Prof. Noam Sobel. Credit: Tomer Appelbaum

How do women’s tears affect men? What happens when we sniff a baby? And can odor be weaponized? Prof. Noam Sobel explains how smells influence the brain.


Talking to: Prof. Noam Sobel of the neurobiology department and the Azrieli Center for Human Brain Imaging, Weizmann Institute of Science, in Rehovot, who studies brain mechanisms related to the sense of smell. Where: His laboratory. When: Sunday, 10 A.M.

How does the sense of smell affect human behavior?

Our primary field of research dealing with that question concerns chemical communication between animals, and in our case, human beings. All animals communicate by means of chemical signals, and we, too, land mammals with an amazing nose, are no different.

As a land mammal, I always thought our nose was quite inferior to that of other land mammals.

You’re wrong. The human nose is an incredible instrument with astonishing abilities. For example, in terms of our olfactory detection threshold. There’s an odor we’ve been working with quite a bit, and its detection threshold is 10 to the power of minus 12 moles. And yet you can still smell the difference between a regular Olympic pool and an Olympic pool into which we’ve introduced a single drop of that liquid.

What type of odor are you talking about?


Thank you.

You asked, I answered. We make the most important decisions of our lives by means of that instrument, the nose.

You’ll have to support that claim very convincingly.

Glad to. Which would you prefer? A slice of gorgeous cake layered with strawberries, whipped cream and chocolate that emits a smell of sewage, or a bland-looking cereal with a vanilla scent? Which would you prefer: a partner who looks like a Greek god but stinks like a sewer, or an ordinary-looking fellow who smells like sin itself? In both types of the truly important decisions of life, eating and mating, we follow the nose and not the eyes.



Because I’d have expected that at this stage of human history, smell would be less critical. Its survival function, for example, is no longer relevant.

Of course, if you force me to forgo one of the senses – sight, hearing or smell – I will choose smell first, because in our Western world, that choice will be least harmful to my day-to-day existence. But it will still be tremendously harmful. Just ask any anosmic [a person lacking a sense of smell]. We do a lot of work with anosmics, and things are very difficult for them. I agree with you that in the past 200 years, even 1,000 years, the importance of the olfactory sense for coping with the world around us has been greatly diminished. But 1,000 years isn’t a long period in terms of the evolution of a sense.

But what exactly is chemical communication between people? Like dogs that sniff each other and behave accordingly?

Just so. Most studies conducted in this field deal with perspiration and with the odors that it contains or doesn’t contain. The smell of fear, for example. Exposure to perspiration collected from people who were in a state of fear affects one cognitively and physiologically.

In both types of the truly important decisions in life, eating and mating, we follow the nose and not the eyes.

In what way?

For example, it makes people who smell the sweat of those people process information more quickly. Much research is being done on perspiration, but I thought it would be useful to examine other odors that we secrete. For example, in tears. That’s extremely interesting material.

It’s also the only material or smell that people spread and don’t try to hide.

That’s almost absolutely correct. Tears are also a product of emotional moments, when it’s critical to communicate. There is no functional explanation for the phenomenon of tears. It’s not clear what purpose they serve. So we asked: Maybe tears contain a chemical signal? We checked: They do. What we’ve discovered so far is that when men are exposed to women’s tears, their testosterone level falls by 30 percent [which makes them less aggressive]. That’s dramatic. We are now concluding a study of women’s reaction to women’s tears. The findings are preliminary, but it looks like there is a drop in testosterone in women, too.

So the idea is that tears emit a signal of vulnerability, they try to reduce the impulse to attack us.

Yes. It’s a chemical signal [that says] stop aggression. It needn’t be specifically related to sexual behavior. Infants, too, I think, try to stop aggression against them by means of that chemical signal, because, obviously they can’t talk. In any event, the significant point here is that a chemical signal is transmitted in tears. That reinforces our hypothesis that we, human beings, are in a state of constant chemical interaction, except that, unlike dogs and cats, we do not engage in mutual, up-close sniffing. It’s accepted to sniff one’s male or female partner, and obviously we sniff babies, so in effect the only context in which it’s not acceptable is between strangers – though we’ve found that it happens even between strangers, and a lot, even if not consciously.

In other words, social conventions prompt us to suppress our sense of smell and the concomitant behavior – it’s considered vulgar and inappropriate – but it continues to exist.

Unequivocally. We showed in a study that handshaking, which is a totally everyday form of behavior, is actually a means for sniffing the person whose hand we shook. Immediately after the handshake, the subjects lifted their hand and in an apparently unconscious movement sniffed it. In fact, we are constantly lifting the hand that touches things to the facial region and sniffing it. It’s a totally automatic movement, done without noticing. We are simply ‘sampling’ one another all the time. I’ve noticed this behavior constantly since we started examining the phenomenon. People sniff themselves all the time, and sniff those around them.

Now, body odor plays a part, which we are unable to quantify, in choosing a mate. It’s apparently connected with a genome segment that’s related to the immune system. In fact, we tend to choose a mate on the basis of optimal immunogenetic compatibility. The funny thing is that this aspect of the olfactory sense is very much affected by contraceptive pills. So anyone who chose a mate while she was on the pill …

… was behaving like a blinkered horse.

When men are exposed to women's tears, their testosterone level falls by 30 percent, making them less aggressive.

Precisely. Again, it’s because we are not conscious of this use of the sense of smell. It’s only here, in the lab, that we understand how problematic it is when the element in the olfactory sense that’s related to social communication breaks down. We have a tool that we use a great deal in research studies called the “sniff response.” The idea is that we have a mechanism in the brain that is constantly checking what enters the nose and determining which sniff to make accordingly. If I were to give you ammonia to smell, you would make do with a very short sniff; if I were to give you cinnamon, you would take a long sniff.

In other words, if it’s a bad smell we don’t bother to sniff.

We receive a signal that it’s hazardous to us: “Stop.” It’s a response that makes use of sensory information that comes from the odor and activates a motoric cycle – the sniffing. It’s known as a sensory motor feedback loop. Like the way our eyes follow something that’s moving.

The idea being to prevent something like mustard gas from entering the body?

Part of the odor’s representation in the brain is related to the duration of the sniff. The same molecule produces a different odor during a strong as opposed to weak sniff. That serves us as a tool in the lab, because I can measure the duration of your sniff response. If you take a long sniff, it’s a sign that either the odor is too weak – or that you find it pleasant. If you stop immediately, it’s a sign that the odor is too strong, or that you find it unpleasant. Those are the extremities. We are currently doing fascinating work with this at Loewenstein Hospital [rehabilitation center, in Ra’anana]. We’re examining this response in comatose people, in a vegetative state. Of course, they can’t say whether the odor is pleasant or not, but from measuring the sniff response we saw that there is a difference between a vegetative individual who still displays this response and one who does not – which might say something about the prospects of recovery. By the same token, I can measure responses to odors among sleeping people or in infants.

Or in children, as in your best-known study to date, which looks at the connection between the sense of smell and autism.

Yes, but to begin with, I want to dissociate myself from headlines like, “Scientists use sniffing to identify autism.” Our hypothesis was that impairment of chemical communication between people is part of the autistic disorder. We took children with autism at the stage of their final tests for diagnosis, children about 5 years old. We built a small device that emits smells, with a new one being coming out every 30 seconds, and we measured the children’s sniff responses.

The results showed a dramatic difference between children with autism and the other children. Those with autism responded more slowly to the odors.

If we had tried to differentiate between the groups of children based solely on this index, we would have achieved more than 80 percent accuracy. In other words, it’s a strong, significant index. Naturally, the clinical value of this finding depends on its predictive ability at a younger age as well. We plan to conduct a lengthy study among siblings of children with autism beginning at the age of four months. Within two to three years we will be able to know whether this index possesses predictive value.

Let’s talk about a truly anecdotal application of odor – as a weapon.

I can tell you that the defense establishment approached us a few times with requests to cooperate, and we refused.

A female mouse that is at an early stage of pregnancy will miscarry if exposed to the smell of a male that is not the father of the offspring.


Because we don’t develop weapons. By the way, they also came to us in connection with the Skunk [a malodorant crowd-control weapon]. They tried to persuade us that, being a nonlethal weapon, it’s actually a good thing. I told them over the phone that, historically and factually, nonlethal weapons simply increased the number of those killed in every locale. They disagreed. I invited them here and showed them data that support what I said, and they took it hard and left.

Were you approached on other issues by military officials?

There were others, though I’m not sure what unit of the state they belong to. For example, they wanted us to create an odor that would help conceal smells from sniffer dogs. I explained to them that even if we were to do that, dogs could be trained to find the concealing odor, and that would be even worse. They asked, “But how will people know about it?” I replied, “When I publish the article.” “But why should you publish an article?” they asked. They didn’t understand that this is my language. I play the science game. I have no problem taking money from the defense establishment for research that will be published according to the rules of science. I would not do private research for them, but I will be happy to receive money for public research. For example, we are funded here, in part, by the U.S. Air Force.

For what?

A project called “The Smell of Trust.” We are trying to see whether it’s possible to change people’s feeling of trust by means of odor. The declared purpose, at least, is to increase people’s trust in robots by giving the robots a body odor, because it turns out that people have a problem trusting them.

But how do you create that kind of odor?

We’re not yet at that stage, we’re still checking to see whether we can create it. There’s a decision-making game called the Trust Game, at the end of which one can arrive at an estimation of how far a person can be trusted. We had many people play the game, and from them we chose those who emerged as inspiring the most trust, and those who emerged as least trustworthy. We then collected body odor from them and asked other people to smell it and on that basis say who was a saint and who was not. People looked at me like I was crazy, but yes, apparently there is something. There is a body-odor fingerprint that’s related to trustworthiness. From there we are already trying to identify the molecules themselves. We are also working on another interesting project, having to do with miscarriages.

Tell me about it.

One of the most dramatic discoveries involving odor is the so-called “Bruce effect,” named after Hilda Margaret Bruce, who discovered it in 1959. She proved that if you take a female mouse at a critical, early stage of pregnancy and expose her to the smell of a male that is not the father of the offspring, she will miscarry. If you damage a certain mechanism in the female’s sense of smell, however, the effect disappears. Together with the recurrent-miscarriage clinic at Sheba Medical Center, we are trying to see if the same effect exists in human beings.

How can you test that?

It’s indeed complicated, because obviously you can’t induce miscarriages in people. Still, it is fascinating ground, because an amazingly high percentage of human pregnancies end in miscarriage. If you look at very early stages, it could be as high as 90 percent. But let’s take the number everyone agrees on: 30 percent. The great majority of the cases that are examined medically have no explanation. No one knows why it happens. We’ve decided to examine the sense of smell of women who suffer from repeated miscarriages. Again, I speak cautiously, because we haven’t yet published this: We have very interesting preliminary evidence about a specific profile related to the olfactory sense of women who experience recurrent miscarriages. These women were significantly better at identifying a certain body odor; moreover, we found a connection between a high ability to identify that odor and a higher number of miscarriages. We are discovering more and more information indicating that a similar effect exists in human beings.

A very disturbing thought.

Yes. Human beings have a sense of smell that is far better than what they tend to think, and we are constantly using it to collect information. That information, along with its collection, remains under the radar of our consciousness. But it constantly influences our behavior, and if it’s damaged, we are in trouble. But it’s less awful to be a “social anosmic,” one who doesn’t sense certain odors, than it is to get an incorrect signal. Like many things related to the brain, an impaired signal is worse than no signal at all. Because if, for example, you are always smelling extreme fear in the people around you, that’s a terrible way to live. And if you take it to the extreme, if the social olfactory sense is indeed a factor in autism – and I don’t have evidence of that yet – and if the implication is that one is constantly receiving wrong social signals, then I think that the right thing to do would be to burn the system and kill the sense of smell. I stand behind that assertion completely.