How Smells Influence Our Hormones, Health & Behavior | Dr. Noam Sobel

Andrew Huberman opens by framing the episode around chemosensation and olfaction, introducing his guest Dr. Noam Sobel, "a professor of neurobiology in the Department of Brain Sciences at the Weizmann Institute of Science," whose laboratory studies olfaction and chemosensation ("our ability to respond to chemicals in our environment"). He then previews a cluster of claims the conversation is meant to substantiate: that "humans can smell things around them as well as dogs can" and are "incredibly good at sensing the chemical world around them"; that at every meeting "you are taking chemicals from that person," from "the chemical cloud that surrounds them" or directly from the body surface, and "applying it to your own body," processing it "at a subconscious level" to gauge "how stressed they are, their hormone levels" and to shape emotions, decision-making, and "who you choose to relate to or not to relate to"; that "tears, yes, the tears of others, are impacting your hormone levels in powerful ways"; that a regularly scheduled "alternation of ease" lets you "breathe through one nostril or the other," reflecting "an underlying dynamic of your nervous system" tied to "how alert or sleepy you happen to be"; and that "we are always sensing our own odors," periodically smelling ourselves deliberately "in order to change your cognition and behavior." He grounds this in a Berkeley observation of people "on their hands and knees," with eyes, hands, and mouths covered and "only their nose was exposed," following a scent trail "buried some depth underneath the earth" "with a high degree of fidelity."

Huberman then asks for "the top contour Or even deep contour" of the "parts list" of smelling and "the various roles they play," deliberately widening the scope to how smell is "mixed in with other senses, like taste," to "pheromones, et cetera," and to "the chemicals that we're inhaling and making sense of without our awareness." The substantive anatomy lesson begins only after this wide framing question is posed.

The batch also contains material that carries no scientific claims: two identical ESPN/NFL sponsor reads ("Monday Night Football," "NFL Live," "NFL Prime Time") that bookend the opening discussion, and a run of personal pleasantries -- Huberman's "extremely excited" fandom "for more than a decade or two," the two having "overlapped at UC Berkeley" without meeting, "although we lived in the same apartment," and a shared apartment "in 2006, I believe." These are conversational scaffolding rather than evidence.

Sobel first corrects the premise that smell happens only "through our nose": "we also smell through our mouth," via "retronasal olfaction," where "odorants come up through the back of our throat and out of our nose, the reverse way," and "a big part of the contribution of olfaction to food and taste comes from that." Primary smelling is "orthonasal olfaction," through the nose, and "sniffing is a big thing." Airborne molecules travel "a distance in the human of about six or seven centimeters" to the olfactory epithelium, which he warns is "not a sheet" but "very convoluted," lined by "about seven million such receptors," restated as "probably about six or seven million receptors in the human, probably of about 350 different kinds," so that "a meaningful percentage of your genome is devoted just to this." He flags the count's weakness explicitly: "that number of six or seven million receptors is probably not very well-grounded. It's hard to count, but it's reasonably grounded."

That caveat segues into a cautionary anecdote about a number that did not hold up at all: the claim of "bloodhounds having a billion receptors in their nose." While renewing the lab's olfaction chapter for "the 'Gazzaniga Handbook of Cognitive Neuroscience,'" Sobel asked graduate student Araya Sharoun ("now a professor at Tel Aviv University") to trace the reference. It led to "an Australian textbook" whose author said she had heard it "at a lecture of an olfaction geneticist by the name of Doron Lancet." When Sobel phoned Lancet, then at the Weizmann Institute, and asked "Did you say that there's a billion receptors in the bloodhound nose?", the reply was "What's a bloodhound?" Sobel concludes "there was absolutely no evidence for that" and that it was "totally made up, and it propagated." Huberman draws the broader lesson about "just how much of what is in textbooks, scientific and medical, is absolutely wrong."

Things propagate, and you cite yourself.

On the signal path, odorants "dock at a receptor" in the epithelium and undergo "transduction," becoming "a neural signal" that is "action potentials, not gradient potentials of any kind," propagating "via the olfactory nerve." That nerve passes "through the thinnest part of our skull," "the cribriform plate, which is perforated. It has a lot of holes," and "synapses at the first target in the brain, which is the olfactory bulb." Sobel notes this anatomy is "an interesting point of sensitivity, because a lot of people lose their sense of smell due to trauma because of that structure."

He corrects Huberman's assumption of a front-of-head blow: "the more common cause for losing your sense of smell for trauma is the back of the head because of what's referred to as a contrecoup injury." Because "our brain is floating in liquid, in CSF, in cerebrospinal fluid," a blow to the back produces forward-and-backward movement and "a shearing motion on the cribriform plate" that severs the olfactory nerve. "If it's completely severed, completely, then yes, you're lost. Forever. ... Because even if you'll have regeneration at the basal cell level at the epithelium, they won't manage to find their way back to the bulb." "If you have partial or something left, or something shows up in a short while after the injury, then you have a good chance of recovery." He offers a hedged rule of thumb: "if you don't get it back within a year to a year and a half, you'll never get it back," while noting "I'm not a medical doctor" and that he nonetheless "get[s] a lot of emails from people who have lost their sense of smell because it's very distressing," now more widely recognized "because of COVID."

The COVID discussion surfaces several evidence-quality disagreements. Huberman reports a one-day, "not total" anosmia with "a remnant of an ability to smell or perceive the smell of a lemon," and an over-the-counter "alpha-lipoic acid" remedy he says "worked successfully for me," while immediately conceding "You don't know if it worked successfully for you or if you would've recovered anyway. You didn't do a controlled study" and "I was not willing to do the control experiment." Sobel is blunt: "the data on the alpha-lipoic acid is eh ... It's not overwhelming," though "losing your sense of smell is overwhelming." He uses the lemon to distinguish the "trigeminal nerve, the fifth cranial nerve" ("It has three branches," with endings "in your nose, in your throat, and in your eye"), so that onion tearing is "a trigeminal reflex," from "pure olfactants" such as "the coffee right here" and "Vanilla" that "have no trigeminal activation." By contrast he calls olfaction "a definite use it or lose it system" and says "there's very strong evidence for success of the training programs, more than the alpha-lipoic acid," framing it as "keeping neurons electrically active ... in order to maintain their connections, because otherwise they will die" (he attributes "fire together, wire together" to "Carla Shatz, not Donald Hebb, folks, or me").

Olfaction is a definite use it or lose it system.

A separate, explicitly unresolved disagreement concerns adult human regeneration. Sobel says "whether the human olfactory system shows the same level of regeneration as it does in other mammals is and was somewhat questionable," then describes a paper "in 'Neuron,' I think somewhere around 2014" that used "postmortem" "levels of C14 in adults who were exposed to atomic bomb experiments" so "you can actually look at these neurons and time them based on exposure to radiation." "That paper suggested that there's not as much turnover in the human olfactory bulb as there is in other mammals. Other lines of data suggest otherwise. So this is kind of a debated question as to what extent of neurodegeneration you have in the human olfactory system as opposed to other mammals."

Sobel then gives what he twice labels "the textbook view of how the system works," promising he will "happily share with you things that pose a problem for the textbook view." It begins with "the most extreme case of convergence in the mammalian nervous system": all receptors of one subtype ("about 350" in humans, "about 1,200 probably" in the mouse) "converge to one location in the bulb" called "a glomerulus or in plural, glomeruli" -- "in fact two glomeruli," two "mirror glomeruli" -- yielding "a map reflecting receptor identity" (his example: "this coffee activates receptors of type one, three, and seven," whose glomeruli "light up," giving "the map of coffee"). Each receptor responds to "a small subset of different molecular shapes" called "odotopes" ("let's say 10"), "each odotope will activate a different subset of receptors," producing "this insane combinatorics of this potentially 350-dimensional space in the human, potentially." From the bulb, information reaches "primary olfactory cortex, is piriform cortex and entorhinal cortex" on "the ventral surface of the brain, the lower portion of our temporal lobe," and also "directly to the amygdala," "probably ... directly to the hypothalamus," "may go directly to the cerebellum" -- "It goes all over the brain." Crucially, "as far as people understand, the map that may exist on the bulb doesn't exist in the rest of the brain," and "the understanding of how coding occurs in the rest of the brain is murky."

the understanding of how coding occurs in the rest of the brain is murky.

Huberman raises the common claim that "the memories that we have of odors are somehow more robust than the memories of other perceptual events in our life," but hedges at once: "I don't know if this is true or not," offering only "I can still remember the smell of my grandmother's hands" as a common saying. "At a minimum, it points to the fact that smell and memory are closely linked," which he ties to "a direct multi-station, but nonetheless, somewhat direct path from the nostrils to the hippocampus"; Sobel's clipped correction is "Two synapses away." Huberman calls this "a remarkably short pathway" compared with hearing, where "sound waves that are transduced into neural signals at the level of the inner ear go through many stations" before being made sense of "as voices or music, et cetera," and asks whether there is "any just-so story or real objective truth to the idea that olfactory memories are formed more easily or maintained longer or more robustly than other sorts of memories?"

Sobel answers "So, yes," but immediately brackets his authority: "I should say that I'm not an authority on olfactory memory. Olfactory memory is a huge field of research, and somehow our lab has never really gone much into that." He cites "the same student I happened to talk about before, Yariv Sharon" ("now a faculty at Tel Aviv"), "a paper we, I think, we published in 'Current Biology' called 'The Privileged Representation of Early Olfactory Associations,'" where "there's something about the first time you experience a smell that generates a particularly robust representation more than other sensory stimuli" and "the first exposure to a smell ... etches it into our being." He invokes "the Proust effect" ("the biggest cliche," the madeleine recalling childhood) as "something very real," stresses "There's a lot of research on it not coming from our work," and repeats "So I'm not an authority."

I'm not an authority on olfactory memory.

Neither speaker claims olfaction is uniquely privileged. Huberman: "it does sound like there's something special about olfaction, and that doesn't mean that there isn't something special about vision or audition," and "Each one has its own unique tone." Sobel: "I'm the last to argue that there's something special about olfaction," explaining that "My students make fun of me because they say ... that I try to explain everything through the olfactory system. I mean, for me, everything is olfactory." Huberman's gloss: "Through the lens of the nose."

Before "even talking about improving," Sobel insists "humans have a remarkable sense of smell" and "Humans have an utterly remarkable sense of smell," a point he says he "has to reiterate" for "non-olfaction audiences" because "we already said, 'Yeah, we know this. This is old news.'" Two tangible thresholds follow. "Mercaptans, which are added to cooking gas so that we smell it" ("the smell of gas is not the smell of gas, of propane. It's an additive") are detected at "0.2 parts per billion": "There's no machine that can really do that that effectively. No gas chromatograph, nothing." The lab's odorant "estra tetraenol" is detected "when we have it mixed at 10 to the negative 12 molar in the liquid phase"; his illustration is "two Olympic-size swimming pools" with "one ml, one drop" pipetted into one, where "you could smell the difference between the pools." His summary: "We're not a bad mammal at olfaction."

There's no machine that can really do that that effectively. No gas chromatograph, nothing.

We're not a bad mammal at olfaction.

The tracking work "started off as a lab bet" at a UC Berkeley lab picnic, when Sobel (then "a faculty at UC Berkeley in the Helen Wills Neuroscience Institute") claimed "humans could truly even track odor like a dog," met with "No way," and blindfolded "a graduate student by the name of Christina Zelano" ("now a professor at Northwestern," "really leading the field of olfaction imaging today") to track "a chocolate bar" "dragged ... across the grass," "which she did very effectively," "I assume she never practiced that in her life before." His students predicted a formal version "would be uninteresting"; Sobel bet "it would be in Nature, which is a bet I won in this case," then corrects himself: "It was Nature Neuroscience, to be fair."

Huberman frames the question as whether "people can improve their ability to track scents quite robustly, especially if we deprive them of vision and somatosensation." The formal study used "naive participants, not graduate students from our lab," "completely deprived them of any other sensory input" ("we blocked their eyes, we blocked their ears, we blocked everything. They were wearing heavy gloves"), with "a consistent odor path in the grass" produced by "burying twine under the grass, an odor-impregnated twine." Behavior was recorded with "aerial photography," a worn "sensor pack" measuring "nasal airflow in each nostril in real-time," and "RTKGPS," "a way to lay a radio frequency grid over the GPS grid so that you have millimeter resolution in space, basically" ("It's used by surveyors mostly").

Findings reported from this setup: (1) "people could just do this right off the bat"; (2) "when we trained them up, then within average of four days, the rate-limiting factor became the speed at which they could crawl" -- "as fast as you could crawl, you could scent track. Of course, you can't crawl as fast as a dog can run"; (3) a "nasal prosthesis" with "two versions" (one "combined both nostrils into one big nostril, centered," one "maintained two separated nostrils") showed "people performed better with two nostrils over one centralized nostril, although the flow remained the same," "taking advantage of the information that comes from your two separate, totally separate nostrils"; (4) the epithelium-to-bulb-to-cortex system is duplicated: "It's a completely unilateral, well, almost completely unilateral system. There are some very small exceptions to that."