task," he says. "For example, if you can walk, and I ask you to walk on a treadmill, or with different shoes, you will learn to do it very quickly, but if I ask you to bike and you never biked before, it will take you much longer."

While Rinberg has focused on initial odor processing in the olfactory bulb, other scientists are looking at what happens when that information moves deeper into the brain—first to the piriform and then the olfactory cortex—and how it may trigger particular behaviors.

Sandeep Robert Datta, associate professor of neurobiology at Harvard University Medical School, made some inroads last summer when he published a study that showed for the first time how relationships between different odors are encoded in these deeper brain areas.

First, Datta and his colleagues employed machine learning to sift through thousands of chemical structures known to have odors. They analyzed each structure's attributes such as the number of atoms, molecular weight, and electrochemical properties. This allowed them to figure out how similar or how different any odor was compared to another, and from there they built three sets of 22 odors each, and with high, intermediate, or low diversity.

Deep-brain imaging with multiphoton microscopy

Last, they exposed mice to various combinations of odors from the different sets and used multiphoton microscopy, which involves shining a strong beam of near-infrared light onto a single point in the brain to create the simultaneous absorption of two photons at the spot where the intensity is highest. This long-wavelength, low-energy light allows Datta to penetrate deeply and safely into the brain tissue of mice to image patterns of neural activity in the piriform cortex and olfactory bulb. Datta's mice had been bred to have a genetically encoded fluorescent calcium indicator within their piriform cortexes so that light can activate odor signal activity there.

Datta and his colleagues found that odors with similar chemical signatures, such as lemons and limes, were reflected by similarities in neural activity and vice versa. They believe this is the brain's way of categorizing odors, perhaps for faster processing.

"All of us share a common frame of reference with smells. You and I both think lemon and lime smell similar and agree that they smell different from pizza, but until now, we didn't know how the brain organizes that kind of information," Datta says. "This is the first demonstration of how the olfactory cortex encodes information about the very thing that it's responsible for."

Scientists still have a long way to go to understand olfaction. Even though smell was perhaps the first sense to evolve in mammals, scientists have spent far more time investigating sight and hearing. That's partly because humans are first visual and then auditory creatures. One-third of the human brain is used for visual processing. It's also easier for researchers to manipulate visual and auditory stimuli. A scientist can alter the shape, size, or color of an object. They can easily lower the pitch on a sound or change the direction it's coming from. But smell is a different animal. "Odorants are nebulous," Verhagen says. "They're kind of hard to control."

Photonics is helping to solve that problem, offering researchers a new way to visualize odors as they move through the air and to see how odor information is processed in different parts of the mammalian brain. Once researchers can understand how a smell moves from the olfactory bulb to deeper in the brain, how it's processed in those areas, and how it consequently gives rise to different behaviors, they have at least a partial roadmap of the inner workings of our gray matter and its tens of billions of neurons.

From that first map may come others. Crimaldi, Verhagen, and other olfactory neuroscientists—who have an admitted bias for the sense of smell—believe that once science has figured out olfaction, they should have a jumping-off point for starting to untangle more knowledge about brain function. Defining how our sense of smell works should also help us better understand why diseases, like COVID-19, may cause temporary or permanent loss of the sense of smell.

"It can be really unsettling when you lose one of your senses," Chrobak says. "I guess you can get by without your sense of smell, but it's a useful thing to have. I use it to tell if my food's gone bad and to experience food...It was quite a relief when I realized it was back."

NANCY AVERETT reports on science, technology, and the environment from Cincinnati, Ohio.