Andrew Huberman interviews nutritional medicine professor Dr Marie Pierre St Onge on the bidirectional relationship between sleep and diet
Andrew Huberman, neuroscientist and host of the Huberman Lab podcast, interviews Dr Marie Pierre St Onge, professor of nutritional medicine at the Institute of Human Nutrition at Columbia University School of Medicine.
Summary
Dr. St-Onge runs one of the few laboratories in the world studying the bidirectional relationship between sleep and food intake. Her research demonstrates that even modest, sustained sleep restriction — as little as an hour and a half less per night over six weeks — increases insulin resistance, reduces insulin sensitivity, and raises blood pressure in free-living conditions. A key finding from her lab is that sleep deprivation affects appetite hormones differently by sex: in men, short sleep raises ghrelin (the hunger-driving hormone), while in women it reduces GLP-1 (the satiety hormone), leading both groups to consume roughly 300 extra calories per day. Moving in the other direction, her controlled inpatient studies show that higher fiber intake is associated with more deep slow-wave sleep, higher saturated fat intake with less deep sleep, and more refined carbohydrates and simple sugars with more nighttime arousals and disrupted sleep architecture. The conversation also covers medium-chain triglycerides, meal timing and fat oxidation, functional foods including ginger, the portfolio diet and its statin-comparable cholesterol-lowering effects, the challenges of publishing null results in nutrition science, and the structural barriers to industry-funded negative results reaching publication.
Key Takeaways
FULL TRANSCRIPT
The bidirectional relationship between sleep and food intake
Andrew Huberman: Sleep impacts how and what we eat, and how and what we eat impacts sleep. That's a different perspective than I think most people take. I think most people are familiar, however, with not getting the best night's sleep, maybe feeling like their impulsivity to eat quote-unquote bad foods is a little higher, and then also hopefully familiar with having a great night's sleep and feeling like they're just in control in a different way. Maybe you could share for us what's really going on beneath that experience — when subtle or not-so-subtle chronic sleep loss, not an all-nighter necessarily, but 45 minutes less here, 90 minutes less there — how that plays out in terms of our nutrition. And then we'll go from the nutrition side to sleep.
Dr Marie Pierre St Onge: Sure. There are a couple of questions in there about the extent of sleep loss and how that influences food intake — what we see in the general population versus what we do in a lab to address causality. Let me start with the population-based studies. When I started being interested in sleep, it was coming from an obesity angle. My PhD is in nutrition. I trained as a postdoc in body composition and obesity research, and we were getting a lot of information from population-based studies that people who sleep too little have a higher body mass index than people who get an adequate amount of sleep. Then it became clear there is a higher prevalence of people with obesity in the short sleep group. Then studies evaluating changes over time showed that people who don't sleep enough tend to gain more weight. There was a famous nurses' health study that I really like to cite when I give talks, published in 2006, where they tracked nurses over 14 years and those nurses who reported sleeping five or six hours had a much higher rate of weight gain over that 14 to 15-year period than the nurses who reported sleeping seven or eight hours per night.
Those are observations from large-scale population studies and cohorts, but what those studies tell us is that things are happening at a point in time or may influence something happening over time — but not necessarily that one causes the other. So I started my work in this field trying to uncover whether sleeping too little actually causes weight gain. Because I was coming from a lab where I trained in the measurement of energy balance — how much energy you eat versus how much energy you burn — I thought, well, if sleep leads to obesity and weight gain, it has to impact energy balance regulation. So it's either that we eat more than we should, or that we exercise less and burn less, or maybe it's a combination of the two. Let's try this out and see.
My first NIH grant, the big R01s, was to look at exactly this. We had people who had adequate sleep and we brought them into the lab. We used a crossover design, so half of the participants started out sleeping adequately — we gave them a nine-hour time-in-bed opportunity — or we asked them to sleep too little, with a four-hour time-in-bed opportunity, very short, but we did this for five nights. We took all sorts of measurements in a controlled feeding condition. For the first three days, we had participants eat the exact same thing regardless of how much time in bed they got at night. We measured appetite-regulating hormones. We did neuroimaging to really isolate the impact of sleep duration on appetite-regulating hormones and neuronal responses to foods. Then on the last day we let them self-select their food intake and measured that in the lab.
From that study we showed that in men specifically, we saw an increase in ghrelin in response to short sleep — the hormone that triggers food intake. In women, we saw a reduction in GLP-1, glucagon-like peptide-1, the satiety hormone, as a result of short sleep. And when we measured their food intake in the lab, they ate 300 calories more in the short sleep condition than when they got their regular adequate sleep of at least seven and a half hours per night. We also looked at neuronal responses to food stimuli and found upregulation in reward centers of the brain in the context of sleep restriction compared to adequate sleep. So altogether, we were really building a case that when you don't sleep enough at night, you have both physiological signals to eat more — for men, or not stop eating in women — that lead to greater food intake, which could also be impacted by pleasurable centers that are activated to a greater extent as a result of insufficient sleep.
Huberman: Amazing. This sex-specific split in the data — if I have it correctly, when men are sleep-deprived, getting four hours per night, the signals that drive appetitive desire to eat are higher. In women, it's more that the brake on eating, on satiety, is reduced.
St-Onge: Exactly.
Huberman: As far as I know, the GLP pathways are not divergent by sex, but of course I'm not deeply versed in that literature. Is there any evidence that GLPs are functioning differently in men and women — circadian-wise or anything like that — or was this just an incidental outcome?
St-Onge: This was an incidental outcome. We really didn't know what to expect. We didn't know at all that we'd see sex differences. There had been prior studies that had shown that ghrelin was increased as a result of sleep restriction, and that leptin was reduced. When we got our data and analyzed it with all participants together, there was no effect.
Huberman: And that was surprising.
St-Onge: People would say, "Don't you know sleep restriction increases ghrelin?" Well, I guess I don't know, because in our study it doesn't. But then we saw these sex-specific differences, and then it made sense — in the full sample, when we had an equal number of men and women, we saw no effect on ghrelin because there was no effect in women, but there was an effect in men, which was reproducing what others had found. Because all the prior studies had been done in men only.
Sleep deprivation, cortisol, and metabolic outcomes
Huberman: Whenever I'm sleep-deprived — four or five hours, which I consider sleep-deprived — I feel like my whole body is in a low level of pain. It's like a central ache. I wonder to what extent people eat to overcome, to quell, the pain of sleep deprivation. Maybe people react differently to sleep deprivation. But what do you think is happening in that short amount of sleep that's missing? What is getting reset? Is it neural? Is it endocrine?
St-Onge: In our study, it was actually a 50% reduction in sleep. When they had a nine-hour sleep opportunity, they slept around seven and a half hours. When they had the restricted condition — these were all people we had screened to sleep at least seven hours, measured by actigraphy — they got on average about three hours and fifty minutes.
Huberman: So it's like staying up late working on a deadline and then trying to catch an early flight.
St-Onge: It's pretty brutal. And that was maintained for five nights.
Huberman: Were they coming unglued mentally too? I think I would feel terrible after that kind of stretch.
St-Onge: By the end they were done. There was no way anyone would want to keep coming back for that. But they were in the lab, under supervision the whole time. We didn't let them go out on their own. No naps either.
What happens is I think there's some subconscious need to eat more when you're sleep-deprived. There's also a thermic effect of food — it gives you a jolt of energy to eat something. People know that eating wakes you up in a way. There's neuronal signaling that enhances pleasurable and reward centers of the brain. And when fatigue sets in, do you really want to have this conversation with yourself about what to choose at the buffet table? Others have also shown that sleepiness tends to correlate with triggers for more pleasurable food consumption with sleep restriction, and it's been reproduced. A meta-analysis showed 250 to 400 calories of overeating, which might not sound like much, but when you layer that day after day — roughly 3,500 excess calories per pound of body weight — and you think about people accumulating that over time in a night-shift condition, as new parents, tending to a sick relative, or during final exams, it's a real thing.
Neymar Coven in 2022 published a paper where they had sleep restriction of about five hours per night versus seven and a half hours per night for two weeks, and participants gained half a kilogram in a two-week period. You do nothing, you just sleep less, and you gain almost a pound in two weeks.
Huberman: It strikes me that for a long time in stress research, the idea was that when people are stressed, they reach for comfort foods — carbohydrate, typically starch-fat or sugar combinations. The just-so story was always that cortisol's main role is to deploy glucose, so people are doing this as a way to bring excess energy. What is the relationship between these forms of sleep deprivation and stress? Is what you're studying essentially stress?
St-Onge: If you're thinking about physiological stress measured by cortisol levels, in that study cortisol wasn't changed in the short sleep condition.
Huberman: Tell me more about that. I'm fascinated by circadian rhythms and cortisol. What does that mean?
St-Onge: There was no difference between the two conditions — adequate sleep or short sleep — on cortisol levels, for five days of sleep restriction at basically four hours a night. Cortisol was still peaking in the morning, still dropping in the evening.
Huberman: Wow. That's very surprising to me.
St-Onge: I don't know. Maybe it's the context of being in a lab where everything is safe and taken care of. There is nothing outside to aggravate. So maybe when you're in the context of sleep restriction but also dealing with your daily life — needing to take care of your kids, needing to get to work, needing to do all the activities of daily living — maybe then that becomes the added stressor.
Huberman: So the message is: if you suffer less than adequate sleep, get someone to take care of everything else. You'd better be in a spa.
St-Onge: Exactly. And in that study also, we didn't see any effect on glucose or insulin. Nothing. The curves were superimposable — while they were eating the exact same food at the exact same time, in the exact same quantity. The only thing we changed was the amount of sleep opportunity they got at night. To me, this means it's a combination of different things that causes the metabolic abnormalities we notice in free-living populations. People aren't isolated. They're not in a box where they're not sleeping enough — they're also choosing to eat higher fat, higher sugar, higher salt, a poor diet that then triggers a worsening, compounded by the lack of sleep, of those cardiometabolic outcomes.
We did a follow-up study to this severe sleep restriction study for exactly this reason — because we did not find any adverse impact on glucose, insulin, or lipid profile. We thought, so what is it then? Why is it that in population-based studies we find that people who sleep too little have higher risk of cardiovascular disease, higher risk of hypertension, higher blood pressure, higher risk of type 2 diabetes? Because we had seen that food choices were different — they ate a diet higher in calories, higher in fat and saturated fat — we thought maybe if you're in a free-living situation, that's when you start to see those cardiometabolic outcomes, compounded by more sedentary behavior and alterations in food choices.
The follow-up study recruited good sleepers — people who sleep at least seven hours per night, verified by actigraphy, who reported good sleep quality on questionnaires. We then said: you're either going to continue your excellent sleep, or you're going to go to bed an hour and a half later, getting an hour and a half reduction in sleep. Because when we screened people to sleep at least seven hours per night, they slept about seven and a half. Reducing by an hour and a half gets to six hours, which is short sleep — insufficient — and on average what people who don't get enough sleep report. They can sustain that for prolonged periods, because that's what people report in population-based studies.
When we did that, we saw that insulin resistance was increased after six weeks of sleep restriction compared to adequate sleep. Insulin sensitivity was reduced — it was worse actually in post-menopausal women compared to pre-menopausal women. Blood pressure was increased. Those cardiometabolic outcomes were adversely impacted by free-living mild sustained sleep restriction for six weeks.
Huberman: Because my mind always goes to: we wake up in the morning because of an increase in cortisol, which is circadian and overlaps with the end of the night's sleep. If that's independent of sleep, and cortisol drives glucose release — at least in the first study you described, glucose levels weren't altered, and it was isocaloric, so people weren't eating more. The hormones driving the desire to eat more are elevated, but you didn't let them eat more. What's the action end of things that causes them to gain weight if they're basically on an isocaloric diet?
St-Onge: I think they're more sedentary during the day. Less spontaneous activity. We also did a study to look at energy expenditure — that's really difficult to measure, in my opinion, because there are multiple components. But we did a study, a small study enrolling only women, using a metabolic chamber at Columbia — a small room in which we keep people and measure minute-by-minute oxygen consumption and carbon dioxide production. We were able to show that energy expenditure is actually increased in the context of sleep restriction in the metabolic chamber, because it's more costly energetically to remain awake than to fall asleep. Energy expenditure when participants were awake was identical in both conditions regardless of how much sleep they got the night before.
Huberman: So it's fidgeting, movement — the non-exercise-induced thermogenesis. It's a big number. People who fidget a lot, bounce their knee — sometimes these people are burning 1,500 calories more per day. People who are very lean tend to have a lot of spontaneous movement. They stand up quickly, walk quickly. These things add up over time in ways that most people underestimate.
St-Onge: For us it was about a 5% increase in energy expenditure, ending up being about 90 calories — nowhere close to the 300 calories more of intake they got in the prior study. So it's still an imbalance towards a positive energy balance when we do the math, but there is an increase in energy expenditure in the confines of a metabolic chamber — which for most people is equivalent to the size of their bathroom. You have a bed, a table, a sink, a toilet. That's it.
Huberman: A little while ago I saw a study saying that if you are one night sleep-deprived — one or two hours less than normally needed to feel rested — it's actually advantageous to exercise because it offsets some of the increase in inflammation. But if you're going multiple nights that way, exercising regularly when sleep-deprived sets up a susceptibility to illness and injury. How much of what you observe under conditions of sleep deprivation do you think is downstream or upstream of inflammation? Is this just a body-wide response where a bunch of systems are dysregulated? Or can we pinpoint what's happening?
I think if women knew their GLPs were down when they're down on sleep — so they should expect to feel less satiety — and if men knew their ghrelin levels were elevated when they're down on sleep, so they're going to feel hungrier — we have pretty big prefrontal cortexes, most people anyway, and we can intervene simply on the basis of knowledge. I think about this sometimes when I'm thinking about my diet. Do I really want to eat this, or is it because I really didn't sleep last night?
St-Onge: If you step back and think that maybe part of it is because you didn't sleep well the night before, then you can make appropriate choices. Say, "Okay, I probably don't need the extra calories right now." Or maybe you say, "You know what, I had a really bad night last night and those extra calories, I don't really care, because they're going to make me feel good and I need a pick-me-up." But those are all choices to make, because mood comes into play as well.
How diet affects sleep quality
Huberman: That brings us to the other direction of the equation — how does what we eat impact our sleep? Most people have heard about trying not to eat too close to bedtime. This is an active debate in many households. Some people seem to be fine eating close to bedtime. Other people find it really disrupts their sleep. I'm interested in both the timing of food intake relative to sleep and the content of the food and how it impacts sleep.
St-Onge: When we started this conversation, I was telling you about population-based studies — cross-sectional data where two things happen at the same time and you don't really know causality. Early on in this field we started thinking about sleep as the promoter of food intake, causing changes in diet and exercise, but didn't really think that maybe it's the other way around, or that the other way around is just as plausible. So I started thinking about that and said, what if we took the other approach? What if we looked at diet and examined how diet influenced future sleep?
My first paper in this field used data from the Multi-Ethnic Study of Atherosclerosis. It's actually kind of hard to find good cohorts that have good nutrition data, good sleep data, and data over years. MESA is one of those great cohorts in the US that has all of the above. I paired up with a colleague of mine, Susan Redline in Boston — she's a principal investigator on their sleep ancillary study — and we asked the question of diet quality and its impact on sleep duration and insomnia symptoms. We found that having a diet that more closely aligns with the Mediterranean diet was associated with better probability of having adequate sleep and reduced insomnia symptoms in this cohort.
That launched a whole field of study to keep looking at this. We've looked at this in different studies and different cohorts. Earlier this year we published data from the Women's Health Initiative — another large cohort with good diet data and sleep information. We took a really nice approach in this longitudinal analysis. Usually when we do longitudinal studies, we exclude people who have the condition at baseline. But insomnia is one of those conditions that's not necessarily static — it resolves. You can have insomnia and then a few years later not have it, or you can not have it now and develop it. So we broke our participants into two groups: people who had no insomnia at baseline and at three-year follow-up, and participants who had insomnia at baseline but not at three years — so they were in the healthful, improving-sleep group. The other group was women who had insomnia at baseline and at three years, and those with no insomnia at baseline but insomnia at three years — the persistent or progressing-toward-poor-sleep group.
We found that women who had a diet more closely aligned to the Mediterranean diet — and we also looked at the DASH diet, the Dietary Approaches to Stop Hypertension — were less likely to have insomnia at three years.
Huberman: And the DASH diet is what?
St-Onge: The Dietary Approaches to Stop Hypertension was developed to reduce and prevent hypertension by increasing intakes of fruits and vegetables, nuts and seeds, consuming low-fat dairy, and a more plant-based type of diet. It has been tested in both a low-salt and regular-salt profile.
Huberman: How did those work out? Do you recall if the low-salt versus high-salt condition made a difference?
St-Onge: There is salt sensitivity, so there are some people who are very sensitive to salt and having a reduced-salt diet will really improve their blood pressure. Others, not so much. But the DASH diet regardless of its salt content did better than the equivalent non-DASH diet — your average American diet, higher in saturated fat and sugars.
Huberman: How do you separate out the variables in a study like that? People who are eating a Mediterranean diet are probably also more apt to walk more, exercise more, socialize more.
St-Onge: In population-based studies, we adjust for a bunch of covariants. We have questionnaires asking about race, occupation, socioeconomic status, different illnesses, depression, physical activity level. We try to take all of this into consideration. Obviously there are always unmeasured factors — social interactions, for example, are usually not captured very well. But one thing we did in my lab was to look at how diet influenced sleep at night in the participants in our inpatient study. We took the nine-hour time-in-bed opportunity phase only — in the four-hour phase, participants were very efficient, there wasn't much variability in sleep duration. In the nine-hour phase, there was variability. Some people got more or less. So we wanted to see if food intake was related to their sleep at night.
That study had polysomnography assessments of sleep every single night. We had a controlled diet initially and then let them self-select their food intakes. We said: first, was the diet they chose different from the diet we gave them? It was. They ate almost 450 calories more, 33% more saturated fat, a little less protein, a little more carbohydrates. Now, was their sleep at night different when eating the diet we fed them compared to what they self-selected? It was different — not in terms of duration, but in time to fall asleep, which was over 70% longer when they self-selected their diet. And their slow-wave sleep was shorter — about 20 to 23% shorter when they self-selected their diet compared to what we had given them.
Huberman: Was timing of food intake impacted? Because when I think of what reduces slow-wave deep sleep, it's eating too close to bedtime.
St-Onge: We did not take that into consideration in that study. We had their food intake profile but didn't specifically look at when their last eating period was. It could have been different than in the controlled feeding condition, because in the controlled feeding condition they had set meals at specific times. But they all went to bed at 10 p.m.
Then the other question was: what was it that they ate that day that impacted how they slept that night? And we found that higher intakes of fiber were associated with more deep sleep, higher intakes of saturated fat with less deep sleep, and more refined carbohydrates and simple sugars with more arousals. When we talk about arousals in the context of polysomnography, it doesn't necessarily mean full-on waking up — it really means going from a deeper to a lighter stage of sleep. You may still be asleep throughout the night, but you're not getting deep slow-wave sleep or REM sleep as much as you would otherwise.
Huberman: Do you create a buffer between your last bite of food and the time you go to sleep? You personally?
St-Onge: Me personally, yes. I personally like to eat my last meal at least three hours before going to bed. I know there's variability — different people have different tolerance. What we know is that eating earlier is better overall for cardiometabolic health. Me personally, I feel better by eating earlier. If I eat too close to bedtime, I get hot.
Huberman: It's the thermic effect of food. We want to be cooling off when we go to sleep.
St-Onge: Exactly.
Sleep duration, sex differences, and sleep apnea
Huberman: There seems to be something asymmetric about sleep requirements in my experience. If I go to bed at 10 p.m., I need about six and a half, maybe seven hours to feel completely rested. If I go to bed at midnight, I could sleep till 9 a.m. and still not feel completely rested. The old adage is every hour before midnight is worth two after. Is there any real data to support that?
St-Onge: I'm not sure there's data to support that. But what I can say is that if you usually go to bed at 9:30 or 10:00 and then all of a sudden you go to bed at midnight, you're kind of out of line with your personal circadian system. It's always harder to get a good night's sleep if you're not going with your internal clock or your internal circadian preference. This is what happens with shift workers — they're not sleeping at night, they're trying to sleep during the day when their melatonin is low. They're fighting their circadian system. They should be getting seven hours, but they're not getting seven hours because their body is not designed to sleep during daytime hours. Plus, you have everything else — the light, the noise, the kids, whatever life happens during the daytime when everybody else is awake and you're trying to sleep.
Huberman: Has your work explored napping at all? I'm a believer in naps and non-sleep deep rest, yoga nidra-type things, meditation. Do naps factor into this diet, nutrition, hunger equation?
St-Onge: We haven't done research on napping per se. What we do know is that you don't want to nap too close to bedtime because you want to build sleep pressure throughout the day. If you're dissipating the sleep need too close to bedtime, you're not going to be able to fall asleep when time comes at your usual hour, and you get into a vicious cycle. But there are some studies that say, if you can't sleep enough at night and you're feeling tired, you should make it a short nap — 30 minutes, no more than an hour — early enough in the day so you can have sufficient time to rebuild that sleep pressure.
There's also this whole question about what a nap is for. If you've pulled an all-nighter, it's easy to know. But if you had sufficient sleep opportunity at night and you're waking up not feeling refreshed, not able to maintain alertness throughout the day and needing a nap, I think you should check to see what's going on at night. Why are you not getting good enough sleep?
Huberman: I'm chuckling because my postdoc advisor sparked a huge debate about this. It was a big lab and we had a couple of people who liked to nap at their desks in the afternoon. He'd walk in, they'd be napping, then they'd wake up and keep working. He had this theory that if you're napping, it's because you're sleep-deprived — that napping is unhealthy. It sparked a big debate. People brought data in. I think what you just described summarizes the takeaway. I'm a believer in the short nap, but I'm one of these people who can sleep anywhere, anytime, which may be reflective of sleep deprivation.
St-Onge: Maybe.
Huberman: What other sex differences are known to exist in sleep requirements and sleep dynamics?
St-Onge: Women tend to sleep a little longer than men across the lifespan. Although if you ask women about their sleep, they don't rate it as very good. More women than men report having difficulties with sleep — insomnia symptoms, difficulty falling asleep, difficulty maintaining sleep across the adult lifespan.
Huberman: Why do you think that is?
St-Onge: There could be some physiological effects, some hormonal effects. Women don't sleep the same across a menstrual cycle. There's discomfort at different times. And then there are different responsibilities, different social roles that come into play. But women are more sensitive to the impact of poor sleep on different metabolic outcomes than men. For blood pressure, for example, at lower thresholds of sleep apnea, their blood pressure would be higher than men's. There needs to be a lot more research in this area.
Last year we published a scientific statement for the American Heart Association about multidimensional sleep health, and we concluded by recommending that clinicians ask their patients about sleep — not a targeted question, but an open-ended one: how's your sleep? Because if you ask only "how much sleep do you usually get at night," you're telling the person that the only thing that matters is the number of hours. Sleep is not just about the number of hours — it's also about regularity, quality, satisfaction, nighttime experiences, daytime experiences. When you wake up, are you feeling refreshed? Are you satisfied with the sleep you got? During the day, are you staying alert? An open-ended question allows the patient to tell you what's bothering them. Then you might get something like, "My spouse keeps kicking me because I'm snoring too loud," and then you can say, "Maybe we should test you for sleep apnea."
Huberman: Does apnea always include snoring?
St-Onge: Yes. You stop breathing and then there's this gasping sound when you awaken from or get aroused by the breathing interruption.
Huberman: I feel like so many people have apnea and don't realize it. It's remarkable how many people I speak to who say they found out they had apnea because they started monitoring their sleep. There are now free apps that can just record you while you sleep.
St-Onge: Weight loss is typically the first-line treatment if someone has excess weight — losing weight might help with sleep apnea. Then there's CPAP, which people don't like, but if they're at a lower weight where the apnea is milder, the pressure may not need to be as high, which might help with comfort.
Huberman: If people think they might have apnea, is it just: get a CPAP, pop that thing on?
St-Onge: I think they should get tested. If you're suspecting sleep apnea because you've been told you snore, because you wake up not feeling refreshed, and you're feeling sleepy during the day, talk to your doctor. Polysomnography is the first line to detect sleep apnea, but there's in-home sleep testing that can be done. You don't have to stay overnight in a lab, and your doctor can prescribe that test very easily.
Huberman: How come we can't just go buy a CPAP on Amazon?
St-Onge: Because you need to have the pressure determined for you. You need to know what kind of pressure to apply and how to set it up. It needs to have the proper settings and someone needs to tell you which setting to use, because otherwise you run into the trouble of having the wrong settings and it's not effective.
Huberman: I just know from doing this podcast a while that if people think they have to go to their doctor, convince them they have apnea, get a script for a CPAP, buy a CPAP — which I'm guessing is not cheap — and sleep with this thing on their face, very few people are going to do it. Somebody out there should come up with an at-home solution to this. Sleep apnea seems important enough — daytime wakefulness, cognitive function, longevity, metabolic health — it wicks out to so many things that it deserves a public health message.
St-Onge: If you use it well and you feel better during the day, that's reinforcing — it keeps you using it and getting treated.
Functional foods: kefir, MCTs, ginger, and meal timing
Huberman: Let's talk about food and nutrients. You've done a substantial amount of work here. First I want to talk about kefir. I love Bulgarian full-fat plain yogurt, but it's right next to the kefir. What's special about kefir and why did you study it?
St-Onge: We studied kefir because it was a fermented dairy product. With probiotics, we figured maybe it would improve cholesterol synthesis based on its impact on short-chain fatty acids. That was the subject of my master's thesis at McGill. We recruited men with mildly elevated cholesterol levels and gave them two cups per day versus just regular milk for a month — two cups like a measuring cup, about 500 mls. We measured the amount of cholesterol they produced at baseline and endpoint in both phases, and there was no effect. It was a null study. It was hard to get published, but we kept at it and got it published.
Huberman: So these fermented yogurts don't do anything for cholesterol levels?
St-Onge: At least in our study, in this population, at this level, with this comparison. But there's so many other things we could have looked at that we didn't — maybe it didn't have any impact on cholesterol synthesis, but maybe glycemic control might be better, or gut inflammation. You pick your outcomes, you study something.
Huberman: Are you a proponent of low-sugar fermented foods in general?
St-Onge: I'm a proponent. Absolutely. I think it's important to feed your gut. The gut microbiome is getting a lot of attention for all sorts of health benefits, and I think that's important.
Huberman: Now I'm going to ask you about ginger.
St-Onge: When I was a graduate student, I was interested in functional foods — foods that provide health benefits beyond their nutritional value. The McCormick company had an advertisement in one of the nutrition journals saying they were going to donate spices for research. I had a grad student and I said, "Take a look at this list, come back to me, say if there's something in there that we should test in the lab based on the things that I do." He did some research and came back and said, "I think we should study ginger." I said, "Okay, to do what?" He said he thought we should look at energy expenditure — the thermic effect of food. So we did this study.
We dissolved ginger powder in warm water, and that was one beverage. In the crossover design, the next time they came in, it was just hot water. This was a one-time consumption period. We looked at the thermic effect of food over a six-hour period — participants were under a metabolic hood, a little bubble, and we measured their oxygen consumption and carbon dioxide production. It was significantly elevated with ginger.
Huberman: Wow. So we think through the capsaicin receptor there's an increase in the thermic effect of food.
St-Onge: Yes. I was interested to see if there are little things we could do, little changes we can make to our diet to boost energy expenditure relative to intake, just to tip the scale. Many adults over the course of their lifetime gain weight, and it's not a big imbalance in calories on a daily basis that leads to 10 pounds of weight gain over 10 or 15 years.
Huberman: Tell me about this paper on weight loss and meal timing — the study where you looked at eating windows and fat oxidation.
St-Onge: We had participants on a controlled diet. They started eating one hour after waking up and had a 10-hour eating window, or they started eating five hours after waking up — a four-hour delay relative to the other condition — also for a 10-hour window. We gave participants the exact same foods, same quantity, same timing between meals, done in a metabolic chamber. The meals consumed later in the day — especially the meals later in the day relative to the earlier version of those meals — led to less fat oxidation.
Someone in the audience at the Obesity Society meeting stood up and said, "Would you then recommend that people eat medium-chain triglycerides in their evening meal as opposed to a different type of fat?" And my eyes just went wide, because my time studying medium-chain triglycerides was 15 to 20 years ago. I thought that was fascinating — timing of intake of different foods and how it influences metabolism is something that's really interesting to me.
Huberman: I confess I'm a first-bite-of-food-around-11-a.m. person. I'm trying to eat breakfast these days and shift things earlier. All it's really done is added a meal because I take my last bite of food usually around 8 p.m. I can't seem to get much earlier. But I and many other people have wondered whether it's best to eat more towards early day or whether it's just overall caloric load. You're saying it does indeed make a difference.
St-Onge: It makes a difference. You want to shift most of your caloric intake to the first two-thirds of your waking day. In that study, one hour after waking up — so basically 8 a.m. to 6 p.m. — is our eating window. It's a 10-hour eating window. It could be 8 a.m. to 7 p.m. That's reasonable. Versus 12 p.m. to 10 p.m.
Huberman: The New York schedule.
St-Onge: Yeah, the New York schedule. There have been studies in Spain that have looked at timing of eating and its impact on weight management. I'm thinking of work by Marta Garaulet, where she showed that in her weight loss program, participants who have their bigger meal — lunch — earlier in the day have better weight loss than those who have their lunch later in the day. So even in those cultures where they tend to eat late, they still find that eating earlier tends to be better for you.
Huberman: I was very relieved when Alan Aragon reassured me that except in rare circumstances where people are really trying to optimize every bit of muscle protein synthesis, 95% of the effect of getting enough protein can be accomplished by having two meals, maybe a little snack. It can be evenly or unevenly distributed. The whole notion that you could only assimilate like 30 grams per meal is totally false — it turns out you can assimilate up to 100 grams. So you could have breakfast and an early dinner with a snack in the middle. What I'm hearing from you is that you really want to avoid the big late dinner. You just don't want to eat too close to bedtime.
St-Onge: Correct.
Medium-chain triglycerides: mechanisms and evidence
Huberman: What are some of the known benefits of MCTs? Where do you find them and what brought you to them as a research topic?
St-Onge: This was a topic for my PhD dissertation. My PI got a grant looking at medium-chain triglycerides. What we did was use purified MCT oil — a liquid oil that contains 8-carbon and 10-carbon chain fatty acids, which are not very common in our general food source. We created a functional oil that also contained flaxseed oil to get some omega-3 fatty acids, and we added plant sterols because that was a big focus of my lab at McGill. But the idea was to evaluate the impact on energy expenditure, because the way we process medium-chain triglycerides is different from how we process long-chain triglycerides — the 12, 14, 16 and up carbon chains. Medium-chain triglycerides travel directly to the liver and get metabolized. We burn them off more readily than long-chain triglycerides, which travel across peripheral circulation and get deposited in adipose tissue.
In both men and women, we found an increase in the thermic effect of food — you burned slightly more calories from the meal that contained medium-chain triglycerides compared to the meal that contained standard fat. For the first study, done in women, we tried to match the saturated fat content of the diets, because medium-chain fatty acids are by default saturated. We used beef tallow as the comparison. It was a lot of beef tallow. Participants were not happy with that diet. We put it on mashed potatoes. Half of the total fat of the diet came from either the medium-chain-containing oil or the beef tallow. There was also the issue of the laxative effect of MCT oil — we had a few participants who initially felt a lot of gastrointestinal discomfort from consuming MCT because it was a lot early on. It resolved after a few days. It was a four-week study and after a few days no one dropped out for GI issues.
Huberman: That's reassuring.
St-Onge: The thermic effect increase was about 45 to 60 calories. It's a small change, but if you're going to use this versus that, you're getting a little boost. If you repeat this a few times in a day — when we measured the thermic effect of food, we measured it only after one meal, but repeated over three meals per day over a certain period of time — we did find changes in body composition, improvements in weight status with medium-chain triglyceride consumption. Lean mass to fat mass. And then we did a follow-up weight loss study with medium-chain triglycerides — this time just purified MCT oil, not added with other types, versus olive oil — and found greater weight loss with MCT.
Huberman: Based on what you're saying, it's reasonable if somebody wants to improve weight loss to shift meal timing to the first two-thirds or so of the day, which sounds like it will also improve sleep, which will also improve appetite regulation and satiety and hunger signals. What does MCT look like for the typical person — a tablespoon or two per day?
St-Onge: Yeah, about that.
Huberman: In place of some other oil, not in addition.
St-Onge: Not in addition. Correct.
Huberman: Some ginger. Are they additive? Are they synergistic?
St-Onge: I think they could probably be additive because the impact is through different mechanisms. Obviously no one's tested that. It makes me think of David Jenkins and the portfolio diet, which actually made the New York Times in December. The portfolio diet was designed for maximal cholesterol reduction. It was initially designed to have four specific foods: high in soy protein, nuts, plant sterols, and soluble fiber.
Huberman: It's going to be a tough one to get past most of the American public. People hear soy. Nuts they like but are easy to overeat. They hear plant sterols and they're somewhere else.
St-Onge: This diet went head-to-head with a lipid-lowering agent — a statin — and had the same cholesterol reduction as a statin. They've since expanded it to be more flexible. It's not just soy protein now — it also includes legumes. They've added monounsaturated fats, so olive oil.
Huberman: When I look at a diet like the portfolio diet, or the current FDA food suggestions emphasizing unprocessed and minimally processed food, the issue that always comes up for me is that in a more plant-based, grain-heavy, nut diet, it's very easy for people to overeat calories. A chicken breast or a couple of eggs is very satiating, whereas we can eat a lot of grains and nuts before we go, "Okay, that's enough." How do you ensure cardiometabolic health while quelling hunger? Do you see where I'm getting at? This is the contour of the debate — are we going to go mostly plants, grains, nuts, with lower saturated fat and improving blood lipids, or are we going to think about more protein and satiety?
St-Onge: I think there's no reason to pit one against the other. Having a diet that's more plant-based has higher volume — it's hard to eat a lot of food. If your food volume is high but doesn't provide as many calories, you'll get satiety from the food volume. Then you put in some nuts, which helps prolong satiety because you get some protein and some healthful fats. I'm not saying animal products are bad. I think they're important for a diet and for health. It's just a matter of portion size and making sure there's not an overemphasis on animal products over plant-based products, because we know that plant-based products are so much healthier in terms of heart health, reduction of type 2 diabetes, cancer risk, and other metabolic diseases.
Huberman: I love fruits and vegetables. I do eat meat — half Argentine, you know. And chicken. I'm not a big fan of fish. I keep working on this. But I don't eat them in excess. The things that are very easy for people to overeat are starch-fat or starch-sugar-fat combinations. The brain and gut respond with signals that scream "more." It's very hard for people to do just a slice of pizza. The stop signals are all pushed down and the go signals are all go.
St-Onge: Reducing white foods as much as possible. White flour, white rice, white pasta. If you're eating a slice of bread and it just dissolves in your mouth, it's not so good.
Seed oils, the Frito-Lay study, and industry-funded research
Huberman: There's a paper on your CV that I could not help but ask about: "Snack chips fried in corn oil alleviate cardiovascular risk factors when substituted for low-fat and high-fat snacks." What?
St-Onge: Yes.
Huberman: Tell me the data.
St-Onge: This was funded by Frito-Lay. At that time they had changed the oil they were using to fry their corn chips — Doritos, Fritos, Cheetos, and Tostitos. They had changed to corn oil, which is higher in polyunsaturated fats. The question was: does it make a difference? Is it going to improve health if people choose those snacks compared to other snacks? We had three arms in that study. Each person went through each of the three arms for 25 days. The question was: if you have a choice for a snack and you're going to go to the vending machine, do you eat a low-fat high-carbohydrate snack, a high-fat higher-saturated snack, or those chips? We gave two snacks a day for 25 days, rotating through four different chips. The better lipid profile was the one from the corn chips. They also had less lipoprotein little-a, which is another cardiometabolic risk factor.
Huberman: In the head-to-head comparison of seed oils with saturated fat, there are many studies showing that when you substitute saturated fat with seed oils, cardiometabolic risk factors go down. I personally avoid seed oils because I like olive oil and butter — mostly olive oil. I avoid seed oils. I love olive oil. When you look at the studies comparing saturated fat to seed oils, you do see better outcomes for seed oils. But then there's this crowd that says, but that's on a backdrop of reasonably high carbohydrate intake. When you start replacing some of those carbohydrates with lower carbohydrate, higher protein intake — not keto, but lower-ish starch and sugar — then maybe that balances out. The big contention seems to be around the processing of these seed oils — this idea that when you take fats and combine them with carbohydrate and heat them up a lot, you create factors that are not good for the body. What is the evidence for or against that?
St-Onge: Different oils have different smoke points. Each oil should be used for its appropriate usage. Some oils, like flaxseed oil, you wouldn't heat up to very high temperatures. Oils that remain liquid at room temperature — that should be your barometer for what's better to use. I'm not saying people should avoid butter like the plague. All in moderation is okay. Is there any reason to replace olive oil with a seed oil? Olive oil has a lower smoke point than other seed oils. Peanut oil, for example, has a higher smoke point, so you can fry in peanut oil. You wouldn't fry anything in olive oil.
Huberman: I wouldn't eat anything fried.
St-Onge: Depending on how you want to use your oil — some are more flavorful and more fragile, and they'll impart flavors to different foods where they're not supposed to be. The seed oil debate has been very contaminated by the issues you mentioned, but also because many processed foods contain seed oil — they're much less expensive than using grass-fed butter or olive oil or even just ordinary butter. So it's important to be nutrition-facts literate. When you're talking about processed foods, cooking at home as much as possible is ideal, but a lot of people don't really know how to do that or feel they don't have the time. Going to the grocery store and looking at the nutrition facts panel, comparing products to one another, and knowing what's relevant for your own health — because what's relevant for my health may not be what's relevant for your health. Some people are very salt-sensitive. Some people are very active and need to replace salt. Being able to know what to pay attention to is important, because otherwise it just gets overwhelming.
Huberman: You mentioned the study was paid for by a company, and earlier you mentioned companies. I think this is an important issue. Anytime I've covered a paper in a solo episode, I would always look at whether there are financial conflicts of interest. What's the difference between a company funding a study and a financial conflict of interest, if any?
St-Onge: I'm glad you're asking that, because people often have this knee-jerk reaction to industry-sponsored studies. There are people who are very vocal against industry-sponsored research. But as scientists, we do research to the best of our abilities. We draft the research question, get the data, analyze it, publish it. Some of the studies I haven't been able to publish have been funded by industry and have had null results. We wrote the paper, wrote the report, provided it to our sponsor out of courtesy. We said, "This is the paper. We're going to submit it for publication." They gave us the green light to submit. That's in the contract — your right to publish. Because otherwise, why do research? There's no point doing research if you're not going to be able to publish it.
That one paper I'm referring to — I must have tried five different journals. The findings are not exciting. They're showing there's no effect on our outcomes. It got rejected, rejected, rejected, rejected. I'm pretty persistent. I ran out of steam. So if I run out of steam, I can imagine so many other scientists who have null results running out of steam much quicker than me.
Huberman: So that's a null-result issue. It's not necessarily unique to industry-funded studies.
St-Onge: Not unique. Yeah. Industry-sponsored studies — I often also say we get NIH reports of scientific misconduct. Reports of scientific misconduct can be found from NIH-sponsored studies where they find that the principal investigator falsified data. So to me, if you're not an honest scientist, it doesn't matter who's sponsoring your research.
Huberman: Doing science for any other reason than trying to find real answers is just insane. These people are legitimately sick. It never ends well — these things always come out in the wash. So I'm hearing that negative outcomes are hard to publish. When you take on funding from a company to address a particular question about a product they sell, it sounds like you don't feel any explicit or implicit pressure for a particular outcome. Why are they funding studies?
St-Onge: They wanted to know if it had a health benefit so they could market a health benefit, probably. And if they don't find a health benefit, maybe they could switch to something else.
Huberman: I'm very sympathetic to the reality that there isn't a lot of research funding coming through NIH and NSF these days. Are you taking money from companies to do this work because it's a great way to fund studies? If NIH had more money to study nutrition, could you imagine a world where you would just take money from NIH?
St-Onge: If you could get an NIH grant, that's the ultimate goal. But sometimes there are specific foods and specific products that would be hard to study without industry support because you need access to that specific food or product.
Huberman: My fairly frequent check-in on the at least stated goals of the now-being-revised NIH includes creating a forum — even some incentive — for publishing negative results, or null results. Jay Bhattacharya, who's been on this podcast, has put that out publicly. We need those results. They're important. They steer people away from certain things. And there seems to be more and more interest in nutrition as a research topic. People are eating every day. There should be more federal funding for these things, and then there's no chance of bias.
St-Onge: I think people assume that if industry funded a study, especially on food, something's not to be trusted. I don't know why for food in particular. If you think about it — food and drug companies — drug companies do research on their own products, most of the R&D done in-house. We don't see the null results from drug companies either.
Huberman: I actually would prefer if it took on a different shape. I think outright scientific fraud — people making stuff up — is pretty rare. But I do think there are a lot of questions about people, because of the incentives to publish, not reporting things that didn't work out. When you run a lab, you want to create a culture where graduate students and postdocs feel very comfortable saying, "There's nothing here." You have a student who comes to you and says, "Hey, this is lower, this is better than this," and you look at the numbers and you say, "Well, it's 25 versus 27 and the standard deviation is 10. No, 25 is the same as 27." You have to make sure you teach them well to know that numerically different effects may not be statistically significantly different.
St-Onge: The ideal situation is when the student or postdoc doesn't believe their own results. They're like, "It's not really..." and then you have to convince them, "Actually, you have something interesting." That's a good situation.
Huberman: That's a good situation. And then eventually they're like, "Oh, okay." I think this whole field of nutrition is contentious for some of the right reasons — it's so very important. And contentious also for a lot of unfortunate and unnecessary reasons. Among students, postdocs, and the general public, when people ask you about nutrition, what are they most interested in? What's coming? What are your antennae picking up?
St-Onge: What should I eat? Or have you heard about XYZ fad? Have you heard that whatever product cures everything in the world? It's always something else. Often very specific to a product.
Huberman: Do you supplement your diet with minerals like magnesium or anything like that, or do you just rely on careful food choices?
St-Onge: I prefer careful food choices. I think it's more pleasurable to eat a complete food diet. That said, I think there are some people who may need to supplement their diets, but I think people should strive to get their nutrients from whole foods.
Huberman: Fiber recommendations are really growing. Many people's doctors are now telling them to take a little bit of psyllium husk. Doctors are prescribing supplemental fiber at a pretty high rate from what I understand.
St-Onge: That's interesting. People don't want to eat their fruits and vegetables, but there's so much more in them — all sorts of polyphenols, all sorts of non-nutrient components that themselves may have benefits for health that we don't fully understand yet, that feed your gut, that may enhance fiber's impact on health.
Huberman: Thank you so much for taking time out of your schedule. You have a very unique research program. Very few people can work on as many different things and find their points of intersection. I'm grateful that you're exploring these things. I appreciate your openness about industry-funded research. Based on your work, I think it's fair to say that we shouldn't just be encouraging people to get great sleep — we should be encouraging people to eat at times and eat foods that allow them to get great sleep, which will allow them to make better food choices.
St-Onge: Yes. I talk often about a vicious cycle where you don't sleep well, you don't eat well, and then that makes you not sleep so well. I'm really hoping for people to get into a healthful cycle — where you get good sleep, where you can make good food choices that then help you get better sleep, to keep propelling this cycle of better health.
Huberman: I love it. It's true integrative medicine and science. I can attest that when you sleep well, you make better food choices. When you eat well, you sleep better. Thank you so much for coming and for taking time out of your schedule. I've learned a ton.
St-Onge: Thank you.