How Dopamine & Serotonin Shape Decisions w/ Dr. Read Montague (Transcript)

What are a few examples of what we know about dopamine and its role in learning?

Dopamine and Learning: From Rodents to Reinforcement Learning

READ MONTAGUE: That’s a world I can’t even summarize in a quick way. People that work on rodents now will take a genetically modified rodent, and they will study the way in which dopamine release correlates with something the animal is learning.

The animal may learn to turn left when it sees a light. It may learn to run toward food. It may learn to run down a maze. All kinds of learning tasks associated with the animal are associated with dopamine fluctuations in your brain.

Now, these aren’t global, they’re all over the place, but there are different kinds of signals that you can find in different spots in your brain. And we’ve begun to understand dopamine as a central player in the algorithms that your brain runs.

And that’s where people like me—computational neuroscientists—have made a connection. And that’s the connection between the kinds of learning rules and learning procedures that are installed in your brain and installed in the brain of every mobile creature on the planet and dopamine fluctuations.

So that’s a strong connection that has been worked out over the last 30 years. The algorithms are well understood. What wasn’t well understood 30 years ago was the kind of remarkable things those same algorithms can learn. I’ll come back to that.

I mean, there have been a bunch of modern breakthroughs in what’s called reinforcement learning. And reinforcement learning’s main biology partner is dopamine. It’s the first big hit.

Now, you know, it’s an area of science. And so what happens when you have a big finding that looks like it explains a lot of things? People come rushing in to sort of beat it up. That’s their job. That’s their job to hack away. “Oh, is it really this? Does it work the same in this context and that context?”

But I think the description of what dopamine does as a learning rule is pretty much true. Let me give you an example.

So psychologists since the time of Pavlov have understood what it means for an animal to generate a prediction and to compare it to an outcome. Okay, the example is—so today is Wednesday. Suppose, and this is Rich Sutton’s example, suppose I make a prediction today that it’s going to rain 2 inches on Saturday. Okay?

Now we’re going to fast forward to tomorrow and I’m going to update my prediction because I have new knowledge. And it’s going to say it’s going to rain 10 inches on Saturday. Okay? There’s been no reinforcing feedback. It hasn’t rained yet because it’s now Saturday. Yet I’m making a prediction about Saturday.

But there’s a difference between this expectation and that expectation. Those differences are encoded by dopamine. That’s called a temporal difference error. And dopamine seems to code that before you ever get to the terminal return.

Imagine that you were playing a game like checkers. You make a move in the game and you might make, I don’t know, 40 moves before you win the game. And suppose winning the game is the reward. Well, you may have some prediction. Your brain makes a prediction when you play board position to board position that you’re going to win the game. And that’s a fluctuating quantity.

That’s a different kind of learning rule. The kind of learning rule that psychologists talk about that you think about in your everyday life is “it’s going to rain 2 inches today.” Okay, how much did it rain? Okay, so that’s a comparison between an outcome and your expectation.

What Rich Sutton and Andy Barto did was said, “Well, what you really want to do is you want to stick between there your next prediction.” So you want successive predictions. Okay, and why is that a good model for animals?

Well, because if you’re an animal and you’re wandering around foraging mainly, you’re not finding anything. You’re going from position to position to position to position, but you’re learning. And dopamine is encoding those signals.

Foraging and Reward Prediction Error

ANDREW HUBERMAN: I’m so glad you said the word “foraging” because I want to hover on the theme of foraging in the context of human decision-making and learning and behavior.

So to stay with your description, Saturday rolls around. Let’s say it doesn’t rain. Let’s say the person doesn’t want it to rain. They’re not a farmer, they want to go to the beach on Saturday. Now we can talk about reward prediction error, right? The difference between the expectation and what actually—

READ MONTAGUE: Okay, let me interrupt and correct that a bit. The reward prediction error that people talk about dopamine representing is the prediction error that you get for every single step whether or not you’ve received reward. That’s kind of diffused out in the psychology literature.

As you have an expectation and you have a reward, it may be positive, negative or zero. And what you do is you make an error there. That was understood in the 60s and 70s. It’s called the Rescorla-Wagner Rule, 1972. That’s how the system should learn.

The fact is though, that doesn’t model reality very well. Reality doesn’t give you feedback like that every time. Reality often gives you long stretches of nothing.

The insight I think of Sutton and Barto in their algorithm was, “Well, a better algorithm for learning continuously is to take successive predictions and to say that’s a learning rule.” Obviously it’s a learning rule of the outcome. When you get an outcome, when it’s not zero, but it’s successive predictions.

It’s like, why that should be such a deep idea is not clear to me. What is clear to me from data is an algorithm based on that is installed in bee brains, sea slug brains, all the way up to human brains. There are these temporal difference reward prediction errors.

And so I guess I’m sitting here trying to backwash the old version of it, which is people say in a kind of vernacular way, “Oh, it’s the difference between your expectations and reward.” Yes, when that happens, but most of the time that’s not happening, in which case it’s the ongoing difference between your expectation and your next expectation.

So it’s fluctuations in your expectation as you move through the world.

The DeepMind guys in London who beat the world Go playing champion and made AlphaFold and won Nobel Prizes—and I mean, they’re starting in 2015—they just had this unbelievable series of hits. They use the Sutton and Barto algorithm. They train those systems where people would make the players, computer players would make hundreds of board position changes before you ever got to the end of the game and update and learn based on that. They threw other tricks into—I’m not going to get technical about it.

So there’s a difference. It’s not just expectation and outcome, it’s expectation, next expectation, current outcome. And that is what rolls through and that is what we see installed in—we have a paper this week coming out on honeybee brains where you can show the same sorts of learning rules in honeybee brains. And honeybee brains, it’s probably octopamine, not dopamine.

But the other thing to say about dopamine is it’s not just dopamine. It’s very clear that lots of neuromodulators like that are fluctuating with learning and motivation and probably the whole symphony of that creates motivation states and things like that.

Dopamine as a Learning Signal

I’d like to take this word foraging and apply it to a real world scenario in humans. And then maybe we can use a combination of what’s known. And you’ll also tell us where it might be conjecture to kind of paint this picture in an intuitive way for people.

I have a friend and she’s on the dating market now. She will occasionally call me and ask me how do I decode this text message or this interaction. I try and offer my support where I can, but the conversations often go something like this: met so and so. They seem really great. They seem really busy and they set a plan for like a month from now. Is that weird? Like, all right, well, you know, and I give my interpretation. I say, well, you know, he’s nice. They’ve set a concrete plan. You know, this and that. Like, person’s busy, you know, this and that.

I also hear the, hey, you know, met someone, they’re really, really terrific. And I say, hey, listen, the last time you said this, like two weeks later, it was, how do I get out of seeing this person again? So, like, go slow, like, collect data slowly. And I’m not going to say I’m always right, but almost inevitably it’s three days later or three weeks later, it’s like, oh, my goodness, how do I get out of this thing? Right?

So in some sense, it’s what you’re saying, right? There’s a foraging for a healthy thing in life, a mate. This has happened since the beginning of time, although not with apps. There’s updating of expectation based on experience and communication. And I think this is a really beautiful example of foraging in the context of updating expectations because, and one could argue, what is the final reward? Is it marriage? Is it whatever? Okay, that’s subjective.

But I think we all can intuitively understand this example either by experience or by observation. So for someone, this person who gets excited about someone they just met, right, then meets them and is increasingly excited, but it’s unclear where it’s going to go, then finds out as life goes that, oh, they’re not perfect, there’s this thing. Can I live with that? So thinking of this as like a sawtooth of dopamine going through their system, is that statement accurate that dopamine and other neuromodulators are encoding the sort of expectation of success or lack of success, without actually knowing what the final endpoint is?

READ MONTAGUE: It’s exactly that. And that’s the insight of Sutton and Bartow. And when I first heard about this, I learned about it from Peter Dan. When I was a postdoc and we both arrived at the Salk Institute together, something about it captured me because all of a sudden, it’s not this, okay, you understand expectation and outcome. I mean, businesses understand that. Yeah, you’re disappointed. You expected to have a quarterly return of X and you had Y. That’s less. You expected it to be low. It was more. That’s, but that’s really rare.

ANDREW HUBERMAN: I studied hard. I wanted an A. I got an A minus.

READ MONTAGUE: Yeah. But the reality is embedded in this little, simple continuous learning update rule. It’s called temporal difference reinforcement learning is the fact that in the world, these expectations are going through their own trajectory. All right? And that’s what dopamine is coded for.

Any learning rule should code for the surprising outcome. You have an expectation for an outcome, and either high or low of that, every learning rule should do that. And the psychologist had that kind of figured out 40 years ago, 50 years ago. But it doesn’t quite work because it won’t account for the way animals learn. It won’t let you chain events.

So, for example, if I show a light and go and train on a reward on that outcome, and I use that expectation, outcome learning rule, it won’t chain back to something that predicts a light. Suppose a sound predicts the light, and we know the light predicted the outcome. Now I ask the question, well, what happens to the sound? Well, we know people learn they’ll associate the sound with the outcome.

ANDREW HUBERMAN: It’s Pavlovian.

READ MONTAGUE: Yeah, but those learning rules won’t do that. They learn the wrong thing. They just do. It’s just not well appreciated. Now, back when we were trying to associate that learning rule with dopamine, there was, we were mainly working on it in a kind of theoretical way. Like, if you had a signal, what would it need to look like? Where might you find this in biology?

I remember our advisor, Terry Sinofsky, who’s been on your show. I think he said something like, there are these diffuse ascending systems. They deliver these transmitters. You guys go work on that.

ANDREW HUBERMAN: Sounds like…

READ MONTAGUE: Yeah. And it just…

ANDREW HUBERMAN: Solitary. And his episode was spectacularly received.

READ MONTAGUE: Oh, great. Well, I mean, it was the most open, inviting environment, but of course, all the problems given out were impossible to solve. And I remember just thinking, what. But the first inroad was realizing that it matched what Sutton had written down not so many years before. Sutton got his PhD, I think in 1984. I think he published the paper in 1988. We were doing this in 1990, and we ran into a guy’s data on dopamine signaling. Wolfram Schultz. We didn’t know him, and we could explain every figure in every paper he published. And we just thought, okay, that’s not an accident.

Okay, fast forward. We’re in generation three now. We’re going to come all the way forward. People doing very fancy, very detailed experiments and rodents where you can control where dopamine neurons are going to fire, when they’re going to fire. You can control reward. Okay, you can just control a lot of things. And so it’s clearly more than that. It’s that and some other stuff. But that central core, I don’t see any good reason to throw away that little explanation there.

Back in 1990, the complaint was, well, that’s really cool. It matches these traces in an arcane Journal of Physiology paper. What good is that? Reinforcement learning like that can’t learn anything. The problem with that was, at the time, it was right. Like there were no systems that had done anything amazing. Now they’ve done everything, and it’s insane how good it was.

The AI Connection

ANDREW HUBERMAN: You’re talking about the AI.

READ MONTAGUE: That algorithm that I just described with my hands waving is the same thing that David Silver and the DeepMind guys did when they made the world champion go playing program, and it beat the world champion. And that particular game had expert advice built into it. Okay. And they removed all that, and then they trained it from scratch. It’s called AlphaGo Zero. And then that game was amazing. This item, it’s never been beaten. It basically beats the history of go.

And so as an example, it’s such an amazing, that’s a breakthrough. Anybody that knows that side, now, that’s the AI side. That’s the AI. But that same algorithm is installed in your head. It’s installed in the head of a songbird. The interesting thing that’s going on now is this kind of convergence, right? There are these little gremlins in your brain stem that run that algorithm.

ANDREW HUBERMAN: Okay.

READ MONTAGUE: They’ve now been externalized and put into a computer program that now does things that supersede us. It’s a little interesting convergence. It’s the only thing I know of that’s sort of crawled out of your mind into a program. And now the program is doing things that we couldn’t imagine before. And it matches the biology. I mean, you can see this in creatures as old as honeybees and drosophila and whatnot. So…

ANDREW HUBERMAN: Okay, so a couple of things. One comment and a couple questions. First comment, I’m just going to say, so that you don’t feel you have to. Everyone should know that when Reid says dopamine is responsible for X, Y, and Z, there are many other chemicals in the brain likely involved as well.

READ MONTAGUE: Other chemicals. And dopamine has multiple functions. Yeah, yeah.

ANDREW HUBERMAN: I just like anything in biology.

READ MONTAGUE: Yeah.

ANDREW HUBERMAN: We should just embed that’s up on the chalkboard now so that if you, to mention it again, you can. But don’t feel obligated to people. We’re talking about dopamine through a narrow cone here, but certainly serotonin, acetylcholine, norepinephrine, peptides we haven’t even discovered or understand yet are contributing. Dopamine is clearly a major player.

I want to step back to a human example, a non AI example, with the understanding of what you just said, which is that the algorithms that AI is running are based on the same algorithms that neurons in our brainstem are using to deploy dopamine, which I…

READ MONTAGUE: Don’t know of an example like that in the world, do you?

ANDREW HUBERMAN: I don’t. I mean, I…

READ MONTAGUE: Where we’ve discovered the nature of an algorithm, once we externalize it, we write code, and then it takes a few very special groups to all of a sudden have giant breakthroughs using that same algorithm. And those breakthroughs are going to end up pumping information back into our head. And so we live in an interest, it’s an interesting recursion there. I don’t know what will come about.

ANDREW HUBERMAN: Yeah, the fact that we took biological learning rules and gave them to a computer essentially, and the computer then can beat our own use of the biological learning rules is pretty spectacular. And I think it’s a little scary, but I want to shelve that for…

READ MONTAGUE: It is a little scary.

Returning to Real-World Examples

ANDREW HUBERMAN: Later in the discussion, I want to return to the dating example.

READ MONTAGUE: You’re going to hang this dating example…

ANDREW HUBERMAN: Around my neck, right? I think that, and we can partner it with another example.

READ MONTAGUE: Dating example is good. You go along in an interaction with somebody, you pick up new knowledge about them on Thursday, you don’t necessarily even see them. It changes your expectations of them. You pick up some new knowledge on Saturday, you run into a co-worker of theirs, they say, oh, I hear you’re seeing so and so, did you know? Blah, blah, blah, blah, blah.

ANDREW HUBERMAN: You get new changes, changes your view. So what I want to know is what is dopamine doing in the context of the constantly updated expectations? We know that dopamine is involved in motivation. Are the changes in expectation modifying motivation to either move forward, become more pessimistic, more optimistic, or stay neutral?

READ MONTAGUE: That’s great. So expectations change, those changes in expectation encoded by positive and negative fluctuations in dopamine. Where does motivation come in? Todd Braver and John Cohen had an idea about that, and I think Matt Botvinick too, and that is those prediction errors are perfect signals for deciding how motivated you should be, how much should you want a thing by measuring across those kinds of signals?

And if you were doing an experiment, you were trying to look at dopamine, depending on the timescale you looked at, you might see little changes in it that correlated with fluctuating expectations. And you might see something as a kind of envelope, a slower changing thing, which is the kind of experiment you might do in an experimental psychology setup. And that would look like it correlated with motivation with all these little wandering things going on underneath. That’s the sense in which it could do both functions.

ANDREW HUBERMAN: We are told that dopamine is what we’re seeking as we go through a social media environment, or we go through a dating environment, or we go through a financial environment that we’re investing or investing time in. But as you mentioned, dopamine is not just at the finish line. We’ve known this for a while now. It’s part of the neural circuitry algorithm that’s driving us forward or causing us to pause.

But is it fair to say that any system, whether it’s a social media platform or it’s another form of business, whether they consciously realize it or not, and they probably do, it’s built on trying to constantly update our expectations so that we keep playing the game, so that we stay in the foraging mode. Because if you think about it, it’s an infinite scroll. There is no final outcome.

READ MONTAGUE: If there was a final outcome, you wouldn’t keep living. You want that system to keep tracking and once it gets to one place, you want it to have another place to which it could go. Otherwise you wouldn’t live. Probably one insight into why it’s in every mobile creature’s brain on the planet.

So if any goal that you achieved, whatever it is, taking a drug, eating a food, getting a partner or whatnot, if that was enough for you, right, then probably be hard. That’s not the way your nervous system works. Your nervous system keeps pushing you forward. That’s what you’re working for. You’re working for this push forward drive.

The mapping that on the dopamine hits is, it’s not wrong, it’s just blunt. It’s just a blunt way to say it. It’s not wrong, but it is blunt. It’s a blunt way to say it. You move around with expectations before you get any sort of big unexpected hit.

ANDREW HUBERMAN: This is why I don’t like the phrase or the words “dopamine hits” because it implies it’s like a reward that gets trickled into you.

READ MONTAGUE: But it is true. You get a hit. It is true that there’s this unexpected reward that your expectations, your series of expectations did not anticipate. And that augments, that is the learning rule. That’s what we think the dopamine fluctuations are encoding. And so it does both jobs. It lets you update and learn and it codes for the kind of motivation you should have. And when you’re surprised, those are extra hits. So it’s not wrong to say that, it’s just incomplete.

Dopamine and Expectation Dynamics

ANDREW HUBERMAN: I’m going to ask you to speculate a little bit here, but speculate within the context of what you know about dopamine, which is a lot. Let’s take any of the different examples that I threw out on the table for us. And we artificially ramp up levels of dopamine with, let’s not say, a drug of abuse like methamphetamine or something. But, you know, we throw a little bit of a dopaminergic stimulant into the picture.

Does that just raise the height of the sawtooth? Does it change any of this? For instance, if this person who goes out on a date, on the second or third date, they go to something that, like maybe a show that’s spectacularly good.

READ MONTAGUE: Okay.

ANDREW HUBERMAN: How does that change the dynamics when, you know, it’s now, it’s now there’s an association with this person, an event. But let’s say that they’re flooded with dopamine. Let’s take a drug out of the picture. The experience generated more dopamine. Does it shape their expectation and motivation around that person?

READ MONTAGUE: If you raise expectations and these code, these are coded by changes in dopamine, then in fact, that’s sort of a tonic question. That’s sort of a tonic phase.

ANDREW HUBERMAN: Can you explain tonic? Most people are going to…

READ MONTAGUE: I will. So slower changing. Okay. So I see a show. It makes me very excited. I have the well fills up with a little more water. Okay. And it’s sitting here. So now the little hits are on top of that. Or I see something that depletes it. I take a drug, and some drugs deplete dopamine.

ANDREW HUBERMAN: Or they went to a play and it sucked.

READ MONTAGUE: Yeah. It’s disappointing or it’s sad.

ANDREW HUBERMAN: He’s got bad taste.

READ MONTAGUE: Yeah, yeah. And it just runs in your mind. And so that can lower the levels and that changes the way in which the fluctuations have an impact on learning. Okay.

Parkinson’s Disease and Dopamine Depletion

Parkinson’s disease is a condition where by the time you show up with symptoms in the doctor’s office, you’ve lost 70 to 75% of your dopamine neurons in your brain stem. Those are the only sources of dopamine in your brain, except for a tiny pathway in your hypothalamus and pituitary couple axis.

ANDREW HUBERMAN: Well, there’s, sorry, retinal biologists in me. They’re doing things totally unrelated to any of this. They’re controlling adaptation at light levels.

READ MONTAGUE: Light level adaptations, and certainly in goldfish. Yeah, those are actually very interesting. I won’t get back to that. I won’t talk about them.

ANDREW HUBERMAN: So you had the dopaminergic brainstem neurons that degenerate in…

READ MONTAGUE: By the time you’re feeling so stiff, starting to have tremors, all the parts of the flat, facey, flat affect, and somebody gets you to a doctor, you’re in the 70 to 75% loss. Okay, so what does that mean now? All of a sudden, these dopamine neurons in your brain stem are maybe 80,000 neurons per side, 160,000 neurons. That’s like nothing. They send dopamine delivering wires, biological wires, throughout your entire brain and down your spinal cord, making hundreds of millions of connections.

But now you’ve shrunk those down. And so the one thing that happens is it’s very noisy. There’s not so many neurons to code for. There’s no smooth changes in it. And the noise floor relative to what you could generate as a signal gets really, really high.

Well, one of the things that we think dopamine is involved in, in terms of information processing, is valuing the world, computing, if you will, the value of taking this action or that action, the value of grabbing this and putting it in my mouth and drinking water, et cetera. Okay? And the Parkinson’s state is sort of like a flat value function. You can’t really see differential value in things. As you look around the world, you expect this system to fluctuate for you to tell you, if I were to do this stuff, or if I were to do that stuff, if I were to look at that, et cetera, it gives you a fluctuation, but you can’t read it. The downstream…

ANDREW HUBERMAN: It’s too noisy.

READ MONTAGUE: It’s too noisy. You can’t read it. The downstream system just has to act as it did before. It says, oh, everything’s of equal value. Just stay put. So I’ve always thought about Parkinson’s as an active freezing disease. The nervous system is doing exactly what it would do if, because it takes energy to transition from where you are to doing the next thing. Why do that if there’s nothing more valuable there? This comes back to the idea of it pushing you through the world. It doesn’t habituate because it has to keep your behavior going or else you’re going to die.

Dopamine, Movement, and Thought

ANDREW HUBERMAN: I don’t think it’s a coincidence. In fact, I know it’s not. That dopamine is involved in learning motivation, feelings, and movement, among a few other more minor roles. Everything about physical movement is intuitive to us. You move forward, you move back, you move side to side, you stay put, okay?

Like movement, the idea that levels of dopamine in a moment, and what you’re referring to as the tonic kind of baseline, what I call baseline levels of dopamine, as opposed to spikes on top of that, predict whether or not you’ll move forward, how much resistance there is to moving forward, these kinds of things.

But I think for a lot of people, it might be useful to think about dopamine in the context of thought movement. And motivation is sort of a version of forward movement. If I think about am I motivated to do something? I no longer like the word motivated. I decided I like the word “a sense of urgency.” You could have a low level of urgency, moderate or high level of urgency.

Urgency I define as sort of a persistent, resilient motivation. Right. And the reason I prefer urgency to motivation is that a sense of urgency is more intuitive, I think, to most people. We kind of know when we feel we have to do something, we really want to do it, or like it’s, we don’t really want to do it, or we’re procrastinating. Whereas motivation is this just kind of like catch all term for how motivated are you?

READ MONTAGUE: Anything?

ANDREW HUBERMAN: Intrinsic motivation, extrinsic motivation. So when I think about a sense of urgency, I think about a sense of a need and readiness to move the body and or move thoughts in a particular direction. Do we think that dopamine is involved in moving thoughts and decision making in a particular direction?

READ MONTAGUE: We exactly think that.

ANDREW HUBERMAN: Okay, thank you. I wasn’t asking you to validate my non theory theory. I think that dopamine is thrown around so much nowadays that we don’t even really understand what motivation is, let alone how dopamine will be playing this game.

READ MONTAGUE: It’s very clear. Dopamine and the other neuromodulators are involved in stabilizing and sustaining brain states. Okay. That’s why they’re thought to be involved in seizures. Right. One thing you have to do with the brain state is kind of hold on to it a bit. It’s got to have a dwell time. Right. Let’s call that a thought. Boom. Okay. And then it goes forward or changes, and then it may come back to that. Okay.

So thinking and sequencing through what you would call thoughts is something that these systems are clearly intimately related to. And there are a lot of great groups now that are exploring this in mice models and theoretical models as well. So I think you tied the words together pretty well in an animal that has to keep moving to stay alive. And that’s all animals. It has to know how valuable is it, how motivated should I be, how much should I want a thing? Right.

The calculations that we think the algorithms are affecting in your brain are exactly those. And so we can have these conversations at the level of these psychology words which are interesting and pertinent to the way looking at an animal behave. But now we’re starting to pull it apart at the level of, what is this computing, how fast is it computing it, how did it update it? And now we can build artificial systems based on that.

And I think there was a paper in 2004 by David Redish talking about addiction as a computational disease gone awry, where you keep feeding the system…

ANDREW HUBERMAN: A…

READ MONTAGUE: Level of dopamine by putting a drug in that blocks its reuptake that it can’t anticipate. Right. And so it keeps chasing that and it never gets there.

ADHD and Dopamine Regulation

ANDREW HUBERMAN: When people have ADHD, even low level ADHD, or they take a drug that increases dopamine, do you think that it makes more things in the world sticky, meaning mentally sticky? Like we naturally just will latch on to more things when our levels of dopamine are elevated. We’ll forage more randomly, or do we forage more narrowly?

Because the whole notion of ADHD is that the whole, like, oh, squirrel, like, that’s the kind of generic example, is that someone with ADHD, the theory is that their dopaminergic systems are dysregulated. These drugs, almost all of them…

READ MONTAGUE: Right.

ANDREW HUBERMAN: Whether or not it’s Ritalin or Adderall or these other drugs, they raise levels of dopamine and norepinephrine.

READ MONTAGUE: Oh, yes.

ANDREW HUBERMAN: And somehow put people into a more narrow trench of focus or give them a little bit more selectivity in terms of what paths they decide to forage. Yeah.

READ MONTAGUE: I suspect if you made me guess, that it’s stabilizing brain states and thought sequences in a way that’s narrow and it doesn’t divert.

ANDREW HUBERMAN: Does that surprise you that increasing dopamine would do that?

The Waggle Dance: Bees as a Model for Exploration vs. Exploitation

READ MONTAGUE: No, bees do this. So when you’re a forager bee, you come back and you do a little dance in the hive, and it tells the hive, other foragers where to go find the nectar source. And it’s a whole language. People have worked that out. It tells you fly this far with the sun here, and there’s a polarization, an amazing phenomenon.

ANDREW HUBERMAN: The bees go back and they literally, they do this dance, the waggle, and—

READ MONTAGUE: They feel the waggle dance on the bee. And by feeling it, they know where to go wild. Well, it’s a language you can decode. I mean, it’s been decoded. It’s very cool to some degree when you look at bees. I know this because I’ve been working with a bee guy, Brian Smith, at Arizona State University for the last few years. I’ve known him my whole career, but I now have some methodology that lets him make measurements of dopamine and serotonin and norepinephrine in bees while they do odor learning.

And he has bees on an axis. Way over here are the ADD bees, let’s call them. And way over here are the concentration bees. And it relates to a chemical that’s related to dopamine called octopamine. But it’s a ratio of octopamine to tyramine. It’s like dopamine and serotonin. If you were talking about primates, the ADD bees, they feel the waggle dance and they start, you know, they start running for the nectar and then they get distracted. You know, they’re the four year old.

ANDREW HUBERMAN: There are a lot of adults like that nowadays too.

READ MONTAGUE: And they can’t. Of course, what they do by being distracted is they explore more. And then the ones on the far end over here, they fly right to the nectar source. So you need both. That’s called exploitation. This one’s exploiting where the nectar source is. It’s going to get it, it’s going to bring it back to the hive. And the sort of ADD guys are the explorers. They’re looking for new information, new nectar sources, et cetera. Well, your mind, kind of, as blunt as that is, your mind plays this dichotomy in the same individual.

ANDREW HUBERMAN: You think that we have this ADHD-like mode and a more focused mode.

READ MONTAGUE: You’ve got multiple bees inside your head. One of them is making you into the explorer, and that’s really, really valuable sometimes. And companies, they keep these people around. These are the lateral thinkers and you just have to feed them enough. And then you have the people that can really follow instructions and follow the best course of action and whatnot. And you need all that, need all that.

And this distribution of abilities is built into all of us, but it’s different across us. You know, if I was looking at an oak leaf and I told you, what about this little wiggle? It’s the wiggle in our software design for motivation and learning. It’s very effective to sometimes be the explorer and other times you have to be able to follow the chain of this is going to lead you to the thing that you want. Stay on course.

The ADHD Mode and Focused Mode Within All of Us

ANDREW HUBERMAN: If you’re a regular listener of the Huberman Lab podcast, you’ve no doubt heard me talk about the vitamin mineral probiotic drink AG1. And if you’ve been on the fence about it, now’s an awesome time to give it a try. For the next few weeks, AG1 is giving away a full supplement package with your first subscription to AG1. They’re giving away a free bottle of vitamin D3K2, a bottle of Omega-3 fish oil capsules and a sample pack of the new sleep formula AGZ which by the way is now the only sleep supplement I take. It’s fantastic.

My sleep on AGZ is out of this world. Good AGZ is a drink so it eliminates the need to take a lot of pills. It tastes great and, like I said, it has me sleeping incredibly well, waking up more refreshed than ever. I absolutely love it. Again, this is a limited time offer, so make sure to go to drinkag1.com/huberman to get started today.

What you’re describing is a sort of ADHD-like mode inside of all of us as well as a highly focused mode inside all of us. You’re also, I think I hear you correctly and thinking that you’re also describing the fact that some people are very strongly ADHD mode and other people are very strongly focused. They’re very linear or task based.

READ MONTAGUE: They can really form a task, hold it in mind. A task stays there. You know, lots of athletes are that way. They set a goal and they set multiple scales of goals. They set some goal. You know, this is where I want to be in two years. To get there I’m going to have to do, you know, have to crawl through, you know, hell to get there in two years and I have to do these things and I’m going to wake up again tomorrow morning and again and again and again.

And these goals have to be reconstituted and pursued. If you wanted to go play in the NBA and then all of a sudden six months into that you decided you want to go do ice hockey, well, that’s a problem. That’s a person who can’t focus on—

Social Media and the ADHD Muscle

ANDREW HUBERMAN: We all know these people. One question I have, and we can only speculate here, is there’s a lot of ideas now that social media, but when I say social media, I don’t want to knock on. I teach and learn on social media. What I mainly thinking about is very short form video. There’s this idea out there that it’s giving everybody ADHD. Now I don’t actually think that’s true, but I could imagine that if we have this continuum of honeybee-like modes in our brains that if we repeatedly engage in a kind of rapid turnover of stimuli like you get when you scroll a TikTok or YouTube shorts or something like that, I mean there’s a very frequent updating of lots of different contexts and information that those circuits might get stronger and that the circuits that allow you to move from node to node and route to a goal.

Updating as necessary understanding and integrating expectations and rewards and failures and all the above.

READ MONTAGUE: Right.

ANDREW HUBERMAN: The athlete example, the academic example, any life navigating relationship, all the stuff that we think of as building a solid life. You could imagine that some of that rapid updating and foraging could undermine the circuitry.

READ MONTAGUE: Yeah, you build your ADHD muscle.

ANDREW HUBERMAN: Is there any evidence maybe from related or other experiments entirely that show that if you give people a task where they have to update very quickly, that you shift the sort of state of the brain toward seeking that more and doing that more easily than you do? Kind of like long haul, a distant reward type stuff.

READ MONTAGUE: I don’t know the answer to that in people, but I do know about training artificial systems to do it. And you have to be very careful to control the mix so that it doesn’t over train on some on one of these two possibilities. If we’re going to divide these two possibilities, chase a goal, chase everything that flies along. Right. And you don’t want to do either one of those things. You have to balance that and sometimes you have to impose constraints to make that happen in an artificial network.

It’s a more complicated problem in people. I mean, I can imagine, I know lots of settings for being ADD is an absolute requirement.

ANDREW HUBERMAN: Can you give me a few examples?

READ MONTAGUE: Combat, combat, rapid decision making, kind of the fighter pilot, situational knowledge. Now what do they do to prepare for that, by the way? My dad was a captain in the Navy. I have lots of combat examples in my head. Well, they practice. They practice. They practice being surprised, they practice being hungry and they put themselves under stress and all so that when that happens they don’t have to run through every possibility. And they’re very effective. But that requires training. That requires an enormous amount of mental training. It’s all about the mental game.

ANDREW HUBERMAN: That’s a good example. We’ve had a couple of experts in ADHD on here and all of them have agreed that children and adults with ADHD, mild or severe, can focus very intensely on things they really enjoy and are interested in. It’s not a lack of ability to focus. It’s that there’s a lot of choppy terrain to get into that narrow mode of focus. Unless it’s something they love. You give a kid with ADHD a video game they love, they’ll drop right in as if it was the most focused you’ve ever seen them.

READ MONTAGUE: Anytime you have to do rapid fire decision making, I think you would want somebody who is able to at least train up to that level there.

ANDREW HUBERMAN: Do you worry about the overexposure to, you know, these frequent media?

READ MONTAGUE: Yeah, these short media. I have a lot of kids and so like every parent, my main nemesis is screen time. I’m trying to figure out how to monitor it, measure it, restrict it. And basically my kids are smarter than me and they’re more nimble and they move faster than I can. I mean, so it’s a battle I’m losing. So I’ve decided that the only way I can combat it is to lose it, but lose it a little more effectively toward my side.

But I have to admit, when I see YouTube shorts, these little, you know, like, oh, look at this person. He built a house out of jello and it’s falling over now. Look at this other person. There’s a parakeet poke. I mean, it’s mind numbing to me, right?

Learning, Effort, and the Passage of Time

ANDREW HUBERMAN: Well, there isn’t a lot of long term learning. You know, one of the things that I define learning by as useful learning is did I reflect on it again at a point later in time? You know, the other day I was on social media and I actually saw a clip. It was on a friend of mine who has a podcast, Stephen Bartlett. And he was interviewing a guest. And this speak gets right to the heart of this conversation.

You know, a lot of stuff flies by. A lot of wisdom type advice, you know, the health advice, all the, you know, it’s constant barrage. But this one stuck with me. It’s interesting. He asked the guy, what’s the meaning of life? People ask this on podcasts. I won’t ask you that today. That’s a Lex Friedman question. When you go on Lex’s podcast, you can answer it to him, but I won’t ask you that.

But Stephen asked this guy, I forget who it was, so forgive me, you know, what do you think the point of life is? And the guy said, it’s to learn, to enjoy the passage of time. And I thought, that’s pretty awesome. I would add to it and also engage in behaviors that buy you more time. You know, I said, make sure you don’t undermine the time piece of it.

But, you know, it was something that flew by on social media, but stuck with me. It is exceedingly rare that a short clip provides entertainment or information that really stays with me that I reflect on it later. Whereas when I read a book, it’s exceedingly rare that I don’t have five or 10 things underlined per chapter that I go back to later.

READ MONTAGUE: It takes a while to read a book. That’s the thing. It takes a deliberative set of intentional actions to read a book. That’s the difference in the modality.

ANDREW HUBERMAN: So one thing that this speaks to then is I’ve wondered whether activities that require effort, that may or may not include reward, but that include effort and that are a little bit slower and effort and slower tend to go hand in hand. Not always. Whether or not that is part of the mechanism that strengthens a circuit. Does effort strengthen an algorithm?

In other words, if I get on social media, it’s very easy to scroll, scroll, scroll, scroll, scroll. Short form video content doesn’t take any effort.

READ MONTAGUE: So.

ANDREW HUBERMAN: And in fact, there’s no learning involved. All you have to do is move your thumb, but there’s really no learning involved. Whereas if I have to do something, if I have to puzzle into do a puzzle to get in, or if I have to solve something or think about something or grapple with something, that is where the learning occurs. What’s the relationship between, if we know, between effort and dopamine?

READ MONTAGUE: There is a good bit of work now where people look at the amount of effort an animal has to do to accomplish a task. Let me just go back to something you just said, which was interesting. When you have to do effort, it’s easier to learn something because it slows you down. I don’t know whether effort is itself the cause or whether the fact that effort is slow and so it slows it down.

ANDREW HUBERMAN: We could design an experiment. Yeah. Maybe slowing it down.

The Rodent Model Limitations and Human Experimentation

READ MONTAGUE: It may immediately give me this idea. So that’s true in simple experiments with rodents. But rodents can’t read very well. I’ve never seen a rodent that I admired that could manage a cell phone very well. And even the rodents that can read are kind of flat affectively and all. I mean, rodent is a terrible model for this, really. I wouldn’t even do the experiment. I do the experiment in a human.

Where you can, with a few words, set a human in a certain state and go after. Or you can make them hungry or you can put a human into a mental state by just asking them to think about X, Y and Z and have various controls to account for that. I have to admit that when I look at the generation we’re concerned about, I’ve just read this book, the Anxious Generation.

ANDREW HUBERMAN: Jonathan was on this podcast and I—

READ MONTAGUE: Was on a MacArthur network, Neuroscience and Law with him for a while and he’s just an extremely clear headed person. Really always made me think about things.

Social Media’s Impact on the Next Generation

On the other hand, I don’t know other than the comparison to others and the speed at which social media lets you do that. And I have four girls and one boy. I don’t know what it’s doing to them exactly. I don’t think anybody does. I think we all suspect there’s features of it that aren’t good and yet it’s like we’re trying to hold back the tsunami. I mean it’s just the water’s going past us and so I think the only way to deal with it is kind of fly by wire.

You know, when a little fire starts over here and somebody says, oh, this really causes a depression and mood and it’s these features of it, then we can go react to that and all. But it’s very hard to know what it’s going to do globally. It’s evolving with its own. It feels like it’s independent of anything we do. And so I think it’s going to have to be sort of a get in front of it reaction.

For example, my kid just got a cell phone. She’s 13. She was the last according to her, and she’s the reporter here. She’s the last seventh grader in her school to get a cell phone. But the raw fact was I’m being left out of all the discussions and whatnot. The answer was that that is true, she is being left out. Their mode of choice is Snapchat now. Well, there’s a lot of downside to Snapchat.

So now my nervous system and my physiology is now hooked up to her blizzard of time requests on my phone. I turned it off before I came in here on the plane flying over the country. I’m denying things and giving 15 minutes and whatnot. Jonathan has real prescriptions for how to fix that. He has good suggestions for how to fix that. But the collective action thing is, collective actions are hard because they’re collections of humans and you just can’t get people to all do something at once. There’s always a defector.

Training Effortful Knowledge Integration

ANDREW HUBERMAN: Well, I think as long as we’re also training the other more slow, effortful type integration of knowledge. I mean it’d be wonderful if social media had settings where I could click entertainment. I would just get entertainment stuff. And then I knew how long I was doing that versus educate me. Because I do learn a lot from social media and I certainly try and learn on social media.

And this, what may sound like kind of a trivial statement the other day and learn to enjoy the passage of time was what sat with me in some way that felt important to me at that moment. And I’ve been reflecting on it through a couple of different lenses. We’re obviously not going to solve this problem.

I am curious about speed versus effort when foraging. Let’s take it back to the dating example. This person’s going to kill me for. I’m not going to reveal who she is. But I said, listen, I noticed this pattern over time. You discount people early or you get very excited and then it always kind of ends up in the same place where you’re like, why did I do that? I was like, well, let’s, so maybe run a different algorithm, maybe start to collect data a little bit more slowly or maybe see them more frequently for like two weeks and then make a decision. So it’s not. You didn’t waste so much time. Still, more frequently means more time, but not over time, so we can change our mode of foraging.

I personally put social media on an old phone and it goes in a supermax prison lockbox that you can’t code out of for 22 hours a day.

READ MONTAGUE: You do that to yourself?

ANDREW HUBERMAN: I do.

READ MONTAGUE: And not like the person that can’t avoid eating chocolate cake?

ANDREW HUBERMAN: It wasn’t that. I just, I’d read this paper that was published recently that said that if your phone is upside down on a table or in your bag in the same room, it lowers cognitive performance. Even if you’re not aware of the phone that it’s pulling resources. It’s pulling resources.

READ MONTAGUE: It’s calling to you.

ANDREW HUBERMAN: It’s pulling resources. If it’s in another room, it seems that your cognitive performance returns to its previously higher levels. So I thought that’s pretty good. So I started keeping my phone in the other room and I thought, how much further can I take this? So I think that the physical distance from things that’s non negotiable feels really good to somebody like me.

READ MONTAGUE: Out of sight, out of mind maybe.

Can Dopamine Encode Reward for Resistance?

ANDREW HUBERMAN: Although I want to bring this back to dopamine. Can the dopamine system learn to get motivation states and pleasure from resisting things? I think of a pathologic version of this might be. We did an episode on anorexia where food is rewarding for most people, but for people who have true anorexia, the reward system seems to enter a state where resisting food becomes the reward.

READ MONTAGUE: Control feels good.

ANDREW HUBERMAN: Yeah, control feels good. But there’s anorexia, obviously the most dangerous and deadly psychiatric illness of all the psychiatric illnesses. But resisting your phone to get other work done and to be more present for people in my life, including myself, but that seems like a good thing. So can the dopamine system encode reward for resisting behaviors as much as it can for indulging behaviors?

READ MONTAGUE: Yes. I mean, I think anorexia is a good example of it. It feels good to resist and they do it pathologically. Such a dangerous disorder, but in a healthy sense.

ANDREW HUBERMAN: Like, I’ll reveal now that you were a fairly accomplished decathlete. So that meant getting to practice doing things. But did you ever feel like, I’m going to bed early when everybody else is staying up late, I’m getting stronger?

READ MONTAGUE: Oh, I relished the whole I’m running this tennis court hill while all those other soft guys are asleep and I’m throwing up on top of the hill. Yeah, that was a thing. And it meant when you got in a tough spot. I was a wrestler all through high school.

ANDREW HUBERMAN: They’re sickos.

READ MONTAGUE: Yeah, they’re sickos. And but you’re never in better shape than when you’re active wrestler. You have to put up with things that are really demanding on you, like having your hair cut off. So the main thing you do and your air is cut off is don’t panic. Well, that’s not, you’re not pre built to not panic. So you have to learn how to do that. That was the most important thing I did in wrestling. Just learn to stay calm, think about where your weight was and all that.

It’s the same thing for people that study a lot. I think people that study a lot want to be better than the people that don’t study a lot. I mean, they want the idea of achieving a goal that’s hard for other people to do. And the most healthy version of that is without any regard to what anybody else is doing. The person who just, this is the life I live and these are my standards and I’m quiet with them and I’m going to go do this thing and it doesn’t matter what anybody else thinks. And you hope that for your children, you hope they get to be a person like that.

Anyway, I can tell you my kids school just to circle back, their collective action is to completely disallow phones during the school day.

ANDREW HUBERMAN: This is junior high school?

READ MONTAGUE: They go to a school that’s K through 12. You have put it up when you get there. And I think 3:30 is when you can activate to call for a ride or whatever. And it’s off. It’s off. Well, I like the head of school, but that’s the best decision she’s ever made. I mean, that’s a great decision. And now they’re wrestling with what do we do with AI in the school? How are we going to let these kids interface with these systems that are smarter than us? More interesting, no less.

Dopamine’s Constant Updating System

ANDREW HUBERMAN: I want to talk about AI, but before we go there, I think you’ve painted a really nice picture of dopamine and the various things it does. And even just this early statement that you made that dopamine is fluctuating according to our constant updating, not just expectation, reward, but expectation, expectation, expectation, expectation. Maybe the reward never comes, maybe it does.

Let’s talk about serotonin, because not in every case, but at least in some cases, my understanding is that serotonin is fluctuating in the opposite direction to dopamine, at least in animal studies.

READ MONTAGUE: These are some interesting in human studies too.

ANDREW HUBERMAN: Great. So educate us about serotonin in this context because I know it’s a huge topic, right?

Dopamine, Serotonin, and Neuromodulator Systems

READ MONTAGUE: A habit that people that work on neuromodulators, I’ll name a few: dopamine, serotonin, norepinephrine, acetylcholine, histamine. Probably on the order of, let’s say 15 to 20, I’d say. And then there are a lot of peptides and all but the big three: dopamine, serotonin, norepinephrine, learning and motivation, active inhibition, attention. That’s what people would say.

Norepinephrine and epinephrine are controlling attentional states. Serotonin tells you to get ready to wait. Like you put an animal, you put a piece of cheese over an area of a table and there’s an electrified grid on the table. The animal knows it’s electrified. They see the cheese, they want the cheese. Rodent, they see the cheese, they want the cheese, but the light is on. That means that the grid is active and they’re not super hungry. So they wait.

But there’s a part of their nervous system that’s making that hard, active waiting. Which also suggests another set of things for serotonin that it’s learning about negative things. Dopamine is learning about positive things or the absence of negative things. Okay, so there’s ambiguity in there because the experiments aren’t all that clear yet.

There’s an enormous amount of work going on in humans. We are the only group who record sub-second levels of dopamine and serotonin in conscious human beings while they do things: reward motivated tasks, social interactions with other people, various kinds of visual perceptual tasks, looking at emotional pictures, positive, negative and neutral and whatnot.

The Opponent Relationship Between Dopamine and Serotonin

The theme that emerges from that is dopamine and serotonin are opponent to one another. When dopamine goes up, serotonin goes down. When serotonin goes up, dopamine goes down. We could talk about those events as being for positive events or anticipation of positive events. Dopamine goes up and serotonin goes down.

An opponent to that at your own institution, Rob Malinka has a set of beautiful results in rodents where the learning that they see in the animal requires that kind of opponency. It’s a definitive experiment in the rodent. It’s harder to do these things in humans because you can do simple things in humans, that’s fine. But humans can sit and have an idea and it can generate these kinds of signals and they can run through the ideas. And so that’s a hard thing to both get our hands around and to do in a controlled setting. And so that’s why it’s been ambiguous.

But the first time we were able to measure dopamine and serotonin concurrently, they look opponent and they look opponent all over the place. They’re old ideas from the 60s and 70s about opponent systems in this sort of affective processing space. Dopamine has now inherited the positive part of that and serotonin, the negative part of that.

Opponency is a theme in the nervous system. In the retina you have color opponency, you have light and dark opponency. These kinds of information channels go all the way through to the visual cortex.

SSRIs and Their Effects on Dopamine and Serotonin

One other thing that’s interesting is that when you put SSRIs on people, you prevent selective serotonin reuptake inhibitors. Prozac, Fluoxetine, Lexapro, it blocks the reuptake of serotonin in the serotonin terminals over a few weeks period. You have a clinical effect. And for some people it’s a life changer. It’s very heterogeneous, but it pushes serotonin into the dopamine terminals too.

This is less well understood, but if you were a system and you thought that the positive juice was dopamine and the negative juice was serotonin, and you put the negative juice in the positive terminals, then the cells that control the release of that are going to chatter for positive things. You might start negatively conditioning on things that you should actually pursue and learn about.

ANDREW HUBERMAN: SSRIs have helped a great number of people. There have also been some devastatingly tragic circumstances where SSRIs have, the theory is that they’ve accelerated suicidality, they’ve accelerated anhedonia. They’ve created a lot of problems. If we were to just take a step back in terms of serotonin as learning about negative things, if you could just summarize these results for me, animals and what the expectation would be in humans.

So let’s say that somebody or an animal is learning a task where they get shocked. If one were to artificially increase serotonin, does that make somebody or an animal more or less likely to code something as negative?

READ MONTAGUE: Well, the idea would be it makes you less likely to code something as negative because you have less serotonin in the serotonergic terminals. And so if they’re communicating this information about more serotonin in the serotonin terminals.

ANDREW HUBERMAN: Gotcha. So if somebody takes an SSRI, serotonin is increased, and they have a tough interaction at work. The idea is that they would encode that cognitively as less bad because there’s an abundance of serotonin or worse than it would be had they not been on this drug.

READ MONTAGUE: When you increase serotonin in your brain because you won’t let it be vacuumed out by the normal mechanisms that clear it from your brain, then it has the opportunity to be there longer, and it has the opportunity to go into the dopamine terminals. This is something we know. The mechanisms that suck dopamine out of the spaces of your brain will also bind to serotonin and suck it out. Not quite as well because it’s tuned. It’s called a dopamine transporter. And so, depending on what the downstream parts of your brain think, that in fact, increasing serotonin could decrease the serotonin in the serotonin terminals by blocking the reuptake.

ANDREW HUBERMAN: I see. So that’s why you said it earlier. I tried to correct you, saying, no, it’s going to increase serotonin because you’re blocking reuptake. You’re saying, no, it pushes serotonin into the dopamine terminals. And this is why people might not get as much reward from a positive event when serotonin is elevated pharmacologically.

READ MONTAGUE: There was a killer paper 20 years ago on that where they showed they gave rodents some common SSRI. They waited this number of weeks and they went in there to say, where is the serotonin? Okay. And what they showed was that the dopamine transporter pathway was the thing that was taking it into the dopamine terminals, because that’s where the dopamine transporters are.

And so I don’t know where that’s gone since then, but that’s a 20 year old result. It’s a very clear result. It was in a journal called Neuron. John Dani was the senior author. He’s at Penn. It was a remarkable paper. I don’t know that people have followed up in humans.

I actually think the only way to follow up in humans is to kind of do what we’ve been trying to do, which is develop methods of measuring these things in humans directly. We’ve been able to do it in people that are having brain surgeries and they have an affliction. They’re going to have an electrode put in their brain for various reasons, and they let us piggyback on that. They consent, obviously, under strict ethical guidelines.

But we can measure serotonin and dopamine when they do a rewarding task or they play a game that has a series of things that go on back and forth with another person, which I like better in the sense that that’s a more natural reward. Somebody does something to you and you do something back to them. Usually in our case, their economic games, this is encoded in money or the expectation of the money that’s going to come. And you can see strong opponency in dopamine and serotonin signaling in the deep structures in your brain.

Measuring Dopamine and Serotonin in Real-Time

ANDREW HUBERMAN: In these experiments, if people positively anticipate because things are, quote unquote, going well for them, you see dopamine going up and you see serotonin going down. And if they’re losing at this game or they feel like the game isn’t going well for them in some way, there’s more uncertainty. Perhaps serotonin goes up and dopamine goes down. Yes, interesting.

READ MONTAGUE: And then there’s state changes in your brain that can be induced by, for example, making somebody hungry. Where we don’t really know how to explain what we’re saying, but they still show opponency.

ANDREW HUBERMAN: What would you say being hungry does to the dynamics for, let’s just take them one at a time. Dopamine. Does dopamine still increase for positive events when people are hungry?

READ MONTAGUE: No, no, not in rodent. In rodent models. I can talk about rodent models. We’re actually in the middle of doing something like that now, where people come in in the morning hungry. In this case, these are people with epilepsy that have wires in their head and we do an experiment on them when they’re right before they’re going to eat and then we repeat the experiment after they’ve eaten.

But in rodents, it’s very clear, I guess at the level of the amygdala. If you make a rodent hungry, then you can show that dopamine will encode something like punishment prediction errors, not reward prediction errors. In other words, it’s like it flips its role. It’s like if you’re in a…

ANDREW HUBERMAN: How hungry do you have to be?

READ MONTAGUE: I don’t know how it feels to be a hungry rat, but imagine that it put it in an emergency state.

ANDREW HUBERMAN: Okay, so it’s not just a kind of like, I don’t do any formal intermittent fasting, but I usually eat my first meal somewhere around between 10 and noon and at 9 a.m. I’m like mildly hungry. I could eat, but I’m not…

READ MONTAGUE: But by 4 you would be hungry.

ANDREW HUBERMAN: I’d be really hungry.

READ MONTAGUE: You’d feel it. You’d feel what we’ve all felt when we’re hungry. And this is a guy called Mark Andermann at Harvard.

ANDREW HUBERMAN: Oh yeah, I know Mark.

READ MONTAGUE: Yeah. So he puts animals in starvation states and he shows that dopamine will encode aversive events, aversive errors, very clear result. So folks, I know this because he called me. I mean, we met.

Hunger and Decision-Making

ANDREW HUBERMAN: Your kid, boyfriend or girlfriend is hungry and you’re going to a show or you’re going someplace. You got to feed them if they want to, if you want them to enjoy the time. I mean, that’s sort of obvious on the one hand, but I don’t think we really…

READ MONTAGUE: Oh, it’s even better than that. There’s an Israeli paper from, I don’t know, about 10 years ago where they looked at judges and the judgments that were made if you hadn’t eaten versus judgments that you made that you had eaten and you really want to judge that’s had a good lunch.

ANDREW HUBERMAN: Very interesting. So general state of stress, because hunger is a form of stress, drives the direction of the dopamine to either reinforce positive things or reinforce negative.

READ MONTAGUE: Yeah, because think about it. If you get to a state where you’re really starving, things have not been going well for a long time. You’ve been making really bad decisions. The creek dried up, some forest fire came through and ruined your foraging area or whatever. Things are going really, really bad. Are you going to really sit around and wait for the rewards? The main thing you want to do is stay alive. You don’t stay alive. It doesn’t matter what rewards you chase.

And so, in a sense, flipping dopamine’s meaning is exactly what you’d want to do. You’re in an emergency state, and you want to use this reinforcement system, this expectation system to stay alive. You want to pay attention mainly to those things. You want to pay attention to them. You want to be motivated by them. You want to be motivated and pay attention and avoid the negative things. But that’s an emergency state.

Negative Feedback and Learning

When I talk to people about how reinforcement learning models, say, have an impact on how you should train an animal, typically in my case, it’s in the laboratory setting, but you could use this. Other dogs, for example. Training animals with really negative feedback is a really bad thing to do. Because what happens when you get really negative feedback? You’re in a mall, somebody shot beside you. That’s negative feedback.

What happens, you have, in the extreme case, you have PTSD, but what you do is you overgeneralize. That was so bad, it’s rational for your nervous system to think anything that looks like the mall, the fear will start to come. It’ll move out to the curb, it’ll egress. This is the whole PTSD cycle. But that’s rational. That’s rational. That was an absolutely unexpected, cataclysmic event. You better. And you don’t know what could have caused it, really, as far as events leading to it. So you overgeneralize and all. So you don’t learn very well like that.

So you know a teacher that instead of when you miss, when you’re trying to add fractions and you don’t get a common denominator quite right. When she takes a ruler and slaps you over the hand. They could still do that. When I was in school, this is a generational shift, a good one. That’s a really bad way to teach me to find a common denominator. Instead you could just say, nudge, nudge, nudge.

ANDREW HUBERMAN: Anytime someone says, “I have a friend,” it’s like code on the Internet for, like, it’s actually you. And that’s like the, I have a friend, but so I have a friend. People might be surprised to hear that I have friends. Perhaps not. I have a lot of friends. And he’s a lawyer. And prior to becoming a lawyer, he studied torture.

READ MONTAGUE: In what context?

# Torture, Stress, and the Inversion of Reward Systems

ANDREW HUBERMAN: Yeah, he was going to become a psychiatrist and he did a rotation with former victims of torture and then it took him down this rabbit hole of political torture and the history of that. He’s a real history buff. He’s a very benevolent person, very kind.

But he told me something interesting that I think tells me that some people figured out this thing that if somebody is stressed enough, it contorts the dopamine reward contingency so that dopamine no longer encodes positive things. It just tries to prevent things from getting worse, is what we’re basically saying here.

He said that the way that people torture people to get information from them is actually pretty surprising. At least it was to me, which is they hurt them a little bit and then they tell them they’re going to hurt them a lot and then they don’t. Rather than hurting them a lot, somehow by hurting them a little bit and then telling them this is going to get much worse unless people give up more information. It’s interesting and I think it speaks exactly to this mechanism.

READ MONTAGUE: Speaks to you hurt a person a lot. That’s cataclysmic. That’s categorical. I’m going to thank you for not breaking my arms again on Tuesday. Right. I mean, if you break limb and you do these cataclysmic things which leave them near death, there’s nothing left. It’s a bad way to train someone.

So I mean, what I’m trying to say is I understand that. And the odd thing is when you stress someone enough, it’s remarkable what becomes rewarding. You know, the incremental removal of threat. Given that you’ve made good on a little promise. Could be a lot. I mean, I’ve seen that. That doesn’t even have to be as extreme as torture. Read Dostoevsky. Family dynamics. Read Dostoevsky and look at the family dynamics and go, you know, they’re dialing knobs on degrees of punishment and it’s very effective.

When Dopamine Gets Too High

ANDREW HUBERMAN: Let’s take the inverse of this, let’s shine some light in the room, so to speak. What if dopamine gets too high and I’m not talking about methamphetamine, which will really skyrocket dopamine. I would say when people are on very high levels of dopaminergic drugs like methamphetamine or cocaine, everything seems like a good idea to them and they become very self-obsessed.

In fact, there’s a wonderful documentary about the Grateful Dead that I watched recently before Bob Weir died, which someone was saying, you know, at some point in the mid-1980s, it got a lot harder to make great music. And someone said, what happened? And they said, cocaine. Said why would cocaine do that? And he said, because it’s a me drug.

Dopamine and cocaine are synonymous with one another, but there are a lot of situations where people are overindulging themselves with food, overindulging themselves with dopaminergic activities. What does that do to the reward?

READ MONTAGUE: If you look at it, on the average, it resets expectations where very few, if any, natural events can exceed them.

The Abused Dog: When Reward Systems Are Permanently Inverted

READ MONTAGUE: So I rescued a dog once back, I mean, when I was a kid. I kept lots of animals when I was a kid. I had my father told me at the end, oh, no, you had over 30 cats.

ANDREW HUBERMAN: Oh my God.

READ MONTAGUE: They stayed outside. That was back when animals…

ANDREW HUBERMAN: Oh, I just kind of…

READ MONTAGUE: 30 cats.

ANDREW HUBERMAN: I had a cat. The cat had a litter. There were seven in the litter. They all survived then.

READ MONTAGUE: Good for you, man.

ANDREW HUBERMAN: I wasn’t really a budding scientist then, but I realized in retrospect that I really watched them, okay, and understood their behavior. And there was this dog, this little dog that I rescued that had been beat up and stuff. And that dog was never right. It had been so abused that basically it started out by biting you, right?

And that’s what happens when you hurt a bit animal, you know, when you take it past the edge. Of course, then you take it even further and you have learned helplessness. So you just sit and don’t do anything. It was tragic. I couldn’t get that dog to lighten up. Those cats and but her world was inverted permanently. Had just been completely inverted. Up was down, down was down, was up.

ANDREW HUBERMAN: Basic safety was reward. Everything else.

READ MONTAGUE: You were in an emergency state. It’s just a lot easier as a behavioral commitment to just start out by biting because you’re going to have to bite at some point anyway.

ANDREW HUBERMAN: Yeah, well, I think we’ve all known some people like this. Yeah, it’s tragic to see hurt people.

READ MONTAGUE: And it’s interesting as we get more and more knowledge about how to hack that and intervene on that. But it’d be nice to be able to fix people like that. I think they would like to be fixed. I think of some people, I know a cousin in particular, drugs of abuse, you know, do this to people. They just, you know, they get people into these states where they just people make decisions that they know are going to lead to, you know, they’ve done it before and they’re just going to go down the hole again. I have lots of family members where that would be true.

Dopamine and Serotonin: A Summary of the Opponent Systems

ANDREW HUBERMAN: I think nowadays we all know or are aware of people that fit that because of the incredible expansion of availability of drugs of abuse, including prescription drugs. I think, if I may, I just want to just to make sure that I’m staying oriented here.

Here’s where we’ve gotten. It seems dopamine encodes positive expectation and rewards. And it’s graded. You can have low levels, medium or high levels, depending on how much positive anticipation. Serotonin seems to encode negative events.

If a human or an animal sadly is raised in conditions or spends enough time in conditions where true rewards aren’t there and survival itself becomes the reward, the dopamine system will adjust its baseline so that it’s just fighting for the survival.

READ MONTAGUE: And it’s important when you’re fighting for survival to recognize and anticipate negative events. And so the Anderman’s work shows that it does prediction on the outcome that’s going to be negative. And it gives you a positive pulse for that. You better go learn about that thing, you better learn about this thing.

Because you’re in a negative state. I don’t know how you got here, but you’re in a real, you’re in an emergency state. You know, this is what stress would put you in. And so positive becomes negative. You need to have positive prediction errors to the prediction of negative events because that’s what’s going to keep you alive by paying attention to that.

The Cat in a Room Full of Rocking Chairs

ANDREW HUBERMAN: I keep going back to relationships, but there’s so many examples from friends in my life. I have a really good friend who had had a series of very, very challenging relationships. I mean, just brutal. And then has entered a phase of his life where things are really good and really peaceful.

And for about the first, I don’t know, three years of that new healthy relationship, hey, how’s it going? And he’d be like, and he used to say, I’m like a cat in a room full of rocking chairs. He’d say, and I go, why? It’s tense. And he goes, no, it’s so calm.

And I would say only recently has that message changed. And now he’s like, really? I mean, he’s really flourishing. The relationship is flourishing. The whole landscape around it is flourishing. And it’s really cool to see, but it’s exactly what you’re talking about.

You know, it takes some time of feeling safe for somebody to stop just thinking they need to fight for survival and safety. And then it seems that the dopamine system can then adjust its baseline so that it can now work for rewards again. Pretty incredible. Makes good sense that the dopamine system will be adaptive in this way and not just for rewards.

Because throughout human evolution, I mean, people have had to deal with tremendous hardship and stress when things are really shitty. What is the serotonin system doing?

READ MONTAGUE: Much less has been done on that. The thing that we know for sure in humans is that in all these probes that we have done, it’s opponent, it’s opponent, it is going the opposite direction. Some of the best data is from humans, not rodents. It’s a minority observation in rodents. I mean, it’s scattered. It’s quite hard to engineer the behavior in a rodent. I think it’s…

ANDREW HUBERMAN: Oh, I love that we’re talking about humans. I mean, most people listening are interested in humans.

SSRIs and the Dopamine System: A Hidden Mechanism

ANDREW HUBERMAN: I am totally fascinated to the point of being blown away by this SSRI thing, that if serotonin reuptake inhibitors drive up serotonin, which they do, they prevent reuptake, that some of that serotonin gets into the dopamine…

READ MONTAGUE: A lot of it does.

ANDREW HUBERMAN: A lot of it does.

READ MONTAGUE: So it prevents…

ANDREW HUBERMAN: Why haven’t we been told this? And then it lowers the rewarding properties of good stuff.

READ MONTAGUE: That’s the best way to explain it. And, you know, somebody doesn’t think that’s the explanation. But the fact is, when you put an SSRI on, the serotonin that’s released due to activity in serotonin neurons, it’s not going back into serotonin terminals. Where does it go?

This paper by John Dani in 2005 showed it goes into the dopamine system. And he knows that because he could block the dopamine reuptake. It was a 40% difference. So there’s all this serotonin sitting there in these terminals. They’re going to be releasing now it’s the negative juice, let’s say.

Let’s say that on the other side of this signaling pathway, electrical activity comes through. You release this transmitter, it has an impact and the receiver goes, oh, I’m getting a lot of negative stuff here. But in fact, it’s because it’s sitting in neurons that chatter for positive things. You would have a hard time learning about positive things.

You might also register negative events as being rewarding and you’d learn yourself into a kind of depression that way. That’s an interesting set of possibilities, physical possibilities. Yeah, that was a fantastic paper. I don’t know why it didn’t sort of catch on. I think they’re I don’t remember, but I know John Dani quite well and he did hard experiments and they took a long time.

ANDREW HUBERMAN: We’ll put a reference to the paper. I think I’ll answer my own question by saying that I think that really good scientific findings and theories need advocacy to get legs.

READ MONTAGUE: They need a shepherd.

ANDREW HUBERMAN: They need a shepherd. I mean, it’s part of the reason I started the podcast and invite amazing guests like you and like Terry on and people who really think deeply about the whole field. And you’re changing the way that I think about serotonin, SSRIs, dopamine. You’re expanding all of it, truly. And I know for those listening, that’s also true.

I know you’re chomping at the bit to talk about learning algorithms and AI. But I want to know first about the experiments where you stuck wires up people’s noses and recorded dopamine signals in their noses. Because these are wild and cool experiments.

Measuring Dopamine and Serotonin in Human Brains

READ MONTAGUE: They are. And they’re not wild and cool because of me. They’re wild and cool because of Christina Zelano at Northwestern. But let me say one thing about how we do measure dopamine in human brains. We do it in very specialized circumstances where you are having a deep brain stimulating electrode put into your brain to treat a movement disorder like Parkinson’s disease or essential tremors.

So when you have that and a minority of patients choose to have a small burr hole put in their head and under careful operating room procedures, it’s put down into different structures in your brain. We won’t name them, and then turned on. And it’s symptom-relieving. Essential tremors is like Parkinson’s disease. You have tremors and you have difficulty with movement and whatnot. I don’t think you have the emotional problems that Parkinson’s patients do, but it’s not Parkinson’s. You have not lost dopamine.

If you give an essential tremor patient dopamine drugs like they do Parkinson’s patients, they get much worse. The tremors are typically just really irritating for people. And so they do elective neurosurgery to have microwires put in their brain. And it’s a very active area of clinical neuroscience and clinical treatment. Parkinson’s disease, also, you can have a stimulating electrode put in.

When you do that, you put a little tiny, and I mean tiny, guide tube down and they drop the electrodes in there. And under those circumstances, we ask to put an electrode in equipped with a neural network model that knows how to interpret electrical signals on the electrode as dopamine, serotonin, norepinephrine, pH, and peroxide fluctuations. This isn’t exactly, you know, you don’t go into Walmart and find this kind of stuff. I mean, the reason nobody’s heard about it is because it’s a very specialized area right on the edge of translational neuroscience.

And so it’s there that we’ve gotten recordings from deep in the brain. What’s amazing is when you ask people, would you let us piggyback on your electrodes? Because the electrodes that they use have research contacts, and we can make measurements of these transmitters without sitting on any of the clinical bandwidth. In other words, we don’t eat up any of the ability of the neurologist to use the electrode output to make decisions about the treatment.

There are a lot of moving parts in that. I’ve been doing this for a while. And so for about the last 12 years, I got very motivated to measure dopamine in human beings and at timescales that were physiological and during cognitive events that we find meaningful. So I thought the method was very clever. I won’t even go into talking about the method. It kind of worked right away, but the entire process didn’t work right away. And it’s taken way over a decade. I mean, it took a lot of work.

So we have sites set up around the world where we do these things. It’s in that context that we have got knowledge about how to use these depth electrodes to, instead of just measuring electrical activity, to do neurochemistry.

The Nasal Electrode Breakthrough

Then I ran into Christina Zelano, who’s an olfactory physiologist at Northwestern unit, full professor at Northwestern University. Very gifted. What Christina was doing was taking these depth electrodes that are FDA approved, normally used to be put carefully down into the tissue of your brain. They’re basically just rubbery little tubes about a millimeter in diameter and snaking it up people’s nose and just laying it up against a region of the olfactory epithelium, the part of your tissue inside your nose, way up high, basically around here, near your eyeball.

ANDREW HUBERMAN: Near your eyeball, above your eyeball?

READ MONTAGUE: Above and northwest of your eyeball if you’re not. And doing electrophysiology, listening to the electrical activity. And she already had rodent model stuff. And I went, I can totally get the chemistry off of that. And why is that important? Well, other than being weird, you can consent healthy people into doing this. You can snake this thing up there and clip it to their nostril, set up the electronics beside them, and then you can do all kinds of stuff, including letting them eat, letting them do mindfulness meditation, breathing exercises, letting them do decision making tasks with and without other people. You can do simple things like just a stimulus and then squirt odor.

ANDREW HUBERMAN: And then a rewarding smell and measure dopamine and serotonin.

READ MONTAGUE: Oh, yeah. And so we’re giddy about this mainly because we can consent healthy people into doing this. One of the complaints, of course, of doing it in people with epilepsy and Parkinson’s and whatnot is they have epilepsy and Parkinson’s. They have an affliction on board.

ANDREW HUBERMAN: Could you just share with us, are there any top contour statements that we can make about brain state dopamine and serotonin as measured through the nose? Like, for instance, if you see a fluctuation in dopamine through one of these nasal probes?

READ MONTAGUE: What we see in the nasal recordings looks very much like exactly what we would expect if we were recording from the neurons in the midbrain, based on what people have recorded on the simple experiments. You know, there’s a cue, there’s a reward, there’s this, it went up, it went down, that kind of thing. This is a positive picture. This is a negative picture. This is positive effect. This is negative effect.

ANDREW HUBERMAN: Okay, so dopamine increases when there’s a positive expectation. Serotonin increases when there’s a negative expectation. And you’re recording that from the nose? Essentially non-invasively. Except somebody has to…

READ MONTAGUE: Apparently, the language is minimally invasive.

ANDREW HUBERMAN: All right, well, cool. I can live with minimally invasive.

READ MONTAGUE: So I haven’t done it yet myself because when I was scheduled to do it, we realized we clipped the age at 65, and I had had a birthday on Sunday and I aged out.

ANDREW HUBERMAN: So how old are you now?

READ MONTAGUE: I’m 66.

ANDREW HUBERMAN: Looking good, man. People are going to be like, what are your protocols? You know, raise five kids, run a big lab.

READ MONTAGUE: Never sleep?

ANDREW HUBERMAN: Never. Do you not sleep? Never sleep?

READ MONTAGUE: Not really.

ANDREW HUBERMAN: Do you not sleep well or you just work all the time?

READ MONTAGUE: You know, my dad, who died at 91 in 2021, he didn’t sleep.

Sleep Patterns and Personal Habits

ANDREW HUBERMAN: I don’t take that. Matt Walker, Brian Johnson. You don’t have to sleep to live tonight. A lot of that’s genetics. I’m just teasing. A lot of that’s genetics. I think I do fine on six hours. I prefer seven, but I don’t need eight. I definitely do not need eight.

READ MONTAGUE: It’s really variable with people. And then there’s the cognitive people, the people that develop an opinion about how well or long they slept, and that idea circulates in their mind. “I didn’t get much sleep last night.”

ANDREW HUBERMAN: How much do you sleep per night? I know we’re taking a tangent here, but people will find this interesting, and I certainly do.

READ MONTAGUE: When I was younger, it would be like four.

ANDREW HUBERMAN: Okay. I know another person like that.

READ MONTAGUE: Now, because I get up really, really early.

ANDREW HUBERMAN: What time do you get up?

READ MONTAGUE: I get up 3:34 in the morning. I really enjoy quiet.

ANDREW HUBERMAN: What time do you go to sleep?

READ MONTAGUE: Well, I go to sleep. I could do the math twice. Yeah. You know, I’ll fall asleep in the evening and then I wake up.

ANDREW HUBERMAN: What time in the evening?

READ MONTAGUE: Like 8, 9, 8 night sleep till 10. But if I do that, I feel good. And then I have to pretend like I need to go to bed and so I’ll lie down and then, you know, when everything’s quiet, I’ll move back downstairs. The way I do science is I have to get quiet. That part I can’t do with other people. I have to do it in dead quiet.

ANDREW HUBERMAN: The data would say that that first round of short sleep, you’re grabbing your deep sleep, you’re getting your growth hormone surge, which is great bodily repair, sufficient to keep you healthy enough. And the second phase, probably you’re getting some REM sleep and enough to. You seem like an emotionally stable guy, so.

READ MONTAGUE: Well, that’s not true. But you know, my mom, who died in 2023 at almost 90, she used to complain to me as a child, “You’re up.” Because I’d wander outside. And back when I was a kid, no one was scared of anything, so I’d walk out in the dark. I might walk a mile away from home or something. I mean, we just weren’t scared of anything then. Now we’re scared of everything.

And I worried about it until I got a little older, until I was 12 or 13 years old. And then I just realized I’m just going to decide how I think I feel and that’s it. And everybody else is different than me. And I was raised in a community where there was clearly something wrong with me compared to everybody else.

ANDREW HUBERMAN: Or maybe something was wrong with all of them.

READ MONTAGUE: Yeah, well, maybe.

ANDREW HUBERMAN: I mean, it’s cool that you learn to trust that because we get a lot of messages about we need X, Y and Z. And I mean, you clearly are, you know, you’re a competitive athlete, your lab’s done spectacularly well. I mean, it works for you. So, you know, if it works for you, it probably also helped with raising kids because having all that energy to raise five kids.

READ MONTAGUE: Well, I’ve had two marriages, so maybe it worked or didn’t work.

ANDREW HUBERMAN: So you have some conditioning too.

READ MONTAGUE: You know what? You can delete that.

ANDREW HUBERMAN: We’ll delete that.

Life Lessons and Personal Growth

READ MONTAGUE: When I hear you talk about people, I just think to myself, oh, you know, I have a person, they have a relationship. The one thing that does for you is you back off people a little bit.

ANDREW HUBERMAN: What does for you?

READ MONTAGUE: Divorce. You kind of go, if you think at all about it, you just go, nobody is one thing. Everybody’s a little complicated. Nobody’s Mother Teresa 99.9% of the time. They were a jerk last Thursday or they were this. Nobody gets through life without making stupid, dumbass mistakes. And it’s easy to be judgmental until something bad has happened to you, like something really kind of soul crushing, you know, like a divorce.

And you have to go because generally there are two people involved in that. And so that’s a learning lesson. That helped me. It’s helped me in many ways. I appreciate my life now for a reason, and it wouldn’t do it if I didn’t have those kind of scars. Also science, I don’t know if you talk about this. Science is a contact sport.

ANDREW HUBERMAN: I haven’t talked so much about this.

Science as a Contact Sport

READ MONTAGUE: Science is a contact sport at the leading edge. Science is a contact sport for it. And, you know, there are a lot of smart people doing science on the world stage and certainly on the American stage. And they’re out there sort of battling at the frontier. And the first thing that happens when you do anything good is, you know, out come the chain mail and the maces and whatnot, and you have to kind of fight for yourself a little bit.

And so then it asks of you to look inside yourself like, is this an important problem? Do I really believe this result? And our job, the reason we’re paid tax money to discover stuff, is our job is to push the edge of what we know, not sit there just getting money to twiddle our thumbs. And so if you’re on the edge, you’re going to make mistakes or you’re going to be wrong or you’re going to be attacked or not popular. And that never ends. We call it the Reviewer Two syndrome.

ANDREW HUBERMAN: Yeah, Reviewer Two is the one that makes your life more difficult. But maybe it makes the papers better in the long run, maybe stronger. It’s the brutally hard coach of our career. I mean, I’m no longer running a lab. I did until a few years ago. And I’ll tell you that the other thing that’s brutally hard about science is that just the work is hard. The culture of it also has some punishing features. But they build us, they make us stronger.

But it’s one of the few professions where there are others. But it’s one of the few professions where you have to work exceedingly hard to get the resources just to do the work. So it’s like having two jobs wrapped into one.

READ MONTAGUE: And I had no idea that’s where knowledge and textbooks came from. When I was 10 and living in Macon, Georgia, I don’t think most people do.

ANDREW HUBERMAN: Part of the reason we started this podcast is that people should interface with scientists, learn from them, understand kind of some of what it’s about. I mean, it’s still an awesome endeavor, right, to discover things. But you’re right, you have to have some real fortitude.

READ MONTAGUE: But you’re not told that when you join the club.

ANDREW HUBERMAN: No, not so much.

READ MONTAGUE: You’re not really told that. You’re generally, science is. You’re an assistant, you’re an apprentice. It’s an apprenticeship training. You go sit by some person who’s great at X, and the main thing you do is you absorb them doing all these little things. It’s not training in X and Y and Z in school and whatnot. It’s not like that at all. It’s not like getting grades in school.

But you do absorb stuff from smart people around you. I’ve benefited from an enormous number of firebrand, intuitive people that are. But you’re typically not paying for yourself then. Then you go out and you try to do your own thing and you’re like, gosh, you know, it’s bracing in a way.

The American system, I think, is. I don’t know how to compare to Europe, even though I’ve had European grants, but I’ve never been plugged in the system here. We’re hard on each other here. And I’ve had people that have been on study review panels from Europe on American study review panels who say, wow, you guys are just. I thought you were slapping everybody on.

ANDREW HUBERMAN: The back and, oh, no, I sat on study section review panels for a lot of years. I was a regular member. And you go in there knowing you’re going to have to eliminate 70% of the grants that you read. So you sort of advocate for the ones that you really like, but you have to come up with reasons why you dislike things. And that’s an unfortunate consequence of not enough funding.

I do think that things are changing somewhat. And there are other sources of funding, fortunately, philanthropy, foundations. Yeah, it’s not for the weak of heart at all. And having a minimal sleep need definitely helps. I mean, I don’t know anyone that succeeds in science without working really hard.

I will also say, if you really want to get your ass kicked, just become a public facing person. But science made all of this feel much easier. Like there’s much easier. I mean, it’s different, right? But science, I mean, I don’t want to get into war stories about long hours because no one’s interested in that. But yeah, science is a thorough ass kicking with the occasional reward.

READ MONTAGUE: Plus it’s biology. Meaning even when you’re right on Tuesday, eventually it’s like, oh well.

Learning to Navigate Rewards and Losses

ANDREW HUBERMAN: But this gets us back to dopamine and rewards, which is one thing I will say is very, in my experience was very valuable about doing a PhD, about working in a lab doing biology experiments is that it teaches you to set up a reward expectation, motivation contingency loop that is based on everyday things and long term goals.

I mean, I think one of the features of being a healthy human is being able to like, oh, like hey, you know, this is a great cup of coffee this afternoon. But also register the serotonergic, like, ah, the experiment failed again. But then when things are working again, you can kind of feel like get some motivation from that and not just think about the PhD as the reward. Right.

So I go through life now not expecting great things to happen every day or even every week because I was trained in a system where the big rewards came every couple of years in terms of publishing papers, sometimes more frequently, but you know, it’s a long term thing.

But what about for the more typical example in people where, you know, you grow up and things are either really easy, really hard, or for most people it’s kind of a mix. Do you think that that’s part of us learning how to navigate life going forward? Like you got to register your wins in order to continue to have motivation. You also need to register your losses in order to not make the same stupid mistakes.

READ MONTAGUE: You have to sustain your losses. Right. And get up again. This is why I like sports for kids. Okay, so I’ve made all my kids do sports and one of them did competitive dance.

ANDREW HUBERMAN: So sports as a means to understand effort, reward, contingency and learning how to lose.

The Value of Sports and Learning to Lose

READ MONTAGUE: Even though you’ve brought everything you could do that day, the best you could possibly do. Yeah, somebody’s better than you, you know, what are you going to do now? You know that that is a template for a lot of lessons.

Same thing for students in science labs, especially mine. Students do things, they come, they show me something and I go, you know, and then they feel sad. But I watch them evolve. They evolve, you know, they all get better at it. And then they do this transition. You know, graduate students. I mean, maybe this is a little academic. Graduate students are completely worthless to you for a long time. And then in my experience. Well, in my world, they have so much to learn before they can do anything. That’s what I mean. No, I mean, as people, they’re valuable. They’re there.

ANDREW HUBERMAN: You mean in terms of data output?

READ MONTAGUE: Yeah.

ANDREW HUBERMAN: Well, they don’t know how to interpret the experience.

READ MONTAGUE: No, they don’t know how to do anything at first. No. They’re very valuable to have around. You want young people around. You want young, fresh people around doing things and thinking great thoughts. But then all of a sudden, they do this transition where they’re literally the most valuable person in the lab. And then six months later, they break your heart. They leave. They leave. Yeah.

ANDREW HUBERMAN: Just like you did to your advisors. That’s what I always say.

READ MONTAGUE: Oh, they were all glad to see me again.

ANDREW HUBERMAN: I was so blessed. I mean, all my students did. Great ones down at UT1’s University of Utah, they’re both kicking butt. Another graduate student in biotech. Another one’s on the job market now. And I’m just glad I’m not competing with any of them because I will tell you, they are phenomenal. I don’t take any credit for it. I did what I could with them. And then, you know, so a couple. One was a postdoc that I just.

READ MONTAGUE: Mentioned, but it was fun to be around to watch it.

ANDREW HUBERMAN: It’s just so cool. I mean, the energy of youth and, you know, and pouring into something with so much focus and not for the money, because Lord knows they don’t pay them very much, even as professors.

What you said about insisting that your kids play at least one sport, I think that also gets back to removing a problem we talked about earlier, which is, at least when you’re playing a sport, you can’t be on your phone.

READ MONTAGUE: Also, if you’re really, really tired, it’s hard to get in trouble. It’s very hard to get in trouble. If you’re a soccer player, you run. I mean, you are shot. At the end of the day, you just, hey, let’s go drive in. I don’t feel like it. So it’s a generic strategy I use.

And I just think sports, it’s not so that they can be champions. I mean, it’s great if that happens. It’s great for them. But they challenge you in ways that other elements of your life don’t. You know, I think of wrestling, and you get your hair cut off. And the main thing you learn when you’re a wrestler is how to manage your rising sense of panic. Don’t panic. Think about where you are. You’re not good at that at first.

I mean, when you go into an office and you face your boss or your co worker or something, nobody comes over and chokes you and says, now think.

ANDREW HUBERMAN: No.

READ MONTAGUE: Okay? That’s the only socially acceptable setting where that kind of thing teaches you. What are you going to do? Losing is such an amazing. Especially when you don’t want to lose and you did the absolute best thing.

ANDREW HUBERMAN: And in front of people.

READ MONTAGUE: Yeah, like track and field, you know, you run, there’s only one winner. You know, my kid’s been going to these giant meets, these sort of Mid Atlantic meets, and they’ll be. This is middle school. I mean, it’s amazing how well run they are. But there’ll be 40 teams there. There’ll be 60 kids in her event, you know, so if you get second, that’s really, really good. But if you want to win, there’s still this little thing that eats at you and you learn how to manage it.

So I couldn’t teach her that. Sport teachers are that. I see those as tests that we don’t get put to in the modern world. You know, it used to be different when we ran around in bands and we literally had to defend ourselves a lot. You may have to do something that requires awful things. You got to be ready to do an awful thing. We’re not put in that circumstance. A lot of modern ills come from, you know, we still have that brain. And civilization forces you to manage the stress in ways that it’s just kind of not designed to do.

Pain, Failure, and Persistence

ANDREW HUBERMAN: It’s funny because when I was growing up, the sport of choice for me was skateboarding. And there weren’t teams or anything like that. I mean, you could get sponsors and some of us did, but that’s not the point. But what I learned from it was pain. Pain, Pain. Pain.

READ MONTAGUE: Fail.

ANDREW HUBERMAN: Pain. Pain. Made it.

READ MONTAGUE: Those guys.

ANDREW HUBERMAN: Pain, pain. Like you just. And I wasn’t, you know, good enough to take a career into it. I had friends that were. And a lot of that is done in solitude. It was a great learning for science where I was alone in the lab. My graduate advisor wasn’t. She was available when I needed her, but I was the only one in the lab, so I worked alone. She said, don’t burn down, don’t kill yourself. Don’t drink the tetrodotoxin, you know, and actually had some pretty good lab accidents from working really long hours late at night.

But it was failure, failure, discomfort, failure, failure, got something. Failure, failure, you know, and it felt a lot like that. And I remember thinking skateboarding was great because as hard as this is, it’s not as hard as falling on concrete. Same thing when I tried to learn to snowboard, everyone was like, snowboarding is pretty tough. I was like, it’s snow. I was like, concrete hurts. Snow is soft. And even the ice pack is softer than. And I was like, you know, so it hurt. It took me some time to get good at it. But like, you’re like, okay, like, I get it.

So I think this actually is directly nested in everything we were talking about before, which is our expectation of whether or not our efforts are worth investing or not. Whether or not we update the keep going or quit depends a lot on how we interpret how many pain episodes or rewards we expect to get before succeed.

READ MONTAGUE: People have cognitive control. People can intend to do something and inhibit your natural instincts to avoid it or to quit or to back away from it. So the aversion signals that you would normally flee from have to be sustained. Right when you’re training to do anything like that. And it transfers when you’re older. It transfers older, yeah.

Trained people that are successful at athletics, people that just tried to do it. Which is why our little school, a little private school in Roanoke, Virginia, has a no cut policy. So I think there were 45 kids on the tennis team last year. It was almost unmanageable. But, you know, they’re all out there and they compete at the level they can compete at and they win and lose and it’s just a great. I can’t teach a kid a lesson that good.

And that’s training these same systems. It is expectations, disappointment, elation, recovery. Do it again. It’s all built in.

# Dopamine, Serotonin, and Time Perception

ANDREW HUBERMAN: I’d like to take a quick break and acknowledge one of our sponsors, Element. Element is an electrolyte drink that has everything you need and nothing you don’t. That means the electrolytes, sodium, magnesium and potassium, all in the correct ratios, but no sugar. Proper hydration is critical for brain and body function. Even a slight degree of dehydration can diminish your cognitive and physical performance. It’s also important that you get adequate electrolytes. The electrolytes, sodium, magnesium and potassium are vital for the functioning of all cells in your body, especially your neurons or your nerve cells.

Drinking Element makes it very easy to ensure that you’re getting adequate hydration and adequate electrolytes. My days tend to start really fast, meaning I have to jump right into work or right into exercise. So to make sure that I’m hydrated and I have sufficient electrolytes, when I first wake up in the morning, I drink 16 to 32 ounces of water with an element packet dissolved in it. I also drink element dissolved in water during any kind of physical exercise that I’m doing, especially on hot days when I’m sweating a lot and losing water and electrolytes.

Element has a bunch of great tasting flavors. In fact, I love them all. I love the watermelon, the raspberry, the citrus, and I really love the lemonade flavor. So if you’d like to try element, you can go to drinkelement.com/huberman to claim a free element sample pack with any purchase. Again, that’s DrinkElement.com/Huberman to claim a free sample pack.

Meditation and Breathing: Effects on Dopamine and Serotonin

I’m going to come back to your athletic career, but I want to ask about meditation and breathing. I think of meditation as some variant on close your eyes, focus on your internal state, direct your attention to your breathing or your forehead. I know there are walking meditations, open monitoring meditations, but I think most laboratory and most people, when they think meditating, they’re doing something like what I just described.

And I think of it as a perceptual exercise first, like you’re deliberately setting your perception internally, not externally. I understand there’s these insights into consciousness, improve sleep, reduce stress, but that’s all secondary and tertiary to me. What is the consequence of going into, of doing the practice of meditation, directing your state inward as opposed to outward, eyes closed, focusing inward, breathing in a controlled way. What does that do to dopamine and serotonin? Or what are dopamine and serotonin just doing when you go from a conversation that we’re having to a meditation?

READ MONTAGUE: So we’ve been doing experiments on this. My graduate student, Nishka Raheja, who went to the Ohio State University and then worked in a guy called Jeff Schoenbaum’s lab at the intramural program in Baltimore, that’s the NIDA National Institute of Drug Abuse, came highly recommended to me. And so she wants to study the neural basis of mindfulness meditation. She’s herself a meditator.

There’s a whole bunch of narrative that you put on top of this thing, and I’m a little bit of a feet on the ground, real simple mind, simple. So I said, I’ll do that, but we’re going to do all these breathing experiments first. And so she’s been in two settings. One is recording from the amygdala anterior cingulate cortex of the hippocampus while people are doing structured breathing. This is her instructing them to breathe. You know, inhale, one, two, three, four, hold, exhale.

Okay. I sit here and I’m not instructed to do anything. I’m just breathing, just free form, freeform breathing. So well, cycles with it. Norepinephrine and dopamine cycle of the breathing cycle. The most interesting on inhale, exhale, or.

ANDREW HUBERMAN: Overall, they change as you do this.

READ MONTAGUE: Well, I love that kind of granular detail. Is waiting on numbers. But the general gist now, this is from deep structures in the brain, is that easy breathing. You can just see the, it is like a metronome. The amplitude of the neurotransmitter fluctuations follows the inhale, exhale cycles. It’s right with it, it’s very easy. You feel like you’re watching the brainstem work.

Okay, so when I tell you, breathe in two, three, four, hold, you know, and now exhale. Okay. All hell doesn’t break loose, but it becomes hard for them to follow. And the transmitters are kind of wiggling and wobbling too. Okay, so that’s in people that are in the epilepsy monitoring unit. This is taking place at Phoenix at Benner Hospital with our colleague Robert Bina.

The most exciting stuff is using this probe that we can put up the nose of healthy people and do the same sort of thing. And we see generally the same sort of thing. The whole instructed breathing, you have to engage cognitive control over it, and you’re differentially adequate at doing that.

ANDREW HUBERMAN: So they’re doing structured breathing. Does dopamine track map onto the breathing?

Dopamine as Currency: The Ultimatum Game

READ MONTAGUE: Yes, but the interesting thing is we have people playing an economic exchange game. It’s called an ultimatum game. Ultimatum game should be labeled take it or leave it. I have $20. I’m going to offer you a split. Eight to you, 12 to me. Okay. Control passes to you. You’re going to either accept that, in which case we walk away with eight for you and 12 for me, or you’re going to reject it, in which case no one gets anything.

What do people do? Well, you know, they tend to, they see the inequity across the players as a signal, and at about 80, 20, you’re indifferent. In other words, 80% to me, 20% offered to you. 50% of the time you’re going to say, I’ll take it. 50% of the time, you’re going to send me a signal and say, you know, go home, I’m not taking that money. So nobody gets anything at a cost to you. When you reject, it’s at a cost to you.

The pattern that we see recording up the nose, the breathing is registered cleanly with the peroxide signal, which is a proxy for mitochondrial function, and the dopamine signal and the norepi signal. So as this pattern of exchange is going on between the two of us, if you’re going to update a model, in other words, if the signal across the two of us is such that you have to do some learning, it’s like your maximum breathing is tracking when you’re going to need oxygen in the mitochondria to produce ATP to make an update that’s going to be allowed by the dopamine signal.

It’s the most amazing looking data I’ve ever seen. And this is, in other words, your breathing and your dopamine signal in your nasal epithelium seems registered depending on the demands of the task and the elements of the task.

ANDREW HUBERMAN: Is this why you refer to dopamine as a currency?

READ MONTAGUE: Yes, I refer to it as a currency mainly for the reason that a currency is used. It’s a way to take dissimilar objects and assign a common value scheme to them. Like if I were going to trade cups for windshields, it’s easier. Unless we’re going to haul a bunch of windshields and cups to the trading site, it’s much easier to say, this is worth something in a currency I understand that you agree to and agree to some price.

ANDREW HUBERMAN: A lot of times I think it doesn’t matter if you’re talking about the US dollar, the euro or Bitcoin. Dopamine is the underlying currency. It doesn’t matter if you’re talking about wins in sport or other kind of more evolutionarily adaptive type examples like dopamine is the currency.

It actually can provide some method for resilience in my experience. You’ll notice in human dynamics online, because I spend a fair amount of time there, that people will try and rob people of their sort of message by taking potshots at them or something. And in science, you see legitimate critique and you see not illegitimate critique. Same thing in human dynamics that you observe online. There’s real lessons to be learned from some of the critique. But sometimes people are just trying to rob people of whatever impact they’re having.

And so you can think of one’s own dopamine, one’s own level of motivation, like are you going to let somebody rob you of this currency? We’re not aware that we’re using dopamine as currency, but ultimately the person who’s winning has more energy to go do more winning. People who are losing sometimes dissolve into a puddle of their own tears or worse. Other times they try and rob other people of their currency.

And this is the notion of zero sum versus non zero sum games. And when I step back now and I look at the media landscape, the political landscape, the social dynamics at large, I always think of dopamine as the currency. If we’re really honest about what’s happening in the world, it’s a battle over resources. And all those resources ratchet back down to this, this one single molecule. It’s really incredible.

READ MONTAGUE: Well, these same systems.

ANDREW HUBERMAN: Systems, yes, thank you. Systems, yeah, because they’re working as a coordinated system.

Dopamine and Mitochondrial Function

READ MONTAGUE: You know, dopamine turns on mitochondria.

ANDREW HUBERMAN: I mean, gives you life. It’s probably the cheat.

READ MONTAGUE: Literally. It literally turns on mitochondria. It binds to the outside of mitochondria to monoamine oxidase and it initiate, it gins up electron transport. I mean, it makes, it’s a signal to make ATP available. That’s a really direct connection.

Now what neuroscience hasn’t understood very well is the connection between the algorithms that the dopamine runs, the computations, and the combustion. So like if you touch your forehead, your forehead is merely warm, right? Now, if you touched a computer and a server center.

ANDREW HUBERMAN: If you could get.

READ MONTAGUE: Your finger near the chip, it would burn your finger. And if you turn the air conditioning off in a server center, within minutes they burst into flame. The whole thing would go up and smoke. They’re mainly generating entropy heat. Okay, there’s a big market play here to make chips that run and do the same computing, but on 40% of the power. I mean, service centers are amazingly, and just our computing machinery is amazingly inefficient. And so there’s great things to come as people take on this problem. But we don’t understand how it is that we get away to run our entire brain on 23 watts.

Sleep, Meditation, and Motivation

ANDREW HUBERMAN: Well, earlier when we were talking about sleep and we talked about meditation, I want to make sure that I at least offer you the opportunity to speculate. What do you think the kind of rejuvenative properties of sleep and meditation are? For instance, however little you need to sleep, if you don’t sleep for two days, you are a different beast altogether.

And sure, adenosine goes up and inflammatory markers go up. There are a lot of reasons for that. But motivation goes way down. Dopamine dynamics change completely, right? So what do you think allows us to replenish this currency, you know, in sleep? Like, what is it? Is it the slow breathing in meditation and sleep that allows.

READ MONTAGUE: Probably it’s a combination of physiological responses and the algorithmic cleaning up. It’s a computational device, at least. We see it is a computational device. That’s the modern metaphor for how we go in and understand it. And it has to erase stuff. You need a time off. You don’t need, you can’t have information streams processing through when you need to be going, I’m not going to save all that or I’m going to consolidate that.

And a lot of it’s about erasure and homeostasis and recovery. I mean, that translates physically into recycling transmitters and rebuilding a whole that.

ANDREW HUBERMAN: Because there’s nothing like the kind of motivation we feel after a great night’s sleep, the way we interpret events and all animals sleep.

READ MONTAGUE: It used to be thought the echidna didn’t have REM sleep, but that’s no longer. That’s false.

ANDREW HUBERMAN: It’s the first time the echidna has been mentioned on this podcast.

READ MONTAGUE: Okay, there we go.

Time Perception and Dopamine

ANDREW HUBERMAN: You may not want to go here, so feel free to say pass, but I’m very interested in the relationship between dopamine and other neuromodulators and time perception. Could we start with some general exploration of this? So really, like in the simplest way, if dopamine levels are artificially increased with a drug, what happens to time perception?

READ MONTAGUE: It changes. One of the things that’s latent in any description of what dopamine is doing, either from a point of view of psychology or algorithms that I focus on is timing. Okay, so to learn something is to suppress the statement. You learn something about what’s going to happen, when and how much, what, where, when and how. And so you have to have a lot of clocks in there.

Okay. As you well know, it used to be thought we had this one area, the super chiasmatic nucleus. It set daylight cycles, et cetera and all. And that was one of the main sources of clocks. Now what we know is every cell in your body has clocks in it and many multiple clocks. This is true for using these dopamine signals too. You not only have to have clocks, you have to be able to register the time that something was happening.

Now, I don’t know exactly how that’s done, but we just know that the system learns particular times and whatnot. And so those almost certainly have to be rejuvenated and reset. There’s a whole literature in rodents called the interval timing literature, where you teach an animal to anticipate something at a particular time in the future, in the near future, few seconds. And dopamine plays a critical role in that.

And there’s a group in London who’s, I forget the PI’s name, who uses that and uses manipulations in humans of dopaminergic drugs to look at time perception changes. But these are whole body. They’re like, what did you perceive?

Dopamine and Time Perception

ANDREW HUBERMAN: Is there a simple statement that we can make? If you increase dopamine pharmacologically and then you, does your perception of time shift to it moving faster or slower?

READ MONTAGUE: No, there’s no. I’m not saying no to that. I’m saying that there’s no clear, there’s no bromide for that.

ANDREW HUBERMAN: All we do know is that people who smoke cannabis oftentimes think a long period of time went by and they find out that a very short period of time.

READ MONTAGUE: Also people on methylphenidate, I mean people on Ritalin, will report that they lose time. Now, presumably they’re concentrating for longer periods of time. And it has the sound of multiple systems, not all of which you’re conscious of. But dopamine’s had a, I could start naming the people. I know this literature in rodents. Interval timing. And they’re beautiful relationships between the dopamine signaling and the timing. But it’s not sort of, they haven’t been experiments where you could open up the modern understanding of it as the key in reinforcement learning that goes on.

ANDREW HUBERMAN: Do you track time well on the order of a day? Like as you move through your day, are you aware of how much time has?

READ MONTAGUE: Terrible. I’m awful. I’m awful with directions and time.

ANDREW HUBERMAN: Can you orient well in space?

READ MONTAGUE: Well, I mean, if you tell me where the sun is.

ANDREW HUBERMAN: Interesting.

READ MONTAGUE: You mean my body? Yeah, like I know where my body is.

ANDREW HUBERMAN: Well, I know you kinesthetically. I mean, we’re a competitive athlete, so that makes it. No, I mean, I am stereo blind.

READ MONTAGUE: Oh, yeah. That’s so weird. Really. Ramachandran came to New York when I was in the audience.

ANDREW HUBERMAN: Your voice perception is messed up.

READ MONTAGUE: Well, apparently.

ANDREW HUBERMAN: But you were a javelin thrower.

READ MONTAGUE: I threw pitched baseballs until I was 15. Did you?

ANDREW HUBERMAN: You probably used one.

READ MONTAGUE: It’s the kinetic depth effect.

ANDREW HUBERMAN: Right.

READ MONTAGUE: Where you shake your head and it motion parallax.

ANDREW HUBERMAN: You move your head from side to side.

READ MONTAGUE: Maybe, but he, Ramachandran was a famous visual psychophysics guy from UCSD came and he was showing all these pictures and he goes, “Can anyone in here not see the thing?” I couldn’t see anything. He goes, “Oh, you’re stereo blind.” It’s about, I don’t know, 5% or something. You probably have a hard time with like barriers and you know, I was a hurdler in high school and college and then and throwing balls and I also pole vaulted and I was left handed until I was eight and my mother made me switch my hands rough. It’s a right handed world. You go into your classroom, you count the number of left handed desks and I came back, I said, there’s one. There you go, you’re going to use your right hand now. And so she was horrified when I cut my food and flipped the fork over in my left hand instead of switching hands and well, you got some.

ANDREW HUBERMAN: Brain plasticity out of it, no doubt. I asked if you track time well, because I think that this is totally, you know, just observation. I think that all the people that I know that are very driven who have more of a, I don’t like to use clinical terms, non clinically, but more of a obsessive nature than more ADHD like seem to not track time well and they’re able to just throw themselves into things and discard with thoughts about the rest of the world.

And I think about this a lot because of a generation of people who grew up constantly being bombarded with information from all over the world all day long. It just feels and sounds so noisy to me. And I saw an interesting article recently that the generation that grew up with social media and smartphones, there’s some interesting data that they may not track time the same way on the order of a day, but also in terms of their life arc and it makes it harder for them to envision long term goals. And I think it’s an interesting but still emerging literature. But it kind of makes sense if the dopamine system is involved in this and if it’s kind of mapped to very short term contingencies.

READ MONTAGUE: I grew up where there were, I was, I would wander forests for hours and hours and hours. There was no, and you weren’t monitored by your parents. It was, you know, you left in the morning and you were supposed to show up by dark or something. And people didn’t worry about their children. I mean maybe this wasn’t the right thing to do, but this is the way it went. If you wanted information, you had to go find it somewhere. And it’s so much calmer than the things that our kids are embedded in.

I think it’s different. I’m not sure it’s also pathological. The adults in the room all share a worry over it but we don’t know really what to do. I don’t think we do. And then, you know, these large language models, I don’t know how much you talk to them but you know, they speak 180 languages.

AI and Large Language Models

ANDREW HUBERMAN: I use Claude.

READ MONTAGUE: I love Claude.

ANDREW HUBERMAN: I love Claude AI, you know, I love the interface. I think the answer is it. I use it for research from time to time.

READ MONTAGUE: Do you ask it to summarize areas for you?

ANDREW HUBERMAN: I ask it to direct me to literatures, I guess I’ve asked it for some summaries here and there, but I’ve asked it to compare and contrast things, which is really cool because I can’t do that in PubMed. I can’t go into PubMed and say compare and contrast, Read Montague’s picture of dopamine to someone else’s. But Claude can do that. It can set up a five person panel around a topic in Claude. I use it more and more these days and I love it. I also really like the interface. It’s very clean and I care about aesthetics and I think it’s awesome.

READ MONTAGUE: The game I’ve been in isn’t the artificial intelligence end or even the neurobiology end. I’ve been at the interface of those two. So I’ve lived in a narrow space that shuttles stuff from one world into the other world. I mean, I’ve used algorithms to organize biological observations. Basically I’m the middleman in a way. I never thought this neural network training would scale the way it has. I just, I would never have guessed the way it does.

And I know there are the, you know, the dissenting voices, the doomsday people. Yeah, it doesn’t really do very well. It doesn’t. Well, I mean compared to who? Does anyone know anyone else that can speak 170 languages, that can translate 170 languages? I don’t know anyone that can translate accurately 170 languages.

ANDREW HUBERMAN: What do you use it for? Things like Claude. Do you use it as a kind of a search engine or?

READ MONTAGUE: Well, I ask it what’s the relationship between the subjunctive mood and the use of complex numbers and non relativistic quantum mechanics? I asked it that recently just for fun. In quantum mechanics, at least not quantum field theory, but in quantum mechanics the ways things might happen influence the probability of the way they actually turn out, okay? Whether or not you traverse that, you have to add up all those possibilities, right? It’s like a counterfactual. It’s like a mathematical rendering of a counterfactual. But it’s based on experiments people have done in the real world for 100 years.

And the subjunctive mood in conditional is the same sort of thing we discovered in language. Once we discovered how to speak language, we discovered how to make reference to the thing that would be if something else had happened. They’re sort of the same. I didn’t say anything, but just what’s the relationship between that and, it wrote this beautiful little essay, as it were. And I just thought, okay, I don’t really care whether it maps onto some notion of consciousness or smart or that’s, I don’t know anybody that could do that. I don’t know any person that could do that. And it’s a better writer than I am. I mean, maybe that’s me, but I’m just blown away by it.

And I’m even more blown away by the reinforcement learning guys. The David Silvers and the Go groups and AlphaFold, and, you know, they solved the protein folding problem. DeepMind, the company that was owned by Google, so I guess Google won 1, 2, 3, 4. They were three Nobel prizes or something. This time. The thing that happened that AlphaFold is the program that takes DNA sequences and predicts protein structure. And this is what Jumper and Hassabis got the Nobel Prize for. That’s a problem that the NIH has probably spent $100 billion on for the last 70 years. Okay.

They’ve also spent money on people crystallizing proteins and seeing where the atoms are and whatnot. And what they showed is they can develop a mapping between the sequence and the predicted protein structure, which is just, I mean, it was stunning to me. Now, it required all that crystallography data, but their general approach was treating it like a game, like they had treated where you do this reinforcement learning thing and you say you take a long sequence of moves and the game ends and you get an outcome, win or lose. And that’s enough to train up the best player that’s ever existed in history.

And then they used AlphaGo Zero to train up to be a grandmaster chess player back a few years ago. It took 30 or 40 days. And I think they’re down to, I mean, from scratch. So, I mean, if anybody’s going to write a history book on that, those are historical breakthroughs, really. And those algorithms are installed in our heads. Biology discovered that this is the way to handle the reality that whatever it is, given the constructs that are generated by our brains and keeps us alive.

That’s just the start, is what I think. In other words, the neural reinforcement learning world is going to continue to grow. It’s going to explode. We’re going to really start to understand that. We may even understand how to engineer it.

Personal Neurotransmitter Monitoring

ANDREW HUBERMAN: Let’s say somebody wants to get better at understanding where they’re at in the whole learning, motivation, reward, contingency, dopamine thing. They’re not going to drop a wire into their brain. They may or may not be able to participate in one of these experiments. But let’s say somebody wants to just reflect on where they are strong and where they are weak at the level of the algorithms they’re running. I’m not suggesting you necessarily have anything for them right now, but aside from telling them to go play a competitive.

READ MONTAGUE: Sport, I have a postdoc that’s making a company that’s going to commercialize these things up people’s noses when it goes from skunk works to kinder and gentler. And you could hack your own serotonin under your cell phone. You could put it up there and you could go do a thing and you could watch it on your cell phone. And we’ve never had anything like that before. Like, I wonder what happens when I do this. I know. I feel, you know, what happens when I solve a Scrabble puzzle or what happens when I, you could do it yourself. That’s his goal, is to take this company and put it into a commercial space where people could make personal use of it.

ANDREW HUBERMAN: Oh, my God. I can give this to this person who’s asking me about their dating life, and they can figure out how their dopamine reward expectation contingencies are running them.

READ MONTAGUE: It would be very interesting to sit and run scenarios through your mind and run them through again and ask yourself, ask whether you saw something like that going on with the mono, the signaling that’s available in your nose. That’s the kind of experiments we’re doing now. We have sentences playing out to people that have, as each word occurs, there’s a probability that there’s going to be a valence change in the sentence. And we’re looking at how it tracks this word by word. People playing social exchange games, thinking about themselves and others.

I can imagine there’s probably somebody out there that has even better ideas about how you could use it. I mean, I sit and work on the other end of it, but so I’m hoping he that’s going to hit the big time for him.

ANDREW HUBERMAN: Could you give me an example of something that you’re particularly excited about that would make one of your kids lives easier?

READ MONTAGUE: Learning how to concentrate. Like if I had a, like this company, this is Seth Batten. His company’s called Nebula Neuro. He, if he had a probe that we could put up there easily, like the little squishy things in your ears, then you could give him that and you could ask him to servo on their neurotransmitter release.

ANDREW HUBERMAN: So they would read a passage, you’re getting real time readout of dopamine and serotonin.

The Future of Neuroscience: AI and Neural Networks

READ MONTAGUE: And then you make a suggestion about how to learn something about it, pay attention to a component of it. Or you could do something as simple as lower this, lower this thing right here, lower this signal right here. We just haven’t had a way to measure that in real humans in settings that are like the real world.

So you take a thousand people and you say, oh look, these people are really comprehending in a way that we want them to comprehend. And these are in the middle and these are, wow, they’re way off beam here. And then you train a neural network who looks at the performance step by step with the transmitters there and it generates a picture of them.

That kind of thing is going to dominate neurobiology coming up. I mean, it’s changed whether or not people realize it or not. So many people are getting trained in it over here. But these are the important problems. These are the human behavior, human mind, human perception problems. That’s what you really want to get at, especially from mental illness and stuff like that. It’s not going to be a simple one thing.

So the fact that these neural networks have had a big breakthrough and how do we train them and how do we. And there’s still a ton of stuff we don’t know. These networks often learn things that the designers don’t know they know.

ANDREW HUBERMAN: And I think that scares a lot of people. But I think there’s excitement in it to be had in it also.

READ MONTAGUE: And they’re very convincing, they can make very convincing arguments and things like that. And so I just think letting it look at data, I can’t imagine a neuroscience experiment certainly on humans where you wouldn’t do that, where you wouldn’t shine these networks on that and feed them the data. So a lot of this is going to be how do you collect the data, how do you feed it to the networks and whatnot?

So I’m very excited about that. I’m excited about it because it was made fun of so much 30 years ago. Oh, reinforcement learning. Can’t learn anything.

ANDREW HUBERMAN: Everything in science was made fun of when. Now these really sound like two old guys talking about. But when I first started going to the annual neuroscience meeting, two things were the dregs. Like, no one attended those very few posters, which were AI and Brain Machine Interface. Those were considered really just the bottom of the pile.

READ MONTAGUE: Now it’s the hot thing. Yeah.

ANDREW HUBERMAN: And for a while, there was the molecular tools and genetic tools, and those are still awesome. But now AI and Brain Machine interface is all the rage.

READ MONTAGUE: We’re going to engineer our way into the brain now. We’re not going to just look for a pill.

ANDREW HUBERMAN: Well, look, the same thing is true. If I may, I’m editorializing here in the health space. Right. And so the same kind of what got knocked on meditation and magic carpet. Is it like a magic carpet ride? You know, mysticism, breath work, meditation.

Psychedelics are making a big comeback now. That needs to be approached with caution. Obviously can set off psychotic episodes, but it’s being looked at seriously, seriously, clinically. Peptide. The GLPs have made peptides super interesting.

I mean, basically, I’ve lived long enough in these spaces of science and health to say, whatever people are beating up on now, that’s going to be the next big thing. It’s just going to take a while, and you have to be discerning in how you go about it. But I think it’s wonderful that guys like Hinton and others kept hammering on this stuff when everyone thought it was, like, kind of backwater. Well, why would you. Why would you do this stuff? Why would you do neural networks?

READ MONTAGUE: Because they can’t learn. They can’t learn anything. And it was sort of true. I mean, they weren’t learning anything. Impressive. And then they transitioned to learning everything. I mean, you can ask it, what’s in that picture? And it’ll answer you. There’s a woman holding a puppy dog with a man dancing in the background. Looks like a painting from a Fellini movie.

ANDREW HUBERMAN: It’s awesome. I mean, it’s proof that whether or not you’re talking about fitness or sport or science, that if you love a certain area of something to just keep going, because eventually the world kind of aligns with you, and then it won’t. Right. Eventually it move on to something else.

READ MONTAGUE: He’s a psychologist, too.

ANDREW HUBERMAN: Yeah, it’s so cool.

READ MONTAGUE: His PhD is in psychology.

ANDREW HUBERMAN: Hinton.

READ MONTAGUE: Yep.

ANDREW HUBERMAN: Well, I’m glad to know that you’re excited I’m excited that you’re excited. And I’m also mostly an optimist about this stuff. I mean, I also think when we talked a lot about social media and reward contingencies and dopamine and stuff, but I also think that the human brain has adapted to conditions over and over and over again.

So this younger generation that we’re like, how could you spend all this time on your phone? We don’t want them to destroy themselves. On the other hand, they’re doing pretty well. Like there are examples of them doing spectacularly well, scrolling super fast and doing homework, playing sports, living their lives.

Q&A: Dopamine Hits and Neuroscience

So you willing to answer some questions from the general public? Yeah, some great questions here. Some of them you’ve already answered. But here’s one I think is worth asking. How much of what the public hears about, quote unquote, “dopamine hits” is neuroscience BS? Meaning it’s probably not real neuroscience. And how much has an evidence base when we hear this thing, dopamine hits.

READ MONTAGUE: Something unexpected and rewarding causes a dopamine fluctuation. And that’s true. Okay, but it’s an incomplete story.

ANDREW HUBERMAN: Do you think it’s an oversimplification to assign a serotonin hypothesis of depression and a dopamine hypothesis of schizophrenia? And if so, what other points would you add?

READ MONTAGUE: It’s a bit of a loaded question in that both of those chemicals are fluctuating in both of those disorders.

ANDREW HUBERMAN: So involved. But that’s not the complete story.

READ MONTAGUE: You know, it was the most conspicuous feature of schizophrenia is the fact that blocking dopamine receptors turns the symptoms down a little bit. We discovered that a long time ago. It was very early on seen as a hyper dopaminergic state. And it is that, I mean, it is that if you block dopamine receptors, you don’t hear voices anymore.

If you take L dopa and you don’t have Parkinson’s and you don’t have schizophrenia. I can find a dose where you will start to hear voices. I can find it as well. You’ll start to feel paranoid. I can make you schizophreniform. And so that’s a rational assignment of dopamine.

The features of schizophrenia, Schizophrenia is pretty ill defined and I think all these words are going to start getting teased apart. Now that we can record things in healthy people, that we can record things in sick people, and that we’re recording these transmitters in people that have these actual disorders.

Grit vs. Sunk Cost Fallacy

ANDREW HUBERMAN: This person is curious about the serotonin to dopamine ratio in quitting decisions at what point does the neurochemical drive to persist, what they’re thinking of as dopamine pursuit, become pathological against the valuation signal that says this isn’t working? In other words, what’s the line between grit and sunk cost fallacy? They want a lot answered in this one.

READ MONTAGUE: That’s a great question. I think it leaves out something that we really don’t know much about, which is for these things, these neurotransmitters, to be released, more or less, you have to set expectations. We know very little about how expectations for now are being set and being updated from state to state to state. And that controls the fluctuations as much as anything.

And so the representations in your brain of how they’re held or gotten from memory, how they control brain states and stuff, that’s not understood very well at all. That’s what AI is going to help us do in the next 20 years.

ANDREW HUBERMAN: I love that. I really appreciate your answer. And guess what, Grok. AI jumped in and answered as well. So we can see what Gro said.

READ MONTAGUE: It answered what? That question.

ANDREW HUBERMAN: Same question, yeah. So these people are asking questions on.

READ MONTAGUE: So he has Grock running over all the.

ANDREW HUBERMAN: Grock just jumped in and answered. This person didn’t say, oh, no, sorry. They tagged Gro. So Grock jumped in and answered.

READ MONTAGUE: I see, right.

ANDREW HUBERMAN: If you ask a question on action, you tag Gro. So I tell you what Gro said. “Great question. Research shows that dopamine drives persistence grit by reinforcing effort and reward anticipation. But high levels can trap us in sunk cost fallacies, ignoring when to quit. Serotonin helps balance by signaling outcome valuation. Low ratios may tip toward unhealthy persistence studies.” And they cite a study link dopamine surges to overvaluing sunk efforts. Worth exploring with the expert.

READ MONTAGUE: Big exclamation mark. You. That’s funny.

ANDREW HUBERMAN: Wild, right?

READ MONTAGUE: Yeah.

ANDREW HUBERMAN: How did Grok do?

READ MONTAGUE: Grok did. Well, if the brain is only a chemical machine. Grok left off the fact that it’s an electrochemical machine and that the electrical activity in the networks set things like expectations, which defines when the release is happening or not. And so that’s half the book.

Common Misconceptions About Dopamine and Serotonin

ANDREW HUBERMAN: What’s the one thing about dopamine the public seems to always misunderstand?

READ MONTAGUE: Dopamine equals pleasure.

ANDREW HUBERMAN: Is not true.

READ MONTAGUE: Is not true. Yeah.

ANDREW HUBERMAN: What’s the one thing about serotonin the public always seems to misunderstand?

READ MONTAGUE: You know, I take drugs to increase my serotonin when I’m depressed. What they don’t understand is that those drugs are really heterogeneous. Often pathological and you know, across decades toxic. So it’s an unfortunate. We’re in an unfortunate moment there to reconfigure those. That kind of treatment.

I bet you. Because I know people who swear by SSRI use, I mean just it transformed their life and they don’t. The side effects are nominal.

ANDREW HUBERMAN: It’d be great to be able to.

READ MONTAGUE: Identify them ahead of time. The candidates for whom that would work. You know, a lot of this is a placebo effect. Psychotropic meds, when you have a good outcome variable, 50 to 80% is a placebo effect, meaning not explained by any. You can’t explain the variance by anything.

But that doesn’t mean it’s a fake effect. Your expectations are set by that. Like if you believe something is going to happen. We have very poor understanding of really how your belief that something is happening, your brain actually marshals it.

Internal vs. External Rewards

ANDREW HUBERMAN: This question I am tempted to relate to meditation, but let’s see what you say. How do dopamine responses change when you remove external rewards and rely purely on internal satisfaction? So I think of an example like meditation, like where you’re not. Maybe you’re not telling people. I’m meditating to get praise, but just going into a state. Or maybe drawing because you like drawing, you’re never going to show your drawings. Or is there any idea of what happens when the.

READ MONTAGUE: No, there’s not. But it would be fantastic to measure that. That’s a fantastic question. To what degree can you, when cut off from the world, let’s say in a sensory deprivation tank, generate internal states that you chase and generate dopamine signals in that context. That’s why I mentioned these not yet published measurement schemes.

ANDREW HUBERMAN: Well, I’m telling Oliver from the UK no info, but he said yours is a superb question and guess what, that’s a lot of external validation. So it kind of runs countercurrent to the question, but. Right on, Oliver. I don’t know you, Oliver, but. Oh, it’s interesting. Many people are asking that same question. What has a greater influence on dopamine levels? Exogenous or feedback or our psychological framework? People are thinking about this a lot.

READ MONTAGUE: Well, they’re linked a little bit. My simple answer betrayed itself in a way. They’re linked. Your ability to generate a clear expectation and hold it in mind. I guess we’re talking about consciously here.

ANDREW HUBERMAN: Right.

READ MONTAGUE: It’s not really clear how good you are at that.

Creating Persistent Motivation

ANDREW HUBERMAN: There are a lot of questions asking about how to create a capacity for persistent motivation under conditions where things aren’t going well. You talked earlier about running up a hill and puking as a self training.

READ MONTAGUE: Friday nights after the football game. On Fridays, I just did it because I could.

ANDREW HUBERMAN: Till you actually vomited?

READ MONTAGUE: Oh, yeah. I mean, I don’t know. Snot and vomit. I would go until I couldn’t go anymore. I would just. I didn’t really lift weights in high school, but I would do. I copied the Russians. I had books, Russian books. Because Caucasian sprinters in the Olympics that won were only from the Eastern bloc countries.

Little did I know they were all these massive steroid campaigns, but they. They did a lot of plyometrics and weight box jumps and stuff. So I would put weight vest on and do it until I just threw up. I worked out this morning. I felt so nauseated, I thought, oh, my God. I don’t know.

ANDREW HUBERMAN: Have you ever done a. Like a mellow workout?

READ MONTAGUE: No.

ANDREW HUBERMAN: Your face says it all. We had a guy.

READ MONTAGUE: It’s my moment every day. That’s why I don’t like to work out with anybody. It’s my moment.

ANDREW HUBERMAN: Yeah, same. Unless I’m working out with Dorian Yates. Well, you know, because he’s—

READ MONTAGUE: He’s going to go, you know, when I am Deland. He was from Deland, Florida. Arthur Jones. I met him at a place in Sandy Springs, which is in Atlanta. And at the same place I used to work out with a guy called Isaac Hayes. You ever heard of Isaac Hayes?

ANDREW HUBERMAN: Sure.

READ MONTAGUE: Called Black Moses. And yeah, he would have this giant gold thing around and I was 12, and I could get in the Nautilus place and do a few things. They would let me in for any. And he was nicest guy to me, Isaac Hayes. He was just like—he just died not so long ago.

ANDREW HUBERMAN: Right. I recognize his name and I can see his face, shape, head. Yeah.

READ MONTAGUE: Yep. He wore dark shades and a giant gold chain around his face. So cool. Yeah.

ANDREW HUBERMAN: I mean, the high intensity work that Arthur Jones encouraged I think is the best way to stimulate hypertrophy and whatnot. The super setting and—

READ MONTAGUE: Oh, I’m sorry you had to carry a bucket there. Oh. Because people would throw up so much.

Serotonin Syndrome and SSRI Side Effects

ANDREW HUBERMAN: High rep, high intensity leg day. You can definitely puke. A lot of questions about serotonin syndrome. I get questions about this all the time. People who feel like because of SSRIs, they are dealing with sexual side effects, anhedonia, motivational issues. What do you think the cause is of all those things?

READ MONTAGUE: Those drugs are binding to all kinds of receptors is what’s happening. And there are all kinds of serotonin receptors. Okay. There’s not that many different dopamine. There are probably what, 80 serotonin receptors or something? There’s a great number of them. And so there’s just a field of dreams of way you can sort of have side effects.

Also, just the idea that you’re on them is itself an effect. I’m on a drug. This is a drug to manipulate my mood state. That has an effect on your mood state and the way you feel.

Closing Remarks

ANDREW HUBERMAN: Thank you for answering those questions. I want to say a couple things. First of all, thank you for taking time out of your very busy family and work and work out to puke often schedule. Despite the fact that you don’t sleep much, you are very busy. And it’s a really wonderful opportunity that so many people can learn about dopamine and serotonin and neuromodulator dynamics from somebody who really understands the science past, present and where it’s going.

These are topics that many, many people hear about and think about. And it’s super important that the conversation be up to date and nuanced. And you’ve done that for us today. I realize it’s far from complete, so we’ll have to have you back.

But also I just want to say thanks for being the pioneer that you’ve been in forging a path that, at least to my knowledge, no one else in neuroscience is tackling all the technical challenges, thinking about the AI and the computational stuff, putting people into scanners, putting wires up people’s nose, putting wires into people’s brains.

From the time I met you 15 years ago, it was very, very clear that you have a goal of solving the answers to particular questions and that you’re going to do whatever it takes to get those answers. And that’s just awesome. It’s the spirit of science at its best. And we’ll put links to your work and I know you’ve written some things and given some other talks, but I’m just so grateful. I learned a ton today and I know everyone else has. So you’ve done us all a tremendous service. So thank you.

READ MONTAGUE: Well, thanks for having me. It’s a blast.

ANDREW HUBERMAN: Thank you for joining me for today’s discussion with Dr. Read Montague. To learn more about his work, please see the links in the show note captions. If you’re learning from and or enjoying this podcast, please subscribe to our YouTube channel. That’s a terrific zero cost way to support us.

In addition, please follow the podcast by clicking the follow button on both Spotify and Apple and on both Spotify and Apple. You can leave us up to a five star review and you can now leave us comments at both Spotify and Apple. Please also check out the sponsors mentioned at the beginning and throughout today’s episode. That’s the best way to support this podcast.

If you have questions for me or comments about the podcast, or guests or topics that you’d like me to consider for the Huberman Lab podcast, please put those in the comments section on YouTube. I do read all the comments.

For those of you that haven’t heard, I have a new book coming out. It’s my very first book. It’s entitled “Protocols: An Operating Manual for the Human Body.” This is a book that I’ve been working on for more than five years and that’s based on more than 30 years of research and experience and it covers protocols for everything from sleep to exercise to stress control, protocols related to focus and motivation, and of course I provide the scientific substantiation for the protocols that are included.

The book is now available by presale at protocolsbook.com. There you can find links to various vendors. You can pick the one that you like best. Again, the book is called “Protocols: An Operating Manual for the Human Body.”

And if you’re not already following me on social media, I am Huberman Lab on all social media platforms. So that’s Instagram, X, Threads, Facebook and LinkedIn. And on all those platforms I discuss science and science related tools, some of which overlaps with the content of the Huberman Lab podcast, but much of which is distinct from the information on the Huberman Lab podcast. Again, it’s Huberman Lab on all social media platforms.

And if you haven’t already subscribed to our Neural Network newsletter, the Neural Network Newsletter is a zero cost monthly newsletter that includes podcast summaries as well as what we call protocols in the form of one to three page PDFs that cover everything from how to optimize your sleep, how to optimize dopamine, deliberate cold exposure. We have a foundational fitness protocol that covers cardiovascular training and resistance training.

All of that is available completely zero cost. You simply go to hubermanlab.com, go to the menu tab in the top right corner, scroll down to newsletter and enter your email. And I should emphasize that we do not share your email with anybody.

Thank you once again for joining me for today’s discussion with Dr. Read Montague. And last but certainly not least, thank you for your interest in science.

Related Posts

• Transcript: Maya Angelou’s Greatest Life Lessons on Oprah Podcast
• Transcript: The Next 50 Years: Humanity, AI, Power w/ Yuval Noah Harari
• How to Stand Out in an AI World by Thinking Better, Not Faster: Sol Rashidi (Transcript)
• Modern Wisdom: w/ Alex Petkas on True Story Of The Odyssey (Transcript)
• Reconstruction: Malcolm Gladwell Interviews Barack Obama (Transcript)