Charan Ranganath: Human Memory, Imagination, Deja Vu, and False Memories | Lex Fridman Podcast #430 | Lex Fridman Transcript

Lex Fridman interviews memory neuroscientist Charan Ranganath on how memory shapes identity, decisions, and happiness

Lex Fridman speaks with Charan Ranganath, psychologist and neuroscientist at UC Davis and author of Why We Remember.

Summary

Lex Fridman speaks with Charan Ranganath about the science of human memory — how it works, why it fails, and what it means for how we live. Ranganath argues that memory is not a replay of the past but an imaginative reconstruction shaped by prior knowledge, emotion, and context, and that forgetting is not a flaw but a feature of a brain optimised for prediction and efficiency. He explains how false memories form through misinformation and social contagion, how propaganda exploits the malleability of memory at a collective scale, and how the act of remembering itself can alter what is remembered. The conversation ranges across episodic and semantic memory, the hippocampus, fMRI, brain-computer interfaces, the neuroscience of time perception, and the relationship between memory and imagination — with Ranganath drawing on his own near-death experience on a paddleboard and his career as a touring musician to illustrate the science.

Key Takeaways

FULL TRANSCRIPT

The Experiencing Self and the Remembering Self

Lex Fridman: Daniel Kahneman describes the experiencing self and the remembering self, and argues that the happiness and satisfaction you gain from the outcomes of your decisions comes not from what you've experienced but rather from what you remember of the experience. Can you speak to this interesting difference — the experiencing self and the remembering self — that you write about in your book?

Charan Ranganath: Danny really impacted me. I was an undergrad at Berkeley and got to take a class from him long before he won the Nobel Prize or anything. It was a mind-blowing class. This idea of the remembering self and the experiencing self — I got into it because it's so much about memory, even though he doesn't study memory.

We're right now having this experience, and people can watch it on YouTube or listen to it on audio. But if you were describing it to somebody else, you could probably summarise the whole thing in ten minutes — and that would miss a lot of what actually happened. The way we remember things is not a replay of the experience; it's something totally different. It tends to be biased by the beginning and the end, and by the peaks — the best parts, the worst parts — and those are the things we remember. So when we make decisions, we usually consult memory, and we feel like our memories are a record of what we've experienced. But they're not. They're a very biased sample — biased in an interesting and, I think, biologically relevant way.

Lex Fridman: In the way we construct a narrative about our past, you say it gives us an illusion of stability. Can you explain that?

Charan Ranganath: Basically, I think a lot of learning in the brain is driven towards being able to make sense of things. Memory is really all about the present and the future — the past is done. Biologically speaking, it's not important unless there's something from the past that's useful. So what our brains are really optimised for is to learn about the stuff from the past that's going to be most useful in understanding the present and predicting the future. Cause-and-effect relationships, for instance — that's a big one.

My future is completely unpredictable in the sense that you could, in the next ten minutes, pull a knife on me and slit my throat. But having seen some of your work and just generally my expectations about life, I'm not expecting that. I have a certainty that everything's going to be fine and we're going to have a great time talking today. And we're often right. If I go see a band on stage, I know they're going to make me wait, the show's going to start late, and then when they come on there's a very good chance there's going to be an encore. I have a memory, so to speak, for that event before I've even walked into the show. There are going to be people holding up their camera phones to take videos. That's like everyday fortune-telling that we do — it's not real, it's imagined — and it's amazing that we have this capability. That's what memory is about.

But it can also give us this illusion that we know everything that's about to happen. And I think what's valuable about that illusion is that when it's broken, it gives us information. You probably know about information theory — the idea that information is what you didn't already have. The prediction errors we make based on memory are where the action is. The error is where the learning happens.

Lex Fridman: Just to linger on Daniel Kahneman and this whole idea of the experiencing self versus the remembering self — I was hoping you could give a simple answer about how we should live life, based on the fact that our memories could be a primary source of happiness. An event, when experienced, bears its fruits most when it's remembered over and over again. And maybe there is some wisdom in the fact that we can control, to some degree, how we remember it — how we evolve our memory of it — such that it can maximise the long-term happiness of that repeated experience.

Charan Ranganath: Well, first I'll say I wish I could take you on the road with me — that was such a great description. Can I be your opening act?

Lex Fridman: Oh my God, no — I'm going to open for you. Otherwise it's like everybody leaves after you're done. I did that in Columbus, Ohio once. It wasn't fun. The opening act drank our beer, we spent all this money going all the way there, everybody left after the opening acts were done, and there was just that one stoner dude with the dreadlocks hanging out. Then we blew our savings on a hotel room.

Charan Ranganath: As a small tangent — I'm a legit touring act. When I was in grad school I played in a band and we travelled, we would play shows. It wasn't like we were a hardcore touring band but we did some touring and had some fun times. We did a movie soundtrack — Henry: Portrait of a Serial Killer.

Lex Fridman: That's a good movie.

Charan Ranganath: We were on the soundtrack for the sequel, Henry 2: Mask of Sanity, which is a terrible movie.

Lex Fridman: How's the soundtrack?

Charan Ranganath: It's pretty good. It's badass. At least that one part where the guy throws up the milkshake. Anyway, we're getting back to life advice.

One thing that I try to live by, especially nowadays — and since I wrote the book I've been thinking more and more about this — is how do I want to live a memorable life? If we go back to the pandemic, how many people have memories from that period, aside from the trauma of being locked up and seeing people die and all of that? We were stuck inside looking at screens all day, doing the same thing with the same people. I don't remember much from that period in terms of those good memories you're talking about.

When I was growing up, my parents worked really hard for us and we went on some vacations but not very often. I really try now to take vacations to interesting places as much as possible with my family, because those are the things that you remember. I really do think about what's going to be memorable and then just do it, even if it's a pain in the ass — because the experiencing self will suffer for that, but the remembering self will be like, yes, I'm so glad I did that.

Lex Fridman: Do things that are very unpleasant in the moment because those can be reframed and enjoyed for many years to come — that's probably good advice, or at least a good way to see the silver lining when you're going through something.

Charan Ranganath: I think it's one of those things where if you've gone through something with someone, that's a bonding experience. It can really bring you together. I like to say there's no point in suffering unless you get a story out of it.

In the book I talk about the power of the way we communicate with others and how that shapes our memories. I had this near-death experience — at least that's how I remember it — on a paddleboard, where just everything that could have gone wrong did go wrong. So many mistakes were made. At some point I was basically away from my board, pinned in a current in this corner, not a super good swimmer. My friend who came with me, Randy, who's a computational neuroscientist, had just been pushed past me so he couldn't even see me. I'm just thinking: if I die here, no one's around, you just die alone. So I said, well, failure is not an option, and eventually I got out of it. Froze, got cut up — the things we were going through were just insane.

The short version is that my wife and my daughter and Randy's wife gave us all sorts of hell about it because they were ready to send out a search party. Then I started to tell people in my lab about this, and then friends, and it just became a better and better story every time. We actually had some photos of the crazy things — like this generator hanging over the water and we're ducking under these metal gratings, going flat on the board. It was just nuts. But it became a great story, and Randy and I were already tight, but that was a real bonding experience for us.

I learned from that that I don't look back on it enough, actually, because I think we often don't have the confidence to believe that things will work out — that we'll be able to get through certain things. But my ability to actually get something done in that moment is better than I give myself credit for. That was the lesson of that story.

Lex Fridman: You're making me realise that it's not just those kinds of stories but even periods of depression or really low points — to me it feels like a motivating thing that the darker it gets, the better the story will be if you emerge on the other side. Maybe if people listening are going through something, one thing that could be a source of light is that it'll be a hell of a good story when it's all over.

Memory and Decision-Making

Charan Ranganath: Let me ask you about decisions. When we face the world and we're making different decisions, how much does our memory come into play? Is it the kind of narratives we've constructed about the world that are used to make predictions?

Lex Fridman: Absolutely. Let's say after this you and I decided we're going to go for a beer. How do you choose where to go? You're probably going to say, oh yeah, this new bar opened up near me, I had a great time there, they had a great beer selection. Or you might say, oh, we went to this place and it was totally crowded and they were playing horrible EDM. Right there — valuable source of information. And then you have this counterfactual reasoning: well, I did this previously, but what if I had gone somewhere else? Maybe I'll try this other place because I didn't try it the previous time.

Even if you think about the big decisions in life — you and I were talking before we started recording about how I got into memory research and you got into AI — we all have these personal reasons that guide us in particular directions. Some of it is environment and random factors in life, and some of it is memories of things we want to overcome or things we build on in a positive way. Either way, they define us. And probably the earlier in life the memories happen, the more defining power they have in determining who we become.

I do feel like adolescence is much more important than people give it credit for. The teenage years are just so important for the brain. That's where a lot of mental illness starts to emerge. We're now thinking of things like schizophrenia as a neurodevelopmental disorder because it just emerges during that period of adolescence and early adulthood.

The other part of it is that the self is an evolving construct — I think we underestimate that. When you're a parent you feel like every decision you make is consequential in forming this child, and it plays a role, but so do the child's peers, and so much else. That's why I think the big part of education that's so important is not the content you learn — think of how much stuff we learned in school that we don't remember — but learning how to get along with people and learning who you are and how you function. That can be terribly traumatising even if you have perfect parents working on you.

Infantile Amnesia and Brain Development

Lex Fridman: Is there some insight into the human brain that explains why we don't seem to remember anything from the first few years of life?

Charan Ranganath: Yes, in fact I was just talking to my really good friend and colleague Simona Ghetti, who studies the neuroscience of child development. There are a bunch of reasons. One is that there's an area of the brain called the hippocampus, which is very important for remembering events — episodic memory. The first two years of life there's a period called infantile amnesia, and then the next couple of years after that there's a period called childhood amnesia. The difference is that basically in the lab, and even during childhood and afterwards, children basically don't have any episodic memories for those first two years. The next two years it's very fragmentary, which is why they call it childhood amnesia — there's some, but not much.

One reason is that the hippocampus is taking some time to develop. But another is the neocortex — all the folded grey matter around the hippocampus is developing so rapidly, and a child's knowledge of the world is just massively being built up. If you trained a neural network and gave it the first couple of patterns, and then bombarded it with years' worth of data and tried to get back those first couple of patterns — everything changes. The brain is so plastic, the cortex is so plastic during that time. We think memories for events are very distributed across the brain, so imagine trying to get back the pattern of activity that happened during one moment, but the roads you would take to get there have been completely rerouted.

The third explanation is that a child's sense of self takes a while to develop. Their experience of learning might be more about learning what happened, as opposed to having this first-person experience of "I remember being there."

Lex Fridman: Somebody once said to me, loosely philosophically, that the reason we don't remember the first few years of life is because of how traumatic it is — basically the error rate when your brain's prediction doesn't match reality in the first few months is probably crazy high. It's just nonstop freaking out. The collision between your model of the world and how the world works is just so high that you want whatever the trauma of that is not to linger around.

Charan Ranganath: I wouldn't necessarily describe it as a trauma. Those first few years — a kid's internal model of their body is changing. Just learning to move. If you ever have a baby, you'll know that the first three months they're discovering their toes. Everything is changing. But what's really fascinating — and this is not me being a scientist, it's one of those things people talk about when they talk about the positive aspects of children — is that they're exceptionally curious and they have this openness towards the world. So that prediction error is not a negative traumatic thing. I think it's a very positive thing, because it's what they use. They're seeking information.

One of the areas I'm very interested in is the prefrontal cortex. It helps us use our knowledge to say, here's what I want to do now, here's my goal, here's how I'm going to achieve it — and focus everything toward that. The prefrontal cortex takes forever to develop in humans. The connections are still being tweaked and reformed into late adolescence and early adulthood, which is when you tend to see mental illness pop up. Then you have about ten years, maybe, of prime functioning of the prefrontal cortex, and then it starts going down again.

Memory researchers always say children are worse than adults at episodic memory, and older adults are worse than young adults at episodic memory. I always thought, God, this is so weird — why would we have this period of time that's so short when we're optimal? But I've come to realise that for most of the human condition, we had a series of stages of life where you have young adults saying, I've got a child, I'm part of this village, I have to hunt and forage and get things done — I need a prefrontal cortex so I can stay focused on big-picture, long-haul goals. Now I'm a child, I'm in this village, wandering around, I've got some safety, and I need to learn about this culture because I know so little. What's the best way to do that? Explore. Play. You don't want a super tight prefrontal cortex — you don't even know what the goals should be yet.

Then you go late in life. Why don't you have a great prefrontal cortex then? I think when you're older, in most societies, your job is no longer to form new episodic memories — it's to pass on the memories you already have, this knowledge about the world, what we call semantic memory, to the younger generations. Pass on the culture. Even now in indigenous cultures, that's the role of the elders. They're respected. They're not seen as people who are past it and losing it. Memory is doing what it's supposed to throughout these stages of life. It's always optimal — just optimal for that stage of life.

Lex Fridman: And for the ecology of the system. Another species that has menopause is orcas — Orca pods are led by the grandmothers. They're the ones that pass on the traditions to the younger generation. Different Orca pods have different traditions — they hunt for different things, they have different play traditions. That's a culture. In social animals, evolution is designing brains that are optimised not just for the individual but for kin.

Charan Ranganath: It's fascinating to think of the individual orca or human throughout life in stages, doing a kind of optimal wisdom development. In the early days you don't even know what the goal is, you figure out the goal, you optimise for it and pursue it, and then all the wisdom you collect through that you share with the others in the system. As a collective you converge towards greater wisdom throughout the generations. In that sense it's optimal. Us humans and orcas have got something going on.

How the Brain Forgets and Remembers

Lex Fridman: How does the brain forget, and why does it remember? What are the different components involved — the hippocampus, the different types of memory?

Charan Ranganath: We tend to think of memories as individual things we can just access, maybe a bit like photos on your phone. But in the brain, you have this distributed pool of neurons and the memories are shared across different pools. What you have is competition — sometimes memories that overlap can be fighting against each other. Sometimes we forget because that competition wipes things out. Sometimes we forget because there aren't the biological signals that would promote long-term retention. And lots of times we forget because we can't find the cue that sends us back to the right memory — we need the right cue to activate it.

In a neural network, there's no single location you'd point to and say "this is the memory." The whole ecosystem of memories is in the weights of the network, and you can extract entirely new memories depending on how you query it. You have to have the right query, the right prompt, to access whatever part you're looking for.

In humans there's a more complex set of ways memory works. There's the knowledge, or what we call semantic memory, and then there's memories for specific events, which we call episodic memory. Different pieces of the puzzle require different kinds of cues. That's a big part of what we call retrieval failure.

Lex Fridman: What are the interesting categories of memory — working memory, short-term memory, long-term memory?

Charan Ranganath: Memory researchers love to cut things up and ask whether memory is one thing or two things or three things. Working memory is a term coined by Alan Baddeley. It's basically this ability to keep information online in your mind — right in front of you at a given time — and to be able to control the flow of that information, choose what's relevant, manipulate it, and so forth. What Baddeley did that was quite brilliant was to say there's this ability to passively store information — see things in your mind's eye or hear your internal monologue — but then there's also a separate component he called the central executive, which is identified a lot with the prefrontal cortex. It's this ability to control the flow of information that's being kept active based on what you're doing.

A lot of my early work was basically arguing that working memory — which some researchers call short-term memory — is not at all independent from long-term memory. A lot of executive function requires learning, and you have to have synaptic change for that to happen.

One of the things I've been getting into lately is the idea that we form internal models of events. The obvious example I always use is birthday parties. You go to a child's birthday party once, the cake comes out, you see a candle — you can predict the whole frame of events that follows, right up to the point where the child blows out the candle. You have an internal model in your head of what's going on. If you follow people's eyes, they're not actually on what's happening — they're going where the action's about to happen. That's a kind of working memory product. It's something you're keeping online that's allowing you to interpret the world around you.

To build that model, though, you need to pull out stuff from your general knowledge of the world — semantic memory — and then pull out memories for specific events from the past — episodic memory. In a way they're all connected. The way we organise information in the present, which is working memory, will play a big role in determining how we remember that information later, which people typically call long-term memory.

Event Boundaries and Episodic Encoding

Charan Ranganath: One of the fascinating things we've been studying — and Jeff Zacks was a big pioneer in this, and I've been working with many other people, Ken Norman, Lila Davachi at Columbia has done some interesting stuff — is this idea that we form these internal models, and at particular points of high prediction error, or points of uncertainty, or points of surprise, or motivationally significant periods, those are the points when it's maximally optimal to encode an episodic memory.

I used to think we're just encoding episodic memories constantly, boom boom boom boom. But think about how much redundancy there is in all that — it's just a lot of information you don't need. If you capture an episodic memory at the point of maximum uncertainty for a singular experience, you've grabbed the most useful point in your experience. What we see is that the hippocampus and these other networks involved in generating internal models of events show a heightened period of connectivity during those breaks between different events, which we call event boundaries — the points where you're surprised, or you cross from one room to another. That communication is associated with a bump of activity in the hippocampus and better memory.

If people have a very good internal model throughout an event, you don't need to do much memory processing — you're in a predictive mode. Then at these event boundaries you encode and then retrieve, and you're like, okay, wait a minute, what's going on here? And maybe you have to go back and remember something to pull out the episodic memory to make sense of whatever's happening. There's this beautiful dynamic you can see in the brain of these different networks coming together and pulling apart at different points in time, allowing you to go into these different modes.

Training and Improving Memory

Lex Fridman: Can memory be trained and improved?

Charan Ranganath: Improvement depends on what your definition of optimal is. What I say in the book is that you don't want to remember more — you want to remember better, which means focusing on the things that are important. That's what our brains are designed to do.

If you go back to the earliest quantitative studies in memory by Ebbinghaus, what you see is that he was trying so hard to memorise arbitrary nonsense, and within a day he lost about 60% of that information. As far as we know, nobody has managed to violate those basics. If your expectation is that you should remember everything, you're already off — that's just not what human brains are designed to do.

On the other hand, the brain is always about less is more. I've seen estimates that the human brain uses something like 12 to 20 watts in a day — that's just nuts. It's all about reusing information and making the most of what we already have. Biologically, neuromodulators — chemicals in the brain like norepinephrine, dopamine, serotonin — are released during moments that tend to be biologically significant: surprise, fear, stress. These chemicals promote lasting plasticity. Attention is a big factor as well — our ability to focus on what's important.

There are different schools of thought on training attention. My colleague Amishi Jha wrote a book called Peak Mind and talks about mindfulness as a method for improving attention and focus. She works a lot with military personnel, like Navy SEALs. Adam Gazzaley, another friend and colleague, has work on training through video games as a way of training attention. It's not clear that you transfer those gains to the outside world — that's very controversial.

The implication is that attention is a fundamental component of remembering something, and attention might be something you could train. We do train it in certain ways. If you're an expert in something, you are training attention. We did a study of expertise in the brain, and one of the things that surprised us — as people became experts in identifying these different kinds of objects we made up — was that we were seeing massive increases in activity in the prefrontal cortex. This fits with studies of chess experts. It's not so much that you learn the patterns passively — you learn what to look for, what's important and what's not. You can see this in any expert. A professional athlete is looking three steps ahead of where they're supposed to be.

If you take memory athletes — people who train in competitions and memorise a deck of cards in a really short amount of time — by all accounts from people who become memory athletes, they weren't born with some extraordinary memory. They practised strategies over and over again. The strategy for memorising the order of a deck of cards might not help you for something else, like remembering your way around Austin, Texas. But whatever you're interested in, you can optimise for that. That's a natural byproduct of expertise.

Lex Fridman: There's something called the Memory Palace that I've played with. And there's also spaced repetition — I use Anki a lot every day. It's an app that does spaced repetition. Medical students use it a lot. The whole concept is that when the thing is fresh you have to remind yourself of it a lot, and then over time you can wait a week, a month, a year before you have to recall it again. That way you essentially have something like note cards — you can have tens of thousands of them and only spend thirty minutes a day actually refreshing all of that knowledge. It's really great.

And then the Memory Palace is a technique that allows you to remember things like the Ikea catalogue by placing them visually in a place you're really familiar with. You walk along that palace in your mind and it reminds you. One of the things I have to solve for myself is how to remember names. I'm horrible at it. I think it's because when people introduce themselves, I have the social anxiety of the interaction — I know I should be remembering that, but I'm freaking out internally about the social interaction in general, and so I forget immediately.

Charan Ranganath: I feel like we've got a lot in common, because when people introduce themselves to me it's almost like I have this blank blackout for a moment and then I'm just looking at them like, what happened?

The reason it's hard is that there's no reason you should be able to remember names easily, because when you say "remembering a name" you're not really remembering a name — you're associating a name with a face and an identity, and that's a completely arbitrary thing. These names are just utterly arbitrary, so you have nothing to latch on to. It's not really a thing our brains do very well — learning meaningless arbitrary stuff. So what you need to do is build a connection somehow. Visualise a connection.

The first thing I think of for you is Lex Luthor. Lex Luthor is this criminal mastermind — and then I just imagine you talked about stabbing or whatever earlier — and that's it. These kinds of weird associations build a richer network. If I've read someone's papers academically and then I meet them at a conference, I can immediately associate that name with that face because I have this pre-existing network to lock everything into.

That's what the method of loci, or the Memory Palace technique, is all about — you have a pre-existing structure in your head, like your childhood home or a mental palace you've created, and you put arbitrary pieces of information in different locations in that mental structure. Then you can walk through the different paths and find all the pieces of information you're looking for.

Lex Fridman: When people were describing this to me a while ago, it seems insane. You literally think of a place you're really visually familiar with and you literally place in that three-dimensional space facts or people or whatever you want to remember, and you just walk in your mind along that place visually. One of the limitations is there is a sequence to it — you can't just go upstairs right away, you have to walk along the room. So it's really great for remembering sequences but not great for remembering individual facts out of context.

Charan Ranganath: Absolutely. And I think the principle here is that memories can compete with each other. Imagine if this were my desk, cluttered with a zillion different yellow Post-it notes, and on one of them I put my bank password. It's going to take me forever to find it — it'll be buried under all the others. But if it's hot pink, it's going to stand out and I find it really easily. So if things are distinctive, if you've processed information in a very distinctive way, you can have a memory that lasts. That's very good for name-face associations — if I get something distinctive about you, that's a great cue.

But the other part is: what if I just organised my notes so that I have my finances in one pile and my to-do list in another? Then I know exactly where to go. The method of loci works because it gives you a way of organising. There's a school of thought that says episodic memory evolved from this kind of knowledge of space — this primitive ability to figure out where you are. People explain the method of loci that way.

But the method of loci is not at all special. If you're not a good visualiser, stories are a good alternative. A lot of memory athletes use stories — if you're memorising a deck of cards, they have a little code for the different cards and they'll make up a story about things that are happening. Songs are a great one too. There's this obscure episode of the TV show Cheers where they sing a song about Albania that a character uses to memorise all these facts about Albania. I could still sing that song to you. I saw it on a TV show once.

Spaced Repetition and the Testing Effect

Lex Fridman: Can you explain spaced repetition?

Charan Ranganath: If I'm trying to memorise something — let's say I have an hour to memorise as many Spanish words as I can — if I do half an hour today and half an hour later in the same day, I won't retain that information as long as if I do half an hour today and half an hour one week from now. The extra spacing helps me retain the information better.

There's an interesting boundary condition, which is that it depends on when you need that information. For me, I can't remember much from college and high school because I crammed — I did everything at the last minute, sometimes studying in the hallway right before the test. That was great because I just had that information right there. Not spacing can really help you if you need the information very quickly. But the problem is you tend to forget it later on. If you space things out, you get a benefit for long-term retention.

In our computational model, one way of thinking about this is that this conversation is associated with a particular context — a particular place in time. All these little cues in the background, these guitar sculptures, that big light umbrella — they're all part of my memory for what we're talking about. Later on, you're at home drinking a beer and you're trying to remember this conversation, but the context is different. Your current situation doesn't match up with the memory you pulled up. There's a mismatch. In our model, what you start to do is erase or alter the parts of the memory that are associated with the specific place and time, and heighten the information about the content. So if you remember this information in different times and different places, it becomes more accessible at different times and different places — because it's not overfitted to one particular context.

That's also why the memories we call upon the most feel kind of like things we just read about. You don't vividly reimagine them. They're just things that come to us like facts. Basically, events we've recalled over and over again — we keep updating that memory, so it's less and less tied to the original experience. But then we have those other ones where you get a reminder of a very specific context — you smell something, you hear a song, you see a place you haven't been to in a while — and boom, it just comes back to you. That's the exact opposite of what you get with spacing.

Lex Fridman: So with spaced repetition, one of its powers is that you lose attachment to a particular context — but then it loses the intensity of the flavour of the memory. That's interesting. But the content becomes stronger.

Charan Ranganath: Yes. And this falls into a category related to what's called the testing effect. The idea is that if you're trying to learn words in Spanish, and you test yourself on the words, that act of testing yourself helps you retain them better over time than if you just studied them. From traditional learning theories, this seems weird — why would you do better by giving yourself this extra error from testing yourself, rather than just giving yourself perfect input that's a replica of what you're trying to learn?

I think the reason is that you get better retention from that error, that mismatch. In our model, it's conceptually similar to what happens with backpropagation in neural networks. The idea is that you expose the bad connections and the good connections, and you can keep the parts of the cell assembly that are good for the memory and lose the ones that are not. But if you don't stress-test the memory, you haven't exposed it to the error fully.

This is a thing I come back to over and over again: you will retain information better if you're constantly pushing yourself to your limit. If you're feeling like you're coasting, you're actually not learning. You should always be stress-testing the memory system. And feel good about it — even though everyone tells me their memory is terrible, in the moment they're overconfident about what they'll retain later on.

What happens is when you test yourself, you're like, oh my God, I thought I knew that but I don't. It can be demoralising until you get around that and realise this is how you learn best. It's like if you're trying to star in a movie, you don't just sit around reading the script — you actually act it out, and you're going to botch those lines from time to time.

Lex Fridman: There's an interesting moment — you probably experienced this. I remember a good friend of mine, Joe Rogan — I was on his podcast and we were randomly talking about soccer. I grew up watching Diego Maradona, one of the greatest soccer players of all time. Joe asked me if he was still alive, and I said yeah. I don't know why I thought that, because he had passed away — I had tweeted about it, how heartbroken I was, all of that, like a year before. I know this. But in my mind I went back to the thing I've done many times — visualising some of the epic runs he had on goal — and so for me he was alive. Part of it also is that when you're talking to Joe there's stress, and the attention is allocated in a particular way. But when I walked away I was like, in which world was Diego Maradona still alive? I was sure in my head that he was. It's a moment that sticks with me. It shows the power of the mind in a positive sense — to erase memories you want erased, maybe.

Charan Ranganath: One of the cool things I found is that some people really revolutionise a field by creating a problem that didn't exist before. One of my former mentors, Marcia Johnson, who in my opinion is one of the greatest memory researchers of all time — she comes up as a young woman in this mostly male field and gets into this idea of how do we tell the difference between things we've imagined and things we actually remember? How do we tell where a mental experience came from?

It turns out this is a huge problem, because essentially our mental experience of remembering something that happened and our mental experience of thinking about something — how do you tell the difference? They're both largely constructions in our head. The way we often succeed is by using our prefrontal cortex and really focusing on the sensory information, the place and time, the things that put us back into when this information happened. If it's something you thought about, you're not going to have all of that vivid detail as you do for something that actually happened. But it doesn't work all the time. It takes time, it's slow, and it's effortful. But that's what you need to remember accurately.

Memory and Imagination

Charan Ranganath: Imagination is exactly the opposite — it's basically saying, I'm just going to take all this information from memory, recombine it in different ways, and throw it out there. Dan Schacter and Donna Addis have done cool work on this, and Demis Hassabis did work on this with Eleanor Maguire at UCL. This goes back to Frederick Bartlett, who was a revolutionary memory researcher. He actually rejected the whole idea of quantifying memory — he came from an anthropology perspective — and he just asked people to recall things. He gave people stories and poems, asked them to recall them, and what he found was that people's memories didn't reflect all of the details of what they were exposed to. They were filtered through this lens of prior knowledge, the cultures they came from, the beliefs they had, the things they knew. He concluded that remembering is an imaginative construction — we don't replay the past, we imagine how the past could have been by taking bits and pieces that come up in our heads. Likewise, when we imagine something and create something, we're creating it from specific experiences we've had and combining it with our general knowledge.

Imagination is fundamentally coupled with memory in both directions. One of the things that's been studied is patients who have amnesia — brain damage to the hippocampus. If you ask them to imagine things that are not in front of them, like what could happen after I leave this room, they find it very difficult to give you a scenario. Or if they do, it's more stereotyped. That's partly because when you have amnesia you're stuck in the present — to get a very good model of the future, it really helps to have episodic memories to draw upon.

One of the most impressive things when people started to scan people's brains and ask them to remember past events was that there was this big network called the default mode network. It gets a lot of press because it's thought to be important — it's engaged during mind wandering, and it only comes on when you stop paying attention to something. People thought, oh, it's just this daydreaming network. But then what people found was that when people recall episodic memories, this network gets active. And if you look at brain images of people imagining possible scenarios of things that could happen in the future — even things that couldn't really be very plausible — they look almost the same as maps of brain activation when people remember the past.

According to our theory, and we've got some data to support this, we've broken up this network into various sub-pieces. Basically it's taking apart all of our experiences and creating these little Lego blocks out of them. You can put them back together if you have the right instructions to recreate experiences you've had, but you could also reassemble them into new pieces to create a model of an event that hasn't happened yet.

Lex Fridman: There's a good percentage of time I personally live in the imagined world. I do thought experiments a lot. I take the absurdity of human life as it stands and play it forward in all kinds of different directions. I suppose I have to be a little bit careful to make sure stuff happened versus stuff that I just imagined happened. Some of my best friends are characters inside books that never even existed. Brothers Karamazov — I love that book. Those characters exist in my mind. I have almost conversations with them. It's interesting to allow your brain to play with ideas of the past and the imagined and see it all as one.

Charan Ranganath: There was actually this famous mnemonist — described by the really famous neuropsychologist from Russia named Luria — and this guy was named Solomon Shereshevsky. He had this condition called synesthesia that basically created these weird associations between different senses that normally wouldn't go together. That gave him this incredibly vivid imagination that he would use to memorise all sorts of things — he would just create these incredibly detailed things in his head. But it also really haunted him. By some reports, at some point he had trouble differentiating his imagination from reality. That's what psychosis is in some ways — your internal signals are being confused with actual things in the outside world. It's both feature and bug. It might be why there's such an interesting relationship between genius and psychosis — maybe they're just two sides of the same coin.

Memory Athletes and Memory Sport

Lex Fridman: Can we talk about memory sport a little longer? There's something called the USA Memory Championship. What are these athletes like? What does it mean to be elite level? Have you interacted with any of them?

Charan Ranganath: There's a guy named Henry Roediger who's studying these folks, and there's actually a book by Joshua Foer called Moonwalking with Einstein where he talks about deciding to become a memory athlete as part of writing the book.

They often have these life events that make them go, hey, why don't I do this? There was a guy named Scott Hagwood who I write about — he was getting chemo for cancer and decided to fight what's called chemo brain, where people lose a lot of their sharpness, by learning these memory skills. He bought a book by other memory athletes or memory experts, and that's the story you hear from a lot of memory athletes — they buy a book, learn those skills, practise them over and over again. They start by winning bets and so forth, and then they go into these competitions. The competitions are typically things like memorising long strings of numbers or memorising orders of cards — pretty arbitrary things, not things where you'd be able to bring a lot of prior knowledge. But they build the skills you need to memorise arbitrary things.

Lex Fridman: I've gotten a chance to work with something called n-back tasks — these kinds of tasks used to load up working memory and test all kinds of things, like how well you do at multitasking, or in the context of driving, if we fill up your brain with intensive working memory tasks, how good are you at also not crashing. But those tasks are arbitrary and usually about recalling a sequence of numbers in some semi-complex way. Do you have any favourite tasks in your own studies?

Charan Ranganath: I've really been most excited about going in the opposite direction and using things that are more and more naturalistic. What we found is that memory works very differently when you study it in the way that people typically remember — it goes into a much more predictive mode, you have these event boundaries, and there's this fascinating mix of interpretations and imagination with perception.

The new direction we're going in is understanding navigation in memory for places. There's a lot of work done in rats — you put a rat in a box, the rat chases cheese, and you find cells in the hippocampus that fire when the rat is in different places. The conventional wisdom is that the hippocampus forms a map of the box. I think that probably happens when you have absolutely no knowledge of the world. But one of the cool things about human memory is we can bring to bear our past experiences to economically learn new ones. If I go to the IKEA in Austin, I could probably find my way to where the wine glasses are without even thinking about it, because it's got a very similar layout to my local IKEA. I don't have to form a brand new map for a new place.

Another thing we're really interested in is this idea that instead of mapping out every coordinate in a space, you form a pretty economical graph that connects the major landmarks together — emphasising the things that are most important, the places you go for food, the landmarks that help you get around, and then filling in the blanks for the rest. Cognitive maps, just like our memories for events, are not photographic. They're a combination of actual verifiable details and a lot of inference.

Hippocampal Ripples and Sleep

Lex Fridman: What have you learned about how people represent locations?

Charan Ranganath: There's a lot of variability and a lot of disagreement. In a world of GPS and physical maps, people can learn from a survey perspective — being able to see everything from above. There's also the way of memorising routes rigidly. And there's another more integrative way — what's called a cognitive map, a sense of how everything relates to each other.

We have a small task we're working on right now. One of the things we're looking at is signals called ripples in the hippocampus — these bursts of activity that are synchronised with areas in the neocortex, in the default network. What we find is that those ripples seem to increase at navigationally important points — when you're making a decision or when you reach a goal. What we found in general in our MRI studies is that the more people can reduce the problem, whether it's space or any kind of decision-making problem, the less the hippocampus encodes it. It's very economical, oriented towards the points of highest information content and value.

Lex Fridman: Can you describe the encoding in the hippocampus and the ripples?

Charan Ranganath: There are these oscillations — waves that you basically see — these points of very high excitability and low excitability. During slow-wave sleep, the deepest stages of sleep, you see these very slow waves where it's very excitable and then very unexcitable, going up and down. On top of them you'll see these little sharp-wave ripples. When there's a ripple in the hippocampus, you tend to see a sequence of cells that resembles a sequence of cells that fired when an animal was actually doing something in the world. It's almost like a little compressed play of the sequence of activity in the brain that was taking place earlier. During those moments there's a little window of communication between the hippocampus and areas in the neocortex. I think that helps you form new memories, stabilise them, and really connect different things together in memory — allows you to build bridges between different events you've had.

So during sleep is when the connections are formed — the connections between different events. It's like you see me now, you see me next week, you see me a month later, and you start to build a little internal model of how I behave and what to expect of me. We think sleep is one of the things that allows you to figure out those connections, connect the dots, find the signal in the noise.

fMRI: How It Works

Lex Fridman: You mentioned fMRI — what is it and how is it used in studying memory?

Charan Ranganath: This is actually the reason I got into this whole field of science. When I was in grad school, fMRI was just really taking off as a technique for studying brain activity. What's beautiful about it is you can study the whole human brain, and you can basically do it without sticking anything into people's brains — very non-invasive. For me, being in an MRI scanner is like being in the womb. I just fall asleep if I'm not being asked to do anything.

You can have people watch movies while they're being scanned, or do tests of memory like giving them words to memorise. What MRI itself is, is a technique where you put people in a very high magnetic field — typical ones we would use would be three Tesla. You put somebody in and you get this very weak but measurable magnetisation in the brain, and then you apply a radio frequency pulse, which is a different electromagnetic field. You're basically using the water molecules in the brain as a tracer.

The part specific to fMRI is that blood is flowing to the brain, and when you have blood that doesn't have oxygen on it, it's a little bit more magnetisable than blood that does — because of the haemoglobin that carries the oxygen. The iron in the blood that makes it red — that haemoglobin, when it's deoxygenated, has different magnetic field properties than when it has oxygen. When you have an increase in local activity in some part of the brain, blood flows there, and as a result you get a lower concentration of deoxygenated haemoglobin, and that gives you more signal.

When you look at fMRI data it's not very impressive — it looks like these very pixelated maps of the brain, mostly kind of white. But these tiny changes in the intensity of those signals — about 1% — can be statistically very large effects for us. That allows us to see an increase in activity in some part of the brain when I'm doing some task like trying to remember something. I can use those changes to even predict whether a person is going to remember something later or not.

The coolest thing that people have done is to decode what people are remembering from the patterns of activity. Maybe when I'm remembering the house where I grew up, I might have one pixel that's bright in the hippocampus and one that's dark. If I'm remembering the car I used to drive when I was 16, I might see the opposite pattern. All that little stuff that we used to think of as noise — we can now think of almost like a QR code for memory, where different memories have a different little pattern of bright pixels and dark pixels. This really revolutionised my research.

You can start to get a whole state space of how a brain area is indexing all these different memories. What we could see is this little separation between how certain brain areas are processing memory for who was there, and other brain areas processing information about where it occurred, or the situation that's unfolding. And the hippocampus is just putting it all together into these unique things that are just about when and where it happened.

Let me try a different analogy. You've got a folder on your computer. I open it up, there's a bunch of files. I can sort those files alphabetically — now things that start with A are lumped together and things that start with Z are far apart. That's one way of organising the folder. But I could do it by date, and if I do it by date, things created close together in time are close. Every brain area or network contributes to memory by having a particular sorting scheme, and these QR codes from fMRI allow you to do that.

We've found that some networks of the brain sort information in memory according to who was there. We showed this in one of my favourite studies of all time, done by a former postdoc, Zach Reagh. He filmed two different people at two different cafes and two different supermarkets. What you could show is that in one particular network, you could find the same kind of pattern of activity every time I saw Alex in one of these movies, no matter where he was. And I could see another pattern that was common every time I saw a particular supermarket. And it didn't matter whether you were watching the movie or recalling the movie — it was the same kind of pattern that came up.

fMRI Limitations and Comparisons with Other Techniques

Lex Fridman: What are some interesting limitations and possibilities of fMRI?

Charan Ranganath: It's very slow relative to the brain. Fifty milliseconds is like an eternity to the brain — so much back-and-forth stuff happens in that time. In fMRI you can measure these magnetic field responses about six seconds after a burst of activity would take place.

One of the interesting things that's been discovered about fMRI is that it's not so tightly related to the spiking of neurons. We tend to think of computation as being driven by spikes — a burst of activity, either on or off. But sometimes what you can have is states where the neuron becomes a little more excitable or less excitable, and fMRI is very sensitive to those changes in excitability.

Most of the cells in the brain are not neurons — they're actually these support cells called glial cells. One big one is astrocytes, and they play a big role in regulating things. If one neuron is talking to another, you release a neurotransmitter like glutamate, and that gets another neuron active. But if you leave that glutamate in the synapse too long, you just excite the other neuron too much and you can start to get seizure activity. So you've got to suck it up. What happens is these astrocytes suck up the glutamate from the synapse and break it down and feed it back into the neurons so you can reuse it. That cycling is actually very energy intensive, and they need to work so quickly that they're working on metabolising glucose without using oxygen — anaerobic metabolism. So what we're really seeing in fMRI may not be the neurons themselves being active, but rather the astrocytes meeting the metabolic demands of keeping the whole system going.

One of the things we found in our work was that when we give people movies and stories to listen to, a lot of the action is in the very very slow stuff. If you're listening to a podcast, you're putting things together and building this internal model over several seconds. We filter that out when we look at electrical activity in the brain because we're interested in the millisecond scale — but there's almost massive amounts of information in the slow stuff.

Every technique gives you a limited window into what's going on. fMRI has huge problems — people lie down in the scanner, there are parts of the brain where you'll see gaping holes because you can't keep the magnetic field stable in those spots, there are parts where there's a vein that produces big increases and decreases in signal, respiration causes changes, lots of artefacts. But at the same time you're getting data you might not be able to get otherwise. Different techniques give you different kinds of advantages.

Major Discoveries in Memory Science

Lex Fridman: What kind of big scientific discoveries have been made throughout the history of the science of memory?

Charan Ranganath: There are so many. When I started writing the book I was like, oh my God, this is really interesting — how did we do all this stuff?

I would say some of the most important include: the first studies just showing how much we forget; showing how much schemas — organised knowledge about the world — increase our ability to remember information, just massively increase it; studies of expertise showing how chess experts can memorise so much in such a short amount of time because of the schemas they have for chess, but also showing that those lead to all sorts of distortions in memory.

The discovery that the act of remembering can change the memory — it can strengthen it, but it can also distort it if you get misinformation at the time of recall, and it can also strengthen or weaken other memories that you didn't even recall. Just this whole idea of memory as an ecosystem — I think that was a big discovery.

The idea of breaking up our continuous experience into these discrete events — that was a major discovery. You walk into the kitchen and you're like, why am I here? You end up grabbing some food from the fridge and then you go back and you're like, oh wait, I left my watch in the kitchen, that's what I was looking for. What happens is you have a little internal model of where you are and what you're thinking about, and when you cross from one room to another those models get updated. Now when you're in the kitchen you have to go back and mentally time travel to this earlier point to remember what you went there for.

In our research, one of the things we found is that as people get older, the activity in the hippocampus at these event boundaries tends to go down. And independent of age, if I give you a test of memory for stories after you've been scanned while watching a movie, you find this incredible correlation between the activity in the hippocampus at these singular points in time — these event boundaries — and your ability to remember a story later on. So it's marking this ability to encode memories, just these little snippets of neural activity.

There's also so much interesting stuff being discovered in sleep right now. And the QR code thing — being able to take fMRI data while you use a joystick to move around in virtual reality, take a rat and record from its hippocampus and prefrontal cortex with really new electrodes and have it move around on a trackball in the same virtual environment, get a person with epilepsy who already has electrodes in the brain and record from that person, and get a computational model — and then relate them all together by virtue of the state spaces you're measuring in neural activity across these different formats, different species, and the computational model. I just find that mind-blowing.

False Memories and Misinformation

Lex Fridman: How does déjà vu work?

Charan Ranganath: Déjà vu — surveys suggest about 75% of people report having had the experience at least once. It's this sense that I've experienced this moment sometime before, I've been here before. There are all sorts of variants of this — the French have all sorts of names for various versions.

There was a researcher named Wilder Penfield — and this goes back even earlier to Hughlings Jackson, this neurologist who did a lot of the early characterisations of epilepsy. One of the things he noticed was that in some epilepsy patients, right before they would get a seizure, they would have this intense sense of déjà vu. It's an artificial sense of familiarity — a sense of having a memory that's not there. What was happening was there was electrical activity in certain parts of these brains. Penfield later on, when he was trying to map out the brain to figure out which parts to remove and which to leave, would stimulate parts of the temporal lobes of the brain and find he could elicit the sense of déjà vu. Sometimes he'd actually get a memory that a person would re-experience just from electrically stimulating some part. Sometimes they'd just have this intense feeling of being somewhere before.

One theory I really like is that in higher-order areas of the brain, they're integrating from many different sources of input, and they're tuning themselves up every time you process a similar input. That allows you to get this fluent sense that you're very familiar with a place. There's a great set of studies done by Anne Cleary at Colorado State where she created virtual reality environments. She would put people in a virtual museum, and then put them in a virtual arcade, but the map of the two places was exactly the same — she just put different skins on them. People would often not have any conscious idea that the two were the same, but they could report this very intense sense of déjà vu. It's a partial match that's eliciting this sense of familiarity.

We think déjà vu is basically a byproduct of our mechanism of error-driven learning as we go through life — becoming better and better at processing things more and more efficiently. It's just an extra elevated version of that, firing for this artificial memory as if it's the real memory.

Lex Fridman: Artificial memory brings to mind false memories. How do false memories form?

Charan Ranganath: I like to say there's no such thing as true or false memories. Johnny Rotten from the Sex Pistols had a saying — I don't believe in false memories any more than I believe in false songs. The basic idea is that memories reflect bits and pieces of what happened as well as our inferences and theories. I'm a scientist and I collect data, but I use theories to make sense of that data. A memory is a mix of all these things.

Where memories can go off the deep end and become what we'd conventionally call false memories is when we fill in the gaps based on things we know that don't actually correspond to what happened. If I tell you a story about a person who's worried they have cancer and they see a doctor, and the doctor says things are very much like what you were afraid of — people will often remember that the doctor told the patient he had cancer, even if that wasn't in the story. They're infusing meaning into it. That's a minor distortion.

But sometimes things can really get out of hand. As I said, the act of remembering can change the memory. You can be exposed to misinformation, and when you remember the event you might remember some original information as well as some information about what I told you. If you're not able to tell the difference, that information gets mixed into the story you had originally. Then I give you some more misinformation, or you're exposed to some more information somewhere else, and eventually your memory becomes totally detached from what happened.

Sometimes you can have cases where — this is very rare, but you can do it in the lab — a significant chunk of people will fall for this: you can give people misinformation about an event that never took place, and as they keep trying to remember that event, they start to imagine, they start to pull up things from other experiences they've had, and eventually they can stitch together a vivid memory of something that never happened. Because they're not remembering an event that happened — they're remembering the act of trying to remember what happened, and basically putting it together into the wrong story.

Lex Fridman: This could probably happen at a collective level — this is probably what successful propaganda machines aim to do, creating false memory across thousands if not millions of minds.

Charan Ranganath: Absolutely. This is exactly what they do. All these foibles of human memory get magnified when you start to have social interactions. There's a whole literature on something called social contagion — basically when misinformation spreads like a virus. You remember the same thing that I did, but I give you a little bit of wrong information, and then that becomes part of your story of what happened. Once you and I share a memory — I tell you about something I've experienced and you tell me about your experience of the same event — it's no longer your memory or my memory, it's our memory. The misinformation spreads. And the more you trust someone, or the more powerful that person is, the more of a voice they have in shaping that narrative.

There's a great example from when John McCain and George W. Bush were in a primary. There were these polls where they would call voters but actually inserted some misinformation about McCain's beliefs on taxation, or something about personal matters — I don't remember exactly. They included misinformation in the questions they asked, like, how do you feel about the fact that he wants to do this? And people would end up becoming convinced he had these policy positions or personal characteristics that were not true, just based on the polls that were being used. It was a case where the people using misinformation were actually ahead of the curve relative to the scientists who were trying to study these effects.

Lex Fridman: It's not just about truth and falsehoods as intelligent reasoning machines — it's the formation of memories where they become visceral. You can rewrite history. If you look throughout the 20th century, some of the dictatorships — Nazi Germany, the Soviet Union — effective propaganda machines can rewrite our conceptions of history, how we remember our own culture, our upbringing. And there's probably some kind of social contagion happening there. Certain ideas initiated by the propaganda machine can spread faster than others. There's something about certain conspiracy theories that makes them really effective at spreading — something about the human mind eats it up and then uses that to construct memories almost as if they were there to witness whatever the content of the conspiracy theory is. Once you feel like you remember a thing, there's a certainty — it emboldens you. It's not just that you believe an idea is true or not. It's at the core of your being that you feel like you were there to watch the thing happen.

Charan Ranganath: People's sense of collective identity is very much tied to shared memories. If we have a shared narrative of the past, or even better if we have a shared past, we will feel more socially connected with each other. I will feel part of this group — they're part of my tribe — if I remember the same things in the same way.

You brought up this weaponisation of history, and it really speaks to one of the parts of memory which is that if you have a belief, and you have a goal in mind, you will find stuff in memory that aligns with it and you won't see the parts in memory that don't. A lot of the stories we put together are based on our perspectives.

I was reading Viet Thanh Nguyen — he wrote a book about the collective memory of the Vietnam War. He's a Vietnamese immigrant who was flown out after the war was over. He went back to his family to get their stories about the war, and they called it the American War, not the Vietnam War. That just kind of blew my mind, having grown up in the US and having always heard about it as the Vietnam War. But of course they call it the American War because that's what happened — America came in. That's based on their perspective, which is a very valid perspective.

The opportunities we can have in memory is if we bring groups together from different perspectives and allow them to talk to each other and allow ourselves to listen. You can see two situations in groups with memory. In one situation you have people who are very dominant who just take over the conversation, and what happens is the group remembers less from the experience and they remember more of what the dominant narrator says. But if you have a diverse group of people — diverse in any way you want to take it — and you give everyone a chance to speak and everyone's being appreciated for their unique contribution, you get more accurate memories and you get more information from it. Even two people who come from very similar backgrounds, if you can appreciate the unique contributions that each one has, you can do a better job of generating information from memory. That's a way to inoculate ourselves from misinformation in the modern world. But like everything else, it requires a certain tolerance for discomfort.

Lex Fridman: I think that's a technology problem that could be solved. If there's a little bit of interaction — kind, respectful, compassionate interaction with people who have a very different memory — that respectful interaction will start to intermix the memories and ways of thinking in ways that slowly move towards truth. But naturally, left to our own devices, we want to cluster up in a tribe.

Charan Ranganath: That's the human problem. I think a lot of the problems that come up with technology aren't the technology itself as much as the fact that people adapt to the technology in maladaptive ways. One of my fears about AI is not what AI will do but what people will do. Take text messaging — it's a pain in the ass to text people, at least for me. What happens is the communication becomes very spartan and devoid of meaning. It's just very telegraphic. That's people adapting to the medium.

When Google started to introduce autocomplete in emails, I started to use it and about a third of the time I was like, this isn't what I want to say. A third of the time I'd be like, this is exactly what I wanted to say. And a third of the time I was saying, well, this is good enough, I'll just go with it. What happens is it's not that the technology necessarily is doing anything so bad — it's just going to constrain my language because I'm just doing what's being suggested to me.

This is why my mantra for some of what I've learned about everything in memory is to diversify your training data. Otherwise, humans have this capability to be so much more creative than anything generative AI will put together, at least right now. But it can also go the opposite direction where people could become much much less creative if they just become more and more resistant to discomfort, resistant to exposing themselves to novelty and cognitive dissonance.

Lex Fridman: Do you think people are happier now than they were 50 years ago or 100 years ago?

Charan Ranganath: I think humans in general like to reminisce about the past and complain about the present. There's so much pleasure in saying life sucks, for some reason. That's why I love punk rock. But ultimately I think on net, on every measure, things are getting better.

Lex Fridman: Life is getting better, but I don't know if that necessarily tracks people's happiness. I wouldn't be surprised if people in hunter-gatherer societies are happier. I wouldn't be surprised if they're happier than people who have access to modern medicine and email and phones.

Charan Ranganath: I don't think there's a question whether you take hunter-gatherer folks and put them into modern day and give them enough time to adapt — they would be much happier. The question is a deeper biological one. Do we want to be like Werner Herzog's movie Happy People: A Year in the Taiga — do we want to be busy 100% of our time hunting, gathering, surviving, worried about the next day? Maybe that constant struggle ultimately creates a more fulfilling life. I don't know.

But I do think this: there was an interesting memory-related thing, which is that if you look at reinforcement learning, you're not learning every time you get a reward. If it's the same reward, you're not learning that much. You mainly learn if it deviates from your expectation of what you're supposed to get. You get a paycheck every month and you're kind of not even excited about it when you get it. But if they cut your salary, you're going to be pissed. And if they increase it — oh, I got a bonus. That adaptation, that ability — basically you learn to expect these things — I think is a major source of the way in which we're wired to strive and not be happy, to be in a state of wanting.

People talk about dopamine as this pleasure chemical, and there's a lot of compelling research to suggest it's not about pleasure at all — it's about the discomfort that energises you to seek a reward. You could give an animal that's been deprived of dopamine a reward and it'll enjoy it — it's pretty good — but it's not going to do anything to get it.

One of the things that we found in our research — I got into curiosity from a postdoc in my lab, Matthias Gruber — is that when we gave people a trivia question that they wanted the answer to, the more curious people were about the answer, the more activity in these dopamine-related circuits in the brain we would see. And that was not driven by the answer per se, but by the question. It was not about getting the information — it was about the drive to seek the information. But it depends on how you take that. If you get this uncomfortable gap between what you know and what you want to know, you could either use that to motivate and energise you, or you could use it to say, I don't want to hear about this, this disagrees with my beliefs, I'm going to go back to my echo chamber.

Lex Fridman: I like what you said — maybe we're designed to be in a kind of constant state of wanting.

False Confessions and Coercive Memory

Lex Fridman: In the legal system, false confessions — I remember reading 1984 where through torture you can make people say anything and essentially remember anything. How much can you really get people to force false memories?

Charan Ranganath: There's a lot of history of this in the criminal justice system. You might have heard the term "the third degree" — if you actually look it up historically, it was a very intense set of beatings, starvation, and physical demands placed on people to get them to talk. There's certainly a lot of work that's been done by the CIA in terms of enhanced interrogation techniques, and from what I understand, the research actually shows that they just produce what people want to hear, not necessarily the information that's being looked for.

One reason is that people just get tired of being tortured and say whatever. But another part of it is that you create a very interesting set of conditions. There's an authority figure telling you, we know you did this, we have witnesses saying you did this. Now you start to question yourself. Then they put you under stress — maybe they're not feeding you, making you cold, exposing you to music you can't stand. They're creating this physical stress, which starts to downregulate the prefrontal cortex. You're not as good at monitoring the accuracy of stuff. Then they start to get nice to you and say, imagine — I know you don't remember this, but maybe we can walk you through how it could have happened. They feed you the information. You're in this weakened mental state, being encouraged to imagine things by people who give you a plausible scenario. At some point, certain people can be very coaxed into creating a memory for something that never happened.

There are actually some pretty convincing cases out there where people confess to crimes they just didn't do, and objective evidence came out later on. The basic recipe is: feed people the information you want them to remember, stress them out, have an authority figure pushing this information on them, motivate them to produce the information you're looking for. That pretty much over time gives you what you want. It's really tragic that centralised power can use these kinds of tools to destroy lives.

Memory and Heartbreak

Lex Fridman: Since there's a theme about music throughout this conversation — one of the best topics for songs is heartbreak, love in general. Why and how do we remember and forget heartbreak?

Charan Ranganath: Oh, that's so hard. Part of this is we tend to go back to particular times that are the more emotionally intense periods. Memory is designed to capture things that are biologically significant, and attachment is a big part of biological significance for humans. Human relationships are super important. Sometimes that heartbreak comes with massive changes in your beliefs about somebody — if they cheated on you, or regrets about things you've done wrong, and you ruminate.

I had this — my first pet, a cat we got as a wedding present, died of FIP when it was four years old. I would just see her everywhere around the house. We got another cat, then we got a dog. The dog eventually died of cancer, and the cat just died recently. We got a new dog because I kept seeing the old dog around and I was just so heartbroken. But I still remember the pets that died — it just comes back to you. There's something about attachment that's just so crucial.

Sometimes it's also not just about the heartbreak but about the positive aspects of it. The loss comes not only from the fact that the relationship is over, but you had all of these good things before that you can now see in a new light. One of the things I found from my clinical background that really gave me a different perspective on memory is that so much of the therapy process was guided towards reframing — getting people to look at the past in a different way. Not by imposing an interpretation, but just offering a different perspective, maybe one that's more optimised towards learning and appreciation, or gratitude. That gives you a way of taking dark experiences and using them as training data to grow in new ways.

Lex Fridman: I often go back to this moment in the show Louie with Louis C.K., where he's all heartbroken about a breakup with a woman he loves, and an older gentleman tells him that that heartbreak is actually the best part — because you get to intensely experience how valuable this love was. He says the worst part is forgetting it — actually when you get over the heartbreak, that's the worst part. I sometimes think about that because having the love and losing it, the losing it is when you sometimes feel it the deepest. It's a good moment to viscerally experience the memories of something you appreciate even more because you don't have it.

Charan Ranganath: After I turned 50 I think of death all the time. I probably have fewer years ahead of me than behind me. I think about what are the memories I want to carry with me for the next period of time. I'm reminded of talking to somebody who's a Buddhist — the whole idea of Buddhism is renouncing attachments, staying out of the world of memory and staying in the moment. They talked about how do you renounce attachments to the people that you love, and they said, well, I appreciate that I have this moment with them, and knowing that they will die makes me appreciate this moment that much more.

Lex Fridman: I meditate on mortality every day. But I also appreciate the full deep roller coaster of suffering involved in life — the little and the big. I'm not sure about the Buddhist kind of removing yourself from the world, or the Stoic removing yourself from the world of emotion. I'm torn about that one.

Charan Ranganath: This is where Hinduism and Buddhism, or at least some strains of them, differ. In Hinduism — if you read the Bhagavad Gita — the philosophy is not one of renouncing the world, because the idea is that not doing something is no different than doing something. What they argue is that you don't want to renounce action, but you want to renounce the fruits of the action. You don't do it because of the outcome — you do it because of the process, because the process is part of the balance of the world that you're trying to preserve. I really think about that from time to time in terms of letting go of this idea of does this book sell, or trying to impress you and get you to laugh at my jokes, and just being more like — I'm sharing this information with you and getting to know you. But it's hard, because we're so driven by the outcome.

Lex Fridman: You're just part of the process of telling the joke, and if I laugh or not, that's up to the universe to decide.

Memory and the Perception of Time

Lex Fridman: How does studying memory affect your understanding of the nature of time?

Charan Ranganath: In some sense, especially the farther we go back — your sense of how different one hour ago feels from two hours ago, you'd probably say pretty different. But if I ask you to go back one year ago versus one year and one hour ago, it's the same difference in time but it won't feel very different. There's this kind of compression that happens as you look back farther in time. That's kind of why when you're older, the difference between somebody who's 50 and 45 doesn't seem as big as the difference between 10 and 5 when you're a child.

During the pandemic, I just decided to ask people in my class — we were doing remote instruction — do you feel like the days are moving by slower or faster or about the same? Almost everyone said the days were moving by slower. Then I asked, do you feel like the weeks are passing by slower, faster, or the same? The majority said the weeks were passing by faster. According to the laws of physics, that doesn't make any sense. But according to memory it did, because people were doing the same thing over and over in the same context. Without that change in context, their feeling was that they were in one long monotonous event. But then at the end of the week you look back and say, well, what happened? You have no memories of what happened, so the week just went by without your noticing. But that week went by during the same amount of time as an eventful week where you might have been going out, hanging out with friends, on vacation. Nothing happened because you were doing the same thing over and over. Memory really shapes our sense of time, and it does so in part because context is so important for memory.

Lex Fridman: That compression is an interesting process. When I think about when I was 12 or 15, I just fundamentally feel like the same person. It makes me feel like it's all connected — not just amongst humans spatially, but back in time. There's a kind of timelessness to life. To me it all feels the same. The details of that time are not useful to understanding the core of the thing.

Charan Ranganath: Maybe what it is is that you really like to see connections between things. When you start recalling the past and seeing the connections between the past and present, you have this kind of web of interconnected memories. In that sense the present is with you.

What struck me about what you said is that your 16-year-old self was probably very complex. When you hear a song that you used to play before you would go do a sports thing or something, you might not think of yourself as an athlete, but once you mentally time travel to that particular thing, you open up this little compartment of yourself that wasn't there before. Dan Schacter's lab did this really cool study where they would ask people to either remember doing something altruistic or imagine doing something altruistic, and that act made them more likely to want to do things for other people. So that act of mental time travel can change who you are in the present. We tend to think of memory in this very deterministic way — that I am who I am because I have this past. But we have a very multifaceted past and can access different parts of it and change in the moment based on whatever part we want to reach for.

Nostalgia

Lex Fridman: How does nostalgia connect into this — this desire and pleasure associated with going back?

Charan Ranganath: My friend Felipe De Brigard wrote about this and it just blew my mind. The word nostalgia was coined by a Swiss physician who was actually studying traumatised soldiers. He described nostalgia as a disease — it was bringing these people extraordinary unhappiness because they were remembering how things used to be.

It's very complex. As people get older, nostalgia can be an enormous source of happiness. Being nostalgic can improve people's moods in the moment. But it just depends on what they do with it, because what you can sometimes see is nostalgia having the opposite effect — thinking those were the good old days and those days are over. America used to be so great and now it sucks, or my life used to be so great when I was a kid and now it's not. You're selectively remembering things, and we don't realise how selective our remembering self is.

I lived through the 70s — it sucked. There were gas lines, people were worried about Russia, nuclear war, all sorts of problems going on. This idea that people have about the past can be very useful if it brings you happiness in the present, but if it narrows your worldview in the present and you're not aware of those biases, it can be toxic — either at a personal level or at a collective level.

Brain-Computer Interfaces

Lex Fridman: What are your thoughts about BCIs — brain-computer interfaces — and the work going on with Neuralink?

Charan Ranganath: I can't say specifics about the company because I haven't studied it that much. But there are two parts of it. They're developing some really interesting technology with these surgical robots and things like that. BCI though has a whole lot of innovation going on. I'm not necessarily seeing scientific evidence from Neuralink — maybe I'm just not looking for it — but I'm not seeing the evidence that they're anywhere near where the scientific community is. There are lots of startups doing incredibly innovative stuff. One of my colleagues, Sergey Stavisky, is just like a genius in this area. Speech prosthetics incorporating decoding techniques with AI, movement prosthetics — the rate of progress is just enormous.

Part of the technology is having good enough data and understanding which data to use and what to do with it. Part of that has really resulted in some real breakthroughs in neuroscience. There are lots of new technologies like Neuropixels, for instance, that allow you to harvest activity from many many neurons from a single electrode.

BCI is much much bigger than Neuralink, and there's just so much innovation happening. The interesting question is — I was talking to Sergey a while ago about this — a lot of language is not just what we hear and what we speak, but also our intentions and our internal models. Are you really going to be able to restore language without dealing with that part of it? And he brought up a really interesting question about the ethics of reading out people's intentions and understanding of the world, as opposed to the more concrete parts of hearing and producing movements.

Lex Fridman: When we talk about language and BCIs, what we mean is getting signal from the brain and generating language — say you're not able to actually speak, it's a kind of linguistic prosthetic that can speak for you exactly what you wanted to say. And then the deeper question is, well, saying something isn't just the words — it's also the intention behind it, the feeling behind it. Is it ethical to reveal that full context of what's going on in our brain?

Charan Ranganath: Our thoughts — is it ethical for anyone to have access to our thoughts? Right now the resolution is so low that we're okay with it even doing studies. But if neuroscience has a few breakthroughs to where you can start to map out the QR codes for different kinds of thoughts — political thoughts, for instance — that's really interesting. I think ultimately the more transparency there is about the human mind, the better it is. But there could always be intermediate battles with how much control a centralised entity like a government has, what the regulation is, what's legal and illegal. Freedom of thought, which is a very important liberty at the core of this country and probably humanity, starts to get awfully tricky when you start to be able to collect those thoughts.

Modifying Memories

Lex Fridman: Do you think we'll be able to modify memories? How far away are we from understanding the different parts of the brain well enough to figure out how to adjust a memory — keep the people but change the place, that kind of thing?

Charan Ranganath: In some sense we know we can do it just behaviourally — under certain conditions you can give misinformation and change the people and the places and so forth. On the crude level, there's a lot of work being done on a phenomenon called reconsolidation — the idea that essentially when I recall a memory, the connections between the neurons in that cell assembly become more modifiable. Some people have used techniques to try to reduce the physical visceral component of fear memories when they're being activated. As an outsider looking at the data, I think it's mixed results.

Part of the issue is that you need somebody to actually fully recall that traumatic memory in the first place in order to actually modify it. And then what is the memory that is the key part of the problem? People can sometimes look at this like behaviourists and say the memory is like I've given you A and you produce B. But I think that's a very bankrupt concept about memory. It's much more complicated than that.

One of the things that was so hard when we started studying naturalistic memory from movies was that if I show you a movie and we both watch it, and you recall everything that happened and I recall everything that happened, we might take a different amount of time, we might use different words, and yet to an outside observer we might have recalled the same thing. It's not about the words necessarily, and it's not about how long we spent — there's something deeper that's there. How do you understand that thought?

AI, Self-Driving Cars, and Human Attention

Lex Fridman: I once saw somebody talking in a discussion between neuroscientists and AI people, and he was saying that the problem with self-driving cars in cities as opposed to highways was that the car was okay at doing the things it's supposed to, but when there were pedestrians around it couldn't predict the intentions of people. That unpredictability of people was the problem in the self-driving car design.

Charan Ranganath: What do you think about that?

Lex Fridman: I spent a huge amount of time watching pedestrians, thinking about what it takes to detect the intention of a pedestrian — really of a human being in the context of having to cross the street. I think it's a window into how complex social systems are that involve humans. I would just stand there and watch intersections for hours. What you start to figure out is every single intersection has its own personality. There's a history to that intersection. Jaywalking — certain intersections allow jaywalking a lot more because there are regulars who get off the subway and start crossing on a red light every single day. Then there are people who don't show up to that intersection often and they're looking for cues of how to behave. If a few people start to jaywalk, they will follow. There's a dynamic to that intersection, a spirit to it. If you look at Boston versus New York versus a rural town versus Austin, there are different personalities citywide, area-wide, and at different intersections.

For a car to be able to determine that is tricky. What machine learning systems are able to do well is collect a huge amount of data. For us it's tricky because we get to understand the world with very limited information and make decisions grounded in this big foundation model we've built of understanding how humans work. AI, in the context of driving, if you just collect a huge amount of data and compress it into a representation of how humans cross streets, it's probably all there — in the same way that Noam Chomsky said AI can't write convincing language without understanding language, and yet large language models without quote-unquote understanding can generate very convincing language.

I think what the process of compression from a huge amount of data does is in fact understand deeply. In order to generate one letter at a time, one word at a time, you have to understand the cruelty of Nazi Germany and the beauty of sending humans to space and all of that — in order to generate "I'm going to the kitchen to get an apple" grammatically correctly. You have to have a world model that includes all of human behaviour. I think AI that drives a car, if it has enough data, will be able to form a world model that will be able to predict correctly what the pedestrian does.

But when we as humans are watching pedestrians, we slowly realise how complicated this is. When you start to self-reflect on driving, you realise driving is really complicated. One of them is determining who around you is an aggressive driver — potentially dangerous. Once you become a great driver, you can see it a mile away. This guy's going to pull a move in front of you. He's way back there but you know it's going to happen. And I don't know what cue it is, but it's like a glowing obvious symbol even in the periphery vision. We're usually just looking forward when we're driving, but we're using peripheral vision to figure stuff out. It's a little puzzle we're usually only allocating a small amount of cognitive attention to. AI just has a fundamentally different suite of sensors in terms of the bandwidth of data coming in, and that allows you to form the representation and perform inference on it.

Charan Ranganath: One of the things that's currently missing — even though OpenAI just recently announced adding memory — is how important and how difficult it is to add some of the memory mechanisms we've seen in humans to AI systems. Superficially, not that hard. But in a deeper level, very very hard, because we don't understand episodic memory.

One of the oldest dilemmas in computational neuroscience is what Steve Grossberg called the stability-plasticity dilemma — when do you say something is new and overwrite your pre-existing knowledge versus go with what you had before and make incremental changes? Part of the problem with things like designing an LLM is that especially for English there are so many exceptions to the rules. If you want to rapidly learn the exceptions, you're going to lose the rules. If you want to keep the rules, you have a harder time learning the exceptions.

David Marr was one of the early pioneers in computational neuroscience, and then Jay McClelland and my colleague Randy O'Reilly and others like Neil Cohen started to come up with the idea that maybe that's part of what the human brain is doing. We have this fairly simple system which just says this happened once at this point in time — episodic memory — and then we have this knowledge accumulated from our experiences as semantic memory. When we encounter a situation that's surprising and violates all our previous expectations, we can form an episodic memory here. The next time we're in a similar situation, we can supplement our knowledge with this information from episodic memory and reason about what the right thing to do is. That gives us this enormous flexibility to stop on a dime and change without having to erase everything we've already learned.

Can you build something like that? Computational neuroscience people say, well, you just record a moment and you're done. But when do you record that moment? How much do you record? What information do you prioritise? These are the hard questions we're still trying to figure out in people. And then you start to think about all these mechanisms in the brain for figuring out some of these things — it's not just one, it's many of them interacting with each other. And then you take not only the episodic and the semantic, but the motivational survival things — the fight-or-flight responses, the reward motivation. Those things are absent from AI. I frankly don't know if we want it. I don't necessarily want a self-motivated LLM.

A friend of mine, Sam Gershman — I might be missing the quote exactly — basically said, if I wanted to train a typical AI model to make me as much money as possible, the first thing it might do is sell my house. It's not even just about having one goal or one objective — it's having all these competing goals and objectives, and then things start to get really complicated.

Lex Fridman: It's all interconnected. Even the thing you've mentioned — recording a moment — it's difficult to express concretely what a moment is. To record a moment you might have to include everything: all the emotions involved, all the context, all the social connections, all the visual and sensory experience, all the history that came before. We somehow take all that and compress it and keep the useful parts and integrate it into our whole narrative. Then each individual has their own little version of that narrative, and then we collide in the social way and adjust it and evolve.

Charan Ranganath: Even if we want to go super simple — Tyler Bonnen, a postdoc collaborating with me, actually studied computer vision at Stanford. One of the things he was interested in is people who have brain damage in areas thought to be important for memory but who also seem to have some perception problems with particular kinds of object perception. He went back to computer vision and said, let's take the best state-of-the-art computer vision models and give them the same kinds of perception tests we were giving to these people. He found that the computer vision models would just struggle with certain images, and adding more parameters or more layers didn't help — the architecture didn't matter. And those were the exact ones where humans with particular damage to an area called the perirhinal cortex were struggling.

Then he found that it only happened if people had enough time — they could make those discriminations. But without enough time, if they just got a glance, they were just like the computer vision models. So maybe let's look at people's eyes. A computer vision model sees every pixel all at once. We never see every pixel all at once — even if I'm looking at a screen, I'm grabbing little points on the screen by moving my eyes around, getting a very high-resolution picture of what I'm focusing on and a lower-resolution picture of everything else. Allowing people to move their eyes and integrate that information gave them something that the computer vision models weren't able to do. Somehow integrating information across time and getting less information at each step gave you more out of the process.

Lex Fridman: The process of allocating attention across time seems to be a really important process. Even the breakthroughs you get with machine learning — "attention is all you need," the transformer — it's about attention. Attention is at the core of what it means to be intelligent, what it means to process the world, integrate all the important things, discard all the unimportant things. Attention is probably at the core of memory too. There's so much sensory information, so much going on, to filter it down to almost nothing and just keep those parts — and then whenever there's an error, to adjust the model such that you can allocate attention even better to new things.

Charan Ranganath: Attention is a weird one. People are terrible at detecting changes that can happen in the environment if they're not attending in the right way, if their predictive model is too strong. You have these weird things where the machines can do better than the people. The machines make different kinds of mistakes than the people do. I will never be convinced that we've replicated human intelligence unless the simulator is making exactly the same kinds of mistakes that people do. People make characteristic mistakes, they have characteristic biases, they have characteristic heuristics. I have yet to see evidence that ChatGPT will do that.

ADHD and Memory

Lex Fridman: Since we're talking about attention, is there an interesting connection between ADHD and memory?

Charan Ranganath: It's interesting for me because when I was a child I was told — I don't know if it came from a school psychologist or just from teachers who hated me — that I had ADHD. This was in the 70s, so basically they said he has poor motor control and he's got ADHD. There were social issues too — I could have been put a year ahead in school but they said he doesn't have the social capabilities. I still ended up being an outcast even in my own grade. My parents got me on a diet free of artificial colours and flavours, because that was the thing people talked about back then.

I've come to appreciate now that I have many of the characteristics — if not full-blown ADHD. Rejection sensitivity, impulsive behaviour — I could tell you about all sorts of fights I've gotten into in the past.

ADHD is fascinating because right now we're seeing more and more diagnoses of it. I don't know how much of that is based on inappropriate expectations especially for children, and how much is based on true maladaptive tendencies. What we do know is that ADHD is associated with differences in prefrontal function. Attention can be both more distractable — you have a harder time focusing on what's relevant and shift too easily — but then once you get on something you're interested in, you can get stuck. Attention is this beautiful balance of being able to focus when you need to focus and shift when you need to shift. That balance seems to be disrupted in ADHD. As a result, memory tends to be poor in ADHD, but it's not necessarily because there's a traditional memory problem — it's more because of this attentional issue. People with ADHD often will have great memory for the things they're interested in and just no memory for the things they're not.

Lex Fridman: Is there advice from your own life on how to learn and succeed given the characteristics of your own brain?

Charan Ranganath: I'm still trying to figure out the flourishing per se, but for education — being in science is enormously enabling of ADHD. You're constantly looking for new things, constantly seeking that dopamine hit. They tolerate your being late for things. Nobody's going to die if you screw up. It's not like being a doctor where you have to be much more responsible and focused. You can just freely follow your curiosity, which is great.

What I'd say is that I'm learning now about how to structure my activities more. Email is the big one that kills me right now — I'm constantly shifting between email and my activities. What happens is I don't actually deal with the email, I just look at it and get stressed because I'm like, oh, I have to think about this, let me get back to it, and then I go back to something else. So I've just got fragmentary memories of everything. What I'm trying to do is set aside time — this is my email time, this is my writing time, this is my goofing-off time. And blocking these things off. You give yourself the goofing-off time. Sometimes I have to be flexible and say, okay, I'm definitely not focusing, I'm going to give myself the downtime. It's an investment in my attention later on, not wasting time.

Lex Fridman: I'm very much with Cal Newport on this — he wrote Deep Work and a lot of other amazing books. He talks about task switching as the thing that really destroys productivity. Switching — it doesn't even matter from what to what — checking social media, checking email, switching to a phone call and then doing work. Even switching between papers if you're reading. Because curiosity and the dopamine hit from the attention switch — limiting that, because otherwise your brain is just not capable of really loading it in, doing that deep deliberation required to remember things and really think through things.

Charan Ranganath: You probably see this in AI conferences, but definitely in neuroscience conferences — it's now the norm that people have their laptops out during talks. Conceivably they're writing notes, but in fact what often happens is you're checking email or working on your own talk. I have this illusion that I'm paying attention, and then I'm going back and forth, and what happens is I don't remember anything from that day. It just kind of vanished. Because what happens is I'm creating all these artificial event boundaries, losing all this executive function every time I switch. I'm getting a few seconds slower and catching up mentally to what's happening. Instead of being in a mode where you're meaningfully integrating everything and predicting and generating this rich model, I'm just catching up.

There's great research by Melina Uncapher and Anthony Wagner on multitasking that talks about just how bad it is for memory, and it's becoming worse and worse of a problem.

Music, Creativity, and Memory

Lex Fridman: You're a musician — how did you get into music? What made you first fall in love with it?

Charan Ranganath: I started playing music when I was doing trumpet in school for a school band. I would just read music and play — I was pretty decent at it, not great. Then in high school I grew up with MTV and started seeing all this stuff. I got into metal early on — I always reacted to things that were loud and had a beat. Everything from Sgt. Pepper's by The Beatles to Led Zeppelin II — my parents had both those albums. Then the Police, Ghost in the Machine. Then I got into metal — Def Leppard, AC/DC, Metallica, went way down the rabbit hole of speed metal.

At that time I thought, why don't I play guitar? I took lessons and stuff like that, but it was different from trumpet because with trumpet I was reading sheet music. With guitar I was learning by looking at tablature — you see a drawing of the fretboard with numbers showing where you're supposed to put your fingers. It's kind of like paint by numbers. I learned it in a completely different way but I was still terrible at it.

It wasn't until I really got into punk. I saw Sonic Youth and it just blew my mind, because they violated the rules of what I thought music was supposed to be. I was like, this doesn't sound right — these are not power chords, this isn't just a shouty verse and then a chorus. It's just weird. And then it occurred to me: you don't have to write music the way people tell you it's supposed to sound. That just opened up everything for me.

I was playing in a band and struggling with writing music because I would try to write like whatever was popular at the time, or whatever sounded like other bands I was listening to. Somehow I morphed into just grabbing a guitar and doing stuff. I realised part of my problem with doing music before was I didn't enjoy trying to play stuff that other people played — I just enjoyed music dripping out of me, spilling out. So I started to say, what if I don't play a chord? What if I just play notes that shouldn't go together and mess around with stuff? What if I don't do four beats — one two three four, one two three four — what if I go one two three four five, one two three four five? I started messing around with time signatures.

Then I was playing in a band with a great musician, Brent Rittell, and he taught me about arranging songs. What if we take this part and instead of making it go back and forth, we make it like a circle? Or what if we make it like a straight line, or zigzag, make it nonlinear in interesting ways? And then the whole world sort of opens up.

Time signatures — we are so brainwashed to think in 4/4. Every rock song you can think of is almost in 4/4. Think of Money by Pink Floyd — you feel like it's in 4/4 because it resolves itself, but it resolves on the last note of the riff, which is actually the first note of the next measure. So it's got seven beats instead of eight. You're thinking in four because that's how we're used to thinking. The music flows a little bit faster than it's supposed to, and you're getting a little bit of prediction error every time. Once I got used to that, I was like, I hate writing in 4/4. Everything just feels better if I do it in 7/4, or if I alternate between four and three. Jazz music does so much interesting stuff with this.

Lex Fridman: What is math rock?

Charan Ranganath: The genre we used to play in was called math rock. Instead of playing four beats in every measure, you might go with five or seven, and you might arrange it in weird ways — three measures of verse, then one of something else, then five measures of chorus, then two measures of something else. You could just mess around with everything.

Lex Fridman: What does that feel like to listen to?

Charan Ranganath: There's something about symmetry and patterns that feel good and relaxing, like home. And disturbing that — it can be quite disturbing. A lot of my style of songwriting is very much about repetitive themes but messing around with structure. I'm not a great guitarist technically, so I don't play complicated stuff. But often what I find is having a melody and then adding some dissonance to it — just enough — and then adding some complexity that gets you going just enough. I have a high tolerance for that kind of dissonance and prediction error. I think some people get scared of that discomfort, and I really gravitate towards it.

Lex Fridman: What's the name of your band?

Charan Ranganath: The cover band I play in is called Pavlov's Dogs. It's a band of mostly memory researchers and neuroscientists. One of your MIT colleagues, Earl Miller, plays bass.

Lex Fridman: What kind of songs do you guys do?

Charan Ranganath: Mostly late 70s punk and 80s new wave and post-punk. Blondie, the Ramones, the Clash. I sing Age of Consent by New Order and Level 42. We have a female singer now — Carrie Hoffman, who does Blondie amazingly well, and we do Gigantic by the Pixies. Paula Croc does that one.

Lex Fridman: Which song do you love to play the most?

Charan Ranganath: One we do with Pavlov's Dogs that I really enjoy is I Wanna Be Your Dog by Iggy and the Stooges. Which is perfect because we're Pavlov's Dogs, and Pavlov of course basically created learning theory. But also — that song just devolves into total noise and I just like fall on the floor and generate feedback. In the last version I think I actually have a guitar made of aluminium that I got made — I thought this thing's indestructible — and I kind of was moving around, had it upside down and all this stuff to generate feedback, and I think I broke one of the tuning pegs. I managed to break an all-metal guitar. Go figure.

What Is Most Beautiful About the Human Mind

Lex Fridman: A bit of a big ridiculous question, but what do you love most about the human mind? When you look at the fMRI scans, the behavioural stuff, the electrodes, the psychology, the neurobiology — when you look at all of it, what is most beautiful to you?

Charan Ranganath: I think the most beautiful but incredibly hard to put your finger on is this idea of the internal model. There's everything you see and everything you hear and touch and taste — every breath you take — but it's all connected by this dark energy that's holding that whole universe of your mind together. Without that it's just a bunch of stuff. Somehow we put that together and it forms so much of our experience. Being able to figure out where that comes from and how things are connected — to me that's just amazing. Just this idea that the world in front of us, we're only sampling this little bit and trying to take so much meaning from it, and we do a really good job — not perfect, but that ability to me is just amazing.

Lex Fridman: It's funny you said dark energy, because in astrophysics you look out there and you look at dark matter and dark energy — this loose term assigned to a thing we don't understand, which makes up the majority of the universe and helps make the equations work. In the same way it seems like there's that kind of thing in the human mind that we're striving to understand.

Charan Ranganath: One of the reasons I wrote the book is that I really felt like people needed to hear from scientists. COVID was just a great example — people weren't hearing from scientists. One of the things I think people didn't get was the uncertainty of science and how much we don't know. Every scientist lives in this world of uncertainty. When I was writing the book I just became aware of all of these things we don't know.

I think of physics a lot. I think of this idea of the overwhelming majority of the stuff that's in our universe that cannot be directly measured. At some point I had this aha moment where I was like, to be aware of that much that we don't know and have a beat on it and be able to go towards it — that's one of the biggest scientific successes I could think of. You are aware that you don't know about this gigantic section, this overwhelming majority of the universe. The more what keeps me going is realising the changing scope of the problem and figuring out, oh my God, there's all these things we don't know that I thought I knew. Science is all about assumptions. Have you ever read The Structure of Scientific Revolutions by Thomas Kuhn?

Lex Fridman: Yes — that's like my only philosophy I've really read, but it's so brilliant in the way it frames this idea of assumptions being core to the scientific process, and the paradigm shift comes from changing those assumptions.

Charan Ranganath: Finding out this whole zone of what you don't know — to me that's the exciting part.

Lex Fridman: Well, you are a great scientist and you wrote an incredible book. Thank you for doing that, and thank you for talking today. You've decreased the amount of uncertainty I have just a tiny little bit, and revealed the beauty of memory. This has been a fascinating conversation. Thank you for talking today.

Charan Ranganath: Thank you. It's been a blast.