 Well, thank you all for coming, and David, thank you for that very kind introduction. So I'm going to be talking about whether we can upload our consciousness to the cloud. I should say that that is not actually the main focus of my research. I'm not actively involved in trying to do that, though as you'll see there are some people who are, but I thought it would be an interesting way to sort of frame a discussion of the kinds of things that I'm interested in. So I really wanted to start with this, which is a picture from Star Trek, which was really one of my favorite shows when I was a kid. And for those of you who don't know, this is the transporter from Star Trek, and the idea is that the characters on the show can stand on these little white discs, and then they slowly dematerialize from the enterprise, from their spaceship, and rematerialize somewhere else. And it actually turns out that the reason that there is a transporter in Star Trek was that the early set designers said it would be too expensive to bring the spaceships to the planet every episode. But it really is like one of the iconic features. It's turning a bug into a feature. It's really one of the iconic aspects of Star Trek. And when I was a kid, I spent a lot of time wondering about, well, what would happen if I dematerialized and then rematerialized somewhere else? Would I really be the same person there? And sort of the premise of the transporter in Star Trek is that it destroys each one of your molecules, and then it puts that molecule in exactly the same place in some other location. And at least for me, even though I was a little troubled by it, I concluded that, yeah, I would likely be the same person with the same thoughts when I was transported to somewhere else. That's not a universal feeling, and I think one of those might be Dr. McCoy, and fans of the show will know that Dr. McCoy did not feel comfortable being dematerialized and was usually worried about whether he was the same person. But the fact that I was convinced I was the same person basically put me in the camp of being a materialist, in that I believe that whatever it is that makes me me, whatever it is that gives me my thoughts and my feelings are contained in my physical body, unless I posit that the soul is transported along with the brain. But that was not something that was covered in the technology. So the idea is that somehow my existence emerges from my brain. And so that actually did, thinking about that actually is part of what set me on the road to being a neuroscientist. The other thing that set me on the road to being a neuroscientist, this again dates me, this is from Star Wars and this is from Terminator, was the idea that it would be, well, if the brain is a physical circuit, then why couldn't we build machines that have the same abilities that we do? And obviously, I'm not the first to think about that. And so I thought that it would be really cool to be able to build machines. And in fact, for my graduate work, what I did was I worked on artificial neural networks from a biological perspective. So I was interested in figuring out how the neural circuits that are actually part of real organisms differ or could be sort of abstracted and made into something that could do computation. Around the time that I finished my graduate work, I had to make a decision was I going to continue to try to build a terminator or was I going to try to understand how the brain worked. And I went with understanding how the brain works, but I never lost interest in that. Interestingly, a lot of my friends who were in that field at the time now work for companies like Google and Facebook. And as many of you may know, this technology actually kind of now work. So you can get if you, some of my families from the Bay Area. And if you go around the Bay Area, you see Google self driving cars all the time, you can buy a Tesla self driving car. And pretty soon they're going to be common. So in fact, Elon Musk, who is the founder of Tesla, and a great inventor who among other things wants to send people to Mars, having spent a lot of time working with his company, developing self driving Teslas is actually worried that artificial intelligence AI will take over. And he is interested in making sure that that doesn't happen. And so none of this though addresses exactly what I'm going to be talking about, which is not whether we can build robots that either serve coffee or drive cars or take over the world. But can we can we achieve immortality by uploading our brain to the cloud? And this is from a movie, Johnny, starring Johnny Depp a couple years ago, where Johnny Depp for various reasons is he's pretty grim about it here. But he's got to upload his brain to a computer. And it's not the first move. So there are people who seriously who take that idea seriously. And last year there was an article in the New York Times called the neuroscience of immortality. And Amy Harmon went and interviewed a bunch of scientists about whether they thought that this was going to be possible sometime soon, and what the prospects were. And there are people who seriously want to achieve immortality by somehow taking their brains and taking the essence of what is in their brains and putting them onto a computer. And so is that possible? Well, I will say that there's one very well known scientist named Winfrey Denk, who is well known to many of the neuroscientists or non neuroscientists at the lab, who is a director of the very prestigious Max Planck Institute who says why I can see he says, I can see within say 40 years that we would have a method to generate a digital replica of a person's mind. And then he says, well, it's not my primary motivation, but it is a logical outgrowth of our work. So reasonable people, and I think everybody who knows him would agree that he's reasonable actually do believe that this is possible. So what would that look like? So I'll just add that there's actually a project funded by a Russian billionaire to achieve it's called the Avatar project. And it's got the milestones here and they literally want to achieve uploading consciousness to the cloud by 2045. So what what what would that entail what would be required and that's really going to be the focus of my talk, which is what what is it that makes us us and how can we figure out what what that is and really what it is that makes us us what is the neural circuit. It's as David mentioned, there are 100 billion neurons and 100 trillion connections between these 100 billion neurons in the brain. And it is the details of of those connections. It is how those neurons communicate with each other that endows us with our our character with our individuality with our ability to do everything that we do to remember things to think to feel to act. All those things. Somehow that circuit is what makes us us. And so what I'll be telling you about is the technology that we've been developing to actually figure out what that circuit is. So the idea that that behavior arises from a neural circuit actually goes back quite far. The earliest demonstration of it that that I've seen goes back to Renee Descartes, who's some 350 years ago now, postulated a mechanism by which we withdraw our foot from a hot fire. Right. So everyone had the experience if you touch something hot, you pull your hand back. So so how does that work? And he came up with really the first mechanistic circuit mechanism for that. And here's what he proposed. He proposed that the flames of the fire displaced the skin, which pulls on a tiny thread. The thread opens a pore in the ventricle. And then the animal spirits flow from the ventricle F through a hollow tube. And the animal spirits inflate the muscle of the leg causing withdrawal. So not everyone here has a medical background. But let me just clarify that that is not actually in detail how we think this happens today. So we know that the animal spirits are actually contained a little bit to the front here. No, it's easy to make fun of him. He had several disadvantages, which is he proposed this before they figured out that a good way to do science is actually to do experiments. So the way he came up with this was he just wrote it down. He was a smart guy, and he wrote down some smart things, but he did not use the scientific method to propose a hypothesis and then test it. But what I like about the hypothesis is that it at least has the form of a physical explanation for what's going on. And in that way, it really is a milestone in neuroscience. I mean, you know, we're making I'm making fun of it. But but it really it you one could imagine building a machine or a physical device based on these principles, right? So in fact, we we would call that a reflex. And the reflexes is worked out in other cases actually it's worked out for that too. But this is probably the the most famous one when you go to the doctor, and he taps your knee. What happens is that the hammer pulls on the patellar tendon, which stretches this right here. And then there's a nerve up here. It's a nerve that goes up here forms a synapse. And then the activity here activates another nerve that activates the muscle. Okay, and that's what it's when the muscle is activated and contracts that you kick your knee. Right? So this is actually a simple circuit that we understand and that explains something that is relevant to all sorts of things. So it's a it's a circuit based explanation for what's going on. And I should I should add that although I'm not talking tonight mostly about neuropsychiatric disorders. Neuroscience circuits disrupted circuitry is at the at the basis of many if not most neuropsychiatric disorders. So right now over the last 20 years, I would say, drug companies have largely stalled in making new drugs for neuropsychiatric disorders for conditions like depression and addiction and bipolar and schizophrenia. They've they've actually they've run out of ideas that many if not most pharmaceutical companies have actually shut down their research because the traditional approaches to to making progress on these disorders do not involve circuits. And because we don't understand enough about these circuits, they these these companies have have sort of given up. So what where has the progress been well actually interestingly most of the progress in the last decade or two in treating severe neuropsychiatric disorders does not come from from drugs. It actually comes from Oh, well, okay, this is a slightly different slide. But basically, so the the usual the pharmaceutical approach is to basically take imagine that the the brain is soaked in a chemical in in some neurotransmitter and that if you the idea is to provide a drug. And if you give a drug that changes the level of some chemical like serotonin that you can ameliorate one condition like depression. And then if you give a different drug, you might ameliorate a different condition. But the idea is that it treats the brain largely as being bathed in this in this sort of pharmacological soup. And what these slides are out of order. But what I was going to say is that the the progress over the last 10 or 20 years in treating neuropsychiatric disorders has not come from drugs. It's come from targeting specific circuits of which we know only a handful. But the ones that we do when properly targeted can actually result in some very impressive improvements. So for especially for Parkinson's for chronic pain, pain for depression, there are some really impressive results. But obviously neurosurgery is is is a is a very invasive kind of treatment. So right now, these these treatments are only used in cases where the traditional pharmacological approaches fail. So that that's my my digression on neuropsychiatric disorders. And I'm going to return to the main path of my talk, which is going to which is focusing on the fact that we need to know the brain's wiring diagram. So that's what's missing. We really don't know the brain's wiring diagram. If we knew it in in greater detail, we would have a better understanding of how we think and we feel and what goes wrong in these various disorders. So those of you who read the news and follow the science paper science times, let's say, might wonder, Well, what are you talking about? We recently actually did figure out the brain's wiring diagram. And so earlier this over this summer, with a lot of fanfare, there was an announcement that we now know the the connectome of the human brain. This was a the result of a very impressive multi year multi center, multi million dollar project called the human connectome project led by David van Essen. And they they carefully went through human brains and discovered they they cataloged the various areas, and they discovered 97 new regions bringing the total to 180 and I'm I don't want to diminish this in any way. This was a real tour de force. And it is important work. Similar work has been done in the mouse, which is the organism that along with the rat that my lab studies, where we know a fair amount about the wiring of the mouse. And I'm just going to make a little a little digression here and explain that the the mouse, you might wonder, you know, why why can we study the mouse if we're interested in people turns out that the the mouse the wiring diagram of the mouse brain is as far as we know very similar to that of the human brain, the monkey brain. And in general, the the the overall organization of the nervous system of mammals is is shared. So here's the mouse and here's the human. And if these look similar to you don't don't be distressed because they also look similar to me and they look similar to any neuro anatomist. And so the idea is that all these different organisms mouse and rat and rabbit and goat and cat and cheetah lion dog and human share the same basic organizational scheme with the obvious difference here of size. And you know, here's the mouse and here's the human. So basically, it seems like what happened is over evolution that mammals figured out how to build a brain. And then they built it on different scales. All right. So all right, we have the mouse connectome, so to speak, the human connectome. Why isn't that enough? Well, basically what we have is something that looks kind of like this. So these are the this is a highway map of the United States. And this is sort of the scale on which we we know the circuitry of the human brain. I didn't count up these little nodes here, but they're probably about 100 or 200, which is about as many areas as they discovered in the in the human connectome at least. But what we really need is something that's more like this on a very different scale. So this is, of course, Cold Spring Harbor, and this is 25a, which is what you came in on to go. And this is Bungtown Road. So we need a much, much much higher resolution map of the connectivity of the brain in order to really navigate through these circuits, right? It's nice to know that if you get on to a major highway, you can end up in Cincinnati. But if you actually want to go if you if you want to understand how signals, or in this case, cars go around, you need much higher higher resolution. So why don't we just get that? Why don't we just go and use whatever techniques were used to get the other atlases and use that to figure out this thing at at higher resolution? The limitation isn't simply that, you know, microscopes don't see well, we can we can see individual cells and synapses. But this sort of illustrates the problem. So here at up close is what an actual bit of cortical circuitry looks like. So what you see here are about a dozen or so neurons in the cortex of a mouse. Each one of this is the cell body of the cell. And this is the dendrites that receives information. And coming out of some of these are axons that send information. And what you can see is that it's a big mess. So if you wanted to know what this neuron, which where this neuron sent its its information where it sent its wires, you can see that you'd be really at a loss, right? You you could follow it up to there and boom, you lost it, right? And if you wanted to know how all these wires where all these wires went and how they were connected, what that spaghetti is like, you would really be you would find that challenging. And that right there is the problem that we face when trying to decipher and tease apart the wiring of a of a real neural circuit in the brain. So what can we do? Well, traditionally, the solution to that the way we figure out neural circuits is we don't worry where individual neurons go. We look at where collections of neurons go. So if we want to know where the neurons from one region send their information, we inject a tracer. There are lots of different tracers, but and some they all have special properties. But in general, what they do is they label neurons that in this case green. And then what what we can do is look at other brain regions and see if we see green. And the way this this experiment is set up, the only source of green is the neurons here that we labeled. So if we inject our tracer here, and we dissect out an area here, a different part of the brain, and we see green, then we know that there's information from these neurons that got sent to those to that location. If we dissect out five different regions, and we see green in three, we know that the information went from each of these to front. We know that the information went from this region to each of those regions, and not to these other two. But unfortunately, what we can't tell from those kinds of experiments is whether it looks like this where each neuron sends its information to exactly one region, or whether, for example, it looks like this where each neuron sends its information to all three of those regions, right? Or, right, so here this neuron is sending its information here and there and there and this to there and there and there and so on. Or maybe it looks like something more complicated. And each of those, each of those different pictures has different implications for how we think about information flowing through the brain. So for example, if your auditory information can simultaneously be must always go simultaneously to the part of your brain that processes fear and the part of your brain that processes music, then you might think that you would always be afraid when hearing music, right? And that would be, that wouldn't make sense. It wouldn't, it doesn't conform with how we think things work. And in fact, there are different neurons that send their information to different places. And so if we're really going to understand how that circuit works, we need to understand how which of these situations, which of these configurations is, is how it goes. So this is just to drive this point home. This is an example taken from one of my students, Eustis, who may be in the audience. So Eustis is a very hard worker, but after working very hard for a while, he decided he was just going to go on vacation. And he decided, all right, he's a spontaneous kind of guy, he's just going to show up at the airport and buy a ticket to somewhere fun. And so he figured, I don't know, he would go to Brazil, because that might be a fun place or if the next flight was to India, maybe he'd go there. And so he showed up at the airport. And he was very disappointed to discover that that La Guardia, which looks like a bus station, only flies domestically, except for this one route, I don't know where that one goes. So that that sort of drives home the importance of routing of information. So in this analogy, the airports are neurons. And if the United States had 100 billion airports, then it would be a perfect analogy. All right, so there is a technique that allows us to trace the flight paths of individual neurons. And it's a very tedious one. You can take each individual neuron and fill it with a tracer. And it's a very laborious method. And as you can imagine, and then after you've you've you filled that neuron with a tracer, you have to dissect the entire brain and trace the processes of that one neuron throughout the entire brain. And that there is an art to doing that. It's been around for 100 years. A lot of what we know actually comes from that kind of work. But as as you can imagine, it's incredibly slow and laborious. So you know, we've been talking numbers, there are millions and billions of cells in the brain. And really, if that's the way we're going to figure out how the brain is wired up, that is serious job security for teams and teams of people. And no, I mean, again, I this is this is a paper that was published in a respectable journal a few years ago. It's one that my lab refers to a lot. But you know, it represents the state of the art, there were 25 neurons successfully traced in 28 rats. And if you do the math, it takes a long time to get 100 million, or even 100,000. So there are some some clever approaches that have emerged over the last decade or so. And this is really the one that is the inspiration for the method that that my lab developed. It's called brain bow. One of the keys when coming up with a new technique is to give it a catchy name. So the idea is it causes neurons in the brain to form the colors of the rainbow, hence brain bow. And using some very clever genetic manipulations, what they did was they tricked neurons into expressing proteins that have colors. And each neuron in this in this specially designed mouse, expresses a random collection of any of these different colored proteins. And so the idea is that it's like painting by numbers, each neuron will have a unique color. And because each neuron has a unique color, these axons, these little wires which otherwise would look completely identical, the idea is that they won't look identical. This one will be blue, and this one will be red, and this one will be purple. So in principle, the theoretical diversity of this approach allows for up to 200 different colors, which is really quite a few. But the problem is that similar colors are hard to distinguish. So although the theoretical diversity based on this genetic trick is over 200, in practice, using real microscopes, you can't resolve more than just a few 10 as the absolute limit, but maybe three or four typically. And that's shown here, based on this randomness, I can be pretty confident that this neuron and this neuron and this neuron are all different colors. But they appear to be the same. So it's a sort of Zen issue. If neurons are different colors, but you can't see them, can you trace? Well, do they make a sound? I'll work on that one. So but that was the inspiration. That was the inspiration for our idea. The idea is that if only you could resolve those colors, if only there were more than than 200, more than five or 10 colors that you could resolve, then then we'd be set. And so because I'm at Cold Spring Harbor, which is effectively ground zero of DNA sequencing, I had the idea that instead of using colors, we could use sequences of nucleic acid. So as as many of you know, our genetic material consists of long strings of of four nucleotides, four letters, AG, ATGC, and any long string of them combined together, eventually produces a unique long string. So if you just put together random strings of letters, you can be pretty confident that those will be unique. So here's the idea in graphical form. If we could just replace those colors with numbers, if we have a way to read out the numbers, we're set. So each neuron, the idea is to express a unique sequence of DNA, a genetic barcode, a unique random label. And the theoretical diversity here is effectively infinite. So here's the core idea. Let's recast connectomics as a problem of high throughput DNA sequencing by labeling each neuron with a unique DNA sequence, a unique barcode. So why sequencing? Well, there are two reasons sequencing is incredibly cheap, and really fast. And so this is a plot of the cost of DNA sequencing over the last 10 or 15 years, 15 years, I guess. Back in 2001, where when the human genome was sequenced, thanks in part to Jim Watson's efforts and a lot of the efforts here at Cold Spring Harbor, the cost of a genome was of order a billion dollars. Yeah, of order a billion dollars. And you can see that the cost drops. So this is a logarithmic scale and the cost was dropping at a very respectable clip. So this line here is called Moore's Law, and that is the rate at which computers have gotten faster and cheaper with time. And so they've gotten faster and cheaper at an impressive rate, as everyone knows. This my my iPhone here has more computing power than the top of the line computer that I used to do my PhD thesis 30 years ago, in fact, considerably more. In fact, I looked it up, and it has as much computing power as a supercomputer from the early 80s, which is a little bit before I did my thesis. So this is a very respectable rate for things to get faster and cheaper. But what you can see is that the rate at which DNA sequencing has gotten faster and cheaper was was keeping up with Moore's Law up here. And then starting in 2007, due to the advent of so called next generation sequencing, it there was a kink and it started dropping at an incredibly, incredibly, incredibly fast rate. And so this curve is not quite as straight as that one, but it's still getting faster and cheaper and cheaper and cheaper. And so if you can convert, what we are reasoning was if we could convert the problem of connectivity into a problem of DNA sequencing, it would now be tractable, there would now be, it would not be expensive to figure out the wiring of an entire brain. Alright, so the idea is that we can randomly, that random sequences can uniquely label neurons. And just to drive that home, there are four different nucleotides. So the number of possible barcodes of length one is four, right, it can be A or T or G or C of length two, it's four times four equals 16, because you can have an A in the first place, and any of the four in the second place, you can have a T in the first place, any of the four in the next place, and so on. And so it's 16. And the overall pattern is four to the nth power. And we use barcodes that are of length 30, which are, which produce an astronomical number of possibilities. So 30 is a very short piece of DNA, we have over a billion nucleotides in our genome. So 30 is nothing. And yet, even by the time we get to a barcode of length 30, we can be guaranteed that we have a unique sequence. It turns out that 10 to the 18 is more than the number of all the grains of sand on the earth. So already by the time we get to even these short barcodes, we have astronomical numbers of possibilities. So the 10 to the 18 grains of sand are all the beaches of the world. And there are only about a few million neurons in the mouse cortex, which means that it says, though, each particular neuron, even if we barcoded every neuron in the mouse's brain with this diversity, every neuron in the mouse's brain would be selecting from 100 billion different possible unique choices and still without overlap. So because we're drawing these random barcodes from such a large pool, the chances that two neurons get the same barcode are infinitesibly small in theory. There are some practical issues that I'm going to skip over. But in theory, that's true. Alright, so how do we do that? Well, the way we do it is we don't actually currently barcode neurons. We actually barcode viruses, which we then used to infect neurons. So a virus, it turns out, is just a little it's nature's delivery system for nucleic acids. And our lab and many other labs use viruses to deliver sequences of nucleic acids that we would like. So what we did was slightly different, which is rather than having all the viruses in our tube be the same, which is usually the way we do it. Instead, we made each virus unique by this is a little picture of how you make a virus. It's a little piece of DNA, and we cut it and then we insert these little pieces in here. And these pieces have random sequences. So we insert these pieces with random sequences in these. That produces what we call a plasmid library and from there we get a virus library. So each one of these little things here is a virus and we've just for for illustration purposes given it a unique two digit number but in fact it should have a unique 18 digit number. Alright, and so then we take those viruses and we inject it into a into one part of the brain. And then we've also engineered a a protein that drags this barcode out to the synapse. And so this picture here summarizes about three years of work by a former graduate student of mine in Paikon along with some help from Eustis. But that is the the the overall scheme is that these viruses express or have this protein that grabs onto the barcode and drags it out to the synapse. So our first application of this was to look at the connectivity of a brain region called the locus ceruleus. So the locus ceruleus is a little brain region that produces a neuromodulator called noradrenaline. It's the sole source of noradrenaline to the cortex. So noradrenaline is a is a neuromodulator that causes neurons to wake up. It's like an alert signal. Alright, so when you get squirted when neurons get squirted with noradrenaline they become active. It says pay attention. There's something that's about to happen. And what we're interested in was how this circuit was wired up. And the reason is that the traditional model of how the locus ceruleus is wired up was that the locus ceruleus it was thought produced a sort of global signal. So the idea was that each one of these neurons just sort of spread its neurotransmitter globally throughout the brain as though there was a single alarm bell that when you hit it caused the entire brain to wake up. But there was another view that seemed sensible to us, which is that well, maybe you don't need your entire brain to become alert all at once. Maybe if you know that there's going to be let's say a baseball coming at you, maybe your your visual system should be should be tuned up specifically. Maybe that should be revved up and maybe the rest of the brain can can be not quite as revved up. So maybe these neurons could could provide more specific inputs to particular places. Right. So if in that view, maybe one neuron projected there and another to another part and a third to a third part of the brain. So the anatomy there as you as you could see if if all these neurons on average project to all places that if you just label these green, you can't tell these apart. What you need for this is to know where individual neurons project. And so Eustace came up with an experimental design in collaboration with our lab next door, Florin Albianu. And what they did was they injected into the locus ceruleus, which is right here. And then they cut the brain into slices and extracted the barcodes from each one of those slices. And so the idea is that if you see a particular barcode in this particular slice, then you know that the neuron here projected there. And so we could we could get all the barcodes from all the slices. And so what we found was in fact that not all neurons projected to all places. So here is a plot of the strength of the projection on this axis as a function of the position along the back front axis of the brain. So this particular neuron, which happened to be barcode 28 projected very strongly to this particular slight area of the brain and almost to nowhere else. So that was supporting the idea that you could have this very specific kind of control over the activation. And other neurons did other things. So this particular neuron barcode 79 projected more broadly. And a third neuron had a more complex pattern. So in fact, we saw a variety of different projection patterns. And out of the 1000 neurons that we looked at, each one had a somewhat unique pattern. But what what we got from that is the idea that in fact there is a degree of specificity that was not expected from the previous notions of how the locus ceruleus operated. Alright, so basically, that model is wrong. The solely specific model is wrong. And really what we think is going on is that there's a mix of different projection patterns, some of which are specific, and some of which are wiggly like this. Okay, so my main interest really is in the cortex, the part of the brain that allows us to think and reason. And so we're now scaling up to a much larger experiment, this is still underway, where instead of making only a single injection in the brain, we make hundreds of injections and tile the entire brain with barcodes, and then section the brain into much smaller pieces. So we get much higher resolution for where the different neurons send their information. And so here is a pilot experiment in which the entire brain, the wiring of the entire brain is captured in a single experiment. Each one of these balls, though, represents over 1000 individual neurons. So to drive that home, you can see here, this is some more significant fraction of the 50,000 or so neurons that were captured in this experiment. And for each one of those neurons, we have information about where so this is we just picked two here for illustration purposes. So this particular neuron sends its information there, this particular neuron sends its information there. And we have 49,998 more that we're thinking about and trying to understand right now. So these this is still early days, but this is orders of magnitude, more information about the connectivity of the brain than than really any other lab has ever had before. So in what can we use this for other than showing pictures of rotating brains? Well, one of the things that we're particularly interested in is understanding what goes wrong when in disorders like autism and schizophrenia that we know lead to disruptions of brain wiring. So we have mouse models of these disorders in which we know that they're behaviorally compromised. And we know that there's something wrong with the brain wiring, but we don't currently have a systematic way to go through the brains of these animals and figure out what exactly is wrong. We believe that if we can figure out what is going wrong in the brain of a mouse that expresses a gene that we know in a human causes autism or schizophrenia or depression, we think that that will provide at least the starting point for understanding what goes on in in in people. So that's that's one direction that my lab is going and then very briefly, I'm just gonna mention that. Well, so most of my lab until I start working on the connectome really has most of the work that I've done here through my career has been focused on understanding behavior, understanding the relationship between behavior and the circuitry that causes it. And so this is just I'm not really going to go through it, but it just it shows how we study that in in rats and mice. We train rats to to poke into a center port here. And then that causes a machine to our computer to deliver a stimulus. And then the rat has to make a decision about whether the stimulus had one property or another. In this case, it's being asked to determine whether the sound was lower, high frequency. And what we do is we try to figure out the neural circuits that enable the animal to make the decision about that. I'll just show you that because it's fun. I don't have the sound here, but go beep, beep. And the point is that it's like a little machine here. So early on I started this this talk talking about reflexes. This isn't a reflex, but it's a it's a bunch of decisions that we've we've trained the animal to perform in a systematic way that we can we can study and understand. And then what we do is we can look at the the activity of neurons using a variety of methods. This is an imaging method where those flashes of light correspond to to neural activity. So I just want to return to the question that I started with, which is so does this all mean that we're going to be able to upload our consciousness to the cloud anytime soon? Well, it saddens me to say that I don't think that anything that I've talked about is going to allow us to do that. I think that we'll understand a lot about how brains work in general. It might help us build build better machines or scarier machines depending on your perspective. But I don't think that we're going to be able to achieve immortality through through these methods or even by just looking at the circuits and to sort of explain why I thought I turned to the first car that I ever owned, which was a Fiat 128. It didn't look as nice as this one. But what I remember about it was that it was incredibly idiosyncratic. It made all sorts of noises. It often failed to start. But in ways that I could understand and sometimes predict, the gas tank only held three gallons, which was sort of like driving electric car, I guess. But it when it came off the assembly line, presumably it was fairly similar to the one next to it. And yet by the time I got it, it was really its own individual with its own history that caused its own dings and bings. And I can't imagine that no matter how much money I paid someone that they would be able to take my car and imagine that this was my car and replicate it and produce a copy of it that captured all the little dings and bings and quirks and idiosyncrasies. And so to the extent that I've been living in my brain now for a few decades, I feel like I'm sort of approaching the level of this Fiat. And it's highly unlikely that if you made another copy of me, you would actually be able to capture all my idiosyncrasies in a way that at least would be satisfying to me. So I'm pursuing more traditional approaches to immortality. I'm trying to leave a legacy by doing scientific work and by leaving children who might, well, we're in the audience early on, but got bored. And so with that, I just wanted to close and share with you this bit of wisdom that I got on a fortune cookie. If the brain were so simple, we could understand it. We would be so simple, we couldn't. And with that, I want to thank my collaboratome. In particular, my students, Ian Picon and Eustace Kepchol, who really drove this work, as well as many others, and the people who funded this work, my fundosome, including the Simons Foundation and Paul Allen Foundation and Cold Spring Harbor. Thank you for your attention.