Making Sense with Sam Harris - #247 — Constructing Minds

Episode Date: April 22, 2021

Sam Harris speaks with Lisa Feldman Barrett about the origins and function of the human brain. They discuss how brains evolved, the myth of the “triune brain,” the brain’s network organization, the predictive nature of perception and action, the construction of emotion, concepts as prescriptions for action, culture as an operating system, and other topics. If the Making Sense podcast logo in your player is BLACK, you can SUBSCRIBE to gain access to all full-length episodes at samharris.org/subscribe.

Thank you. of the Making Sense podcast, you'll need to subscribe at samharris.org. There you'll find our private RSS feed to add to your favorite podcatcher, along with other subscriber-only content. We don't run ads on the podcast, and therefore it's made possible entirely through the support of our subscribers. So if you enjoy what we're doing here, please consider becoming one.

As always, I never want money to be the reason why someone can't get access to the podcast. So if you can't afford a subscription, there's an option at SamHarris.org to request a free account. And we grant 100% of those requests. No questions asked. Okay, some housekeeping today. I have a new podcast to announce, a single episode, which we will be dropping, I believe, Friday of this week, if all goes according to plan. So look for it in your feed on the 23rd of April. the title of this episode is Engineering the Apocalypse. And it was produced by my friend Rob Reed, who is a podcaster and author, also a tech entrepreneur.

I met Rob at the TED conference some years ago, and then he started his own podcast, the After On podcast. And he interviewed me, I think, for the first episode there. I thought it was probably the best interview anyone had ever done of me, so we aired that here on Making Sense. I believe we titled it the After On interview. Anyway, in the intervening years, Rob has gotten very interested in existential risk, and in particular the risk posed by advances in synthetic biology, which could very well lead to an engineered pandemic. But everything he says in this podcast is relevant to a naturally occurring pandemic, like the one we are currently suffering. Anyway, this is a deeply researched and,

by turns, harrowing and hopeful look at advances in synthetic biology. And it's broken into four chapters, which are separated by interstitial conversations that I have with Rob. Anyway, I thought the job he did was fantastic. Pandemic preparedness has to be a huge priority for us going forward. And this is our best effort to argue that it really must be. COVID has been a dress rehearsal for something far worse. And as such, it has been pretty much an unmitigated disaster. We may have lost sight of this, given how successful our vaccine production has been and how the rollout has ramped up, but our response to COVID, in particular our failure to organize a globally coherent response, was just a terrifying failure.

Terrifying given how much worse a pandemic can be, and how much worse it's likely to be if it's ever consciously engineered. So anyway, this upcoming podcast will be dropped as a single episode that's nearly four hours in length, and again the title is Engineering the Apocalypse, and needless to say we'll be releasing that as yet another PSA, which is to say the whole thing will be freely available. But of course, if you find this work valuable, the way to support it is to subscribe at SamHarris.org. And to coincide with the release of this podcast, the Waking Up Foundation will be giving two significant grants to relevant organizations that are working on the front lines of pandemic preparedness.

As many of you know from my conversations with the philosopher Will McCaskill, I've been thinking more about how to effectively do some good in the world, in addition to just talking about what is good to do. So we formed the Waking Up Foundation for that purpose. And at least 10% of the corporate profits of Waking Up go there, as does a minimum of 10% of my own income. And the foundation works as a pass-through to other organizations. So 100% of the funds leave it and go elsewhere. And so these next donations are focused on this problem of pandemic preparedness. And in this vein, we're supporting the Center for Communicable Disease Dynamics at Harvard University, which focuses on improving our methods of understanding the data around infectious disease. And it engages

policymakers to improve their decision-making, which often leaves a lot to be desired. And the second organization is the Coalition for Epidemic Preparedness Innovations, the CEPI, whose mission is to accelerate the development of vaccine technology. They're funding new platforms so that we can develop vaccines even more quickly than we did for COVID, and really do it just in time in response to a novel pathogen, which is precisely what we're likely to face in the case of a synthetically engineered pandemic. Now, neither of these organizations are set up to take small individual donations, but if you're a philanthropist and you want to come along with us in helping to improve our pandemic preparedness, I would certainly encourage you to support these organizations.

Once again, that's the Center for Communicable Disease Dynamics at Harvard University and the Coalition for Epidemic Preparedness Innovations. And I should say that the Waking Up Foundation is getting great advice on this front from Natalie Cargill of Longview Philanthropy. This is an organization that advises individuals and foundations who want to deploy significant funds to solve long-term problems. And I was introduced to Natalie through Will McCaskill, and I've been extremely impressed with the research that they've done at Longview and the clarity of their advice, all of which is given free of charge. Longview is independently funded. So if you're running a

foundation or you're a wealthy person who wants free advice about how to give most effectively, I highly recommend that you get in touch with the people at longview.org. Again, this is not a recommendation for small donors. I believe you need to be giving away at least a million dollars a year before Longview can help guide you. But for those of you who are in the philanthropy space, I recommend you get in touch. But if you are an individual donor and you want to ride along with me, we will be detailing all the orgs we

support at the Waking Up Foundation once that website is launched. And on that point, I want to say that the Making Sense audience has been fantastically generous in the past. On the occasions where I've discussed specific non-profits on this podcast, the people who run them always come back astonished at the result. To give you just a couple of snapshots here, GiveWell.org reached out recently to say that just by my mentioning their organization a few times on this podcast, this is the group that does exhaustive research on the effectiveness of charities and recommends what they consider to be the most effective ones in several categories,

my discussing their work a few times, once with Will McCaskill, their work a few times, once with Will McCaskill, resulted in you guys donating $1.8 million through them directly and pledging another $1.8 million in recurring donations. So that's $3.6 million through the end of this year. And Will McCaskill's organization, Giving What We Can, And Will McCaskill's organization, Giving What We Can, which was started by Toby Ord, who's also been on the podcast, has told me that in response to my discussing their pledge, this is the pledge to give a minimum of 10% of one's lifetime earnings to the most effective charities, which you can do at any level, whether you're making $30,000 a year or $30 billion. level, whether you're making $30,000 a year or $30 billion, I'm told that my discussing this pledge with Will caused hundreds of you to take this pledge yourselves. And after Waking Up became the first company to take the pledge, 10 more companies soon followed. Now, I don't know how

much money to the most effective charities this represents, but it's surely many, many millions of dollars. I believe giving what we can just passed the two billion dollar mark in lifetime earnings that have been pledged. Anyway, my point in mentioning this isn't to brag about the influence of this podcast, but rather to convey my gratitude and astonishment, frankly. I mean, it's just amazing to see the knock-on effects of discussing these things. Anyway, I will keep you all informed about this, but this is just to let you know that over at Waking Up and here at Making Sense, and here at Making Sense, we have transitioned into doing more than just talk about specific problems. We're marshalling our own resources to try to do some good directly ourselves. Okay, today I'm speaking with Lisa Feldman Barrett, who is one of the most cited scientists in the world for her research in psychology and neuroscience.

She's a professor at Northeastern University with appointments at Mass General Hospital and Harvard Medical School. Lisa was awarded a Guggenheim Fellowship in Neuroscience in 2019, and she's a member of the American Academy of Arts and Sciences and the Royal Society of Canada. And she's the author, most recently, of a very enjoyable book, Seven and a Half Lessons About the Brain. And we cover a few of those lessons in today's podcast. We talk about how the human brain evolved, the myth of the triune brain, which has been all too influential.

We discuss how the brain is organized into networks, the predictive nature of perception and action, the construction of emotion, concepts as prescriptions for action, culture as an operating system, and many other topics. And now, without further delay, I bring you Lisa Feldman Barrett. I am here with Lisa Feldman Barrett.

Lisa, thanks for joining me. It's my pleasure. So you've written this wonderful little primer on the brain, seven and a half lessons about the brain, which I think will be the focus of our discussion, although we'll probably wander to other topics. But I just want our listeners to know that this is a marvelously accessible book and a short one.

It's only 130 pages or so. And we need more of this kind of thing. There's this kind of awful property of the brain and neuroscience generally, which is when you get into the details, it becomes just a catalog of anatomical names that are certainly not written by writers, especially ones who wanted to write books for a general audience. And it becomes this blizzard of mnemonic challenges for a reader. And you've managed to avoid all of that and still deliver a very interesting discussion about the brain and the mind. So congratulations. Thank you so much. So before we jump in, perhaps you can summarize your background intellectually. What kinds of questions have you focused on as a scientist?

Well, I started my training as a clinical psychologist and then very quickly went through a series of retrainings in physiology and then in neuroscience and more recently in engineering, learning something about systems theory and in evolutionary and developmental aspects of neuroscience. So the questions I really think about now relate to, you know, how is the brain, how is your brain in constant conversation with your body and the other brains and bodies, you know, that surround you? How is it conjuring the features of your mind? How does it control your, the internal systems of your body at the same time as it's controlling your behavior and giving you memories and thoughts and feelings and so on?

That may sound like too big of a question to answer, but I would say I'm really interested in understanding a systems level kind of approach to brain function. And that encompasses a lot of things. So I have a large, a large lab and we have a lot of different research projects going on. So it's really hard when someone asks me, so what are you what is your newest research project? And I'm like, well, we have like probably 40 of them going on. So it's hard to it's hard to summarize in one sentence. And you're currently a professor as well, right? So do you spend some time teaching or is it all research at the moment?

I mean, I know we're talking in COVID land or at the tail end, one hopes of COVID, the COVID pandemic. So nothing seems normal, but what is your general life like as a professor? normal, but what is your general life like as a professor? Yeah, so I run a lab which has 25 full-time people in it. And then usually we have, not during COVID, but usually at other times, we have about 100, 150 undergraduate researchers in the laboratory in any given year. And the lab is spread out across two different places. So I have personnel at two different places, graduate students, postdocs, and so on, postdoctoral fellows. I teach one course a year for undergraduates. It's a lab course. And then occasionally, I will also teach, formally teach graduate seminars, but I also run a weekly or now biweekly seminar that I've been running for, I guess, about eight or nine years that I don't get any credit for. We just do it out of the love of doing it with engineers and computer scientists and other neuroscientists and psychologists.

And so I and another and my colleague in engineering, we run this seminar for all of our peeps. So it's about 25 people who attend this seminar. And it's been going on, like I said, for for quite a number of years. And then I also run other reading groups that people attend on particular topics, depending on what we're interested in. So for example, on predictive processing or on energetics, which is a word that we use to refer to brain metabolism and the way that the brain is regulating the metabolic functions of the body. So one of the things you do throughout this book, especially at the outset, is debunk a few myths and bad metaphors we've relied on to understand the brain or seem to understand

the brain. And this seems like a very useful thing to do. Perhaps we should just start where you start with the larger context of evolution and what we think we understand about the evolution of the human brain. And perhaps this is a good place to part company with Paul McLean. So how do you think about the brain in evolutionary terms? I love this question. I think this is one of the most fun questions, really. It occurred to me at one point, why do we even have a brain? It's a really expensive organ, right? That three-pound blob of meat between your ears costs you about 20% of your entire metabolic budget. So it's pretty expensive. I'll just point out, depending on what you do with it, it can cost you much more than that. It certainly can. Especially on social media.

It certainly can. That's absolutely right. And so I'm very fortunate in that I've been meeting really weekly with Barbara Finley, who is an evolutionary and developmental neuroscientist. And she's basically, you know, to use her words, she's like downloading all of her knowledge into my brain, which really means that she repeats herself frequently and has to explain things often more than one time. And this is pretty, pretty, you know, not to make a bad pun, but like pretty heavy stuff. It's pretty complicated. You know, I had to learn embryology and I, you know, barely understand what I'm reading, but I understand a little bit now at least.

But the really cool thing I think is that if you go back, you know, 550 million years ago to a time in the Earth's history called the Edicarian, animals didn't have brains. And so I was just really interested to try to understand, well, why, you know, why did brains evolve? And Sam, you know, you know, you can never really answer the why question very easily in evolution, but you certainly can answer what questions. So like, what is the brain's most important job? What is a brain really good for? And you can look at the evolutionary story that molecular geneticists and anatomists and so on, ecologists have crafted. And it's a really cool and interesting drama. And what it suggests is that your brain's most important job isn't thinking or seeing or even feeling. So these are

characteristics. These are features that the brain performs or computes, but they're not actually the brain's most important job. Its most important job is regulating the systems of your body, your heart, your lungs, your immune system, your, you know, endocrine system and so on. And of course, you know, we don't experience every delight and or, you know, every drama in our lives this way. We don't experience every hug that we get or used to get before COVID or every insult that we bear. We don't we don't experience things this way, but this is actually what is going on under the hood. what is going on under the hood. And when your brain thinks and decides and sees and hears and feels, it's doing this in the service of the regulation of your body. And that turns out to be a really important insight. I would add one piece here. I know you, I don't recall if you put it this way in your book, but it does strike me that

just by the logic of evolution, the motor behavior is in some ways primary here, because if you can't move, if you can't do anything with a brain, if there's no way that it can influence the differential success of an organism in the contest for mates or survival, then there would have been no evolutionary pressure in this direction. So it seems to presuppose an ability to do something with respect to the environment. I don't think there's a bright line between that story and the story of regulating the internal states of the body. I think we'll get to that. But don't you see an ability to actually act in some way as being the necessary context for this evolutionary pressure? Absolutely. In fact, really, you know, I guess I'm very persuaded by

work in motor neuroscience and certainly in philosophy, the idea that motor action is primary and all sensory processing is in the service of motor action. I think that's absolutely right. The one thing I would say though, is that, you know, in vertebrates, in all vertebrates, certainly, and in, I would maybe hazard to say, all animals who have limbs that move or parts that move, there's usually an internal set of systems that support that movement. Now, invertebrates, you know, like us, that's, a cardiovascular system and a respiratory system and so on. You know, not all animals have the kind of viscera that we have, that vertebrates have.

So invertebrates, you know, have their own systems. But there is no external movement of bodies without internal systems to support that. And in motor neuroscience, as much as I respect that work, and I really do, I think they're really ahead of the curve in certain ways. They tend to ignore the internal systems of animals' bodies. And I really think that that's an important part of the story that is missing. So when I say, you know, that the brain is regulating the body, I really mean everything motor about the body. That would include what we call visceral motor, which means the beating of your heart and the, you know, contraction of your lungs and so on. But it also means the movement of your skeletal

motor system, your muscles, the voluntary movements of your muscles. And in fact, your muscles, the voluntary movements of your muscles. And in fact, if you look at, for example, primary motor cortex in a monkey brain and macaque brain, it has visceral motor maps in it. And some of the regions that are considered to be, you know, sort of association regions for the motor system are actually the primary cortical controllers of visceral motor regulation, meaning regulation of the viscera of your lungs and your heart and so on. So in your brain, the internal systems of your body, the neurons that are controlling the internal systems of your body and the neurons that are controlling your skeletal motor system, your voluntary muscle movements, are really intertwined.

That's not well documented in motor neuroscience work, but it's present in the anatomy. You can just see it. It's there. Yeah, but we'll talk about emotion, but I tend to think about emotion now as a kind of covert behavior, right? So that the line between emotion and action that is commonsensical, I think, can break down if you follow that framing. But let's not leap to emotion just yet. The evolutionary story we have told ourselves for a long time has been summarized by this concept given to us by Paul McLean of the triune brain. And so people refer to their

lizard brain, or they think of a stepwise evolution from reptiles to mammals generally, and then to primates as having kind of climbed up from the brainstem to the cortex. What's wrong with this picture? Well, what's wrong with that picture is that it doesn't really match the best available scientific evidence for how brains evolved. I mean, if you look at a lizard brain and say a mammal brain, like say a rat or like a rodent brain, say, and you look at a monkey brain and a human brain, you know, they look different to the naked eye. It looks like the rat, or I should say, it looks like the lizard doesn't really have much of a cerebral cortex. It looks like the rat has, you know, maybe a little bit of kind of old

cortex, and that the monkey and the human have quite a bit, and the human having, you know, substantially more than the monkey. That's how it looks to the naked eye. And this, you know, led Paul McLean and others, you know, guided by, I think, certain cultural beliefs to describe brain evolution in much the way that you just described it. Although your description, Sam, is slightly more lyrical than maybe what McLean wrote. But, you know, the idea that a lizard brain is mostly has parts for instincts, you know, like freezing and fighting and fleeing and copulating, which, you know, neuroscientists make a funny joke, you know, like they refer to it as the four Fs. So that's neuroscience humor for you. And then layered on top of that evolved what's called a limbic system, limbic meaning border, bordering this,

you know, these lizard parts for emotion. And then what evolved on top of that is the cerebral cortex or the neocortex, the new part of the cortex, which you only see in what are referred to as higher mammals, you know, like us. And the idea is that, you know, your lizard brain contains your instincts, your limbic system contains your emotions. And then these are these make up your inner beast. And they are constantly in battle with the more rational side of yourself, which resides in your cerebral cortex. So your brain is a battleground between your inner beast and your rational self for control of your behavior. And the idea is that, you know, when your cortex wins and you behave rationally, you're a moral person and you're healthy. And if your inner beast wins to control your behavior, then you're either immoral because you didn't try hard enough, or you're sick, because it didn't work, you know, that there's something wrong with your rational cortex.

And the problem with this, even though it makes a lot of sense in terms of the stories that we tell ourselves about what it means to be moral and responsible for behavior, and it's very consistent with Western views of the self. The problem is that it doesn't actually match the evidence that when you peer into neurons, and you look at their molecular structure, in particular, you know, the, the genes that guide the formation and function of, of those neurons, you see a really, really different story. And the story is that really all mammals whose brains have ever been studied, actually their brains follow the same developmental plan. Their neurons actually, there are no new neurons, really no new neuron types. And remarkably,

the stages of development, and I'm talking about, you know, embryological development forward, the stages of development in all of these mammal brains that have been studied, different species, proceeds in exactly the same order, pretty much. What changes is the duration of each stage. And there's this really interesting observation that George Streeter, the neurobiologist, made about brains in his book on brain evolution, by the way, excellent book, if anyone wants a primer on brain evolution, it's a really fantastic book. He says, brains reorganize as they grow larger and so it can look like there are new structures there just because there are more of certain neuron types but actually the you know there's nothing new in terms of the neurons it's just there they look like they're

reorganized and they look like there are miraculously new parts there but there are really no new parts it's just that certain types of neurons have certain stages in development have gone on for longer. And so there are certain types of neurons, there's just more of them. And if you go back even further and you look at other animals, other vertebrates, you see that many of them have also really striking similarities to mammalian brains. So for example, birds don't have a cerebral cortex, but they certainly have neurons that are the same as the neurons that make up our cerebral cortex and that seem to perform some

very similar functions to what the various functions our cortex performs. So basically, there is no lizard brain. You don't have an ancient beast lurking inside your brain, and the only animal who has a lizard brain is a lizard. Are there any exceptions to this? I had thought that von Economo neurons were an exception, that von Economo neurons were an exception, that they were present in great apes and I think cetaceans and elephants and a few other charismatic vertebrates, but were not found in reptiles or birds? So von Economo neurons are very contentious. I mean, there are some anatomists who will tell you that von Economo neurons are not a special class of neurons. They're just really big honking

pyramidal cells. So, you know, you find them in large brained animals because, you know, as brains get bigger, sometimes the neurons also get bigger. And, you know, one thing that's happened, for example, in large brain animals, what often happens is that there are certain parts of the cortex in particular, that as they grow, what happened, you know, evolutionarily, but also in development, what happens is not that they develop more neurons, but they develop fewer neurons that get much bigger, and they have much more connectivity. And the reason for that is, I don't know the reason for it, but the functional consequence of

that is that, which is something I explained in essay seven, which is that it means that the animal's brain can summarize information much more efficiently and maybe even do some abstraction, meaning can find similarities in things that look and feel and smell and taste different, find functional similarity. So this is abstraction. This is what we call abstraction, right? And that's really, you know, maybe what these very large pyramidal neurons are for. But there are some anatomists and some neuroscientists who look at von Economo neurons and say, well, these are just ordinary big, you know, neurons. They're not, there's nothing really special about them. And you find them in animals who have large brains relative to their body size. Right, neurons. There's nothing really special about them. And you find them in animals who have

large brains relative to their body size. Right. Right. So what is the appropriate picture of the structure of what we have in there if it's not this cartoon of descent from reptiles? cartoon of descent from reptiles. What picture of complexity and now leading the witness network complexity should we have in our heads? Yeah, I think I'm going to ask you a question, but I just want to take one step back for a minute and say that we live in a world where we see objects and we see boundaries between objects. And, you know, like here's a book, here's a purse, here's a computer, here's a glass, whatever. And so we have a tendency to think about things in terms of objects instead of in terms of relationships between features. relationships between features. And so for a really long time, people have thought about the brain as having these distinct parts, you know, like there's this group of neurons called the

amygdala, which performs emotion. And there's this other group, you know, called the basal ganglia, which performs, you know, movement. And then there's this other part called the cerebral cortex. And the prefrontal part of that really performs decision-making or rationality or what have you. And that's just, I mean, there are people who still hold to that view and, and it's certainly people have built their whole careers on such notions and, and been very successful, but I think there's also a growing understanding that that's really not how the brain works. It's not how the brain is structured. There are no objects, you know, there are no kind of mental organs in your brain.

That's just not really the way, that's just not really the best way to understand the anatomy or the function. And that instead we should be understanding neurons in terms of their relationships to one another and the features that they compute. And so they're really that this can take many forms in published papers on neuroscience, but one that's very popular at the so if you think about, you know, instead of thinking about neural signals as being passed from one, you know, region to the other, like a baton in a race, you can think about neural activity and the patterns that are created more like weather patterns or something where, you know, many, many, many neurons are participating in computing an event that has a set of features. And some of those features

are, you know, very close to the data that you get from your sensory surfaces, like your retina and your cochlea and all the sensory, all the sensors inside your body. So, you know, like a line, for example, or color, like the color red, your experience of the color red is a feature that your brain computes. It doesn't detect, as you know, and it's computing it using information from not one color detector, you know, like so-called cones. And, you know, you have three, you have cones in these cells in your retina that register three different

ranges of wavelengths of light. And you need all three to see red or green or any color. And so your brain computes these features. And it also computes features like seeing a face. It computes features like threat. It computes features like novelty. It computes features, all kinds of features. And in a given event, your brain is sort of computing sequences of events. And in computing an event, what it's doing is computing features in the service of regulating the body, regulating action and all the visceral changes that will support that action. And so the way to think about it is your brain is a single structure with 128 billion neurons, give or take, and it can take on trillions of patterns. or take. And it can take on trillions of patterns. And these patterns are, you know, helped along by the chemical bath that surrounds these neurons. So your neurons are bathed in a chemical system.

And your brain is basically dynamically along a trajectory from one pattern to another pattern, to another pattern, to another pattern to another pattern to another pattern, and trying to understand what launches those patterns, what maintains those patterns, what features your brain is computing. That's really the goal of understanding brain function. Yeah, I would also just point out that the methods we use to understand brain function, like increasingly functional neuroimaging, can also give a false picture of the modularity of the brain and therefore the mind. Because just by the nature of the tool that we look at the data in terms of these pretty pictures of certain regions of the

brain, so-called lighting up in response to stimuli or tasks. And it can give a sense, you know, not to actual neuroscientists generally, but perhaps in a more subtle way, it can even corrupt their thinking. But it certainly can give a sense to the general public that this is a question of other areas of the brain actually not doing anything when they're not part of the illuminated map of what is most active during a certain function. So it can just give this false picture of separate organs in the brain that are, albeit connected, are really independently responsible for an emotion like disgust, say, or a certain kind of perceptual task. And you just can't visualize the network behavior and the fluctuating network behavior and the

weighting between nodes in the network as easily as you can just aggregate the data by subtracting two states of the brain and showing one where these regions were more active than in the other? Yes and no. I think I mostly agree with you, but I would probably just push back maybe a little bit on a couple of points. One, I would say it's not the fault of brain imaging techniques. It's really the fault of the analysis techniques that we use and the sample sizes we have. So I would say that with fMRI, fMRI has its problems for sure. It has limitations in terms of its temporal resolution and also even some spatial resolution issues. But really, it has much more to do with the kinds of designs that scientists use and the kinds of analytic techniques that they use. And I'll give you a

really good example. There's what I think of as a really brilliant paper that was published in the Proceedings of the National Academy in 2012. The first author is Gonzales Castillo. And it's this really nice paper where they, you know, compare the sort of standard, you know, experimental design really for a very, very simple task, which is, I believe it was a visual, visual perception task, maybe visual orientation, I think it was, but very, very straightforward task, so visual attention task. And when you run, you know, some subjects, and you have maybe, you know, 40, 50 to 100 trials,

where a trial is, you know, you show something unexpected to the subject and then they, you know, they have to make a judgment of whether, you know, lines are pointing in the left direction or the right direction or what have you. What you see in the way the analysis is done, the way that choices, analytic choices are made to separate signal from noise and so on, are made to separate signal from noise and so on, you see a couple of islands of increase in activity that are depicted on a brain image as spots that light up, like the light bright sort of brain. And it's important to really understand here that these images that we see in magazines and in journal articles and so on are curated by

scientists. They don't just pop out of the data on their own. They're made contingent, these images are contingent on a bunch of analytic decisions that are made. Now, if you expect that there are islands of activity because different parts of your brain are responsible for different specific psychological functions. And that's what you expect. And you've designed your study that way. And you've only, you know, tested your subjects on 50 to 100 trials and you threshold, that is, you make decisions about signal versus noise in particular ways. what you get are a couple of islands of activity. However, what this paper showed is that if you run 400 trials for each subject, so you bring them back for multiple scanning sessions, and you analyze the data in a slightly different way by,

by, instead of assuming that every part of the brain has, that the shape of the response is the same, and instead of assuming that, you model, you know, the variability in how the different parts are responding. What you see is that 85% of the brain shows an increase in activity. that 85% of the brain shows an increase in activity. That means 85% of the brain is showing a change to make a very, very simple decision that is considered. Yeah, so the point is that if your studies are designed in a way that is underpowered, you're not going to realize that you're making what we would call a type two error,

which is that you're missing a lot of important activity that's there. Because, you know, you're expecting to see blobs, and what you get are blobs. And so, you know, if what you expect is islands of activity, you'll perform your studies, you know, with something I used to call blobology, which is that, you know, you'll identify these blobs of activity. I think people have to realize that these images are really curated by humans who have a set of assumptions. I'll just give you one other really quick example. And that is, you know, when people started looking at networks in the brain, so this is regions that are, have correlated where the brain response is correlated. So you take a brain and you divide it up into lots of little cubes called voxels. And so you

look for sets of voxels that have a similar change in blood flow during an experiment. And you call that a network. And it turns out, you know, this actually does reveal something about the underlying structure of the brain. But when you look at the way that scientists mostly study these networks, they look like Lego blocks, like they're completely unrelated to each other and like, you know, pieces of a puzzle, and you put them all together, and you get a brain. But you know, that's a decision, those are computational decisions that are made on based on analytic, you know, choices that are guided by certain assumptions. If you do the analysis slightly differently, which is what we did. So we took, you know, almost 1000 subjects, and we,

instead of asking, you know, using kind of standard way of looking for signal and noise, we said, okay, anything which replicates from one subject to another is signal by definition, and anything which doesn't is noise. And so let's just try to parse the, you know, networks in the brain by doing this. And what we found was, you know, we found that the sort of networks that people often talk about, but they're really, they overlap. They're not, they're not disconnected. They're actually overlap and they overlap in, in particular regions of the brain, which are known to be, they're called hubs or rich club hubs, meaning densely connected regions that are responsible for really coordinating activity across the whole brain. They're called, you know, these Ritz-Carlton hubs are called the backbone of neural communication

in the brain. There's a really nice paper by Olaf Sporns and Vanden Heuvel Sporns. I think it's Vanden Heuvel and Sporns in 2013 in the Journal of Neuroscience. 2013 in the Journal of Neuroscience. And so my point is that these images that you see, they're beautiful and awe-inspiring, but they're curated by humans who have a set of assumptions. Yeah. And it's also easy to see the temptation to think in those terms because we have something like 170 years of neurology attesting to the fact that highly focal lesions, you know, brain damage, can lead to very specific deficits. Again, this can be understood in network terms, but it is in fact descriptively true that you can have a small region of the brain damaged,

and that can dissect out a very specific mental capacity, you know, language use or an ability to recognize faces or even to recognize specific classes of objects like, you know, tools versus animals. objects like tools versus animals. And that does give you this sort of jigsaw puzzle-like, Lego-like intuition about the modularity of the mind. Yeah, you're right. But even there, it's more complicated than it first appears, right? Because when you damage a part, when you damage tissue, you don't really know whether what you've damaged, the critical part, you know, to the function that you've lost are the neurons that are damaged or what are called fibers of passage, which means, you know, axons that run through that area,

which are really important. And I just learned about this really, this phenomenon that I just, this is the kind of stuff I just love, honestly, where you can lose, if you damage one part of your primary visual cortex, so this is in animals, they'll ablate a part of the primary visual cortex and the animal will lose the ability to see. And so obviously you think, oh, well, okay, this region must be super important to seeing.

And it is important, except that you can recover some of that function by a second lesion in the superior colliculus in the midbrain. So there's information that could make it from your retina to your primary visual cortex, but it's being suppressed by the colliculus in a regular fat, in a regular neurotypical brain. So you can recover function by a second lesion. And so it's just things like that, right? That make you, or here's another example, another, you know, example, which I find just absolutely fascinating. i find it slightly horrifying as a person but because of what happens to the animals but as a scientist it's really fascinating so they took these rats and trained them to run on a wheel and you know recorded directly from neurons in visual cortex, primary visual cortex.

And then they ablate the damage, the retinas, destroy the retinas of these animals. So they can't see. And V1 neurons, primary visual cortex neurons, quieten down. And then over 24 hours, they ramp up again and start firing at normal rates. So what's causing these neurons to fire? You know, you put the rat back on the wheel and its neurons, the pattern of firing looks really similar to what it looked like when the animal was sighted. So what is it exactly that's driving the activity in these neurons? And the answer probably is regions of the anterior cingulate cortex, which have direct connections to V1. And the reason why this is interesting is

that this region of the brain is a primary regulator of the systems of your body. Both it is a primary motor area for the viscera of your body, and it's an association region for your skeletal motor system. And what this activity is, essentially, what you can think about it is, are a set of visual predictions that are coming from past experience that these motor regions are able to reinstate. And so it's just trickier, Sam, than, you know, I mean, if you start to just poke at it a little bit, modularity starts to fall apart. Yeah, yeah. Well, I think we found the seminar you can teach at Esalen one day, ablating brainstem nuclei so as to recover a proper vision of the world. Yeah, I really wouldn't recommend that people try that at home. It's not advised.

Yeah, I really wouldn't recommend that people try that at home. It's not advised. So let's talk about prediction and just this uncanny circumstance we're all in, which very few people realize, and those of us who realize it, I think, rarely think about, which is we have this venerable philosophical thought experiment of the brain and the vat, and this is a kind of device to think about many things in the philosophy of mind, but rarely is it pointed out that we really are brains and vats already. The vat is our skull, and we do not have direct contact with the physical environment, much less reality itself, in any straightforward way. It's not like our senses are windows through which we're peering or hearing or sensing directly.

very active and even anticipatory, to use your term, predictive activity that is producing a visionary experience, a dreamlike experience of the world. It's exactly like a dream, except for the ways in which, in the waking state, our envisioning of the world is constrained by sensory input to a different degree. So how do you think about the situation we're in, just epistemologically, existentially? We are, and this is a phrase you use at some point in the book, we are experiencing a kind of controlled hallucination. It's not to say that nothing is veridical or that no statement about the world as it is is better than any other or more convergent with facts that we could intersubjectively find credible, but it's much more like the matrix than we give it credit for most of the time. And so perhaps that can get you going in the direction of how you think about the mind and the brain as a predictive computational system

and not one that's merely passively encountering the world as it is. Well, I think you just did a beautiful job describing it in very poetic terms, actually. Calling it a dreamlike, calling the brains, you know, or describing the brain's function as conjuring a dreamlike state is actually something that I just came across in this really wonderful book by Carlo Rovelli. It's his new book called Helgoland. I don't think it's available yet in the US. I had to order it from the UK. And I, and, you know, he's really, what he's doing, he's explaining his understanding of quantum mechanics for a civilian like me, you know, I'm not, I don't, I'm not a physicist. And, but, you know,

and with, with very, very little math and, and then, you know, as often seems to happen, you know, everyone wants to take a shot at explaining what the brain does and, you know, what consciousness is. It doesn't matter if you trained as an, you know, a physicist or what have you, everyone takes their shot. And, but his shot, you know, a physicist or what have you, everyone takes their shot. And but his shot, you know, he's describing, trying to describe prediction based on, you know, I'm imagining what he what he read from the literature and visual neuroscience, where a lot of this work has taken place. I think, though, there's much, there's a lot more work, which is very consistent with, you know, your description. There's a really, really nice paper that was written, actually, which was my review. I was I reviewed this paper,

actually, for behavioral and brain sciences, which is a really great journal. And this is what alerted me to this growing literature. This was back like in 2010, I think, maybe, or 2011, this growing literature on what's called predictive coding or predictive processing. It's a paper by Andy Clark. Philosopher, right? Philosopher, but also just writes beautifully about, very intuitively and beautifully about the brain as a predictive organ. Yeah.

But for me, I don't know about you, but I am inherently skeptical person. I don't even believe my own data necessarily. It takes me a really long time before I don't jump on bandwagons typically. And I also really don't, I mean, scientists, I think in general, wouldn't you agree? We don't really like to use the F word, you know, fact. That's a really scary word. So we try to avoid it. And, but, you know, if you look in the literature, if you look at anatomy and you look at any number of literatures in neuroscience, and you look at signal processing literatures in engineering, and so on, what you see is that exactly the same discovery is being

made over and over and over again by literatures that don't talk to each other. And I found this really compelling. And that is this idea that your brain is trapped in a dark, silent box called your skull. And it is constantly receiving sense data from the world, you know, through its sensory surfaces, your retina, your coch cochlea whatever and also in inside your body so it's it's it the world to your brain is everything outside of the skull and it's receiving these this sense data that it has to make sense of and this is an inverse problem because it these sense data are the effects that they're the outcomes of some set of changes but your brain doesn't have access to those changes it only has access to the outcomes the the consequences of those changes so how does it you know if your brain if your brain is exposed to a loud bang how does your brain know what that loud bang, how does your brain know what that loud bang is? How does your brain know what to do about it?

You know, you would do something different if it was a slamming door or a dropped box or a gunshot. And similarly, you know, when you feel a tug in your chest, how does your brain know? How does your brain know when it detects a tug, right? Whether when it's sensing a tug, whether that's, you know, anxiety or, you know, that there's some uncertainty or that you just ate a big meal and you're having a little trouble digesting it or the beginnings of a heart attack. It has to guess. And what does it use to guess? It uses the only other source of information that it has, which is past experience that it can re-implement, reinstate in its own wiring. So colloquially, we would call that memory. So when a brain remembers, when your brain remembers,

when my brain remembers, brains don't store memories and then call them up like files in a file drawer. Basically, remembering is reassembling, reassembling the past in the present for the purposes of making sense of sense data. And for a number of reasons, some of which are metabolic, your brain is sort of doing this predictively. So it's not waiting to receive the input and then trying to make sense of it and there are lots of ways to demonstrate this to people sometimes when i'm giving talks you know i'll use a baseball example and i'll kind of walk people through the timing of the baseball example you know baseball couldn't exist as a sport no actual ball related sport could exist if we had reactive brains.

There just isn't physically enough time for a batter to wait to see a ball before he swings and actually hit the ball. And there are lots of really cool, interesting examples from everyday life. But the point is that, metabolically speaking, it's much cheaper for the brain to use past experience to guess what's going to happen next, where the guess is not some abstraction. It's actually your brain changing the firing of its own neurons to prepare you to see and hear and smell and feel and do something in the next moment. And then it checks those predictions against the incoming sense data from the body and from the world. Scientists call this, you know, running a model of the world. But really what your brain is doing is it's running a model of your body. And it, it's the model of your body in the world,

but it only knows the world by virtue of the sense data that it gets from the sensory surfaces of your body. So essentially, every feature that your brain computes, it's computing in relation to your body in a particular moment in time, in a particular context or location relative to or related to the particular shape of your ear and the particular distance of your two eyes from one another and the particular state of your mitochondria and so on and so forth. It's all relative. That doesn't mean some kind of post-modernist morass. But what it does mean is that we really have to realize that everything that we experience,

we experience from a particular perspective. And there is nothing really called objectivity. The best we can hope for, according to the historian of science, Naomi Oreskes, is that a bunch of people with their own subjectivity, you know, with different histories and different backgrounds and different experiences in the world, that they can come to consensus over a scientific set of observations. And that's about as close to objective fact as we can get. And it's a pretty, it's pretty darn good. It's worked out pretty, pretty well for us, you know, but the idea that there are universal facts that can be objectively adjudicated by being rational or something, it's a fiction that, interestingly, that brains tell themselves,

even though brains are completely incapable of doing such things. Well, to say that there's no true objectivity is not the same thing as saying that it's not possible to be wrong, right? And we know certain things are wrong. Oh, for sure. And it's also not saying that anything is possible, right? So, I mean, sometimes when I say, well, there's more than one, you know, there's more than one, you know, when I talk about, you know, variability is the norm, right? That in many places in biology and in psychology, there's much more variation than we often acknowledge or would like. But that doesn't

mean that anything is possible. You know, it means that there's just more than one possibility. And similarly, I would say, look, you know, we can all agree, right, that we're going to have ground glass for dinner, but that doesn't necessarily translate into the objective reality that we can actually eat glass, right? It doesn't really matter what we believe. We could all agree that COVID is not infectious and that we don't have to wear masks, but, you know, the virus doesn't care about that. I mean, viruses don't care about anything, but really all the virus needs is a wet set of nice wet set of lungs. It doesn't, it doesn't matter what the, you know, what that person's brain believes, but there are many, many, but I think, you know, there,

there are many, many cases where what we believe really matters to what we experience. But even if you want to take belief out of the equation, what you experience, what your reality is, how you experience the world is very much relational. It's in relation to the body that you have. And you don't experience yourself that way. I certainly, I don't mean, I can't tell you what you experience. I don't experience myself that way. And if I wasn't a scientist and somebody just told me that, I'm not sure that I would believe it actually. But it is, that is the best available evidence that your brain is constantly cultivating your past for the purposes of predicting your future, which will become

your present. Yeah, let's see if we can make this concrete for people, because this is really ground upon which the scientific framing of what's going on can unlock a kind of psychological freedom to just change one's sense of what one is as a subject in the world. And I think it can relieve certain kinds of suffering. In the simplest case, just to take this predictive piece, which can sound spooky, you take something like a voluntary motor action. So I can decide to reach and pick up a cup on my desk. And this does relate to this controversy that I keep resurrecting for myself over the reality or lack thereof of free will. I don't know if you know how far down that rabbit hole I've gone. Oh, yes. I've enjoyed, I guess, following you down that rabbit hole.

So we can talk about that if it interests you, but people have a sense that they are subjects that have this capacity to freely initiate behavior, and that's different. I would certainly agree that voluntary behavior is different from involuntary behavior, but I just don't think we need the concept of free will to differentiate the two. So one way they're different is when I'm doing something of my own volition, reaching and picking up a cup, that feels a certain way. And it feels a certain way because there are certain implicit processes that we know must be going on neurophysiologically there that do follow this kind of predictive mapping of things. So when I'm reaching, and I'm not consciously aware of it, but I can be made

consciously aware of it, certainly when anything goes wrong. So I'm not aware that I'm a prediction machine when I'm reaching to grasp this cup, but if I reached and my fingers passed through it, right, if it was a hologram of a cup and not a real one, or if it felt, you know, squishy, if it was made of, you know, rubber and I wasn't expecting that. All of those occasions of surprise are built on some set of expectations that I wasn't aware of having until I became disillusioned. So I was not aware of expecting solidity, though of course I was. I mean, everything about the grasping behavior of my hand was anticipatory in a certain way. And you can make that predictive program consciously felt,

certainly in the moments in which it's violated, but it's just simply neurologically the case that we are comparing in order to... The only way to detect anomalies in the environment is to have this background modeling going on of what's likely to happen in each moment based on what I'm doing now and what I'm doing next. And this question of what to do next really does cover so much of what we're about as minds. We're constantly deciding what to do next on some level. Oh, absolutely. There's so much to say about, there's so much to unpack that's interesting

about what you just said. I mean, first of all, I would say, it seems to me that, you know, because for whatever reason, we've talked about why, how people don't think about the situation. But because for whatever reason, we've talked about why. If you'd like to continue listening to this conversation, you'll need to subscribe at SamHarris.org. Once you do, you'll get access to all full-length episodes of the Making Sense podcast,

along with other subscriber-only content, including bonus episodes and AMAs and the conversations I've been having on the Waking Up app. The Making Sense podcast is ad-free and relies entirely on listener support, and you can subscribe now at SamHarris.org.