Latest / Incredible Mind / Knowledge and the Brain (& AI): Yoed Kenett

Transcript

Knowledge and the Brain (& AI): Yoed Kenett · Incredible Mind

• speaker-0: No. speaker-1: You were made for creativity, learning, problem solving, and of course, action. On the Incredible Mind podcast, you'll hear from neuroscientists, psychologists, doctors, artists, and others whose work expands our understanding of how the brain works, where things can go wrong, and what simple everyday steps you can take to make the most of the unique intelligence you were born with. The Incredible Mind I'm talking about is yours. Where does knowledge live in our brains? How is it organized? And how do we mentally connect different areas of knowledge together or mix and match them to solve problems and come up with new ideas? In this fascinating conversation with neuroscientist Yoid Kennett, he answers these questions and more, sharing insights into creativity and knowledge, aging in the brain, and how the limits of AI's knowledge impact its users, adults and children alike. If you like what you hear, Please subscribe wherever you're listening and share this episode with friends and family. Let's get started. Welcome everyone back to Incredible Mind. This is the start of season two, which is all about learning and knowledge and the different ways that we can grow our own our own knowledge banks for use in things like creativity. So today we are speaking with Yod Kennett. Yoid is an associate professor at Technion, also known as the Israeli Institute of Technology. Welcome, Yoid. speaker-0: thank you. I'm really happy to be here open the second season. That's a real honor, especially since all of my work is exactly on on what your podcast is about. So I'm really, really excited to to talk to you today. speaker-1: Me too. So so tell us a little bit, what are the general themes of your research? speaker-0: Great question. I am a cognitive neuroscientist. I like to understand how people think. And especially in what we sometimes call high level cognition or complex cognition. There's not a really good, clear definition, but it usually is about the things that make us interesting, like creativity, intelligence, aesthetics, etc. And my research centers around in principle the role of knowledge in complex behavior. And most of my research is is on creativity, memory, associative thinking and question asking. And I study these topics at the cognitive and neural levels in typical and clinical populations and also lately in AI. speaker-1: Fantastic. Now, given the breadth of your research, I'm just gonna let the audience know that we are probably gonna hear from Yoid twice. We're gonna cover two major topics, the first one being knowledge and the second one being question asking. Right. So we're gonna kick things off with knowledge. My first question for you, Yoid, is how does our brain receive and encode knowledge with some talking at the cognitive level? speaker-0: That's a really really deep long question. A really good a really good question. But a question that there's actually still I would say a lot of lengths to it. I wanna if I go sort of zoom out as much as I can in in in research on memory in general, there's different types of of memory that we can talk about, but in what is called declarative memory. This is memory that we can talk about. Then we can talk about two different types of memory or memory systems. This goes the way back to decades ago for cognitive theory on on different types of memory systems, where in principle we can talk about episodic memory, which is the memory of the experiences that what we are experiencing in real time. Like we're talking right now and there's multi sensory perception and things that are going on. And then the other is called semantic memory, which is about knowledge that is stored that classically has been considered a void of context and time. The typical example is I know that Paris is the capital of France, but I don't remember who said that to me or where I read it or when I knew this fact. Static memory. There has been a revision in that sort of perception and the the the line between these two sort of systems is blurring over the years. And the argument is that we experience things in real time and those memories slowly become more and more what some people call semanticized. They become more and more knowledge focused. I use the word knowledge because semantics is studied in linguistics and is studied in memory kind of psychology. And I I've been constantly been struggling with that ambiguity and because I my work is cuts across fields and a lot of times different people are asking me questions about these terms. So my perception and my interest is knowledge in general and I think that our memory systems are multidimensional and we can talk about how to study that aspect. My postdoc mentor Sharon Thompson Shield has a fantastic paper where on this exact topic. When we think about something, then all these different types of signals or dimensions are activated. Some of it is the knowledge semantics of it. What is it? But then there's also how do I use this to X or sort of more like emotional attachment of the memories that I have for that information. So when we think about an object or an idea or a concept. A lot of magical things are happening very quickly without our understanding of what's going on. So this is why I think knowledge is a better term, the information or what I'm interested in studying. But a lot of time people will talk about this as semantic memory. And then how is that encoded in the cognitive and neural levels? So there's a lot of theory on semantic memory, but the most classic theory. proposed by Collins and Luftus in 1975, argues that our knowledge or semantic memories organize like a network or a space or a map of knowledge where concepts or ideas are organized according to similarity. So things that are related to each other or shared information, they will be closer in that space. Already feeding you things I want to talk about later and how we can study that. But this is the idea. And then when we think about something, whatever it is, then that in is somehow activated in our memory system. And then there is some sort of activation of information. It's called spreading activation. This is where associations kick in. So if I think about a table, then I will think about a chair, and I will think about a desk, and I will think about a lot of things. So because all of these concepts are related to each other and close to each other in that space. get this sort of diffused magic activation, cognitive activation. That's the theory in cognition. At the neural level, we have similar ideas, how the brain stores information across the brain and how our frontal parts of the brain have like a search engine or it's called semantic attention or semantic control. That can sort of retrieve that information. But these two levels, cognitive and neural, they don't necessarily align yet. This is still work to be done, how the brain realizes our information. So there's a lot of questions on how our brains and cognitive systems store memory. Okay. There's a lot of frontier there. speaker-1: My basic understanding, and this is super basic because I am not a neuroscientist, is that our information is or our knowledge and memory is stored in our gray matter, especially on the outer part of our cortex. Is that accurate? speaker-0: Well, everything is in a way stored in what you call the gray matter, which is just neurons on in the cerebral cortex. Yes. Which looks like a gray matter, especially in MRI scans of our brain. But everything is stored there. And then we have deeper brain systems that are also relevant for memory. Then you cannot say the word hippocampus, which is the most important brain region for memory. And that's deep inside the brain. And we talked before I mentioned episodic memory. So every aspect of information about an episode that we are experiencing goes through, enters this hippocampus, and there there is some sort of neural stamping of that memory. So the hippocampus is critical for connecting different sources of information into a single event. And then that information leaves the hippocampus and is stored across the brain in in gray matter or neurons. So there's a whole complex dynamics of information. This right, the brain is a complex magical system that we don't really understand. We still understand very little of the complexity. And this goes back to the title of my lab is the cognitive complexity lab. So I'm really much interested in complex behavior, complexity of systems and dynamics and how we can study this because these are the most exciting, interesting things. And our brain is a really complex system, but our also our software, our mental capacities are really, really complex. And I use methods that come in from different computational fields from mathematics to computer science to try to better quantify that complexity and then study those in empirical research to reverse engineer how people are thinking and try to look at the brain when or what sort of neural dynamics support retrieving of information or constructing of new information and how that maps onto individual differences and creativity. I speaker-1: I so appreciate both your explanations. It's not every day you have a neuroscientist in front of you who you can ask these questions to. So thank you. But one of the questions that I did have is association. So I think most of us sort of intuitively, based on our knowledge of language, can understand what association is. But how does it work? speaker-0: This will be probably recurring thing. The word association also has multiple meanings in cognitive neuroscience depending on the domain that you're looking at. So if you talk to people who study learning, especially how rats learn, but also how everyone, the humans, children and adults learn, there's this term that's called associative learning. And that's a very simple idea that things that go together are somehow sort of at the brain level related to each other. And there's this famous saying that what fires together, wire together. Associative learning is one thing that we're not talking about at all. This is really important to say. Freud mentioned this and and William James, the father of psychology, talked about the train of thought or the stream of thought, right? That we just constantly think of stuff. And it we're like a moving train or a river that flows. One thing leads to the other. That is the association that we that I want to talk about. The way that we explain that or the mechanism that Potentially leads to that is that idea of spreading activation I mentioned before. It's not a good mechanistic explanation because it doesn't really explain how that information flows across concepts in your memory map or network. But the idea is that there's like this sort of random automatic spread. So you can't help it. But there's other views that this sort of association flowing or ebbing is actually directed. This is on the basis of Current research for the last several years now study what we can say cognitive search or creative search, exploration, how we search our minds when we try to come up with good ideas. And all of that is in a way based on this idea that we have an associative process that allows us to spread and move away. My dear friend and colleague Roger Beatty, we published a paper in twenty twenty three. The title of it is Associative Thinking as a mechanism of creativity. So It is a critical aspect that allows us to be creative, our ability to move around in a space. And in a recent paper with another colleague, Thomas Heals, we published a paper that says the way that creative people do this is by just constantly moving differently. They throw a different net when they're searching for ideas. It's not a linear transition. It's actually very random like. So there's a lot of things that are happening now on trying to understand this mechanism that is considered, or I would say Because I'm biased, an essential aspect that allows us to be creative. And that goes all the way back to the nineteen sixties. speaker-1: One last question I had, sort of another neuroscience basics con question is field of cognitive network science. Can you just give us a high level explainer of of what is it? What makes it a special area, a special field of study? speaker-0: Wow, this is something I care greatly about. And I actually am actively constantly involved in trying to build and promote this. And I humbly have contributed to it. It's a very small field, or it's an approach more than a field. One thing that has really, really revolutionized a c a lot of different fields in science is a an approach that was developed from mathematics, but also from different fields. That sort of branch in mathematics that is called graph theory. That is the basis of what we call network science. And network science is a computationally quantitative formal way to study complex systems. Okay? And it allows to simplify the complexity into a graph or a network or a map, whatever term you want to use, and then use quantitative measures to study the structure and dynamics that operate on this system. That's like the the quickest version of this. This field that was really started in the late nineteen nineties, but had has its origins in seminal work of the social psychologist Stanley Milgram, so it has to be mentioned. Talks about communication patterns and organization of societies and structures. So it's actually this approach is now used in all fields of science. So any system that you w may think of road systems, airline systems. Computer systems, right? We know the word the term social networks. Everybody talks about social networks, right? Computer networks, biological networks, the brain. Using these tools or methods to study the brain have revolutionized how we understand the brain as structure and function. And there is a subfield in neuroscience today called network neuroscience, which is just mind-blowing the things that are done. I'm biased, but almost every paper, cognitive neuroscience paper. That talks about functional MRI, how our brains talk to each other, communicate with each other, will use the words brain networks. So it's just been integrated, this idea that we can measure complexity of systems through network science. This is a mathematical approach to represent systems. Okay. What is cognitive network science? Cognitive network science is a very the small group of people. That use network science to study cognition or cognitive systems, mostly language or memory. A lot of my claim to fame is on this work on representing semnets, semantic memory networks. I want to represent knowledge as a space or as a network, and then I use quantitative methods to study how different memory networks relate to any behavior, creativity, intelligence, language, aesthetics, humor. Anything. So once we have a ways to create quantitative representation of one's internal organization of memory, then that is a really important scaffolding of behavior. So all of this is great, but I, as I said before, I'm a cognitive scientist in training and I want to understand the mind. And when I started my PhD, I very randomly read a paper on how insight is a result of changes in one's memory network structure. That was a very theoretical paper and one that I till this day I think is one of the best papers that I've read. And the argument there is that when we learn things or when when we have an insight experience, when the light bulb flashes, right? That leads to or is a result of creating of shortcuts between concepts and memory. And that facilitates this I as I said before, mechanism of spreading of activation or information that is really, really important in creativity. That got me really, really excited because that is exactly when I was starting my PhD and was trying to figure out how to study empirically the most widely known but extremely hard to study theory on creativity that was proposed in the nineteen sixties. This theory that was developed in nineteen sixty-two. All right. Sarunov Mednik, who passed away a couple years ago. He was not a creativity researcher and he came up with this idea that is called the associative theory of creativity. And it's one of the most widely accepted theories of creativity, but it was really, really hard to study for a very long time. This theory has two arguments to it, and it's very similar to things I talked about till now. The first, so what he was interested in is what is the difference between a more creative and a less creative individual. And it is really, really important to say at this point that creativity is about a lot of very different things. So we're just going to focus on knowledge and memory and associative thinking. Okay, there's a lot of complexity about what creativity is. But his theory is that higher creative individuals have something that is different about how their memory is organized, and that difference allows them to associate wider and deeper in memory. And to create connections between concepts that are not related to each other originally. And the more remote these concepts are from each other, the more creative the combined idea is. So in principle, his theory has two elements to it. One is the structure of memory that somehow allows a process to operate on it. And if that structure is flexible, then that allows this sort of diffused process to search broader and deeper in memory and make new connections. And in all different theories on creativity, and there's quite a few of those, embedded is the idea that creativity is about being able to move away from pro typicality, to extract remote ideas. Novelty, originality, are about uniqueness, are about sort of distant ideas that somehow make sense. It is also really, really important to say that a creative idea is original, but also about an appropriate an idea that makes sense. But a lot of the idea is that memory structure, if it's the scaffolding of knowledge, then that allows us to search through it and create new combinations. And search and explore and all these terms. Everything I'm saying right now, we nowadays use more formal mathematical ways to actually study what they mean. But for a very long time, these terms were very much opaque and subjective and really hard to study in experimental settings. In my PhD, I set out to study this theory from an empirical settings, and I was thinking ways to do that. Then I read that paper on how insight is a consequence of semantic memory network restructuring or changing. And I thought, network science, that sounds like a cool idea. And there were a lot of different random things that fell into place, stars aligned, and that really allowed me to study that idea. And we used a cognitive network approach to represent the semantic memory networks of people that are more creative versus people are that are less creative. And what we found was that the semantic memory network of higher creative individuals is more random like. It's less organized. Everything's like closer together in a way. It's more condensed. Which we inferred or interpreted as meaning that it's more flexible. Versus things that a system that is very much organized. So things are different from each other. So there's more distance or space that you need to traverse to get from one point to the other. And that makes it harder to get to that destination because these processes decay very quickly, these searching activations, association processes. So when everything's closer to each other, it's easier to just run around randomly and just fall on random ideas. When everything's really, really organized, then it's much harder. And one of the themes in my research is on this sort of tension between structure and chaos, with paper with that title. Because The more everything's random, that i is not always a good thing. You want people also to focus and understand. You're asking me a question, you expect me to understand and to reply. So if everything is in my life chaotic, then you wouldn't understand me. But if everything is on the flip side, everything is overly rigid, then if you ask me a question and I don't understand it because to me everything is so much specific in its meaning, then I can't really communicate as well. So one of my themes in my research is on how our cognitive systems constantly try to find a balance between one, we need structure. Our communication systems need efficient processing, understanding, and communication. But we also need flexibility. We need randomness in our lives to constantly create new connections. And how our system constantly navigates between these two states is really fascinating. For example, work that we have done over the last several years on and aging. Is that memory systems that are overly organized are really rigid and inhibit the ability to be creative. So we need a flexible system. And that also leads to some potential applied research on how can we help people better enrich their systems, better make them more flexible. But we can only do that when we have a good way to r study that and good mathematical terms that allow us to really objectively operationalize what that means. This was a really long answer to say the associative theory of creativity was proposed in the 1960s and for a very long time it was only indirectly studied and then we used network science to study that and ever since for the last decade and half we have repeatedly showed across different methods, across different labs, across different languages, different contexts how Creative people represent a more flexible memory structure. And how did they get to that? And that's a question that we don't have any causality yet. But recent research from my lab, a cognitive neuroscience study from a fantastic PG Sulanahead showed that when people are acting in a creative way, when you're trying to come up with creative ideas, that facilitates active changes in one's memory. So if you are trying to constantly be creative, that could help. create these bridges or new connections and that will make your memory more flexible. There's a circular relation between knowledge and creativity. There's also costs, but in principle, acting creative is good. Learning new things is good. But you want to learn things in a way that constantly enriches your memory and not overly structures your knowledge base. And there's work on expertise effects that shows that experts are really good in what they're doing, but that really mitigates their ability to be creative in that field. So there's a cost and balance between these two aspects. speaker-1: So expertise can impact your ability to be creative in that specific area of because you you have it so very structured, your speaker-0: That's right. speaker-1: knowledge. Okay. This is so interesting. When you studied aging, what did you see there about this associative type of thing? speaker-0: With the aging stuff, we have several papers that came up with a fantastic collaboration I have with aging scientists in the US. But in principle, one thing that is fascinating about aging, they are the only growing population in the world. So we and we should help improve the quality of life of people as they age more and more. So there's a lot of cognitive capacities that are suppos supposedly declining or changing perception, motor, understanding, etc. Attention. One thing that's continues on throughout our lifespan is learning new knowledge, which is amazing. We'd never stop learning new things or we never stop understanding and and require and acquiring new information. And that's like a paradox in aging research. What we set out to do was not try to understand why people continue to learn things across the age while they sort of decline in other aspects, but what effect that has on their memory system. Semantic memory network. So one thing that we find is that indeed as they are learning more things, their memory systems become more and more organized. And in principle, would mean that they should exhibit lower creativity. But in fact, in the few studies that have compared older and younger adults on creativity tasks, there is either no difference between them or higher more advantage for the older adults. So they are more creative in this. Exactly, right? This is great news. And then That creates a question, what's happening? What we think is happening is that there is a balance between they know more things. It's harder for them to spread and and search and expand, but they know more about things in the world. So that a mitigating strategy in that. So there are a lot of complexity, as I said, and the the role of knowledge and searching through it to be creative. And as a side note, large language models, AI chatbots, are seemingly really creative or more creative. But the only two papers that have come out so far using this cognitive network science approach on language models and looking at whether different AI models, entropic, OpenAI, meta, etc., what their memory networks look like. And this has consistently shown that their memory networks are very much organized and rigid in a way. Which explains why some argue they're not creative. And maybe again, that's not a surprise because you train these models on all of the knowledge that we have that is available. So o immediately they go all the way to the end of the lifespan of humans. They already know everything. So it's harder for them to be flexible and learn things. A lot of the magic is happening as we acquire new information and how that changes our knowledge. That process is missing, I would argue from From AI models, especially after they're trained. And that I think is part of of the uniqueness of humans that we constantly update our information, our understanding. And and that is a good thing. If we do it in the right way, if we ask the right questions, that will allow us to get better information and that will help enrich our memory systems, that then will allow us to search better through it and will allow us to be creative. So there's a complete process here that that one thing feeds to the other that feeds to the next. speaker-1: So yoid, I know many people who lean into AI as a source of jogging their own creativity or getting creative ideas by prompting chat or asking Claude. I saw that your lab just released a paper on the impacts of artificial intelligence on creativity. So can you share us what you found? speaker-0: It's really important topic and and one that I care greatly about. first, the question of if machines are creative seemed science fiction only three years ago. It's unbelievable how much things have advanced. And almost no week goes by without a new scientific paper on creativity and AI. It's just one of the hardest topics, also in creative research. And a lot of these papers argue that AI is already more creative than humans. Which is like, no. And the way that they do that is by having AI models and chatbots do our creativity tests, like the alternative test. And when that happens, the results that the AI generates seem to be more creative. That's something that we have been processing for a while now. They seem to I wanna yeah, exactly. I wanna really unpack that. Because a couple years ago, a scientist from Australia did this fantastic work that was published. And what he did, he used another task that has now become popular in creativity research that's called the divergent association task. All you have to do is come up with 10 words that are just unrelated to each other. And it seems really easy, but actually it's a very hard to do. And people that are more creative are better at it. So what this scientist did, Dr. David Cropley, he asked different AI models and humans to do this task. And when he compares the creativity of their ideas or originality of of the responses, the AI models, there were two of those. They were more original in their responses. But then he looked at the variability of the responses that are generated multiple times by these A AI models. And what he found at the time, this was published in twenty twenty three, right? So it's really like ancient history now, is that as the AI model was more advanced, the diversity of the responses that it generates are shrinking. So the AI is generating response that seem to be more creative, but actually it is acting in a less creative way. That to me is really, really alarming. And then about a year ago, I was approached by the c fantastic computer scientist from Duke University, Dr. Emily Winger, and she studied this topic on her own and she just asked me to help her with the creativity stuff. And I just blew my mind. So what she did, and this turned out into the paper that we published about a half a year ago now. Was she looked at all the AI models that she could at the time and they were doing different creativity tasks compared to humans. And what she shows and what we argue for is that this is the most comprehensive evidence that AI models are what we call homogeneous in their creativity. So if you ask a lot of AI agents to do the same task over and over, then they will repeat their answers. And they will be very similar to each other. And this is not how humans operate. Each of your friends or colleagues, when they do this on their own, they prompt an AI to be creative or to come up with a creative task. AI gives it what looks like a creative task, but across multiple users, it will give very similar type of response. So that means that the creativity is narrowing or shrinking in how these AI are generating output, which is very alarming. But it doesn't stop there. The question, this is follow-up work that we are now writing on with Emily, was the question, okay, maybe a language models are not creative because that they are homogenized in the process of retrieving information. Maybe that doesn't matter at all for humans. So what we've done over the last year is a series of lab studies where we have people interact with AI agents doing repeated association tasks and testing their creativity before and after this human AI interaction. And we find in principle that the human creativity is also narrowing. So overly using AI models, getting more and more down this rabbit hole could also drastically impact human creativity. And that is the danger because our world is becoming more and more complex and requires of us more and more creativity in the next years and decades. Think of climate change. There's so many problems this we have to constantly Strengthen that muscle of our younger generations to be more creative. But these AI models are potentially pushing them to like this very much regression to an aggregate, shallow, mean. And that is the danger. It doesn't mean that these tools are bad. These tools are extremely powerful and they could really help us be a much better version of ourselves. All we need is to use them correctly. But no one tells us what to do or how to use this. That is the danger. Because this technology of is evolving so quickly and people are more and more attributing sentient like yes dimensions to these tools, there's a a lot of assumptions, theory of mind, perspective, perception that these tools are actually like human like, but they're not. They are only appearing to be that. I really am concerned about what that means. also recently work that was a MRI, functional MRI study. With a colleague of mine from the Technion and and will be published hopefully in the coming months. We scanned under MRI children talking to an AI agent, building or opening a shop together. It's a very creative task. And what we find is that the brain networks that we know that are critical in how they interact for creativity are less communicated with each other. And that's a very red flag that we should all be very concerned about. What this is happening. just need to rethink on how these tools can be used. And one term that I think is important to say is cognitive offloading or de skilling. We can't have creativity be offloaded. That will be a really dark future for for humanity. So we have to change the trajectory of where all of this is taking us. And there are ways to change that. We just need to have more awareness and more conversation. And possibly speaker-1: Possibly more control because right now the door is so wide open. speaker-0: I agree again. I think this is a really complicated, long discussion. speaker-1: Yeah. Yeah, we won't get into that. But okay, one last piece. I noticed that one of your evolving lines of research is how semantic memory evolves and is shaped by problem solving, learning, and time. Can you just give us a little highlight reel on where these topics are going? Great. speaker-0: This is such a great conversation. I'm really enjoying it. And so I don't know. Hopefully haven't bored. But I mentioned this paper that got me into all of this space of network science and cognitive networks that argued that insight is a result of memory structuring. But that was a theory. And then with a colleague of mine from Paris, he's she's just a fantastic neurologist, Emmanuel Volet, we actually set out to study this using our powerful methods into studying this. So what we do and We have now replicated this already a couple of times. It's a very simple idea of a study. We have people come into the lab and just trying to solve a riddle. The riddle, it's like a classic riddle that was studied early 1990s. It's called the bar riddle. A man walks into the bar, looks at the bartender, and asks for a glass of water. The bartender pulls out a shotgun. The man says thank you and walks away. What happened? That's that's the riddle. Wow. So yeah. speaker-1: He would. speaker-0: I'm not telling you the answer. Don't tell me. But they have ten minutes to try to solve this riddle. Now, before and after these ten minutes, we use our network science methods to represent their own semantic memory structure for words or concepts that either relate to the solution of the riddle or and or not. And then we're using mathematical approaches to compare these networks before and after for people. We start off with two groups. Right? People that solved the riddle and people that did could not solve the riddle. The majority of people don't solve this riddle or similar riddle. So don't feel bad. It's actually very hard to solve. That was one of the limitations. We got very low solution rates. That was really hard to argue that what we find is actually awesome. And this I think really converges on on the theme of your podcast as well. What we find is that people that solved the riddle, their memory structure changed, and we know how to quantify that in a way that facilitated their feeling of insight and then also help them solve similar problems. Which I think is so awesome that we can formally quantitatively actually measure this, a change or memory structuring. But then reviewers were not satisfied because reviewers are usually not satisfied. And they said, how can you actually show that this is active learning or process that really happened? So we added a third group where we give the solution immediately to the people. So we're doing everything exactly the same, but now they they see this solution. So they know what's happening. So there's no active attempt at solving. What we find is that people that we give them the solution, their memory supposedly superficially changes, but in a way that does not facilitate subsequent problem solving. speaker-1: Yeah, 'cause it was handed to them. speaker-0: Because it was handed to them, right? don't give a fish, learn how to fish, right? It's all these things. And and it really is we w we can actually demonstrate that. That took forever to publish. But that is a way that we are now using to show how we can study how memory changes with problem solving, with aging. I mentioned that before. With learning, we have studies looking at how successful learning classes relate to. Changes in one's memory structure relating to the theme of the class, et cetera, et cetera. And another cool expansion of this riddle study with my colleague in Paris was done in a sleep lab. People are trying to do something very similar, but in between they undergo different types of sleep, either deep sleep or light sleep or no sleep. And we find that we can facilitate or inhibit this memory structuring. Really? through this type of sleep that they do. So we can have Cognitive networks, brain activated sleep, we can have like pa designs that allow us to really measure things. And we have theory that we can now go and predict what will happen and we use formal mathematical measures to actually demonstrate that these things change in an active way. And if you actively learn, if you actively make an effort, then that has downstream positive effects on your memory system. So it's so important to make an effort. It is so critical. And I think going back to the AI discussion we just had, that lowers your motivation to make it. Right? Because everything is given to you. And so it's not a surprise that people are not creative because why would they just ask that to generate for we have to change the conversation about how to constantly facilitate interest, motivation, curiosity. But yeah. So this is the type of research of studying how memory changes. These are things I'm invested in. Then speaker-1: Is so cool. And and I have to say that when I need to learn something or make a decision, I must, as the saying goes, sleep on it. speaker-0: Right, there you go. speaker-1: I need to learn at least the day before. And after a night's rest, things will suddenly make sense or a connection will be made that I hadn't considered the day before. speaker-0: That's interesting research on creativity and sleeping, but there's mixed evidence on what's happening there. What we offer is a way to try to better represent what is changing or what is the effect of sleep on on memory, on on knowledge. speaker-1: Yeah, specifically knowledge. Yeah. I it especially yeah. Yeah. thank you so much, yo. This was this was super cool conversation. speaker-0: This was awesome. So really nice conversation. Thanks. speaker-1: This episode is dedicated to the people I've known who treasure knowledge and use it to make the world a better place. Incredible Mind Podcast is written, produced, and edited by me, Michelle Cross Smith. Peter Aronson is consulting producer. Angela Ferrero designed the cover art, and the music was created by Alex Gruhl. Did you know the Incredible Mind Project also offers direct education on critical thinking? For a taste of our short form courses, check out the Incredible Mind Podcast channel on YouTube. And our website incredible dash mind.org.