In this Essentials episode of the Huberman Lab podcast, Dr. Noam Sobel explores how the sense of smell shapes human behavior, social interactions, and hormonal responses in ways that operate largely below conscious awareness. Sobel explains the neuroscience behind olfaction—from how airborne molecules bind to receptors in the nose to how these signals create distinct patterns in the brain. He reveals that humans engage in subtle olfactory sampling during everyday interactions, including after handshakes, and that body odor similarity can predict friendship formation.
The episode also covers chemosignals—chemical signals humans emit that influence emotions and hormones. Sobel discusses how fear odor can spread fear contagiously, how molecules from babies' heads affect parental aggression differently in men and women, and how women's tears reduce testosterone and aggression in men. Finally, Sobel describes his work on digitizing smell through algorithms that can predict odor perception and transmit scents over the internet, with potential applications in medicine and health monitoring.

Sign up for Shortform to access the whole episode summary along with additional materials like counterarguments and context.
Airborne molecules travel up the nose to reach the olfactory epithelium, a sensory surface containing approximately seven million receptors spanning about 350 different subtypes. This diversity reflects a significant portion of the human genome dedicated to olfaction. When an odorant binds to its receptor, the binding event converts into neural signals through a process called transduction.
Neural signals propagate through the olfactory nerve, passing through the thinnest part of the skull to reach the olfactory bulb. In the mammalian nervous system, this stage displays the highest degree of convergence: all receptors of the same subtype send their signals to one specific location called a glomerulus. This creates a unique pattern of glomerular activity that maps and identifies particular smells. Each odorant molecule contains various molecular aspects that activate specific receptor subsets, creating distinctive patterns of activation.
From the bulb, olfactory information projects widely across the brain, including to the primary olfactory cortices, amygdala, hypothalamus, and potentially the cerebellum. Unlike the clear convergence at the receptor-to-glomerulus level, spatial mapping is not preserved in higher brain regions, where olfactory coding is less well understood.
Despite not showing obvious olfactory behaviors like dogs, humans possess a remarkable sense of smell. Much of human olfactory sampling occurs subtly and without conscious awareness, significantly contributing to daily interactions and experiences in ways that don't resemble other mammals' overt sniffing behaviors.
Noam Sobel's research reveals that humans engage in nuanced olfactory behaviors during social interactions that play an underappreciated role in forming impressions and friendships.
Sobel's team observed that after handshakes, individuals often bring their hands to their nose or face. Through video recording and nasal airflow measurement, they confirmed that after a handshake, participants markedly increased hand-to-nose movements that coincided perfectly with sniffing. Additionally, people often smelled both the shaken hand and their other hand, which researchers believe may be a mechanism for comparing self-odor with other-odor.
Sobel's research examined body odor similarity among "click friends"—same-sex, non-romantic pairs whose friendship began with an instant, mutual bond. Using controlled conditions and collecting body odor via t-shirts, the team measured similarity with an electronic nose. Results showed that click friends' body odors were significantly more similar to each other than random pairings, and human participants could also reliably detect this similarity when smelling the t-shirts.
To test causality, researchers measured strangers' body odors before having them participate in a nonverbal interaction. Post-interaction ratings of friendship potential and likability were strongly predicted by pre-interaction olfactory similarity measured by the electronic nose. This demonstrates that smell unconsciously guides human social bonding, even before any conversation takes place.
Chemosignals—chemical signals released by humans—play a powerful, often subconscious role in emotional and hormonal responses, shaping social interactions and family dynamics.
Sobel explains that humans in a state of fear emit a distinct body odor. Experiments show that exposure to this "fear odor" raises autonomic arousal and sympathetic nervous system activity in those who smell it, making fear contagious through chemical signaling.
Studies using the molecule hexadecimal show opposite effects in men and women: it lowers aggression in men but increases aggression in women by comparable magnitudes. This is hypothesized to relate to parental behavioral roles, with increased maternal aggression enhancing protectiveness while decreased paternal aggression reduces risks to infants.
Sobel's research reveals that hexadecimal is the most abundant semi-volatile molecule found in babies' heads, suggesting a strategy where infants boost maternal protectiveness while reducing male aggression. Neuroimaging shows that hexadecimal exposure alters brain patterns affecting motivation, face processing, and aggressive decision-making.
Sobel's lab found that exposing men to women's emotional tears prompted a 14% drop in free [restricted term] within 30 minutes—a remarkably strong hormonal effect that was independently replicated by a South Korean research team. Brain imaging revealed tears reduce activity in the hypothalamus and fusiform gyrus, supporting the idea that tears act as a chemical signal that subconsciously decreases aggression. This evolutionary mechanism frames tears as a nonverbal chemical defense to reduce male aggression risk.
The digitization of smell is advancing through breakthroughs in algorithmic prediction and molecular mapping. Noam Sobel and collaborators are pioneering an algorithmic framework that can recreate, analyze, and transmit scents using molecules and computers.
Sobel's team developed an algorithmic framework that predicts with high accuracy how similar any two molecular mixtures will smell. This enables the design of "olfactory metamers"—distinct molecular mixtures with no shared components that nonetheless smell identical. For instance, they produced a scent metamer for Chanel No. 5 comprised entirely of different molecules yet perceptually indistinguishable from the original, defying classical perfumery wisdom.
The framework uses a map of approximately 4,000 molecules with established perceptual characteristics to predict how new mixtures will be perceived. This allows for identifying "odorous primaries"—minimal component mixtures that can be blended to reproduce the spectrum of human-smellable scents, advancing the infrastructure for full digitization of olfaction.
In a landmark experiment conducted with Jonathan Williams at the Max Planck Institute, the team measured molecular makeup of odors in Germany and transmitted this information digitally to Sobel's lab in Israel, where a device recreated the scents. While rose and anise failed to reproduce accurately, and gasoline had modest success, fifteen of sixteen participants accurately recognized the transmitted scent of violets—the first odor successfully sent and identified over the internet.
The long-term vision is to achieve "4K-like" resolution for transmitting and recreating smell. Such high-resolution olfactory analysis could revolutionize medicine, as every disease produces identifying metabolites with characteristic odors. Sobel envisions electronic noses embedded in household fixtures like toilets to provide real-time health monitoring through odor analysis, making preventative healthcare far more accessible. While this vision remains decades away, foundational advances today set the stage for a future where smell is as quantifiable and transmissible as sight and sound.
1-Page Summary
Airborne molecules, or odorants, travel about six or seven centimeters up the human nose, reaching a sheet-like structure called the olfactory epithelium. This sensory surface of the olfactory system contains approximately seven million receptors. These receptors are not all the same; there are about 350 different subtypes. This diversity indicates that a significant portion of the human genome is dedicated to encoding olfactory receptor subtypes. When an odorant binds to its receptor in the olfactory epithelium, it triggers a process called transduction. In this process, the binding event is converted into a neural signal in the form of action potentials.
After transduction, neural signals propagate through the olfactory nerve, passing via the thinnest and most perforated part of the skull. The signals reach their first brain target: the olfactory bulb. In the mammalian nervous system, this stage displays the highest degree of convergence. All receptors of the same subtype from across the olfactory epithelium send their neural signals to one specific location in the bulb called a glomerulus. For example, if coffee odor activates receptor types 1, 3, and 7, the corresponding glomeruli for these receptors will become active, effectively mapping the coffee’s scent as a unique pattern of glomerular activity. This mapping system allows for identification of particular smells.
Odorant molecules contain various molecular aspects, or "ototopes," each of which activates a specific subset of olfactory receptors. Therefore, each ototope creates a distinctive pattern of receptor—and thus glomerulus—activation.
From the bulb, olfactory information projects widely across the brain, including direct projections to the primary olfactory (piriform and entorhinal) cortices located on the ventral surface of the temporal lobe. In addition, signals are sent directly to the amygdala, hypothalamus, and potentially the cerebellum. Unlike t ...
Neuroscience of Olfaction: Anatomy, Receptors, Brain Processing
Noam Sobel’s research reveals that, although humans do not overtly sniff each other as dogs do, they engage in nuanced olfactory behaviors during social interactions. These subtle actions play a role in how people form impressions and friendships, highlighting an underappreciated dimension of human social life.
Sobel’s team observed that after handshakes, individuals often bring their hands to their nose or face, touching areas like the cheek, nose, forehead, or chin. To investigate, they conducted a study in which participants were invited to an experiment room, sat alone, and were secretly video recorded. Researchers first measured baseline facial touching (to the nose, forehead, chin, etc.) over three minutes. Then, an experimenter entered, sometimes offering a handshake (handshake condition) and other times not (no-handshake condition).
After a handshake, there was a marked increase in participants bringing the shaken hand to their nose. The study not only documented this behavior with video, but also measured nasal airflow without participants’ awareness. The data confirmed that the hand-to-nose movement coincided perfectly with a sniff, irrefutably proving it as an act of olfactory sampling.
Additionally, people often not only smelled the hand that participated in the handshake, but also brought their other hand to their nose. Researchers believe this dual-hand sniffing may be a mechanism for comparing self-odor with other-odor, giving insight into the nature of social olfactory sampling.
Sobel’s research extends beyond handshakes to examine body odor similarity among friends, specifically “click friends”—same-sex, non-romantic pairs whose friendship began with an instant, mutual bond. To control for external factors, participants used odorless shampoo and soap for several weeks and slept alone in a t-shirt for two nights, which was then used to collect their body odor.
The similarity of body odor among click friends was measured with an electronic nose, a machine with sensors responsive to volatile compounds. The results showed that click friends’ body odors were significantly more similar to each other than to those of random pairings.
Humans are also able to detect this similarity. In a follow-up experiment, participants smelled t-shirts worn by click friends and non-friends, reliably judging click friends’ odors as more similar.
Both electronic nose measurements and human judgments converge on the finding that friends selected by “clicking” share more sim ...
Unconscious Olfactory Behavior in Human Social Interactions
Chemosignals—chemical signals released by humans—play a powerful, often subconscious role in emotional and hormonal responses, shaping social interactions and even family dynamics.
Noam Sobel explains that when humans are in a state of fear, they emit a distinct body odor. This fear-related body odor differs recognizably from body odor collected when the same individuals are in calm states. Experiments from several labs, including those led by Denise Chen, Haviland Jones, and Sobel himself, show that exposure to this "fear odor" raises autonomic arousal and sympathetic nervous system activity in those who smell it. This means the emotional state of fear can be transmitted from one person to another purely through smell, making fear contagious through chemical signaling.
Sobel details studies using the molecule hexadecimal and the Taylor Aggression Paradigm to test its effects on aggression. In men, exposure to hexadecimal lowers aggression significantly. However, in women, hexadecimal exposure increases aggression by a comparable magnitude. This opposite effect is hypothesized to be related to parental behavioral roles: increased maternal aggression enhances protectiveness for offspring, while decreased paternal aggression can reduce risks of harm to infants.
The research reveals that hexadecimal is the most abundant semi-volatile molecule found in the heads of babies. This prevalence suggests a strategy akin to "chemical warfare" enacted by infants: boosting maternal protectiveness by heightening maternal aggression and simultaneously reducing the aggression of their fathers or other males. Such dual modulation benefits the offspring by reducing their risk of harm from males (who are more likely to direct aggression or infanticide at infants) while increasing protection from their mothers. Neuroimaging within these experiments shows that hexadecimal exposure alters brain patterns, affecting circuits related to motivation, face processing, and decisions about aggressive acts, further supporting the role of chemosignals in regulating social behavior at a neural level.
Sobel’s lab investigated whether emotional tears might also function as chemosignals. Collecting and exposing men to women's emotional tears—comple ...
Chemosignals: Effects on Behavior and Hormones
The digitization of smell, long considered a science fiction notion, is rapidly advancing thanks to breakthroughs in algorithmic prediction and molecular mapping of odors. Noam Sobel and collaborators are pioneering an algorithmic framework that can recreate, analyze, and transmit scents using the language of molecules and computers—ensuring the possibility of future scenarios where the aroma of blueberry pancakes or a rose can be experienced digitally as vividly as an image or sound.
Sobel’s team has developed an algorithmic framework that can predict with very high accuracy the perceptual similarity between any two molecular mixtures. Given two mixtures, the algorithm determines how similar they will smell to a human observer. This mathematical approach goes beyond mere analysis: it enables the design of new mixtures with targeted olfactory similarity. By applying this solution, the group generated “olfactory metamers”—distinct molecular mixtures with no shared components that nonetheless smell identical to human noses. For instance, they produced a scent metamer for Chanel No. 5 comprised of entirely different molecules, yet it was perceptually indistinguishable from the original. This finding defies classical perfumery wisdom, proving that zero-overlap mixtures can achieve identical smells, and laying the groundwork for digital olfaction.
The framework's utility relies on a robust map of approximately 4,000 molecules, each with established perceptual characteristics. When a new mixture is proposed, its “location” can be projected within this map, and its distance to reference points—such as lemon or tomato—indicates how it will be perceived. This predictive capacity allows for the recreation of virtually any odor by identifying a set of “odorous primaries”: minimal component mixtures that can be blended to reproduce the spectrum of human-smellable scents. Using these odor primaries, any desired smell can theoretically be synthesized, advancing the infrastructure necessary for full digitization of olfaction.
In a landmark proof-of-concept experiment conducted with Jonathan Williams at the Max Planck Institute, the team demonstrated online odor transmission. Williams measured the molecular makeup of specific odors in Germany, transmitted this information digitally to Sobel’s lab in Israel, where a device mixed the appropriate primaries to recreate the scent. Four test odors were attempted: the system failed to reproduce rose, with the result perceived as bubblegum, and anise, which was mostly interpreted as cherry. Gasoline was recreated with modest, above-chance success. Most notably, out of sixteen participants, fifteen accurately recognized the transmitted and reproduced scent of violets, making it the first odor successfully sent and identified over ...
Digitizing Smell: Algorithms and Molecular Analysis to Recreate Odors
Download the Shortform Chrome extension for your browser
