Humans have a remarkable sense of smell. We simply do not give it the credit it deserves.
We tend to think of smell as the ability to appreciate coffee, recognize a perfume, or notice that something in the refrigerator has gone wrong. But olfaction also participates in social interactions, emotional responses, and hormonal changes. Some of those effects occur without any conscious experience of an odor.
My conversation with Dr. Noam Sobel explored this less obvious side of smell: how we sample other people’s body odors, why friends may smell similar, how chemical signals can influence aggression, and what it might take to transmit an odor over the internet. The central idea is straightforward, but its implications are substantial: our noses are collecting information that can shape our behavior without our realizing it.
Key Takeaways
- Humans sample social chemical information through ordinary behaviors, including bringing a hand toward the nose after handshaking.
- Body odor similarity predicted positive social impressions between strangers in a controlled interaction study.
- Chemical signals can influence arousal, aggression, and hormones without producing a consciously recognizable smell.
- Digital smell technology has demonstrated early successes, while household medical applications remain a distant research goal.
Table of Contents
- How the Brain Turns Molecules Into Smell
- Are We Smelling Other People Without Realizing It?
- Body Odor and the Chemistry of Friendship
- Human Chemosignals and the Smell of Fear
- A Chemical From Babies’ Heads Can Influence Aggression
- Emotional Tears, Testosterone, and a Possible Chemical Shield
- Can We Digitize Smell?
- Why Digital Smell Could Matter for Medical Diagnosis
- Smell Is More Than a Conscious Sensation
- Sponsor: LMNT
How the Brain Turns Molecules Into Smell
The most familiar form of smelling is orthonasal olfaction: airborne molecules enter through the nostrils as we sniff. They travel roughly 6 to 7 centimeters into the nose, reaching a folded sensory surface called the olfactory epithelium.
That surface contains approximately 6 to 7 million olfactory sensory neurons, expressing roughly 350 different receptor types. A meaningful portion of our genetic machinery is devoted to building this diversity of odor receptors.
When an odorant interacts with a receptor, the chemical information is converted into electrical activity. This process, called transduction, ultimately produces action potentials that travel along the olfactory nerve. Those nerve fibers pass through tiny openings in a thin part of the skull and reach the olfactory bulb, the first major processing station for smell in the brain.
A pattern of activity, not one receptor per smell
The organization of the olfactory bulb is extraordinary. Sensory neurons expressing the same receptor subtype converge onto specific structures called glomeruli. In the simplified textbook account, neurons of a given receptor type converge onto a pair of these structures.
An odor such as coffee activates a combination of receptor types. Imagine that it activates receptor types 1, 3, and 7. Their corresponding glomeruli become active, creating a pattern that reflects the input from those receptors.
In principle, that pattern is something like a map of coffee. It is not that one receptor means “coffee.” Rather, a combination of receptor responses contributes to the experience.
From the bulb, information travels to regions including the piriform and entorhinal cortices, as well as directly to the amygdala. There may also be direct projections to the hypothalamus and other structures. Smell information is distributed widely, including to brain systems involved in emotion and physiological regulation.
Importantly, the relatively orderly receptor map in the olfactory bulb does not appear to be preserved throughout the rest of the brain. Exactly how those later networks encode odor remains much less clear.
Are We Smelling Other People Without Realizing It?
Dogs make their interest in smell obvious. They approach one another and sniff. Humans generally do not walk up to strangers and deliberately smell them, although smelling a baby or a loved one is perfectly familiar behavior.
But the absence of an obvious sniff does not mean we are failing to gather chemical information.
Sobel’s laboratory began examining handshakes from this perspective. The familiar explanation that handshaking evolved to demonstrate the absence of a weapon is not supported by strong evidence. His group asked a different question: might a handshake provide an opportunity to sample someone else’s odor?
The handshake experiment
In a study led by Idan Frumin, participants sat alone in a room, supposedly waiting for an experiment to begin. During the first three minutes, researchers measured how often they naturally brought a hand toward their face.
A new experimenter then entered and delivered a standardized explanation that preparations were still underway. In half the encounters, the greeting included a handshake. In the other half, it did not.
After a handshake, participants became substantially more likely to bring a hand toward their nose.
That observation alone could have several explanations. So the researchers also measured nasal airflow. The hand movements were accompanied by sniffing, providing evidence that this was an olfactory behavior, not simply face touching.
Participants sampled not only the hand that had shaken the experimenter’s hand, but also the other hand. One possibility is that they were comparing the other person’s chemical signature with their own.
That comparison remains an interpretation. The stronger conclusion is that an ordinary social greeting can be followed by unconscious odor sampling. It does not establish that smelling is the sole purpose, or evolutionary origin, of handshaking.
Body Odor and the Chemistry of Friendship
Most of us have met someone and felt an immediate connection. You click before there has been much opportunity to discover shared interests or experiences.
Inbal Ravreby, working in Sobel’s laboratory, investigated whether similarity in body odor might contribute to these “click friendships.”
She recruited same-sex, nonromantic pairs of friends who reported that their friendship had begun with an immediate connection. To collect body odor under controlled conditions, participants used unscented washing products and slept alone in designated T-shirts for two nights.
The team then analyzed the samples with an electronic nose, a device containing sensors that respond to airborne molecules. This is not a mechanical version of the full human olfactory system. It is a much simpler instrument that can compare patterns in chemical samples.
The electronic nose found that click friends smelled more similar to one another than randomly paired people did. Independent human assessors also judged the friends’ body odors to be more similar.
Does friendship create similar odors, or do similar odors precede friendship?
There is an obvious alternative explanation: friends may eat at the same restaurants, spend time in similar environments, or share habits that influence body odor.
To address that problem, the researchers recruited strangers, measured their body odors before they interacted, and then had them play a nonverbal mirror game. Partners stood close together and tried to mirror one another’s hand movements. Everyone interacted with everyone else.
The researchers made two predictions, and it is important to distinguish the outcomes:
- Movement coordination: Similarity in body odor did not predict how well partners mirrored one another’s movements.
- Social connection: Similarity in body odor did predict ratings of potential friendship and other positive impressions after the interaction.
Because the odor measurements came before the interactions, shared friendship habits cannot explain that result. It supports the possibility that body odor contributes to an initial sense of connection, although it does not establish odor as the sole cause of friendship.
The published friendship study adds an intriguing dimension to social chemistry: some of that chemistry may be literal.
Human Chemosignals and the Smell of Fear
Discussions of chemical communication often get tangled in the word “pheromone.” Sobel generally avoids that term in his human research because it brings definitional disputes that can obscure the actual findings.
Chemosignal is a useful alternative: a chemical emitted by one individual that influences another individual’s physiology or behavior.
Fear provides one example. Researchers can collect body odor from the same person under fearful and nonfearful conditions, then examine how other people respond to those samples.
Work across several laboratories indicates that fear-associated body odor can influence recipients, including increasing autonomic arousal and sympathetic nervous system activity. In this limited but meaningful sense, fear can be chemically contagious.
This also puts a familiar claim about dogs into perspective. People have long said that dogs can smell human fear. Sobel noted that experimental work eventually provided evidence for that ability. But humans, too, can respond to chemical information associated with another person’s fearful state.
The larger lesson is not that every emotion has a simple, recognizable smell. It is that body odor can carry information about internal state, and other people’s nervous systems can respond to it.
A Chemical From Babies’ Heads Can Influence Aggression
One of the more surprising findings concerns hexadecanal, a molecule abundant among the semi-volatile compounds emitted from babies’ heads.
The work began with a different question. Eva Mishor, then a graduate student in Sobel’s laboratory, was developing experiments on chemical signaling and aggression. Meanwhile, researchers studying mice had identified hexadecanal as a signal involved in social buffering, the beneficial effects of being in a group.
Evidence that its receptor was highly conserved across mammals raised the possibility that the molecule might also have effects in humans.
Opposite effects in men and women
Mishor tested participants using the Taylor Aggression Paradigm, a standardized laboratory procedure for measuring aggressive behavior. Participants completed the task while exposed either to hexadecanal or to a control condition.
The results went in opposite directions:
- In men, hexadecanal reduced aggression on the laboratory task.
- In women, hexadecanal increased aggression on the task.
The experiment was repeated during functional MRI scanning, allowing the researchers to examine brain activity as well. The behavioral pattern appeared again.
At first, the opposite effects seemed difficult to reconcile. Mishor proposed an explanation centered on offspring survival. Across mammals, male aggression can pose a threat to offspring, whereas maternal aggression can be protective. A signal that reduced aggression in males while increasing it in females could therefore benefit an infant.
That is an evolutionary hypothesis, not proof that every exposure to a baby’s head produces a particular parenting behavior. The direct finding is narrower: exposure to this molecule altered aggression differently in male and female participants under controlled experimental conditions.
Still, the possibility is remarkable. The comforting smell of a baby’s head may be more than a pleasant sensory experience. Some of its chemical components may participate in communication.
Emotional Tears, Testosterone, and a Possible Chemical Shield
Why do we shed emotional tears?
Darwin wrestled with that question in The Expression of the Emotions in Man and Animals. He could explain functional antecedents for many emotional expressions, but tears were harder to account for.
Sobel’s group approached the question chemically. Perhaps tears do more than create a visible signal. Perhaps they also carry information that acts through the nose.
An odorless substance can still have an effect
The researchers collected emotional tears from women and exposed male participants to them. The tears had no detectable odor. Participants could not consciously identify a smell.
Nevertheless, exposure was followed by an approximately 14% reduction in free testosterone within about 20 to 30 minutes. Sobel described an independent South Korean study that subsequently replicated the testosterone effect.
Brain imaging also revealed reduced activity in the hypothalamus and fusiform gyrus during an arousing condition. The significance of every affected region was not clear, and Sobel was explicit about that uncertainty.
The original human tears study illustrates an important distinction: a substance does not need to produce a consciously recognizable odor to influence physiology through nasal exposure.
Why describe tears as a “chemical blanket”?
Later research in mice found that tears from mouse pups reduced aggression directed toward them by adult males, through an olfactory mechanism. Researchers also identified a tear component responsible for that effect in mice.
Together, these findings suggest a possible protective function for tears: a chemical signal that changes the state of a nearby individual.
But the human and mouse findings should not be collapsed into one claim. The human work discussed here examined men exposed to women’s emotional tears and measured hormonal and brain responses. The mouse research directly examined aggression toward pups.
The idea of tears as a chemical blanket is therefore a useful hypothesis about protection, not a complete explanation of why humans cry.
Can We Digitize Smell?
Imagine searching for a blueberry pancake recipe and receiving not only the instructions and an image, but also the aroma.
The difficulty is that a digital description of an odor is not itself an odor. A device at the receiving end must physically release molecules that recreate the intended perceptual experience.
Sobel’s group developed an algorithmic framework to predict how similar two molecular mixtures would smell. It also allowed them to design mixtures with particular perceptual similarities.
Different molecules, the same perceived smell
One important demonstration involved olfactory metamers: mixtures with no molecular components in common that nevertheless smell the same to people.
The researchers even generated a perceptual match for Chanel No. 5 without using its components.
This matters because transmitting smell may not require reproducing every molecule in the original source. It may be enough to create a different mixture that produces the same perceptual experience.
The group’s framework used a map of approximately 4,000 molecules with characterized perceived odors. A new mixture could be placed within that perceptual map and compared with known odors. A mixture positioned near lemon, for example, would be predicted to smell lemon-like.
The next challenge is identifying a useful set of odor primaries: components that a device could combine to generate a broad range of smells. That remains a research objective, not a finished universal smell printer.
Sending odor information over the internet
In collaboration with atmospheric chemist Jonathan Williams, the team measured odorants in Germany using gas chromatography and mass spectrometry. They transmitted the chemical information over the internet, fed it into their computational framework, and attempted to recreate the odors using a device that mixed primary components.
The proof-of-concept results were mixed:
- Rose: The recreated odor was generally perceived as bubblegum rather than rose.
- Anise: The recreation was generally described as cherry.
- Gasoline: Recognition was slightly but significantly better than chance.
- Violets: Fifteen of sixteen participants identified the recreated odor as violets.
The failures are as informative as the success. This was not a demonstration of a machine that could faithfully reproduce any smell. It was an early demonstration that measured odor information could be transmitted and used to recreate at least one recognizable target odor.
Sobel compared the goal to an early black-and-white television, not a high-resolution modern display. Recreating something broadly recognizable as wine would already be a major achievement. Reproducing the distinction between a cabernet and a merlot would be a much more demanding problem.
Why Digital Smell Could Matter for Medical Diagnosis
The most consequential application may not be scented entertainment. It may be health monitoring.
Diseases involve changes in biological processes, and those processes can alter the metabolites a body produces. Some metabolites are volatile and may contribute to measurable odor signatures.
That creates the possibility of electronic noses that detect chemical patterns associated with disease. Sobel’s long-term vision includes sensors integrated into a bathroom or toilet, repeatedly sampling chemical information and potentially identifying changes in health.
It is important to keep the timeline and uncertainty clear. His suggestion that nearly every disease might have an odor was explicitly an extrapolation. A distinctive metabolic process does not automatically provide a reliable diagnostic test.
Likewise, recognizing violets in an early odor-transmission experiment is very different from accurately distinguishing disease signatures. High-resolution chemical sensing, interpretation, and diagnostic validation would all be necessary. The household system he described remains a distant possibility, not an available clinical tool.
Smell Is More Than a Conscious Sensation
What makes this field so interesting is the gap between what our noses do and what we notice them doing.
We may sample a stranger’s chemical signature after a handshake. Body odor similarity may contribute to an immediate social connection. Fear-associated odors can affect arousal. An odorless sample of emotional tears can alter a hormonal measure.
These findings do not mean smell determines our relationships or overrides every other influence on behavior. They mean that chemical information is one part of the biological conversation occurring between people, often outside conscious awareness.
The science becomes especially compelling when surprising results are paired with careful controls, replication, and a clear distinction between what has been demonstrated and what remains a hypothesis. Our sense of smell deserves both that rigor and a great deal more attention.
Sponsor: LMNT
LMNT is a sugar-free electrolyte drink mix containing sodium, magnesium, and potassium. I use it mixed into water in the morning and during exercise, particularly when sweating heavily. The offer discussed is a free sample pack with a purchase of LMNT drink mix.



