 A big picture about how different parts of the brain interact with each other to generate coherent behaviors during courtship and aggression, and specifically the flexible, the cockpit aspect of the behavior, we still don't know exactly what is the neuro-maximism for that, and that's something that my lab is particularly interested in. The valence and the rousal, the two building blocks of elementary emotions can actually be accessible in Drosophila as well, so in addition to all the cognitive studies that we're doing right now. And I would say Drosophila, I consider it as a hydrogen model for social cognition. What's up everyone, welcome to Simulation, I'm your host Alan Safian. We are on site at the beautiful Westlake University in Hangzhou, China. We are now going to be talking about the neural mechanisms of social cognition. We have Dr Yi Sun joining us on the show, hi Yi. Thank you so much for coming on the show, really appreciate it. I'm very excited for this conversation, for those who don't know Yi's background. He's a neuroscientist, physical chemist, engineer, and professor at Westlake University pioneering novel methods in functional brain imaging, focusing on the neural mechanisms of social cognition. You can find the links in the bio below. Let's start things off with one of our favorite questions we like asking our guests. What are your thoughts on the direction of our world? Yeah, that's a very good question. I think the direction is moving towards inflammation explosion, and I think part of the world is about the dynamics of information flow. So a key question is how information are integrated, and those involve integration of information through, say, computers in the physical world, or information in our brain. And that's why I'm interested in the social brain. Information flows, and the dynamics of information flows, yeah, exactly. Okay, and then specifically with, especially biology has this social behavior between creatures that then enables information flows and to do specific objective functions like, yeah, like reproduction or gathering food and resources. Go over here. There's plenty of food over here. Faraging behavior, yeah. Interesting. And then we, as 8 billion creatures, also have our dynamics of information flow that enable us to say, well, maybe you should move to the city because there's more opportunity, there's more efficiency, there's more wealth to be made, that type of stuff, and we tell these stories about people going to the city and finding opportunity and stuff like that. Do you frequently make these abstract leaps from biological communicative processes for social cognition, which we're going to dive really deep into, and abstractly give metaphors for how we as a civilization of humans behave? Ah, that's a pretty tough question. So these days I'm generally concerned with, you know, the internal representations and things like that in animal brains. How do we make metaphors between abstract aspects of how a brain works compared to the operations, information flows, things like that in the real world? I think, I think my study in the simple systems like small brains like Joseph LeBlanc and Wayne Steady eventually helped us understand much more complicated real world, world of human brains, things like that. Yeah, yeah, that part to your research is probably the most interesting for me is once you understand the neural mechanisms of social cognition and the internal representation, then you can maybe begin to start abstracting up from a fruit fly to the mammals and human brains. And then all of a sudden we have a better understanding of how we work and how our society works. And so that's my favorite part. And let's jump into your journey. So where were you born? Who were you as a kid growing up and how did you get interested in science? I got interested in science in the family because when I was a kid my parents were educators and they gave me a lot of different toys to play with. Importantly they asked me to create toys and I can play with the toys that I created. And through that process I can come up with things that I'm interested in which are the toys and the technologies to build those toys. And I think that really eventually lead me to my career in science and technology. Wow, and this was in the Jiangsu province? Yeah, that's the Jiangsu province very close to the province that we're living in now. So once you started making your own toys and tools and stuff that kind of led you into your interest in engineering and then that got you into that was that your bachelor's was at the School of Mechanical Engineering and Automation and Beijing University of aeronautics and astronautics. So walk us through going to doing your bachelors and getting interested in engineering further and then also going starting to get into your PhD. Let's start going on the track. Yeah, that was a pretty long journey. So I was originally interested in engineering because I thought engineering was cool. Actually I still think engineering is very cool. So after graduation I actually went to the industry. I was working in companies that were producing information technologies. I was bringing a lot of programming. And at certain point I realized that it's very important to study instead of mechanical things or like physical things. I want to get better understanding of ourselves. So I went to graduate school. I was studying things that have some connections with like building tools to help us better understand how we are, what we are and how we do. So that was about my PhD. It was actually in physical chemistry we were building nanotechnology tools that eventually can help us to study biological systems as one of the applications. But by the end of my PhD study I became more and more interested in the brand. Part of the reason was because we were devising technologies to record the activity from electrical activity. And one of the applications was of course to record from the activity of the brand cells. So from there I became interested in, I started to read neuroscience papers and started to read neuroscience textbooks and I went to neuroscience courses. And I find this is much more interesting than I wanted to study. So eventually I moved on to neuroscience and I did a co-stock in Howard Hughes Medical Institute and then I was working on the research specialist there for about six years time. Really got me to dive into understanding the brand and technologies for understanding the brand of course. Yeah, it's interesting hearing about how you had this engineering background that then was like well I want to know how life works. I want to know how the life works. And so that kind of gave you this really good footing for dissecting and diving deeper into biology with an engineering mindset and then this fascination with reading brain cell activity and yeah, I like this a lot. So then that PhD happened, physical chemistry and nanotechnology, the National Center for Nanoscience and Technology, the Chinese Academy of Sciences. And then you went to do this post-doc at Howard Hughes Medical Institute. Which actually I was recently learning is just one of the most well-funded research institutes in the world. And it's really cool that there's lots of people actually from Westlake that got a chance to be there and do their research there. So let's start talking about that. So you were doing optical probes that report electrical activity through fluorescence. Okay, let's walk us down what you were doing there and what that is. Apparently it's very important to record the activity of the brain. That's the way you really understand how the brain works, right? And one of the first technologies was to record through electrical recordings. The very famous patch clamp recording technology, they really allow us to get a sense of the activity of individual cells. At a very high temperature resolution, of course. The problem is the brain doesn't work as individual cells. The brain works as a neural network, composed of many, many, actually millions, billions of neurons, right? So really to understand how does neural network generate activity patterns and eventually to behavior, we need to understand the activity of a population of cells simultaneously. And that requires technologies to recording, not just one cell at a time, but many, many cells at a time. So that was why we were looking for optical technology, because for electrical technology, there are certain ways to record more than one cell, but those technologies doesn't tell exactly which cell we're recording from. To hear the sound of many, many cells is like a cocktail party, but it doesn't really tell which one is which. So we designed these probes, which are built upon the fluorescence technology, the green fluorescent protein, which is a protein that fluoresces all the time. So we did some protein engineering so that the protein wouldn't fluoresce all the time. Instead, it fluoresces whenever there's neural activity. And the way we translate neural activity into the fluorescence is through calcium, because we know from other studies that neural activity changes leads to calcium concentration changes within individual cells. So we designed probes that can detect calcium changes real time. So that really is two-step process. One is from electrical activity to calcium. The other one is from calcium to fluorescence changes. It all happens with pretty high temporal resolution. And with the help of imaging technology, we can do that recording for many cells simultaneously. Wow. Okay, so from electrical activity to calcium changes, from calcium changes to fluorescence, and then being able to image not only a single neuron, but a network of neurons that then give you an idea of behavior. Right. Okay, cool. Okay, now where else along the way to Westlake were you starting to think about this pioneering these novel methods in functional brain imaging? What were you thinking of when you were at Howard Hughes? What were you trying to understand at deeper level of these neural mechanisms? And then how did you end up transitioning to this professorship here at Westlake? Yeah, so when I was at Howard Hughes, actually at the Genealogy campus, I was interested in two things. One is the optical probes for recording the activity. The other one is about how we use these probes to really understand the operation of the brain. So one of the things I was really interested in is about the selection process. We make the selections, we make decisions all the time in our lives, right? So I was trying to understand this from some really simplified power lines. So we were studying how the fruit fly, Joseph Lamelana Gasker, which is a great model system because it's a more accessible brand, so a rich wrapped bar of behavior, how do they make decisions? So I was using very simple visual stimulation and recording the activity in the brain using the probes that we developed, not just the green calcium indicator, G-CAM 6, G-CAM 7, but also red calcium indicator, say JARC ECHO 1, JARCAM 1. We were doing this simultaneous two-color, two-photo imaging so that we can record from different channels that are from different populations and neurons that are spatially intermingled. That tells us about the representations and the transformations of information of that representation of the brain that eventually lead us to the fly to make the decision to, say, dedicate, engage in a specific object, amount out of several. So the study I was studying in a simplified paradigm and I wanted to study this decision-making process in more natural realistic paradigm, the scenario, and I think the ultimate natural situation is our society. We all live in a society. We have to make decisions. We make choices all the time in our society. And I think Joseph is still a great model system. And at that time Westlake was establishing, it was actually, it was still under active construction. So I have experience of working in newly established institutions. I got my PhD from nanoscience and technology, National Center for Nanoscience and Technology, which was established about a little over 10 years ago. And that was also the time when a junior research campus of Howard Heath Medical Institute was established in Ashburn, Virginia in suburban areas, D.C. And I think I have a great experience of working in those newly established institutions because of the excitement and the fear on the campus. And that was one of the driving forces why I was choosing Westlake out of other places. Yeah, the magic of all the different departments here that are coming together to push the ed frontiers of science. Right, the interdisciplinary environment here. It's a small but still highly diversified institution. Yes, yes. Okay, so the focus now is on neuro mechanisms of social cognition, internal representation. Internal representation has a couple of subsections, behavior, anatomy, imaging. Start walking us through what this is. Yeah, so the way we study the neuro mechanism of social behavior, first we have to have behavior paradigms that help us to dissect specific aspects to basically define the questions, right? To translate behavior, which is highly variable into very specific questions you can address with technology, with tools in your science. And that's about behavior. So we have paradigms that are involving free behaving flies. The flies are free to do whatever they want to do. And from there we record their behavior, record their all aspects of the behavior, their sound, their emotion. And we do a lot of analysis. And from there we get a specific perspective. We find clues about how they make decisions, how they learn from each other, say, and how they recognize each other. And from there we do more controlled behavior experiments. Those more controlled behavior experiments would tell us something about how they perceive others. And that's really a hallmark of indications of internal representations in their brain. To get at the internal representations in the brain, we first need to get a picture, a static picture of the brain because we first need to know the basics. If you want to understand the behavior of Hangzhou, we know about the dynamics of the traffic. We first need to know about the roads, right? The network of roads in Hangzhou. Before we dive into the dynamics of the traffic. So in a similar way, we first need to get an anatomical basis with the static pictures of the brain before we get to the more dynamic aspect, which is about the physiology. So there we apply the technologies, the optical probes that we developed early on to record the activity of the brain. We try to do all these experiments in one animal. So the animal is doing specific behavior tasks and then records their activity. And from there we can find the neural correlates of internal representations that are kind of modeling their perception of others. Okay, so let's continue this example where we have the dress off of the fruit fly that is in a, you guys have chambers, right? So let's explain these chambers. So you are doing things like watching it with a camera and you're recording that data. And you're also recording the audio of the fruit flies, okay? And then you are also then having another fruit fly, join it, okay? And then you're observing how they interact. So we have two fruit flies just like we have on this couch, two humans, and there's a camera watching us if there would be a camera watching fruit flies. And then there's a recording of the audio happening. And so now what are the fruit flies saying to each other? What are they talking about? And how do you make neural correlates of specific audio frequencies and wavelengths and et cetera to what they could be talking about? That's a good question. So our communication is a response to specific situations, right? So we say specific things because we want to deliver a specific message to change the situation. Say, we sit here and I say hello because we want to make better connections, right? So the approach we study, this problem is to categorize recordings, the patterns of the sound. And then we find the context where specific types of communication, of patterns generated, right? You say what I mean? So first we core the audio of the flies and then we categorize into different types because first we need to know what they're communicating. And then the second question is why they're communicating that. They're communicating that response to specific context because, say, another fly was giving them an active or positive information or there's something they need to communicate about the environment. So we take out the specific time stamps of specific types of communication pattern and then we study through the videos and audios about the context and we try to see if there is a correlation between the context and specific patterns of communication. That's how we understand what the messages are actually being delivered because we record the sound, we can only know the physical, the acoustic property of the pattern but we don't know exactly what the message, what the meanings of those communications are, right? We try to understand, derive the meanings by looking at the context, the recent history, yes. So then we do something like we start in a sense building out a library of the different audio recordings and video analogs to those audios. So now I have a library and there's type A, type C, type Q, audio. Library or dictionary. Yeah, I like that. Okay, so there's this library or dictionary happening and then there's an audio and a video component that goes to every type in the dictionary. And then now you're built out a catalog. So this catalog is happening and then how do you then say let's map type A to specific neural activity or type D to neural activity so we can understand what social cognitive processes are. Right, right. So that's about how we move free behaving animals to more restrict hazard preparations where the flies are flying or it's walking on a ball but it's restricted, it's hat fixed. It's important to have fixation because we can do physiology, we can do calcium imaging, right, using the optical props we developed to look at their brand. And then we try to assimilate the environment that we just discovered, right. So from the behavior studies I mentioned before we know the context of specific patterns. So we try to replay the context and then we look at the activity of the brand. So we try to see what was going on or the inter-representations for the context and then what are the inter-representations that are predictive of the patterns of sound that they're going to deliver. So it's very important that we study the behavior and physiology at the same time. Yeah. Okay, so you're doing a couple things. You have a chamber for the fruit flies where you're doing video and audio recordings in a free moving space and you enter in second fruit flies. Do you ever enter in third and fourth fruit flies? How many at most do you guys look at? I mean the paradigms. How many total fruit flies do you let hang out in the chamber to see what happens with their group dynamics? Oh, we have collective behaviors composed of many, many flies. So we have a whole range of different studies from two flies to many, many flies. So there's a dictionary for two flies, there's a dictionary for three, four, five. I think the dictionary for two flies is probably kind of general for arbitrary numbers of communications. We don't know yet. Yeah, although there are interesting dynamics that change when there's a third person in this conversation or a fourth person. Then the four-person conversation could break into two-person conversations. And that two-person combination could change over time. Yeah, yeah, yeah. Exactly, they'll start talking to the other two. Yeah, exactly. Okay, so there's a chamber one that's happening. And then there's also this, you call it partial tethering. Okay, so let's talk about the anatomy of the fruit fly. Yeah, it's very important to actually skip the anatomy part because it seems like after studying the behavior we go directly to physiology, which is not really true. So after studying behavior, we basically have no idea what's going on in the brain. So I have to know the near substrate of specific behavior because not all the brain is used for all the behavior. So we do some behavior genetic study to find potential candidate circuits for specific behavior. And then we use anatomical approaches, which you can call it neurocircuit study, whatever, connectome, things like that to understand a static picture of the brain that probably underlines this behavior. Yes. So with the anatomy we're doing things like, let's see this visual brain, then thorax, which is connected to the six legs and the two wings. And you tether it specifically to the thorax to make it so that the fruit fly, in a sense, can still move on this ball. So this is again, this is different than the chamber. So it can move on this ball, but that the brain, it's not moving so much that you can't properly image the brain over time in space. So it's enabling you to give a, and this is where I wanted to also ask you before, we used to just take a dead piece of brain tissue and image it. And now we're keeping the brain tissue in the organism while it's alive, while it's tethered and imaging it so it's giving us a more real understanding of social cognition rather than just a dead piece of tissue. Right. So the early studies about the brain, basically neurophysiology was done to a large extent on disactive brains. Actually, people also do culture, dissociated culture neurons for cellular mechanism. So those studies are great for studying a lot of molecular and some of the synaptic physiologies, but they're very difficult to recapitulate the actual environment the animals are engaged in. And also when it's dissected, a lot of connections are cut, right? So the physiological states of the brain is different. We don't know exactly how different they are. So people moved from dissected brains to in vivo studies, but people anesthetized the animal so that there's no movement. It's very important to get rid of the movement because movement or detrimental for a lot of physiological studies say electrophysiology or imaging. But still the animal is not in a normal physiological state, anesthetized is not a normal physiological state. So people start to do a wake, but not behaving. People try to tether the animal as much as possible. The animal is awake and we can do the recordings, but we don't know exactly the animal is thinking about. They could be daydreaming about whatever they're daydreaming about. So after that we start to do physiology in behaving animals, tethered behaving animals. So that really gives us a context to study the dynamics of the brain because the animal is attending to very specific features in the environment and they're processing those features in the environment and they're generating specific behaviors. So we can look at the dynamics of the brain when the behavior results are correct and we can also compare those to when the behavior effects are wrong. So that really tells us about how does the brain solve the problems that animals are trying to figure out. Okay, so where does this dictionary or library catalog that we were talking about before with audio and video representations of social behavior, how does that begin to be related to when you have the second, these second experiments that we're talking about where the fruit fly is tethered and it's on the ball and then you have a couple other very interesting additions to the experiment happening. You have a virtual reality that the fruit fly is basically in with a display that is... If the fruit fly moves on the ball, the display changes and so it gives a false representation of control of it actually being in a fake environment and then how does then the fruit fly in that tethered secondary environment, how does that give you when you're actually looking for these neural correlates and you're doing the microscopy on the brain, how does it then in that environment have the relations to that first environment where you were doing the catalogs of audio and video and social behavior? Yeah, so the first one is more about in a more natural context. We try to think through the possibility space of the communication, the vocabulary, the dictionary of the fly. And from there, we pick some of the things that we think are more interesting, some of the patterns of communication that are critical for specific behaviors. They try to deliver a message like a male fly saying to a female fly, I love you, or a male fly saying to another male fly, that I hate you. We take those pieces and we focus on those pieces and we try to devise simulations that tell us more about the details of that communication and how are those decisions like love or hate made in the brain? So this is a much more controlled experiment that we do tethered behavior. So the first one is about generating hypothesis and finding all kinds of phenomenology and the second one is to specifically look at the mechanism for a few critically important, some of the most important key points during social communication and social interactions. Okay, so if in the chamber you see an interaction that you can run pattern recognition on a lot of people are using AI for these types of purposes, they're looking at tons and tons of image or tons and tons of audio and they're finding, okay, here's a pattern to a specific thing in our dictionary to type A. We hypothesize that that could be an I love you, let's say, okay, a mating call of sorts, okay, for reproduction and maybe that can even be better seen by if they decide to have sex afterward, the fruit fly sex afterward, then you're like, okay, that was probably I love you, that was probably let's reproduce. Okay, so then when you take it into the second environment you want to then do some sort of simulated I love you to be able to then map what's happening in the brain during that I love you. So maybe it's like optogenetic stimulation for the fruit fly where it can be simulated for that fruit fly that, it's that I love you and then you're mapping what's happening in the neural architecture in that. Yeah, that's one of the things we do, we use optogenetics to activate or assign specific populations and neurons that give us clues about or give us controls over the state of the animal. But we also deliver carefully designed quantitative stimulations, visual stimulation, auditory stimulation, things like that. And we look at, we quantitatively record, measure the activity of the brain, we try to build correlations between those sensory stimulations and the neural activity and we try to see at which stage what are the representations for those sensory inputs and what are the transformations between different stages of representations and how those sensory, the percepts are gradually transformed into categories that facilitate decision making. So we look at, the important aspect of this is that everything is quantitative. So actually the first stage of free behaving behavior we talked about earlier on it's also involved a lot of quantitative analysis about the videos, actions, the movement of the animals and the audio, the communications, the sound of the animals. We have to do analysis and some of the analysis are done online so that we can have close loop feedback, real time feedback. What would you say are the most common social, cognitive communications happening between fruit flies? Like if you had to say that there were certain things that were being communicated in the dictionary or in this catalog, what are the most popular types of communication that are happening? Courtship. Courtship is the most frequent communication. Because reproduction is just top priority. I think that's probably the same across many different species. So funny, courtship is a... Because you can do faraging, you can find food on your own, but you cannot reproduce on your own. I guess in a sense a fruit fly can just reproduce and then maybe can not necessarily have to care so much about its offspring. Oh, am I financially stable enough? Do I have enough food to support my offspring? Maybe it's a little bit more just like I'm just going to have offspring a lot and just not necessarily have to worry about if there's going to be enough food for the offspring and just kind of let them go. What would be a second most popular communication that's happening? Fighting, I think. Fighting? So between males, male to male, and female to female, wow. They fight. They fight for food, they fight for mates. They fight for a lot of things. But the fighting is much more dynamic than courtship. And it's actually, compared to courtship, it's pretty rare. Oh, interesting. So it's sort of graded. They don't burst into fierce fighting immediately. There's a building up process. There's a building up with courtship too, right? Yeah, there's a building up with courtship. It seems like courtship is more robust. There's courtship between males as well. Okay, so there's obviously a lot of biological hierarchy of importance of prioritization happening across different species. Humans, fruit flies, doesn't matter what species. There's always this kind of like hierarchy of priorities. And for a lot of the time it's just like reproduction is a big one. But if a fruit fly has a limited lifespan, then it needs to prioritize reproduction. Whereas we have an 80-year lifespan, we can reproduce in a longer window in the mid-part of our lives or early mid-part. Okay, okay. Okay, let's talk about when the most common behavior is courtship. So what visual and auditory signal processing are you doing that is able to say, oh, look, the male and female are about to, how can you predict courtship for happening in an audio signal or a video? Right, so when they're starting to show signs of engagement, the first they orient to each other. The male will be orienting towards specific females. So basically the tension of the male will be directed towards the female. And then start to catch up with it, start to chase the female, and the female would just flee. And then they try to keep doing this chasing game. And if the female is interested in the male, the female will slow down at certain stage. So become receptive to the male. Give the male an opportunity. Then the male will start to sing a song. Then the female will make a decision if she likes the song, she likes the song of the male and she's ready to accept the male and she will just come to a stop. And eventually they're engaged in courtship, a copulation. Wow. It's a pretty dynamic process. Wow. There's a lot of visual and acoustic communication during this process. Interesting. Orientation first and then chasing. And then attempting copulation. But the song. The singing, yes. That was very interesting. The singing, yes. Wow. So how do you guys know it's like singing? It's pretty office. You can record the sound. And it's very typical. It's called Pao Song. The Pao songs are a strongly indicative of a courtship process. Wow. Okay, so you're literally singing and then the females deciding whether or not the song was good enough. And then copulation. And then offspring. How quickly do they end up like, do they end up staying together or do they end up going and reproducing with others? They will do the latter. Yeah. Yeah. Diversify. Drosophila is not monogamous anymore. Yeah. The Drosophila is not monogamous and it's just constantly trying to swirl its genes with other genes instead of just the specific genetic code. Yeah, I think in birds and some of the other higher order species you'll see monogamy, synchromagamy or Sasha's monogamy more frequently. Yeah. Yeah. Well, because there's also completely different biological dynamicism to that process where you have us which has to raise a child for such a long period of time and we can't just birth the child and then go and try and swirl our genes with other of our same species because that child can't just go off and try and find food itself. It has to be raised. Yeah. Yeah, yeah, yeah. There's different timelines. There's different... There's so many other factors that go into this. Okay, well... Yeah, so it's very important that we try to study the various factors that get into the decisions during social interactions and try to see how does the brain weigh different factors and eventually, you know, make a choice to dedicate to a specific decision. Yeah. So there's information integration and decision-making and some of the scientific questions that I were terribly interested in at the moment. Likewise. I'm so fascinated by that. It loops us all the way to what you were saying at the beginning. So I'm constantly taking in my environment and I'm figuring out, you know, I have, like, a ledger of my environment and my internal ledger and I'm constantly trying to figure out, like, you know, if my stomach ledger is low, maybe I'm going to change my environmental ledger so I'm closer to food so I can get food. And I like this a lot, this, like, dynamic game of, well, okay, if my internal ledger is, like, I'm getting closer to the age of 35, maybe it's more and more important for me to find a mate to reproduce with. And it's so interesting that you guys ended up identifying these neural mechanisms and internal representations of a courtship to be the most, like, common and now you're stuck. So studying courtship is the lab, is your lab's most? Well, actually, we're specifically interested in social cognition. There's a community of scientists that are interested in the biological basis and neural basis of social behavior like courtship aggression itself. And in addition, on top of that, there are cognitive control during courtship and aggression. Say there's attention, there's memory, there's decision making, all those things there's learning during those very dynamic social behaviors, social interactions. And my lab is specifically interested in the cognitive aspects of social behavior. So we don't focus on the basic social interactions, the hot-wired circuits of social behavior interested in the dynamics, in the modulation, in the cognitive aspects during social interaction. Okay, so how do you then replay, how do you simulate a male, do you, like, literally put up on the screen like the female fruit fly and then the male fruit fly when you're doing the second part? We do a whole range of stimulation from realistic physical flies of different states. And we also do simulated more controlled or computer simulated stimulation, yeah, for the other flies. So simulations that then stimulate the neural activity so that you can map correlates of social cognition. Okay, and you're, like, what, I guess, what areas of the neural architecture of the Drosophila do you see most engaged with courtship? Well, at this stage I would say pretty much, you know, many, many parts of the central brain is engaged. There are indications for many, many different types of neurons, but we're still lacking a big picture about how different parts of the brain interact with each other to generate coherent behaviors during courtship and aggression. And specifically, the flexible, the cognitive aspects of the behavior, we still don't know exactly what is the neural maxims for that, and that's something that my lab is particularly interested in. Looking for the neural mechanisms of courtship and aggression and other social interactions. And other social interactions. Okay, okay, cool, cool, cool. Right, so basically the courtship, the social behavior of flies are not stereotyped. Not as stereotyped as we thought. There's quite a bit of flexibility. But flies do make choices between say, different potential mates. Yes. And the aggression process, the fighting process is also highly dynamic. It's not something that is very reflexive. So that's something that that's the kind of behavior that we're particularly interested in. And neural mechanism for that is where we're really interested in. So that's the kind of behavior of me looking at potential mate number one and potential mate number two. And then me looking at them and trying to figure out what the dynamic activity in the brain of like is that one a better mate right now or is that one a better mate right now? Yeah. Okay, and how I switch between those two in the making process. Yeah, what kind of circuit architect or support that? Yeah. And then it could potentially extrapolate up to our own brains when we are close. Exactly, yeah. So it's very important that the reason we study Josephla which is entirely different from our self. People wonder why we study social cognition in such entirely different species. The reason is because it's a much different system. So from these studies we can know the circuit architecture and the activity patterns that generate the behavior of say attention, of say memory or decision making, learning things like that. And these knowledges can actually directly translate into algorithms that help us to design more intelligent systems directly. We don't have to study the human brain to do that. Any kind of biological model system in Josephla, I would say is probably the most convenient system that we have access to at the moment. At the same time the results that we learned the principle that we learned from this simple model system will help us to understand much more complicated system because if we look at the history of biology where we see a lot of these fundamental principles were actually in the molecular level were actually discovered in single cell model algorithms such as E. coli, yeast, things like that. And if you look more and more broadly across the entire scientific field, say physics, the quantum mechanics we spend most of the effort to study hydrogen. We use hydrogen model as the model for a lot of studies in quantum mechanics. So that's the basics of the whole architecture of quantum mechanics. And I would say Josephla, I consider it as a hydrogen model for social cognition. Okay, okay. I like that. Especially with the long time scale of biological evolution and just how the neurology, the nervous system inside of creatures has been developing for such a long time there is almost inevitably going to be correlates in the nervous system across species. Especially since in the hierarchy of decision making like reproductionists is it 100% common across like pretty much almost every species has to have reproduction in its hierarchy and likely the top thing in its hierarchy along with foraging. So then it's very possible that when you look at how a fruit flies determining mate number one versus mate number two and switching between those to find out which one it could almost kind of be like when a woman or a man is looking at potential mates themselves as well and like switching what architecture in the brain is lighting up and also what just like a male has a song that is played in a sense we have our songs that are played with our like online profiles of ourselves when someone looks at our online profile in a sense it's a song that's being played to them about who we are our work our travel or our interests or whatever and also we have these little micro songs that are played in terms of like when we meet each other in person or when we send each other messages or emails or you can add little smiley faces or little winky faces or little all different types of to our social communication that can then be related in some way to people deciding whether or not we are a good mate to reproduce with yeah emotion comes together is cognitive process yes that's the very important part of social study social cognition as well so when we talk about cognition it's not just limited to classic definition of cognition it also includes emotion even though the emotion in flies is a little bit elusive compared to emotion in high order species but the flies do display some elementary some very simple forms of emotion like or dislike likes and dislikes some of them are innate some of them are learned and these responses are generally considered as valences which is according to some theories is one of the building blocks of emotion so flies do have valences they have avoidance and approach and those things can be steady in flies as well that's a very very primordial look at it like or dislike go towards or go away from yeah flies also have another aspect of the emotion which is about arousal so some of the flies can be highly aroused some of them can be more like carved and those things we do have a lot of behavior avoidance for that and those things can be studied in Joseph as well so valence and arousal the two building blocks of elementary emotions can actually be accessible in Drosophila as well so in addition to all the cognitive studies that we're doing right now throughout this conversation I'm just frequently just trying to see how your analysis of the Drosophila is related to my ideas about how potentially those neural correlates to human behavior and my mind has just been racing about how is it possible that we also do these courtships or these aggressions or these emotions these like or dislikes and how can we start saying this is hard because how many neurons again 100,000 right in the Drosophila? 100,000 brain that's our rough estimation at the moment 100,000 neurons in the Drosophila about 70 to 80 million in the rodent in the mouse and then 80 billion or more in the human brain and we maybe think that it's possible that a circuitry of like or dislike something as simple as that could have even though it may only be a network of like a thousand or a couple thousand neurons in the Drosophila there could be an analog to like or dislike of maybe tens of thousands or hundreds of thousands in the human brain yeah yeah so I'm just like this research is so cool and there's obviously great importance to this also this catalog like if there's a literally like a type a social interaction that leads towards like higher success in courtship and you can like watch that through video and hear it in audio like there's a like there's a specific song style that leads to greater success in courtship then you can simulate that song style to when the Drosophila is tethered and then watch as the neural activity yeah maybe towards maybe towards a higher propensity for desire for copulation yeah and then we can maybe say that there is a possibility that that same thing exists when like a female identifies a human female identifies a human male that is maybe a little bit higher up on the hierarchy maybe has a little bit more financial success or has advanced a career in a certain degree there's maybe a greater amount of desire for mating with that then if the male is lower on the hierarchy and so you can maybe map a neural correlate of a greater amount of activation that happens yeah I'm so fascinated with this subject this is such an interesting one okay so let's have you explain also the the overall like differences in communication modalities can you list them one more time for us you're doing visual, auditory somatosensory somatosensory meaning touch we didn't talk about this one too much but so the fruit flies also come up and touch each other shake hands they shake hands with their little insect legs exactly they do shake hands wow what but if there's friendship or a courtship they will do a little like handshake literally like a like a lake shake yeah they do yeah exactly this is very frequently happens okay so okay visual auditory somatosensory there's also olfaction and gastatation we don't specifically study those things but there is smell olfactory smell and taste what you call just gastatation it's interesting okay so we have these sensory modalities yeah these are basic sensory modalities that have been evolutionary conserved across many many different species and it's very important that we study information integration in the central brain so my lab doesn't study visual processing and peripheral river processing say in the retina which is optic lobe in the fly brain or periphery auditory or olfactory processing we don't study that we just study the higher order processing in the central brain especially information integration and decision making things like that a main takeaway it seems like from this conversation is also that it's important to in a sense humble ourselves with how many different modalities of communication we use visual auditory somatosensory olfactory et cetera with us and then basically out of the 10 million species on the planet that there's also different modalities of communication between all of those species and with specific hierarchies of decision making involved in why they're communicating in those specific ways and so it's basically this massive open a catalog that we have yet to really investigate into that you're at the frontier of investigating into and you're doing it with one species the Drasophila and there's so many other species of course the Drasophila is one of the best scientific species for us to invest in a model system a model system like the zebrafish that we use to convince our model systems wow they're like little like gifts of nature like little hacks that make it easier for us to yeah yeah it's easier because there's a community of genetics study for more than 100 years for Morgan time that really allow us to dissect the brain of a already relatively simple brain almost one cell at a time so this is really unparallel out in any other species say zebrafish or rodents because you don't have such a great reptivar of genetic tools yeah the once you have like the entire cell lineage of an animal and also the or an insect and then also the entire connectome of the nervous system then it becomes a model a more of a model and also it has high transgenic abilities high reproduction rates all this type of stuff which make it another thing that I think is really important for you to teach about is that a lot of people are wondering ok like if you're going to be recording all of this video this audio all of the somatosensory interactions trying to make catalogs of all of them it's obviously a tremendous amount of data you're probably trying to store petabytes of data of fruit fly interactions yes yes so the question is how do you in your lab most effectively store that data and most effectively analyze that data so you have to do all this digital signal processing of like audio and video so how do you guys do that part? yeah we use the ladies tools from machine learning to do video analysis audio analysis we invest a lot on the hardware infrastructure computational storage infrastructure and many of the graduate students actually come from engineering background say computer science electrical engineering things like that so really the way the modern neuroscience system neuroscience is studied takes a shape that's slightly different from traditional necrobiology so neuroscience system neuroscience really focus on information processing is it like assigning variables to different fruit flies let's make it easy fruit fly A, B, C in the chamber B, at a very specific time stamp makes an audio transmission to fruit fly C is that how you're approximately that's exactly what we do so we try to record the videos and audio simultaneously and then we do the analysis and we do the synchronization so that we know at each specific time what are the states for specific flies so among the different flies in the chamber say we pick one of the flies as a fly of interest and by doing this analysis we know all the sensory information the context the social context that available to this fly of interest right and that's what the brain is about to process and if we get the statistical signature of that state and we can try to recapitulate that statistical structure with desert, with more controlled program stimulation so we can generate hypothesis by doing that okay so I observe fruit fly B making an audio transmission to fruit fly C and then you record that exact transmission and then maybe you also saw that that transmission was the song and then B approach C and C slow down enabling B and C to go through a copulation process maybe then you take one of those two fruit flies into the tethered environment and then you re you simulate what you initially logged as that process of B making a song maybe you take C into the tethered environment and you simulate B's audio stream and then see how C's neural architecture responds and then that's maybe what you're trying to develop as these correlates okay and there's 10 people right now in the lab and all of the 10 in the lab are analyzing their own data building their own data analysis tools which is really cool I thought when you were teaching me this that's a pretty good way to run the lab so everyone's gaining this experience of knowing how to do the data analysis which is such an important part is doing that that way they can all go and also do their own experiments more easily in the future and you're training people really well yes that's an idea yeah and then will you tell us what's available from your lab like if other people from around the world are interested in collaborating with you what parts of your techniques are available for like non-commercial use yeah we don't have any specific commercial interest at the moment we develop and use tools for behavior for anatomy and for physiology including functional imaging and actual physiology and many of those tools actually are available for other people especially for people who are here on trust of course so then scientists that are using that are trying to study for non-commercial purposes are able to get in touch to potentially we also distribute say transdandy flies reagents for recording activities say GCAM7 flies we distribute it across many many places people get that prior to publication actually okay so your GCAM7 is what you have to transgenically add GCAM7 and that enables modify the genome of the fly make the transgenic flies that express the GCAM7 modify the genetics of the fruit flies so that they express GCAM7 which enables the electrical activity to fluoresce for when you image it so genetics can do that for very specific neurons say one or two neurons in the brain we can do that with as little as just one or two of the neurons out of the 100,000 to fluoresce that's really great so you can distribute that so people can ask for those cool okay what would be your ideal neuroscience tool of the future maybe like 50 or 100 years down the line what would be an ideal tool that would let you do the reading of the whole connectome do the manipulation to the connectome like what would be an ideal neuroscience tool for you yeah I think along the line of functional imaging there's a lot more to do in terms of the spatial and temporal resolution and the spatial and temporal scale so calcium imaging the the GCAM7 series allows us to do recordings across many cells a pretty high resolution across a pretty long time but that's not enough we want you to go into say more than one kilohertz a second in terms of temporal resolution we want to record from every cell ideally from every branch of the dendrite across the brain and we want to do that say across the entire life of the animal so that's the ideal tool I mean that probably is not just one tool it's probably a whole combination of tools including optical instrumentations including the molecular tools and the probes and including technologies for data analysis yeah I think those are well I would like to see in say one or two decades later yeah how do you think we can inspire more people around the world to work together education I think I think the people across the world they come from very different cultural background religious background there's a lot of barriers between different groups of people but there are also some general principles general beliefs that are held by everyone in the world and those things can be educated and people could be educated and understand to realize that actually people from entirely different background actually they have a lot more in common than they thought I think this is a message that should be delivered you know probably by the mass media than what they're doing now which is more emphasis about the differences between different communities yeah I like that answer a lot what do you think is a skill that young people should know as we go into the exponential technology age the skill there's a whole lot of different skill I wouldn't emphasize specific skill browsers but I do believe that information technology is something that are critically important for next generations as most rapidly developing one of the most rapidly developing direction and information technology I would consider neuroscience is part of this direction of information disclosure what would you say is the meaning of life this big human experiment that's happening the meaning of life yeah how to make sense of the world and to make sense of our self I think that's probably the meaning of life I like that answer making sense of our reality and our selves I like that a lot what do you think is the role of love in our world love is like a glue right it's kind of sticking putting things together like your first question about how people across the world could collaborate more and love is an important ingredient because you know love as we say is actually part of the emotional system right so it can modulate our brain and modulate our behavior in a way that in the absence of love will be entirely different yeah do you think that this is a simulation which one our reality the kind of definition of reality and simulation for a neuroscientist we do video reality all the time we trick the animal to believe what's actually not reality virtual reality so I think it really is just about information processing information in the brain in the outside we perceive the outside world through simulations in our brain so you can say simulation is reality and reality is simulation and last question is what do you think is the most beautiful thing in the world the brain is the most beautiful thing in the world and the society is that are composed of the brains if you the one that comes after and why it's just amazing no reason I love it this has been so mind expanding thank you very much for coming to our show thank you thank you so much holy cow wow everyone thank you very much for tuning in we greatly appreciate it we'd love to hear your thoughts in the comments below on the episode let us know what you're thinking have more conversations with your friends families co-workers people online about the neural mechanisms of social cognition about all of these different things that you was teaching us reporting electro activity through fluorescence pioneering novel methods and functional brain imaging optimizing ways of imaging physiology the neural mechanisms social cognition in general internal representations and the different communication modalities for biology just really think about this these topics more and have more conversations with your friends about them check out the links in the bio below to Yi's work in Westlake University and also support the artists the entrepreneurs the organizations around the world that you believe in support simulations we continue doing cool things like coming on site to Westlake University in China interviewing great professors and principal investigators like Yi and go and build the future everyone manifest your dreams into the world we love you very much thank you for tuning in and we will see you soon peace wow thank you oh my gosh it's so mind-blowing yeah you ask a lot of tough questions that I wasn't prepared at all