Andrew Huberman interviews neuroscientist Dr Scott Kanoski on how food shapes memory, cognition, and appetite
Andrew Huberman speaks with Dr Scott Kanoski, professor of human and evolutionary biology and neurobiology at the University of Southern California.
Summary
Dr. Scott Kanoski, whose lab studies how hunger and memory interact, how the gut communicates with the brain, and how dietary choices — particularly during development — affect cognitive function across the lifespan. Kanoski presents evidence from rodent models showing that early-life consumption of high-sugar or ultra-processed diets produces hippocampus-dependent memory deficits that persist into adulthood even after the diet is corrected. He also discusses the gut microbiome's role in these deficits, identifying specific bacterial populations that can either impair or rescue memory performance. The conversation covers GLP-1 drugs in depth — their mechanisms, limitations, and open questions about whether they act centrally or peripherally — as well as the role of acetylcholine, oxytocin, and the vagus nerve in connecting gut signals to hippocampal memory function. The episode also addresses the influence of the sugar lobby on food research and the broader relationship between the food industry and public health institutions.
Key Takeaways
FULL TRANSCRIPT
Introduction and the GLP-1 System
Andrew Huberman: You work on a topic that people experience every day, which is hunger, feeding, and hopefully satiety. But we also have this obesity epidemic, which seems to be partially mitigated by the GLP drugs. There's a lot to this whole thing that we call hunger and food and appetite and weight. Today you'll educate us about all of it. But to start things off, because it's on everybody's mind — what are the GLPs? What is glucagon-like peptide-1 in its natural form in the body? And then perhaps we can talk about the drugs that increase their levels, because either everyone's taking one of these now, coming off one, thinking about taking one, hates the idea that anyone's taking them, or has someone they know who is taking one. So GLPs are here.
Dr Scott Kanoski: In its natural form, it's an incretin hormone. So it's secreted from the distal intestines around food ingestion. There's a little bit of secretion before the meal actually starts, then additional secretion during the meal, and this gets into circulation. There's also a source of GLP-1 that's produced in the brain — in the brain stem, the nucleus tractus solitarius. There's a population of neurons that make GLP-1 and then extensively send GLP-1 signals throughout the brain.
Huberman: What do those do? Are they related to hunger as well?
Kanoski: The peripheral GLP-1 and the central GLP-1 have very similar effects. They're involved in hunger suppression, but they seem to be operating independently. There doesn't seem to be a strong connection between the peripheral GLP-1 physiological system and the neurons that make GLP-1 in the brain. But they both have appetite-suppressing effects.
Huberman: But presumably they're doing other things as well, right?
Kanoski: Yeah, they're helping to regulate blood glucose levels, to stimulate insulin secretion when there are high levels of glucose present in the blood. In fact, the first application of GLP-1-based drugs was for diabetes treatment, not for obesity. But because they also suppress appetite and food intake, which leads to body weight loss, they're now probably more widely known for their weight loss capacity. But they started off as diabetes treatment.
My understanding is that the GLP-1 drugs — semaglutide, Wegovy, and on and on — increase levels of GLP-1 thousandfold more than we ever experience naturally.
Kanoski: That's correct. Super-physiological for sure.
Huberman: Is there anything known about the potential downside to that, or upside?
Kanoski: Well, most of what we know is an upside. People have been taking GLP-1-based drugs for over a decade now, and the clinical trials support the efficacy of these drugs. The side effects — nausea and emesis, emesis being vomiting — are the most common, and this can lead to discontinuation in a lot of people. They tolerate it for a certain period of time and then at some point discontinue the treatment. So if there's any primary limitation of these drugs, it's trying to mitigate those side effects and to have more long-term efficacy beyond the initial weight loss.
Microdosing GLP-1 Drugs
Huberman: I hear that some people nowadays are taking far less than they are prescribed — so "microdosing," a term borrowed obviously from the psychedelics community — in order to offset some of the side effects, still get some appetite suppression, and perhaps also reduce their cost, because some of these are quite expensive, although many are moving to generic forms. Are there any animal and/or human data that you're aware of on microdosing, or is this just a natural experiment that people are doing?
Kanoski: I'm not aware of any data on that. Placebos are very effective, so there could be some of that involved. I'm old-fashioned and I like to see large clinical trials — randomized controlled clinical trials — before I start to believe something's actually working. And we just don't have those studies to date on microdosing.
Huberman: It seems like they probably won't be done. There's not really an incentive for the drug companies anyway. People taking less and buying less seems like not a good reason to start. These trials are so expensive to conduct that it's unlikely anyone's doing that right now or will in the near future.
Kanoski: So we're kind of left with anecdotal reports in that zone.
Huberman: I also hear about some non-weight-loss-related benefits of the GLPs. In fact, I know people who are of healthy body weight who are taking quote-unquote microdoses or even full doses in order to glean what they claim are some benefits on cognition, on energy. A friend who was diagnosed with chronic fatigue syndrome has been microdosing it and she loves the effects. She's working with a physician, obviously. She doesn't like the appetite suppression because she's already very lean. But it speaks to the possibility that elevating levels of GLP-1 could be beneficial for things unrelated to weight loss.
Kanoski: Absolutely. The receptor for GLP-1 is not exclusively expressed in regions of the brain that control appetite and food intake and glucose regulation. They're pretty densely expressed throughout the brain, including regions that are in the reward circuitry — the mesostriatal pathway. There are receptors in the midbrain VTA, the striatum, the nucleus accumbens. One of my favorite brain regions is the hippocampus, which has a lot of GLP-1 receptors. So there's definitely the possibility of applications for these drugs, or different GLP-1 receptor agonists, for substance abuse disorder and potentially even memory disorders and dementia. There's a lot of preclinical data to support that. The clinical data are just ongoing and starting to come out now.
Part of the idea with the applications of GLP-1 receptor agonists to things other than appetite suppression has to do with the capacity of these drugs to get into the brain. This is somewhat of a controversial topic in our field. From 2005 to 2012, we collected some preclinical evidence in rodents suggesting that a large part of the mechanism of action for food intake suppression was through brain action — the drugs would get into the brain and activate receptors. Whether these newer-generation GLP-1 drugs like Wegovy are getting into the brain is less clear. Really, again, controversial.
Some studies — mostly preclinical, exclusively preclinical — will tag the peptides with a fluorescent molecule and try to look at where it accumulates in the brain, and you see a lot of it clump into what are called the circumventricular organs, where there's no blood-brain barrier protecting the brain from the circulation. So there's a school of thought that these drugs are acting primarily in those circumventricular organs. But those studies are tricky because fluorescent molecules fade and the brain has autofluorescent signal, so we could be missing something.
There are better ways to do this — expensive and complicated — where you radiolabel the peptides with carbon-14 and can actually determine if it's getting into the brain and where. Elizabeth Rial, Bill Banks, and colleagues at the University of Washington are starting to do some of those, and it looks like some of the early-generation GLP-1 drugs readily get into the brain, and some of the newer ones — less clear. Maybe it takes a little bit longer, but the jury is still out. Where they're getting into the brain is not something we have a clear understanding of at this point.
How GLP-1 Suppresses Hunger: The Vagus Nerve Mechanism
Huberman: I'm sort of veering us toward reward and reinforcement and memory, because even though those are topics that come up in the context of drug-seeking and pleasure-seeking, in the context of food, what we consider rewarding and what we crave are very closely related. And memory for what we think is going to satisfy us at the level of food also seems to be a strong component. But as a way of getting directly there, I'm very curious how the GLPs are actually suppressing hunger. If we assume that these drugs aren't getting into the brain — if I take a full dose or a microdose of a GLP-1 drug, what is the sensation that people report? I've never actually taken one. I hear about reducing "food noise" in air quotes. What is this at the level of biology?
Kanoski: The normal physiological GLP-1 that we secrete in the body likely has a paracrine mechanism of action.
Huberman: Could you describe paracrine?
Kanoski: Yeah. The vagus nerve innervates our GI tract, and there are receptors for GLP-1 on the terminal — right where the nerve is contacting the gastrointestinal tract. The paracrine mechanism is where the GLP-1 is released from these L cells in the intestines and immediately acts on nearby receptors expressed on the vagus nerve. So that's one mechanism through which GLP-1 — both the GLP-1 drugs as well as our physiological GLP-1 — may be producing this feeling of satiation.
The vagus nerve communicates a lot of within-meal signals that contribute to us stopping eating. At some point we get full. This process is called satiation. GLP-1 is a big player in that process — not exclusively through the vagus nerve, but probably largely through it via this paracrine mode of action. This sends a signal to the brain stem, which is the first site in the brain where the vagus nerve signals that the animals have had enough food. The stomach is starting to expand, there are nutrients entering the intestines, and there's a symphony of signals — some acting through the vagus nerve, some not — that eventually lead to this feeling of fullness and meal termination. GLP-1 is one instrument in that symphony.
Huberman: Does it turn off the desire for food as well? People talk about less food noise, and I know many people who are taking these drugs. They all talk about reduced food noise, which tells me that before there was noise. I'm trying to imagine, having never taken one of these, what that noise is. When I see blueberries I want to eat them. But when I see bread or a steak or something, I may or may not want to eat them depending on when I ate last, time of day, this kind of thing. So what is food noise?
Kanoski: I am not someone who experiences food noise either, so it's a bit of a strange idea to me. But I've certainly heard, same as you, anecdotal evidence of people who are taking GLP-1-based drugs and saying that it really canceled the food noise. Whether that's based on receptor binding in the brain in higher-order brain regions that are involved in memory and cognition — hippocampus, prefrontal cortex — we don't know. I think that's one of the big gaps in this field: understanding where the drugs are acting. Are they getting to these regions in the brain that are involved in higher-order cognitive processes that may mediate the food noise? Or is this basically just a gut-vagus satiation mechanism? There's a lot to uncover in that area still.
Huberman: What's the evidence that the vagus nerve is crucial for the action of GLPs in suppressing hunger?
Kanoski: If you do a selective lesion to the sensory vagus nerve as it comes up from the GI tract to the brain and then inject GLP-1, it's not effective in reducing food intake. The native GLP-1, not a long-acting GLP-1 drug — that's a different story, since they're in circulation longer and can get into the brain. But if you inject the short half-life native GLP-1 and the animals have a sensory lesion of the vagus nerve, they don't show the food intake reduction. This is work from Wolfgang Langhans in Zurich and his colleagues. They also knock down using short hairpin RNA interference — you can dial down the expression of GLP-1 receptors on the vagus nerve — and those animals have a phenotype. They eat larger meals.
Huberman: Gotcha. So that's all consistent with the body telling the brain when we've had enough, because of fullness of the gut and the presence of enough nutrients.
Artificial Sweeteners and Brain Development
Huberman: Artificial sweeteners and low-calorie sweeteners are a really big topic. There seem to be three camps. One believes these things are terrible — they alter blood glucose, mess up the microbiome, and maybe even cause brain cancer. The second group says this is a non-issue, that these are inert in reasonable quantities. And then there's the third group that kind of wonders, well, it's human-made and maybe this stuff isn't good. What is the evidence — the research on consumption of low-calorie artificial sweeteners — on development of appetite, development of the brain, development of the system that you work on?
Kanoski: This is a very controversial topic, and I don't think we have definitive answers to support any of those three camps. When we consume sweet taste, there's what's called a cephalic phase response, where we're releasing insulin before the nutrients actually get through the GI tract — just from the simple taste of sugar in the mouth. This is part of our metabolism. If this was a sugary drink, you take a sip, you would have some insulin being released to prepare the body for that perturbation in blood glucose levels.
There is some evidence in pre-clinical animal models that if you disrupt the relationship between sweet taste in the mouth and the caloric consequences of consuming something, you can have perturbations in appetite and in glucose regulation. Kind of like you're tricking the wisdom of the body — sweet taste means calories are coming, but then it doesn't. There's not a lot of definitive evidence in humans. There have been randomized clinical trials, and there's no definitive evidence from those trials that it's good for or bad for weight loss — there's some evidence in either direction.
In the pre-clinical animal model world, there's some evidence that if you give animals habitual exposure to artificial sweeteners, this will disrupt their capacity to metabolize sweet taste later on and can potentially lead to increased appetite and weight gain.
Huberman: So this is early-life daily exposure to artificial sweeteners.
Kanoski: Yeah. We've done one study in my lab, and I can speak more confidently about that. We had animals in the juvenile-adolescent stage of development — the period of life we've studied a lot in terms of dietary factors on adult cognitive processes. We gave the animals a little bit of artificial sweeteners to consume each day, kept within the FDA's acceptable daily intake levels. So this wasn't a sledgehammer — it might be something equivalent to six of these a day. The animals consumed it voluntarily; we didn't put it in the drinking water.
When the animals were adults, we tested neurocognitive and metabolic effects of having had that exposure during development. At that point, they were no longer getting the sweeteners — just consuming healthy rodent chow and water. We found that the animals had some memory impairments, which was an unexpected finding. We don't really know the mechanism for that. But they also had some glucose regulatory changes. If you infuse glucose and bypass the mouth, there's an insulin spike, and the animals that had this exposure showed a larger insulin spike. They also showed changes in sweet taste receptor expression on the tongue.
We found some changes. This is not like whopping sledgehammer effects, but at least in that model, with rodents consuming artificial sweeteners daily during that developmental stage, there is potentially some impact on neurocognitive and metabolic processes when they're adults.
Huberman: Interesting. Do you recall which sweeteners?
Kanoski: We had stevia, acesulfame potassium, and saccharin. Saccharin is not one that's common in human beverages and foods these days, but it's one that rodents like quite a bit. So a lot of the rodent model artificial sweetener literature is based on saccharin.
Huberman: Thanks for clarifying why. I always wondered. Nowadays stevia, aspartame, and sucralose seem far more common. I realize it's one study. Is there any work that's going to try and reconcile the conflicting findings in this area? Do you actively avoid artificial sweeteners yourself?
Kanoski: I don't think we have the evidence for recommendations at this point. The literature's all over the place. There's some evidence that if you replace habitual sweetened beverage consumption with habitual artificially sweetened beverage consumption, that could have some beneficial effects on weight loss. But again, there's no reliable trend line where you can say this is good or bad. I consume Coke Zero regularly.
Huberman: So I guess the takeaway is that people will have to decide whether they want to be in the experimental or the control group given what they know — which is that the state of evidence is mixed, somewhat unclear, and somewhat driven by preclinical models.
Memory, Food, and the Hippocampus
Huberman: The memory system in the context of food — I'm fascinated by this. What is the evidence that our memory of a food experience can play an important role in how we frame that food and our behavior around it?
Kanoski: Our memory of eating occasions can have a powerful effect on how much we eat subsequently. In one study from Suzanne Higgs and colleagues, they would have people consume a pizza lunch and then have them come in a few hours later and prime them to really think about the memory of that pizza lunch. Another group was primed to think about something else. They were given a biscuit — or cookie — taste test, and they looked at whether that memory priming had an effect on how much they ate. If they were primed to remember the lunch they had just consumed, they consumed fewer of these cookies. So in general, the memory of recent eating will contribute to eating less at the subsequent eating occasion.
There's another interesting study by Jeff Brunstrom and colleagues where they had participants eat tomato soup from a bowl that had a peristaltic pump either putting more soup into the bowl or pulling it out — but the participants were not aware that was happening. So they saw a certain volume of soup and consumed the whole bowl, but there could be a mismatch: they might see 300 ml and consume 500, or see 500 but actually consume 300 because the soup was being pulled out. This is a way to manipulate the memory of the meal — how much did you think you consumed?
If you asked them to rate their hunger right after consuming the soup, it didn't matter what they saw — it was based on how much they actually consumed. That's the wisdom of the body. But then a couple of hours later, that flipped, and their hunger ratings were more driven by what they saw, what they thought they consumed, versus what they actually consumed.
Huberman: That's really interesting. The wisdom of the body piece is great — you've eaten until you're satiated. But then the memory of how much one thought they consumed impacts what you eat at your next meal.
Kanoski: The study didn't actually have them eating after that — it was a self-reported hunger rating. But presumably it could extend to how much you would eat at a subsequent meal.
How Diet Impairs Memory: Western Diet and Developmental Windows
Huberman: Let's talk about how food impacts memory. If I want my memory system to be as robust as possible, what should I avoid?
Kanoski: When I started studying how diet impacts neurocognition and memory function, we were working with adult animals and using what we refer to as a western diet — about 35% of calories from fat, 15% coming from sugar, and a lot of that simple sugar, not complex starch carbohydrates. In animal models, that can lead to hippocampus-dependent memory impairments if the animals are on this diet for even a week, though generally these studies take a few months. The hippocampus-dependent memory processes seem to be more susceptible to the negative impacts of this diet than other kinds of memories that don't rely on the hippocampus. So the hippocampus is somewhat of a canary in the coal mine.
When I started my lab here in Los Angeles, we shifted away from adult models and looked at whether giving animals individual components of a diet during the juvenile-adolescent phase of development would have a longer-term impact on their memory systems. The first model we studied was a high-sugar model. The animals basically had ad libitum access to a sugar solution — about the same percentage of calories as a soft drink — and otherwise had healthy food and water. With this model, the animals show hippocampus-dependent memory deficits when they're older, and these seem to persist later into life even though they're not consuming the sugar anymore.
There may be a role for the microbiome — there are some changes in the bacteria in the gut from the sugar model that may have a functional connection to the memory impairments. We've also looked at models where the animals are consuming a total junk food diet. We give them chocolate peanut butter cups, Ruffles potato chips, a rodent high-fat, high-sugar diet. They show these memory impairments that last even after you switch them to a healthy diet.
That's not to say we couldn't reverse these memory impairments with some other intervention besides just putting them back on a healthy diet. But at least in rodents and during this developmental window, these kinds of foods have a long-term impact on neurocognition.
The Maillard Reaction, AGEs, and the Microbiome
Huberman: You mentioned a preprint about food processing.
Kanoski: We have a preprint — not a peer-reviewed publication yet — where we were trying to look at food processing from a mechanistic, individual-variable perspective. The way food processing is studied in humans is through the NOVA classification system, which has four different categories, where ultra-processed food is basically just poor-quality food and the other extreme is completely unprocessed food. That's not really mechanistic in terms of what types of processing matter, because there are a lot of processes that happen to food in the modernized food industry. So we wanted to take it one variable at a time, and we ended up heat-treating the food. This is an otherwise healthy food that we baked for an hour at a high temperature. The idea behind this is that it induces something called the Maillard reaction.
Huberman: Do you want to describe the subjective experience of it? Like if I took a carrot and just ate it — most people know what that tastes like. But if I took that carrot, sliced it, basted it with some butter, maybe a little olive oil, some salt, and stuck it in the oven and cooked it until it was soft — it comes out with almost a meat-like quality. A little tangy but savory. That's what you're referring to, right?
Kanoski: Yeah. The actual flavor components that result from the Maillard reaction are not well understood — we can't point to certain compounds and say this is what makes it taste like this. But in general, if you bake a casserole, it's going to taste better than before you baked it.
What happens when you heat this food up is there are chemical byproducts of this process called advanced glycation end-products — the acronym is AGEs. These chemicals are actually in our normal physiology. We have AGEs circulating, we have receptors for AGEs — the RAGE receptor, which is a pretty cool name for a receptor. This system is generally associated with diabetes and aging. When you heat food, depending upon the chemical composition of the food, the temperature, etc., you get these AGEs introduced into the food itself.
So we were looking at whether animals consuming this AGE-enriched diet during development would show an impact on cognitive processes when they're adults. And it turns out — again, just the caveat that this is unpublished work — we see memory impairments with this AGE-enriched diet. We think there may be a role for the microbiome. There was a population of bacteria that was eliminated through the heat treatment process, and when we reintroduced it to the animals, we were able to reverse these memory impairments.
Huberman: When you say memory impairments, could you describe the type of memory impairment? Are these big effects?
Kanoski: We use a variety of memory procedures in our lab. Some of the more common ones involve a maze about the size of this table, where there are holes on the edge and the animals are using visual-spatial cues surrounding the maze to remember where things are located. This is a hippocampus-dependent type of spatial learning and memory. Generally, they're trying to remember where food was consumed — they crawl into a little tunnel, eat a meal, and then the next day we ask whether they remember where that tunnel is. We also do memory tasks that do not involve food, where the animals are placed in an arena and have to remember which objects they've seen in which locations before.
In terms of effect size, it's not the equivalent of having a lesion to that brain region. But in some of these memory tests, you have above-chance performance and then you have not that. Generally, we see the control treatments will have above-chance performance, giving us some indication that the animals remember where that food was located or which object they've seen in which location. With some of these diet manipulations, the animals are performing basically at chance — they're not showing any evidence that they remember where they got the food from before.
Diet, Dementia, and Alzheimer's Disease
Huberman: I'm intrigued by this result because for the last five years or so, there's been this unofficial naming of Alzheimer's-like dementia as "type 3 diabetes," which I know drives certain physicians and neurologists absolutely bonkers because that's not an official diagnosis term. But we do now well appreciate that diet impacts the brain, and that really disrupted metabolic function — like chronically elevated insulin due to elevated blood glucose — is problematic for the brain. A lot of people are trying to offset their potential memory loss in the future by saying, at a minimum, we can control food intake. Some are trying ketogenic diets, some are trying to manage blood glucose using metformin or berberine. And now there's the discussion about the GLPs — maybe these will improve longevity of brain function or improve memory. What are your thoughts on this notion that maybe some major component of age-related dementia, maybe even Alzheimer's, is food intake related?
Kanoski: There's pretty strong evidence from prospective studies — they're correlations, but from prospective studies showing that midlife dietary patterns and metabolic status can be associated with an increased risk of dementia later in life, including visceral adiposity. So the connections appear to be there.
In animal models, we have transgenic models of Alzheimer's where animals have gene mutations that lead them to develop symptoms analogous to human Alzheimer's at a later point in their life. In the pre-clinical world, there's evidence that consuming a western diet — again, a high percentage of calories from fat and simple sugars — can exacerbate that pathology, essentially leading the animals to develop the symptoms of that Alzheimer's transgenic model earlier than they would if they were consuming a healthy diet.
I think it's more prospective than something you can deal with once symptoms arise. It's more important to have a healthy dietary pattern throughout life as opposed to switching once there's a problem.
Huberman: It seems these days that people feel really good talking about the fact that exercise is one of the few things we can do to maintain or improve brain function. The question I can't seem to find an answer to is: are the positive effects of exercise on cognition and memory the consequence of improved metabolic function, end of story? Or is there something else there?
Kanoski: I think the answer is probably both. Getting back to this diet model in rodents where we have a western diet, there is some evidence that exercise can reverse those memory deficits, and in part that's happening through changes in the brain.
Huberman: So for the animals that were fed an unhealthy diet and later put on a healthy diet but still exhibit these long-standing memory deficits — if they exercise, they can rescue those deficits?
Kanoski: That's been shown in rats. Yeah.
Huberman: When you look out on the landscape of dementia and Alzheimer's research, do you think we're going to get to the place where what we call Alzheimer's is a reflection of a bunch of bad choices plus some genetic preloading of the system? Or do you think there's really a disease that is independent of all that?
Kanoski: We haven't come very far in terms of finding effective treatments for Alzheimer's, as I'm sure you know. One thing we've been doing in my lab is trying to think outside of the box and think about how the gut might be connecting to this pathology, as opposed to just looking at what chemical processes are changing in the brain.
One example is we've done some work with the vagus nerve. If you lesion the vagus nerve as it's coming out of the gut, the gastrointestinal-innervating components of the nerve, animals show hippocampus-dependent memory impairments. So there's a connection between the vagus nerve and the hippocampus, and we're interested in understanding whether that connection is relevant to Alzheimer's and dementia pathology.
There was one study in a population in Taiwan where, in the 1980s and into the '90s, people would get vagotomies — a truncal vagus nerve lesion for stomach ulcers. That was a somewhat common treatment at the time. In other countries — not so much the US — they have impeccable medical records kept for decades, and you can look at whether an individual who had a truncal vagotomy for a stomach ulcer had a higher probability of developing dementia later on. There does seem to be, in one small study, a positive correlation — some risk factor for vagotomy in humans and developing some kind of Alzheimer's pathology later on.
We're interested in looking at that in a different cohort in the Swedish National Registry, to find out whether there's any risk in humans of having vagus nerve disruption and Alzheimer's pathology. In animals, we can study this a little more directly. We have a transgenic Alzheimer's model in a rat — they basically develop the amyloid and tau pathology in the hippocampus in an age-dependent fashion. And we have devices called vagal nerve stimulation devices that we can implant. The one we're working with relies on mechanical distension and contraction of the stomach — it's sutured to the stomach and only sends electrical stimulation to the nerve when the animals are eating, because that's when you get the natural expansion and contraction of the stomach.
We're interested in knowing whether these devices can potentially improve the outcome in these animals. That's something that's ongoing right now. We have shown that these devices can rescue the memory impairments from an unhealthy diet. So we think there's some kind of functional benefit potentially of vagal nerve stimulation in models of memory impairment, but we're still in the baby stages of figuring that out.
The Gut-Brain-Hippocampus Axis
Huberman: I can make up a simplistic model in my mind that the memory systems of the brain — which clearly involve the hippocampus — evolved first to keep us alive as opposed to make us remember sweet memories from junior high school. And nothing is more fundamental to survival than our ability to keep ourselves nourished and hydrated from day to day. So a huge component of our hippocampal function is nested in our gut's ability to communicate with our brain. Am I on solid ground there?
Kanoski: Yeah. And not just the vagus nerve either. If you look at hormones that are involved in appetite and metabolism that get into the brain — like insulin and ghrelin — there's a lot of receptors for these hormones in the hippocampus. So that seems to be an alternative pathway through which the hippocampus is getting insight into the metabolic state of the animal through these circulating humoral signals. Leptin is another one that gets into the brain from circulation. The hippocampus is really receiving both the neural connection from the vagus nerve and the humoral, longer-term energy status connection from these hormones.
It's similar in many ways to the hypothalamus in terms of the receptors that are expressed there. And of course, the hypothalamus is the classic feeding center in the brain — that's where the AGRP and POMC neurons are. That's the traditional feeding center. But we've been making a case for a number of years now that the hippocampus is really important for this too.
Feeding, eating, appetite is not a reflexive response. It's a very complex decision that has to be made when an animal decides whether to forage, where to forage, what to eat. These are not simple behaviors. So it makes a lot of sense that these complex behaviors are not exclusively being mediated by brain substrates that control more basic homeostatic regulatory processes like the brain stem and the hypothalamus. You also have these higher-order inputs that are very much tuned into this behavior as well.
Huberman: So here's the simplistic model I'm coming up with: the hypothalamus has these switches — find food, eat, stop eating — and we pretty much know what those switches are. And then there's this other system that's a major topic of focus: this body-to-brain, vagus-to-hippocampus kind of memory system. And it's so interesting that when an animal or a human doesn't eat well — meaning not enough micronutrients, not enough essential fatty acids, vitamins and minerals, protein relative to the calories they're taking in — when that ratio is skewed in the wrong direction, the memory system suffers. The gut is constantly updating: the body is getting what it needs, therefore if you're a neuron in the hippocampus, you can basically stay functional and alive. Otherwise, neurons start cashing out. This is a different picture of nutrition that I think is so important.
So if this system from the body to the brain is really a core feature of keeping memory online, how do you personally think about this in the context of your own food intake?
Kanoski: I don't make specific recommendations about what people should or shouldn't eat, but I do share our results with people who have kids who may be thinking about what to feed them. I always make the very important caveat that we're studying rodents.
That said, there is evidence in humans that supports some of what we're doing in rodents. Dick Stevenson and colleagues did some studies in healthy adults — undergraduate students approximately — and looked at their habitual dietary patterns and then how well they did on hippocampus-dependent memory tasks. These were otherwise healthy young adults, not diabetic. There were connections between habitual sugar consumption and saturated fatty acid consumption and how well they did on these memory tests. That's not to say these can't easily be reversed by consuming a healthier diet. But there does seem to be evidence in humans to support what we're doing in rats.
Not making precise dietary recommendations, but: excess sugar — not good. Excess saturated fatty acids — not good.
Huberman: I wonder why we've gotten so good at creating highly processed food that's designed to taste really good so people consume more of it, but we can't engineer foods that have robust amounts of micronutrients. It seems that the engineering can be used to push the quality of food in either direction, and I don't understand why this isn't happening.
Kanoski: I think the technology is there. I think it's just that the sea change hasn't happened in terms of where the demand is, where the production is. The food industry — over 60% of calories that many people consume come from ultra-processed foods in the US. That's a lot. And that's a direct product of the food industry producing these kinds of foods. If there's a sea change in demand or some top-down regulatory change in the food industry, I think that's certainly an optimistic possibility.
Huberman: Even if people aren't eating enough fruits and vegetables, if you can spike an otherwise unhealthy diet with like two pieces of fruit a day and one vegetable — and if you overload those with more micronutrients — it just seems so straightforward to do. The ultra-processed food industry has clearly hacked both the "get people to desire more" piece and the marketing piece. It just seems obvious. There's also an issue in the science field itself. Years ago I would occasionally come across an article about blueberry intake correlated with improved cognition in rats, but I noticed that kind of work got pushed aside. There seems to be more emphasis on finding relationships between the bad stuff and deficits.
Kanoski: I'm very much guilty of that too. Part of the problem in the pre-clinical world is that a standard rodent diet is very healthy — about 10 to 15% of calories from fat, with complex carbohydrates as the carbohydrate source. So we're starting with an incredibly healthy baseline. That's not to say there's not room for looking at certain extracts, supplementation, and whether they can have beneficial effects. But with the standard rodent diet, the animals aren't showing memory impairments. It's sort of easier to make something go wrong than to take an already healthy baseline and improve upon it.
Huberman: Lacking a strong scientific literature in this area, anything goes. And I think by now we all know what we shouldn't do.
GLP-1 Drugs, Micronutrient Deficiency, and Muscle Loss
Huberman: I've been reading that with the advent of the GLPs, we may be entering the first time in nearly 75 years where we might start to see thiamine deficiencies and micronutrient deficiencies in a fully modernized culture, because people aren't eating enough. And when they are eating, they're having a little bit of a donut and a little bit of a salad — and most salads don't actually provide that much nutrition. It could be that we end up with a large population of people who are kind of skinny-fat because of the loss of muscle with the GLPs if they're not exercising, plus micronutrient deficient. In terms of hippocampal function, it seems like that would be maybe as bad as obesity.
Kanoski: That's not a concern that I have, and part of the reason is that people don't in general take these GLP-1 drugs for a long period of time. A majority will discontinue treatment within two years of starting — because of side effects, cost, or in elderly populations, lean mass loss. When you lose weight, you lose both fat and lean mass. But people just aren't staying on these drugs long enough to — to my knowledge, there have never been micronutrient deficiency issues in any of the clinical trials, and most of what we know is from those clinical trials.
Huberman: Those are about a year or two as I recall.
Kanoski: Typically. But those trials are done in individuals where everything's controlled — diabetic and/or obese populations. Now it's a whole new world where people can get their hands on these who don't need them. So I think there are a lot of potential issues with GLP-1-based drug use in lean individuals that we just don't have any grasp on, because we haven't done those clinical trials and probably won't.
Huberman: Any weight loss program should have some nutrient and physical activity components embedded in it. Simply taking a pill is not the optimal way to lose weight.
Kanoski: And the clinical trials where we have data do include nutritional guidelines and physical activity guidelines embedded within those trials.
The Vagus Nerve, Acetylcholine, and Memory Encoding
Huberman: I'd like to know more about the nature of the signal from the vagus to the hippocampus. My understanding is that the ascending, from-the-body-up-to-brain vagal connections are mostly excitatory — mostly about creating more arousal in the brain, not more calm. My friends who do neurosurgery tell me that if a patient is drifting under a bit, you can do vagal stimulation to make them more alert again. So what is the nature of the signal from the gut to the brain? Is it a bunch of gates — if you have enough fatty acids, enough protein, enough glucose, then three gates have to be open and then boom, there's a signal?
Kanoski: It's not one common signal. There are mechanoreceptors in the stomach picking up on distension when food is in the stomach — that's a vagal signal, but it's distinct from chemoreceptors in the intestines. So you get this different timing of different signals as the meal is being processed. The neurons of the vagus nerve and the nodose ganglia express different receptors — they're very heterogeneous, not all the same kind of neuron responding to the same thing. We don't know a lot about what the signal is when it gets to the brain stem, but we do know that there are many different things contributing to that signaling pathway.
Huberman: So it's a song, not a gong.
Kanoski: Yeah, exactly. We've identified how this information gets to the hippocampus. From the brain stem, there's an input to the medial septum, and then acetylcholine-producing neurons in the medial septum are innervating the hippocampus. So it's a three-tier pathway through which the gut — at least from the vagus — gets this information to the hippocampus. Acetylcholine is the currency that we've been studying most recently in this gut-brain-hippocampus axis.
Huberman: Maybe we could drill into that a little bit. We had Michael Kilgard on the podcast — he trained with Mike Merzenich, one of the granddaddies of modern adult neuroplasticity — and the data on increasing acetylcholine and the effects on adult plasticity are still kind of jaw-dropping. If you get acetylcholine levels in the brain acutely elevated, you ungate incredible plasticity. So what is the acetylcholine doing for the hippocampus when the gut sends this signal?
Kanoski: Using fiber photometry recording, we can look at the activity of neuromodulators like acetylcholine when the animals are awake and doing something like eating a meal. What we found is that if you track the acetylcholine release into the hippocampus when the animals are actively eating — holding food, chewing it, swallowing — the acetylcholine levels are ramping up when that's happening. And then when they stop, take a break, look around a little bit, it starts to ramp down. It kind of jigsaws like this throughout the meal.
But one important finding is that at the end of the meal, the baseline levels are higher than they were at the beginning. So it's like a staircase in some ways — up, down, up, down, but then eventually it's higher. We think that elevated post-prandial level of acetylcholine in the hippocampus is to allow the animals to remember what just happened — just ate this meal, where am I, can I get back here and do this again?
Huberman: That's definitely how I felt when I went to Portland some years ago and had the best croissant I've ever had in my entire life, including in comparison to France. I have a memory map of how to find that place, assuming it's still there.
Kanoski: I can see why that would be important for survival reasons, but I could also see why if this system was hijacked by highly processed foods, why we're in so much trouble nowadays. We see the same acetylcholine profile whether the animals are eating a healthy food versus a high-fat, high-sugar diet.
Huberman: This really does have an analog to TikTok scrolling versus reading a challenging chapter in a book.
Kanoski: One thing you said made me think of something. You mentioned taste and nutrients. We don't see this acetylcholine release pattern with taste alone. If the animals consume stevia — a non-nutritive sweetener — we don't see that acetylcholine release pattern in the hippocampus. We haven't done it the other way around, where we infuse the nutrients without taste — that's something in our queue to do. But my guess is that you need the combination of taste and post-oral nutrient sensing to drive this vagal-acetylcholine signaling to the hippocampus.
Huberman: How much is being done to really parse the character of the signal from the body to the brain? The body-brain connection some years ago was considered really hippie-dippy, and now everyone talks about it. The vagus is the superhighway champion of this body-brain thing — we talk about it in the context of breathing, meditation, trauma therapy, psychedelic therapies, learning, food intake. Do we know what the nature of the signals are?
Kanoski: Unlike the visual system, interoceptive awareness is a very messy process. It's not a precise process. We can discriminate between "I'm extremely hungry" or "I'm extremely full," but in between that, this is very much unlike the visual system. The sensory inputs coming in are some neural, some endocrine — it's a symphony of chaos in some ways. I don't think we're ever going to get — or at least in the near future — to the precise level of really understanding the gestalt of that signal.
One thing I should say since I've been talking about the hippocampus a lot: it seems to be where this interoceptive awareness is being processed in the brain. If you take a human who is densely amnesic — who has bilateral damage to the medial temporal lobe including the hippocampus — these individuals have episodic memory impairments, their memory span is between 10 and 40 seconds depending upon the individual. If you have them eating food, the experimenter goes away, comes back with another meal, they'll eat the next meal, and this will happen even a third time. They don't remember eating the meal. But what's always fascinated me about this study — I think Paul Rozin did the first one — is that if you ask them to rate their hunger and satiety level on a scale of 0 to 10, they will be a seven out of 10 after eating three meals in a row. So of course they're getting this gut feedback to the brain, they're probably even feeling full, but they don't know how to interpret that signal. The hippocampus is needed to understand what that signal means.
There are analogous studies in rodents. If you do a lesion to the hippocampus, a normal animal can learn that a high level of energy restriction is different from not being food restricted — they can use that as a cue to predict something's about to happen. If you lesion the hippocampus, they can't do that. So it's interoceptive agnosia — they seem to be unaware of what this interoceptive signaling means. Even if they're detecting it at the sensory level, it doesn't allow them to appropriately guide their behavior based on that signal.
Huberman: It seems that some people have so much interoceptive awareness and other people do not. I wonder how that relates to food intake.
Kanoski: That can be impaired with a high-fat, high-sugar diet as well. So that's a process that is plastic and susceptible to what we eat.
Reward, Dopamine, and Food Cravings
Huberman: Reward, reinforcement, enticement. Most people are familiar with the word dopamine. They think of it as pleasure. Thanks to Anna Lembke's book Dopamine Nation and the book The Molecule of More, most people now understand that dopamine is about craving, wanting, motivation — at least as much and probably more than it is about pleasure. In the context of food and food intake and nutrient intake and memory, what is the role of the reinforcement pathway — the mesolimbic reward pathway?
Kanoski: This system — at its most basic, a midbrain ventral tegmental area to the nucleus accumbens, the classic mesostriatal dopamine projection — is engaged by palatable foods like liquid sucrose or a high-fat sucrose combination. But there's a little bit more to it than that. You can get the system engaged by a bland-tasting healthy food when an animal is food-restricted, or by water when an animal is water-restricted, or even a sodium-depleted animal will show a dopamine release in this pathway to sodium consumption. So it's very much dependent upon the animal's state.
That said, palatable foods tend to tickle that pathway more so than what we consider to be more healthy and bland foods. And then of course substances of abuse tap into the same system — albeit to a different degree, but it's definitely a common brain currency for things that we enjoy.
Huberman: When I hear you say that if we are food-restricted or very hungry, bland foods can activate the dopamine reinforcement pathway, I wonder whether eating highly palatable foods changes our threshold for what we consider reinforcing. I try to eat healthy and I notice that I really love and crave blueberries and other foods. If I eat healthy foods and I like the way they taste, I find that I crave those foods and don't tend to crave other foods. Is there a way that we can recalibrate this system?
Kanoski: I think so. The examples you're saying make a lot of sense. Individuals who consume a lot of sugar and then stop consuming it or consume much less — they change their preference. Something that they normally would have enjoyed is now too sweet. So there's a lot of plasticity in terms of what we want to eat, and that's based largely on what we choose to eat. We can change that, as you've described.
Huberman: It's different with food than it is with other reinforcing substances, because if somebody who really craves alcohol or stimulants is deprived of those and then gets a small amount, they usually really like that experience. But here it seems we can rewire the circuitry toward healthier choices and maybe even to the point where what used to be rewarding is now aversive. The other day I had a sip of a milkshake expecting it to be delicious, and I felt like it was unbelievably sweet — I needed a palette cleanser afterwards. But there was a time when I really enjoyed a vanilla milkshake, and I bet if I kept drinking it I would start to enjoy it again.
Kanoski: I don't think those careful studies have really been done. There are ways to measure, with PET, a proxy of dopamine binding in the human brain. In animals, we have really nice tools — I mentioned the acetylcholine sensors with fiber photometry, and we also use dopamine sensors in my lab. We can record dopamine release when the animals are awake and doing things. We haven't looked at the question you're getting at — how plastic is that system.
The most extreme example of the plasticity of that system is with a conditioned taste aversion. If you have an animal who normally consumes sucrose and gets a dopamine release in that pathway — either when exposed to a cue that predicts sucrose is forthcoming or during actual consumption — and then you train the animal that sucrose consumption is associated with malaise by injecting a compound that makes the animals uncomfortable for a few minutes, they attribute that feeling to consuming the sucrose. With that kind of training, you will eliminate that dopamine response to sucrose. That's just the most extreme example of the plasticity of this system.
Humans have tried to use the aversion approach — basically giving people things that make them feel sick or electrically shocking them to try to rewire everything from smoking to, in the old days, conversion therapy for sexual orientation, which was obviously terrible and is fortunately illegal now. I'm not aware of any really good studies showing that people will quit smoking if they're electrically shocked when they pick up a cigarette. My understanding is that people basically just figure out how to endure the shock — that's how rewarding nicotine is.
Huberman: I wonder if there's something going on with the processed food aspect too, because I don't know that people really enjoy those foods. When we hear that they're addictive — I don't know that that's the proper word — but there seems to be this thing where people get into a kind of mindless eating, not getting the thing they're looking for. That looks a lot like addiction.
Kanoski: It's tricky making the comparisons because there's a common currency in the brain for substance abuse and palatable foods, but you have to eat every day — this isn't a choice like drug use. There's a lot of controversy about whether there can be an addiction to food. I'm not going to take a side on that one. But if you look at the DSM criteria for substance abuse or addiction, you can make an argument that there is food addiction based on those criteria. That doesn't mean it's the same thing.
The Sugar Lobby and the Food Industry
Huberman: Late last year, there was a really interesting set of videos that were authenticated. The discussion was whether food stamps should be allowed to be used to purchase soda, and someone from the American Heart Association was asked on this panel whether soda is healthy. They did an incredible but really dismaying tap dance around the question to avoid answering it. Are these health associations really in bed with the sugar lobby?
Kanoski: The sugar lobby has been very powerful historically.
Huberman: Is it really the case that our medical industry is in bed with the processed food industry? Hospitals have some of the worst food in the world.
Kanoski: It certainly seems possible. As someone who's researched the negative impacts of sugar consumption, I've had a few brushes myself with the sugar lobby early in my career.
Huberman: They've been giving you basically —
Kanoski: Yeah. Early in my career, we had a result — it wasn't even published yet, it was from our early-life sugar model — there was a press release. It wasn't through me trying to get it noticed; it just got picked up. And when you get press, that's when the sugar lobby, for lack of a better term, may have a forward public statement about that particular study to try to protect themselves. Generally, it's not a "that's a great study" kind of comment. But that was a long time ago.
Huberman: Don't worry, you're safe. Actually, by saying this here, you're actually safer now. Also, the world has changed and nobody trusts big institutions anymore, especially medical-related institutions. I do think we've experienced a tide change. The United States is really on board that things need to change. This week we have an episode out with Michael Pollan, who's written extensively about the food industry and plants and health. He actually said that in order to move the health of this country forward, it needs to be a bipartisan effort. He tipped his hat to some of the things that the administration has been doing, but also pointed out some things that were not good — like related to liberating the use of pesticides. He has a very balanced view, and he makes the point that we need to get away from the politics if we ever want to move forward and be an actually healthy country.
Kanoski: We do. But there's hope.
Oxytocin and the Social Facilitation of Eating
Huberman: Before we wrap, I have a question about oxytocin. Most of us think of oxytocin as the lovey-dovey hormone and neuromodulator. Educate us on the role of oxytocin in the context of food intake and nutrition.
Kanoski: It's been known for quite some time that oxytocin can reduce appetite and food intake. Through careful studies from Harvey Grill and others, we have identified the brain stem as being the most robust target of oxytocin signaling to reduce food intake. Oxytocin is released or created in the PVH in the hypothalamus, released into the circulation, but there are also projections from oxytocin-producing neurons throughout the brain. Some of those neurons project directly to the brain stem. That system seems to be an important system for reducing food intake mainly through the level of meal size — making meal size smaller, increasing satiation.
We had a study recently where we were looking at its role in the hippocampus, and we found some results that were a little bit surprising. There's a phenomenon in humans — that rodents also show — called the social facilitation of eating. If you're eating alone, you might eat less than if you're eating with a group of people. And if those people are familiar to you, you're going to eat more than if they were unfamiliar. We can model this in rodents. We found that that phenomenon requires oxytocin signaling in the hippocampus. If we genetically turn down the signal, the animals don't show that effect.
So this is a different role for oxytocin, because under these conditions in this brain region, it's actually promoting intake in a social context. Historically, the direction has been that oxytocin in the brain reduces food intake. But some recent findings from our lab and others are saying, well, maybe — but it might depend on the social context and where the signal is getting to.
Closing Remarks
Huberman: Well, Scott, thank you so much for coming and sharing your mechanistic data. I love that you're highly mechanistic. It's clear that you're studying a really important system for the health of all humans — especially right now with the highly processed food issue, but also the GLPs and this explosion of discoveries on the microbiome and the positive explosion of people who are interested in doing what they can for their brain and bodily health. You're clearly in the right area. Evidenced by the fact that when I asked what you would study if you had a vault of money, you said what you're doing right now. So you're clearly in the right laboratory, which you run. Thank you for doing the work that you do. It's not lost on me — and it shouldn't be lost on the audience — that you do that work with taxpayer dollars. So you're working on our behalf for our health. Thanks for coming here today.
Kanoski: Thank you very much. Appreciate it. It's been fun.