New study reveals how the brain organizes information about odors

The premiere of the film Scent of Mystery in 1960 marked a singular event in the annals of cinema: the first and last film debut “in glorious Smell-O-Vision”. Hoping to wow moviegoers with a dynamic olfactory experience alongside familiar sight and sound shows, select theaters were outfitted with a Rube Goldberg-like device that carried different scents directly to the seats.

The public and critics quickly concluded that the experience sucked. Burdened with technical issues, Smell-O-Vision came under fire and became a gag that occupies a unique place in entertainment history. However, the failure of Smell-O-Vision failed to deter entrepreneurs from continuing to pursue the dream of offering smells to consumers, particularly in recent years, through digital scent technologies.

Such efforts have generated news headlines but little success, due in part to a limited understanding of how the brain translates odor chemistry into smell perceptions, a phenomenon that in many ways remains opaque to scientists.

A study by neurobiologists at Harvard Medical School now provides new insights into the mystery of odor. Reporting in Nature On July 1, the researchers first describe how the relationships between different odors are encoded in the olfactory cortex, the region of the brain responsible for odor processing.

By delivering odors with carefully selected molecular structures and analyzing neural activity in awake mice, the team showed that the neural representations of smell in the cortex reflect chemical similarities between odors, allowing the brain to classify odors into categories. Furthermore, these representations can be re-wired for sensory experiences.

The findings suggest a neurobiological mechanism that may explain why individuals have common but highly personalized experiences with smell.

“We all share a common frame of reference with smells. You and I think that lemon and lime smell similar and agreed that they smell different than pizza, but until now, we didn’t know how the brain organizes that kind of information, “said the study’s lead author, Sandeep Robert Datta, an associate professor of neurobiology at the Blavatnik Institute at HMS.

The results open new avenues of study to better understand how the brain transforms information on odor chemistry into the perception of smell.

“This is the first demonstration of how the olfactory cortex encodes information about what is responsible, which is the chemistry of odor, the fundamental sensory signals of smell,” said Datta.

Computational odor

The sense of smell allows animals to identify the chemical nature of the world around them. Sensory neurons in the nose detect odor molecules and transmit signals to the olfactory bulb, a structure in the forebrain where initial odor processing occurs. The olfactory bulb mainly transmits information to the piriform cortex, the main structure of the olfactory cortex, for more complete processing.

Unlike light or sound, stimuli are easily controlled by adjustment characteristics such as frequency and wavelength, it is difficult to investigate how the brain constructs neural representations of small molecules that transmit odor. Often subtle chemical changes, a few carbon atoms here or oxygen atoms there, can lead to significant differences in smell perception.

Datta, along with first study author Stan Pashkovski, a researcher in neurobiology at HMS, and his colleagues addressed this challenge by focusing on the question of how the brain identifies related but distinct odors.

“The fact that we all think that a lemon and a lime smell similar means that their chemical composition must somehow evoke similar or related neural representations in our brains,” Datta said.

To investigate, the researchers developed an approach to quantitatively compare odor chemicals analogous to how differences in wavelength, for example, can be used to quantitatively compare the colors of light.

They used machine learning to observe thousands of chemical structures that have odors, and analyzed thousands of different characteristics for each structure, such as the number of atoms, molecular weight, electrochemical properties, and more. Together, these data allowed researchers to systematically calculate how similar or different any one scent was from another.

From this library, the team designed three sets of smells: one set with high diversity; one with intermediate diversity, with odors divided into related groups; and one of low diversity, where structures vary only by incremental increases in the length of the carbon chain.

They then exposed the mice to various combinations of odors from the different sets and used multiphoton microscopy to image the patterns of neural activity in the piriform cortex and olfactory bulb.

Odor prediction

The experiments revealed that the similarities in odor chemistry were reflected in similarities in neural activity. Related odors produced correlated neuronal patterns in both the piriformis cortex and the olfactory bulb, measured by overlaps in neuronal activity. Weakly related odors, by contrast, produced weakly related activity patterns.

In the cortex, related odors led to more strongly clustered neural activity patterns compared to patterns in the olfactory bulb. This observation was maintained in all individual mice. The cortical representations of odor ratios were so well correlated that they could be used to predict the identity of a prolonged odor in one mouse based on measurements made in a different mouse.

Additional analyzes identified a diverse variety of chemical characteristics, such as molecular weight and certain electrochemical properties, that were linked to patterns of neuronal activity. The information obtained from these characteristics was robust enough to predict cortical responses to an odor in one animal based on experiments with a separate set of odors in a different animal.

The researchers also found that these neural representations were flexible. Mice repeatedly received a mixture of two odors, and over time, the corresponding neural patterns of these odors in the cortex were more strongly correlated. This occurred even when the two odors had different chemical structures.

The cortex’s ability to adapt was generated in part by networks of neurons that selectively reshape odor relationships. When the normal activity of these networks was blocked, the encoded cortex smells more like the olfactory bulb.

“We present two odors as if they were from the same source and observe that the brain can rearrange itself to reflect passive olfactory experiences,” Datta said.

Part of the reason why things like lemon and lime smell the same, he added, is likely because animals of the same species have similar genomes, and therefore similarities in smell perception. But each individual has personalized perceptions as well.

“The plasticity of the cortex may help explain why odor is, on the one hand, invariable between individuals and yet can be personalized based on our unique experiences,” said Datta.

Together, the study results demonstrate for the first time how the brain encodes the relationships between odors. Compared to relatively well-understood visual and auditory cortices, it is still unclear how the olfactory cortex converts information about odor chemistry to smell perception.

Identifying how the olfactory cortex maps similar odors now provides new insights that inform efforts to understand and potentially control the sense of smell, according to the authors.

“We still don’t fully understand how chemistry translates into perception,” Datta said. “There is no computer algorithm or machine that takes a chemical structure and tells us what that chemical smells like.”

“To build that machine and to someday be able to create a virtual, controllable olfactory world for a person, we need to understand how the brain encodes information about odors,” said Datta. “We hope that our findings are one more step on that path.”