Episode Transcript
Ana Marija Jaksic [00:00:00]:
Or you can automate the whole process and have a robot pick flies for you. When I say that we're testing flies to do a learning task, we're actually having a robot do this experiment for us, and then waits for a fly to solve a task. And then it picks each individual fly, remembers and learns which fly is a good learner, which one is a bad learner, and then selects from this really large pool of flies the best performing ones.
Mizter Rad [00:00:36]:
Welcome to the Mizter Rad Show, where I talk to the most interesting global personalities about the future of humanity. Hello, beautiful humans. Today, I have the absolute pleasure to have Ana Maria Jaccic in the show. Ana is a researcher and evolutionary neurobiologist at the Ecole Polytechnique de Lausanne in Switzerland and is focused on. Listen to this. Making flies smarter. Ana, how are you doing? And why are you making flies smarter?
Ana Marija Jaksic [00:01:17]:
Hi. Hi there. Thanks so much for inviting me here. Why do I make flies smarter? Well, I'm making them smarter to be able to understand how a brain works and eventually how our brains work. We're trying to kind of dig into the nuts and bolts of the fundamentals of the function of a brain. And the reason why we use flies here is because they are a simple model system to study on multiple levels, going from the genetic level to molecular level to structural brain level. And in the end, we can study cognitive phenotypes or cognitive traits such as learning and memory and all of the different things that we call intelligence nowadays. And so a fly has a numerically tractable brain that we can go through and study and take apart piece by piece.
Ana Marija Jaksic [00:02:23]:
And this is kind of a traditional way that people have been trying to understand the mechanisms behind the brain function. But what I'm trying to do is trying to flip things around. So instead of breaking the brain piece by piece and looking at how different components, like what are the essential components, I'm rather using forces of evolution that have already built up a brain from scratch in natural populations, including the human one. And I'm trying to use these forces to build the brain from something that is relatively simple in a fly to something more advanced. And with the process of building the brain from bottom up from scratch, I kind of plan to understand what are the small components that we might miss by just breaking the brain top down.
Mizter Rad [00:03:23]:
So when people, when scientists, researchers break their brain top down, how do they actually do it? Do they use a real brain or. This is a very simple and dumb question, but I think it's important to clear this out. They don't open up the heads of humans?
Ana Marija Jaksic [00:03:41]:
No. So I know it's a fantastic question. So to kind of break a brain, we don't use human brains. We use model systems, such as the fruit ply, which I work with, or the mouse. Mouse brains or rat brains. And what you normally do, you would use different sets of transgenic tools. So these are bioengineering tools that are available to scientists and neuroscientists to, for example, silence a particular neuron or silence particular groups of neurons in the brain and see what this does to the behavior of individual fly or an individual mouse or whichever system you're working with.
Mizter Rad [00:04:29]:
Okay, so you basically have the whole system, and then you start, bit by bit, analyzing different kinds of mechanisms based on whatever you input or how you alter the function of it.
Ana Marija Jaksic [00:04:42]:
Exactly. Exactly.
Mizter Rad [00:04:44]:
What you're doing here is you're going from the bottom up, and you say, okay, the fly is a very simple organism with a very simple brain. I just want to understand how the simplest brain works.
Ana Marija Jaksic [00:04:57]:
Yeah. And not just that. Not just how the simplest brain works, but how do you enhance a brain that is simple into something that is more complex?
Mizter Rad [00:05:08]:
Okay, so that's.
Ana Marija Jaksic [00:05:09]:
I think that's the essential component of it. We're not trying to understand the simple brain, but we're trying to see how do you make a simple brain more complex, and what is required to make a complex brain.
Mizter Rad [00:05:23]:
When you mean complex brain, you mean like a smarter brain?
Ana Marija Jaksic [00:05:27]:
Yeah. So that's a really good way of thinking about it. You can either make a smarter brain, you can either make a brain that has more neurons and see whether this makes it smarter or not. But what I'm interested in, I'm trying to make, really a smarter brain that can solve intelligence tasks more efficiently.
Mizter Rad [00:05:50]:
Okay, so explain me the process. How do you make a smarter fly, for example, in this case, yeah.
Ana Marija Jaksic [00:05:57]:
So you start from a normal fly. Obviously, you take flies from nature that have lived their normal fly life. And we know that in nature, flies are very diverse. So there's a whole distribution of flies that can learn well or learn less well than the average. And so what you do is you take these flies, the whole set of diversity of flies, you put them in a lab setting, and you design a task that tests their ability to solve problems. So, obviously, for a fly, a hard problem is not the same thing as a hard problem for a human. So what we do is we create learning tasks where a fly has to pick between different colors or between different patterns. And we ask a fly to remember this pattern and pick it in the next step.
Ana Marija Jaksic [00:07:02]:
So you're testing basically the memory and the learning ability of a fly in the lab. And so now many flies will perform well at this task. Some will perform less well, but they will have some sort of.
Mizter Rad [00:07:18]:
Hold on 1 second. Something pops in my mind. How do you ask a fly to select a color on the first place?
Ana Marija Jaksic [00:07:25]:
Oh, yeah, that's a difficult challenge, experimentally and technically. So we have designed this really cool platform where a fly walks on an lcd panel that projects colors on the floor. And it actually walks in what we call a Y maze. It's a y shaped arena, y shaped room where a fly goes from one arm of the letter y, let's call it from one arm to another. And every time it picks one of the arms, it makes a decision. It makes a binary choice to go left or right. And so it can go and go around this maze in a loop. So every time it turns around, it has a new left and right choice.
Ana Marija Jaksic [00:08:13]:
And because we project colors on the floor, each arm will have a different color. So once a fly makes a decision to go left into a green arm of the y maze, it actually picks the green color. And if it picks a color that we don't want it to pick, we condition it using really, really mild electric foot shock that makes it uncomfortable for the fly to stay around a particular color. So the fly associates this discomfort with one color and then learns to avoid this color over time. And so how fast it learns to avoid a particular color or a pattern or whatever you project to it is our readout of how smart this fly is. Now, this is a very specific type of smartness. Obviously, if it can pick different colors, it does not mean that it will be able to pick other types of stimuli. So, for example, it might pick between odors differently.
Ana Marija Jaksic [00:09:23]:
It might do different tasks better than it does a color picking task. But we start very simple with this simple memory learning essay, and then we go and complexify this particular task further on as we identify smarter flies.
Mizter Rad [00:09:41]:
Okay, so the first easier, let's say the level zero for all the flies is this white maze with the different colors. And if they choose the color where there's not electric shock, that means they are smarter, let's say, after several trials. So those matter flies pass on to the next round.
Ana Marija Jaksic [00:10:06]:
Let's say yes, yes.
Mizter Rad [00:10:08]:
And what's the next round?
Ana Marija Jaksic [00:10:10]:
So the next round is pretty much the same, the same task. We just make it a little bit harder. We can make a small delay between, between the shock and when the fly experiences the color, we can go from picking between two colors to picking between gradients of colors. So we complexified the task as the flies pass different rounds. And so here what happens is that a new round means a new generation of flies. So the flies mate and they produce their little fly babies. And here the cool part happens is that all the flies that go into the next round are better than the previous, than the previous set of flies. So every single genetic component that they have that contributes to their ability to solve a task well gets combined in their progeny.
Ana Marija Jaksic [00:11:06]:
So instead of having an average fly that has one set of genes, that makes it solve a task well, now we have a new fly that has a set of genes of one smart fly and another smart fly. And you now recombine these different components that make both, both of the flies smarter into a new fly that will be just a tad bit better than its parents because it has a completely new set of gene combinations, or gene variant combinations.
Mizter Rad [00:11:39]:
So we already proved this, that this happens, that the offsprings are smarter than their.
Ana Marija Jaksic [00:11:45]:
Yeah, yeah. So this has been done not only by my lab, but by other labs where people have done selection for any type of cognitive behaviors, for learning and for memory, specifically in flies. And we have shown that there is a proportion of learning and memory, or ability to learn that is heritable and can be transmitted from one generation to another. Now, we're using this to kind of combine the best out of the best in flies to produce something that is cognitively or in the form of learning more extreme than any fly that you might find in nature.
Mizter Rad [00:12:27]:
So what happens now? Something pops in my mind that is a bit, a bit on the risk side. What happens if these flies go out in the nature, actually, for any reason? There's an accident in your lab, flies go out. What would you expect that happens?
Ana Marija Jaksic [00:12:42]:
Yeah, I mean, there's a reason there are no extremely intelligent flies in nature. That is because with every trait in a population, including intelligence or learning ability, there is a particular cost to it that gets selected out in natural populations. So I'm not very much worried about smart flies.
Mizter Rad [00:13:07]:
But what do you mean with that? Can you explain that? I'm not sure I understand.
Ana Marija Jaksic [00:13:11]:
Yeah. So, for example, I think humans are a very good example. Humans have these super big brains, super big heads, that kind of allow us to do all sorts of cool things, like record this podcast and talk about science and all the cool stuff like that. But having a big head comes with a cost. For example, it comes with an evolutionarily great cost, which we call fitness cost, darwinian fitness cost, where big headed babies are really difficult to give birth to. And this comes with a disadvantage where you have really risky childbirth because we have big heads. And so this is very easily selected against in nature. The more riskier it is for you to give birth to the next generation, the more selected against are you in the population.
Ana Marija Jaksic [00:14:14]:
So, for example, the super smart flies in nature might have a great disadvantage, which prevents them to become super smart in the nature. So that's one way of thinking about it. And also, I mean, when I say super smart flies, you also have to understand that there is a limit to their smartness. They will not become masterminds of some sort. It's more like we try to make them a little bit better than the current population in terms of their learning ability to understand what gives them this ability.
Mizter Rad [00:14:51]:
There's matter in a very specific task, right? We cannot know if there's matter in general.
Ana Marija Jaksic [00:14:57]:
Yeah, exactly. So we test them for this color learning task, which. Yeah, absolutely. Exactly. As you said, does not make them are generally better learners, though. We're trying to figure out whether this is really true. Like, if you become a true expert in a cognitive task that is very specific, does this make you a better learner in other tasks? Are there some fundamental processes that are shared across different learning abilities that can be transferred to other tasks? We actually don't really know whether being great at one task generalizes to other tasks. And this has been like a big question in AI research recently.
Ana Marija Jaksic [00:15:43]:
How do you get to general artificial or generalizable artificial intelligence? By making a neural net really great at recognizing images or something like this.
Mizter Rad [00:15:57]:
Now that you touched the topic, I would like to understand. I listened to one of your interviews before, and you mentioned that one day we'll be able to use biological chips to power our computers. What do you mean with that?
Ana Marija Jaksic [00:16:13]:
Yeah, I mean, this is, I'm a Star Trek fan, so I always think about, like, crazy utopian futures where we use all sorts of technologies that we have today in a crazy way. I think one of the appeals of knowing how a brain works is knowing how to build better computers. And this comes with all sorts of challenges. So, for example, the way we build computers nowadays is based on binary code. So either one transistor is turned on or off. You encode the whole computation in terms of zeros and ones. And this is really powerful. I mean, we all know how powerful it is, but this is, for example, not how a brain works.
Ana Marija Jaksic [00:17:08]:
The neurons in the brain, even though they are a network that turns on and off, in a sense, does not turn on and off in a binary way. So there is, like, a gradient of activity in every neuron that makes it turn on or off. So the way we're building computer chips nowadays. Is not necessarily the best possible way to support artificial intelligence or artificial neural nets. So I'm kind of thinking that by learning how a brain works, we might be able to design chips or components of the computer that allows us to go from this zero one paradigm to something that is more gradual or even going back to something like more analog way of turning parts of the computer on and off to mimic how a brain works. So that's one way of thinking about it.
Mizter Rad [00:18:11]:
I want to pick your brain more on this because I want to understand exactly how you like. I really want to get it. What do you mean with that last part of more analog ways of computing or what exactly you said there?
Ana Marija Jaksic [00:18:27]:
Yeah, I mean, these are kind of like we're going into the weeds of how computers work, which I'm definitely not an expert on. But let's say we want to mimic how a brain works in an artificial system. We would ideally want to replicated in a way that is as true as possible to an actual brain. And so the way that we encode any sort of patterns in a computer is very much binary. And the way the brain encodes information is not binary at all. There are gradients of activity that account for past experiences that account for memories. And so a single neuron is not a single point that encodes zeros and ones. It encodes the entire complexity of a cell, a biological cell.
Ana Marija Jaksic [00:19:31]:
There's a lot of regulation within these cells that cannot be encoded by just zeros and ones, I think. And that's how I'm. I mean, this is really kind of higher level of thinking about what we can do by knowing how a brain works. So it's far away in the future kind of thinking.
Mizter Rad [00:19:57]:
No, I like it. I like it.
Ana Marija Jaksic [00:19:58]:
But, yeah, I have no idea how we will achieve this. I'm kind of hoping the computer scientists are gonna. Engineers are gonna be able to solve this component. But from our point of view, we're trying to understand the basic components of how a brain works so that we can advise engineers and computer scientists in the future on how to maybe construct or engineer new types of computers.
Mizter Rad [00:20:26]:
How advanced have we gotten already in the understanding of how a human brain works in your perspective, how a human brain works?
Ana Marija Jaksic [00:20:38]:
Yeah, I think we're far, far away.
Mizter Rad [00:20:42]:
Very far away from understanding 5% 10%, 2%. You would have to put a number.
Ana Marija Jaksic [00:20:50]:
If I have to put a number. I don't know. We know a bit, but not much. So we know, we know fundamental components of it, like how. How individual neurons work, more or less even that is a bit of a stretch, I would say. It's really difficult for me to put a number on it. I don't even know.
Mizter Rad [00:21:17]:
Right. But very. It's very, very unknown, basically.
Ana Marija Jaksic [00:21:20]:
I think it's very unknown.
Mizter Rad [00:21:21]:
I think people very small, very little.
Ana Marija Jaksic [00:21:24]:
Yeah. When we think about a human brain, it's always like, kind of a final frontier of biology. It's almost mysterious at this point. We know big, big basic components of it and chunks of it. But, like, for example, we don't even know how memories are really formed in a human brain. This is like a fundamental concept that we don't even understand how it's encoded in the brain. We're trying to understand.
Mizter Rad [00:21:54]:
And do flies have memory?
Ana Marija Jaksic [00:21:55]:
The flies absolutely do have memory. Yeah, yeah, yeah, they do.
Mizter Rad [00:21:59]:
So are you trying also to understand how it works in flies memory?
Ana Marija Jaksic [00:22:03]:
Yeah, yeah, absolutely. So it's actually quite, quite fun. One of the first ever discovered genes that kind of regulate or contribute to formation of memories has been described in the fly, in the fruit fly. And this is where the whole field of neurogenetics, or how genes contribute to building a brain, its structure and its function, came about. So flies absolutely do have a memory. They can form different types of memories, short term, long term memories. They have a working memory as well. And so we're using their ability to form memories to understand the basic concepts behind it.
Mizter Rad [00:22:48]:
So memory is correlated to smartness. In this case, if a fly gets electroshocked from walking on one of these colors, they will remember next time and walk up. Well, if they're smart enough, walk on the other color instead of where they got the shock first.
Ana Marija Jaksic [00:23:06]:
Yes.
Mizter Rad [00:23:07]:
So that would mean they're smarter, let's say, because they have a good memory.
Ana Marija Jaksic [00:23:11]:
Yeah, exactly. So one of the main components of a learning task is the ability to form short term memories or have a working memory. And so we have. So with flies, it's actually really fun to work on memory because we can turn on and off different genes that we know are essential for memory formation. And we can ask whether a fly that has no ability to form memories at all, like amnesiac fly, that is also learning impaired, can it really solve a task that we are asking it to do? And the answer is no. If you ablate or if you turn off genes that are essential for memory formation. The flies cannot learn and cannot do the tasks that we ask them to.
Mizter Rad [00:23:59]:
Interesting. What other kind of use cases do you see from what you're getting in your lab? The kinds of results, the kind of discoveries, the kinds of things you're learning, what kind of use cases in the real world you think we can see soon, or we already may be seeing.
Ana Marija Jaksic [00:24:23]:
So, for example, one of the big unresolved challenges in neuroscience or neurobiology are the neurological diseases that affect cognitive ability in humans. So we have all sorts of diseases that neurological diseases, such as Alzheimer's disease or Parkinson's disease, all sorts of neurodevelopmental developmental disorders that break down the cognitive ability of an individual. And so how this cognitive ability breaks down or understanding how it breaks down, really relies on our understanding on what cognitive ability is and how it's actually built up and structured. So one of the applications of my research and research of the entirety of neuroscience is understanding how can we fix cognitive ability once it starts breaking down due to all sorts of neurodevelopmental or neurological disorders. That's one very direct application of knowing how a brain works. It's kind of knowing how to fix it when it breaks down. That's one way of thinking about it. I mean, the experiments that we are doing are technologically really challenging.
Ana Marija Jaksic [00:25:52]:
You have to be able to test behaviors in many, many individuals in a high throughput way. So one of the side effects of doing basic science or fundamental research, like what I'm doing, my lab's doing, is this development of technology that can be used not only by fundamental neuroscience, but also by pharmaceutical industries that are trying to figure out what are the drugs that we can actually use to ameliorate behavioral disorders or cognitive disorders that we find in a human population. This is something that has been really difficult for pharmaceutical industries or biotech industries to tackle. And that's because screening behaviors in mice that are usually done in drug discovery research, that are usually used in drug discovery research, is really tricky. You need to be able to test behavior in multiple individuals at once. And this is a great ethical issue. The reason why this is more tricky in behavior is because behaviors are very plastic or very sensitive to environmental conditions. So inherently, the output that you find from testing for behavior is really noisy.
Ana Marija Jaksic [00:27:27]:
So the way people have been trying to tackle this noise coming from behavioral data is to just test more individuals. And this becomes very tricky with using mammalian model systems or vertebrate model systems, because it becomes a bit more unethical to just scale up and test thousands or hundreds of thousands of animals at once.
Mizter Rad [00:27:55]:
But with insects, it's not unethical or it's not considered unethical.
Ana Marija Jaksic [00:28:00]:
It's not considered.
Mizter Rad [00:28:01]:
You find it weird in some sense that that's like that because, I mean, insects are also, you could say they're part of. I mean, they are definitely part of the ecosystem.
Ana Marija Jaksic [00:28:12]:
Yeah.
Mizter Rad [00:28:13]:
You know what I mean?
Ana Marija Jaksic [00:28:14]:
Yeah, I know.
Mizter Rad [00:28:14]:
I know. Probably they also suffer. You cannot really hear they're crying, but they may suffer or not. You know what I mean?
Ana Marija Jaksic [00:28:22]:
Yeah, exactly.
Mizter Rad [00:28:23]:
Is it also something that you think about sometimes?
Ana Marija Jaksic [00:28:25]:
Yeah, I do quite a bit, actually. And I think it's a real challenge. How do you address figuring out how learning and memory or any type of behavioral trait works without testing a live animal? So, normally, in bioengineering, nowadays, in biotechnology, people have been moving away from testing on live animals to testing organoids that are grown in a petri dish using cell lines that are not full organisms. And this is really, I mean, it's been a tremendous progress in that sense. But for behavioral research, for figuring out how cognition works, you really need a fully functioning live organism that can perform.
Mizter Rad [00:29:19]:
You need the species to move the species to react with the live, real time action kind of thing.
Ana Marija Jaksic [00:29:30]:
And so now, one way of solving this, I mean, it's not a solution to a problem, but amelioration of the problem is, instead of working with a cognitively complex system, such as, like mammals, such as mice or rats, which can produce complex behavior behaviors, if you want to test something in a super high throughput way, you would rather want to go to a more cognitively simple species where the ethical concerns are not as great as in mammalian species?
Mizter Rad [00:30:07]:
Let me ask you something. Do you think that because of ethical concerns, we went from studying more complex species like monkeys or rats, and now because of, again, because of ethical concerns, we're studying insects like flies. And therefore, some scientists may say that we went back in a way, because now the simpler organisms are, let's say, less alike even than us.
Ana Marija Jaksic [00:30:36]:
Yeah, it's. You would think, like, if you think about it very simplistically, you would think, oh, yeah, we're studying something that is, that we think is very much a human trait, like cognition, and something in a species that might be so far away from humans that we would never be able to understand it. But I think the reason why we use insects is not only an ethical concern, it's also a technical opportunity or methodological opportunity where we. I mean, I'm going back to what I've mentioned that a fly has a numerically tractable brain where we can count up all the neurons. This is something that cannot be done in humans. And most of the, for example, the neurogenetics of neurogenetics of learning and memory, has found its fundamental components, really, in fly research, in figuring out how the simplest models work and function. And I think it's not something that. So, using insects is not something that we think about going back to something less convenient or less generalizable.
Ana Marija Jaksic [00:32:04]:
It's more like using a model system that is simple enough for us to understand and translate it into more complex systems. Picking apart a human brain or a mammalian brain is a huge challenge. A huge, huge challenge. And knowing the fundamentals is actually easier to understand than a simpler model in a more deeper way. So we can pick apart, really, the essential mechanisms easier in an insect model compared to a mammalian model. So that's the appeal of it.
Mizter Rad [00:32:44]:
This thing of going back to the basics, let's call it like that, to understand the. Like you said, a brain like the fly brain, that is. That is numerically trackable. Is this something that is new in this, in the field of sciences? Is this, like, a new trend? Were we doing this 20 years ago, 30 years ago? Or were scientists more focused on doing these studies on brains of mammals or other kinds of animals?
Ana Marija Jaksic [00:33:17]:
This is.
Mizter Rad [00:33:18]:
I'm trying to understand if this is. If we went. If we went from studying bigger animals to now going to the basics, and because we realized we're not ready to open up a monkey brain and try to understand how it works, we have to go back to the basics.
Ana Marija Jaksic [00:33:36]:
Yeah, I think this is a very, very old concept of studying simple models to understand both fundamental and complex mechanisms and how to translate it to systems like, there's million systems. But I don't think it's in terms of, you know, science and research. We haven't started with a simple model and then built it up to a more complex model. It's more like we're using everything that is available for us to understand the complexity of a behavior. So, for example, while flies can give us a really clear insight into the fundamental concept behind behaviors, behind how a brain works, we still need, and we rely on systems like the rats and mice and monkeys and humans even, to understand more complex behaviors and what are the fundamental principles behind those. So it's more like we're doing everything in parallel and just using the. The best opportunities that any individual system would be able to give us. So it's not like we're going back and forth between the models.
Ana Marija Jaksic [00:34:53]:
I think it's just different neuroscientists or different scientists are focusing on their model organism to tackle very different questions that are more approachable or more pliable using one model system versus the other. For example, you cannot really study language in flies. This is something that you cannot really do, but you can do this in humans. Of course, neuroscientists studying language and language processing would go towards human neuroscience or human cognitive neuroscience.
Mizter Rad [00:35:31]:
So, yeah, when you study the flies, can you understand that flies also suffer from neurological diseases? Let's say, maybe caused by. Because there's. There's a potential to. Some neurological diseases are caused by toxins and the air, you know, chemicals that are used to spray what we eat, stuff in the water. Do you also see that in flies as well?
Ana Marija Jaksic [00:35:56]:
Yeah, absolutely. In fact, flies have been used quite generally and extensively as neurological disease models. So, for example, we can use genes that we know are risk factors or contribute to human neurological diseases, such as Parkinson's disease. We can take some of these genes and transgenicly transplant them, or put them inside of a fly's genome and then test on the fly which drugs can be used to ameliorate the problems that come with Parkinson's disease, as well as Alzheimer's disease, and all sorts of neurological or neurodevelopmental diseases. So flies have been used quite extensively in that sort of research.
Mizter Rad [00:36:50]:
How would you like your research to affect or bring benefits to society? What part of society do you think will benefit from the work you're doing?
Ana Marija Jaksic [00:37:02]:
Yeah, it's a great question. I mean, I'm mostly curiosity driven. When I study cognitive behaviors in flies, I want to understand how the brain works. But obviously, this sort of curiosity driven research has all sorts of future implications on, most directly, on how we treat cognitive disorders, how we treat cognitive diseases or neurodevelopmental diseases, and how we understand the biology behind the brain. So, one way that I think about how my science and how my research can contribute to the society is having this fundamental understanding of the organ that makes us who we are, that makes us as a species who we are, and as individuals as well. So this is one thing I'm trying to kind of contribute to, in a sense of broader understanding of cognitive behaviors. But I also think that there is a really true technological advancement that comes with the research that I'm doing. We're trying to build tools to understand some fundamental biological concepts.
Ana Marija Jaksic [00:38:28]:
And by building these tools, we can support more translational research and development that is more industrial, for example, figuring out which drugs to use to treat behavioral disorders.
Mizter Rad [00:38:45]:
So what kind of tools, for example? Can you give me an example?
Ana Marija Jaksic [00:38:48]:
Oh, yeah. So, for example, we've built a robotic system that helps us screen through very complex behaviors, such as learning and memory in flies in a super high throughput way. And the reason why we've done this is to be able to evolve flies, because in evolution, you need extreme numbers of individuals that you screen through and test and choose. And so this is directly a technological advancement that was done as a curiosity driven, through a curiosity driven research that can now be used for drug screens, super high throughput and low cost drug screens for behavioral disorders or learning disorders in flies. It can be used, for example, in pharmaceutical industries. So this is one way of thinking about it. While you're trying to figure out fundamental concepts in science, you are kind of implicitly building all sorts of technology around it to be able to understand it. It can be translated and used later on in a more direct way by the society.
Ana Marija Jaksic [00:40:02]:
Now we're kind of in this flurry of AI advancement, and we're trying to figure out what intelligence really is. How does it generalize? How do we make it better? I think knowing how it does get better in a biological system can give us really deep insight into how we can reproduce it artificially and build better AI, build better computers, build better software in the future. Of course, this is far away into the future, but still quite important, I would think.
Mizter Rad [00:40:39]:
Do you mind explaining me a bit more of that machine? Because you said, maybe I got it wrong, but you said that the machine accelerates the evolution of flies, or did I make this?
Ana Marija Jaksic [00:40:51]:
Yeah, basically you can phrase, phrase it this way, but what we're doing it is. So evolution relies on selection, on diversity.
Mizter Rad [00:41:05]:
So we have flying throughout a long period of time. I'm also in magic.
Ana Marija Jaksic [00:41:10]:
Exactly. So evolution is a slow process, but it's a persistent process. So once it starts, you cannot really, you cannot really stop it. You can just kind of model its trajectory and force it towards something that you wanted to achieve. So let's call it this way. And in order to do this, you have to be able to select many flies from the population that have a learning ability that you think is correct or better in some sense. And to do this by hand is basically impossible, because you would have to pick thousands of flies every day, the whole day, 24/7 nonstop. So this would either require an army of people picking different flies, or you can automate the whole process and have a robot pick flies for you.
Ana Marija Jaksic [00:42:07]:
So when I say that we're testing flies to do a learning task, we're actually having a robot do this experiment for us. So a robot actually takes each individual fly and it places it into this y maze arena that I mentioned earlier, and then waits for a fly to solve a task. And then it picks each individual fly, remembers and learns which fly is a good learner, which one is a bad learner, and then selects from this really large pool of flies the best performing ones. And so this is done in a loop. So the robot keeps on selecting flies over and over again, over time and over multiple generations of flies. And when I say multiple, it's more like dozens or hundreds of generations of flies that are being selected for this trait. So this is something that you. You cannot do by hand if you're a single person or a small lab, but you can do it with a robot.
Ana Marija Jaksic [00:43:12]:
So this is one of the technological breakthroughs that we've had in our lab. We've built this robot, which can be now very interesting.
Mizter Rad [00:43:23]:
And how many flies do you work.
Ana Marija Jaksic [00:43:26]:
With at the moment at the momoha? I never know exactly the number. We have thousands of flies in our lab, either in experiments or as stock flies, we call it.
Mizter Rad [00:43:42]:
So do you have, like, top performers? Like 10% top performers, 5%, 1%?
Ana Marija Jaksic [00:43:50]:
Yeah. Yeah. We have sort of a leaderboard on which flies, the best flies. So we have some genotypes that we know are really, really good learners. They tend to outperform all the other flies. They have their own problems, but generally they are just from the outside, they look just like any other fly, but when we give them learning tasks, they tend to be really good at it. Now we're trying to figure out what in their genotype or what in their genome is actually providing this ability for them. So we're trying to figure out, do.
Mizter Rad [00:44:26]:
They get more attractive to other flies when they get smarter?
Ana Marija Jaksic [00:44:31]:
Not that we have noticed, actually, no. They're very plain and simple in terms of any other traits that we've tested them on. For some reason, these flies tend to learn really well, but they otherwise seem to be like very standard flies. So figuring out what is it that makes them special, it's going to be really cool to figure out. We already have some clues, like which genes are, are contributing to this, but, yeah, this is to be seen in the future. What comes out of it.
Mizter Rad [00:45:05]:
Cool. Tell me something, Ana, who is funding all this?
Ana Marija Jaksic [00:45:10]:
Yeah, that's a great question. So right now, I'm funded through this, through this superb program called the Elsir program, or else year scholarship. A very unique thing, very unique to EPFL, the university I'm in, it's a philanthropically funded scholarship. So some people were very much interested in investing, investing in young emerging scientists, let's call it this way, who have only just started their careers, but have crazy, wild ideas that might not be easily funded otherwise. But because of our track records, people believe that we can achieve the goal of pursuing such crazy, wild and interesting scientific projects. So this is an EPFL based funding system that is really generously and philanthropically supported. So I invite everyone who is interested in this sort of research and this sort of science to talk to us. We always need philanthropists that are interested in the future of neuroscience, but not just neuroscience, any type of fundamental, quality research.
Mizter Rad [00:46:39]:
So before I forget, then, where, if there's a philanthropist or someone wanting to support it, a young scientist like yourself, where can they reach out?
Ana Marija Jaksic [00:46:47]:
Oh, they can just send me an email where you can go to my website. It's jakshitslab.com. So jaksiclab.com. And you can find my email there. You can shoot me an email and then we can talk and see what people are interested in. Yeah, that would be great.
Mizter Rad [00:47:12]:
I'll add your website on the footnotes of the episode on Spotify and Apple podcasts and all this.
Ana Marija Jaksic [00:47:17]:
Perfect.
Mizter Rad [00:47:18]:
So that people can find you as well and can reach out to help out on this super interesting path of discovery. Anna, before we close, I would like to know more about your background. How did you get into being the queen of flies?
Ana Marija Jaksic [00:47:35]:
Yeah, my background is actually really a winding path. It wasn't a linear path towards neuroscience and the queen of flies. When I was a kid, I was very much attracted to biology in general. And I had this idea that I wanted to be a farmer. I wanted to have a dairy farm. So I ended up studying agriculture and specifically animal sciences and genetics and breeding in farming. And it turns out most of the, the modern agriculture time relies heavily on population genetics and genetics generally. And so I kind of moved from the farm.
Mizter Rad [00:48:23]:
Can you explain me a bit that. Sorry to interrupt, but what do you mean with population genetics? I'm not familiar.
Ana Marija Jaksic [00:48:29]:
Yeah. So, for example, when you want to have a herd of cows, if you want to have a population of cows that is really yielding a lot of milk or has really good meat production, or in another case, if you want to have a crop of corn that is high yielding, you have to apply all sorts of selection processes, breeding methods to improve a crop, to improve milk yield, in cows, or, I don't know, grain yield in corn. And so these sorts of methods are really nicely described by population genetics. So you take a population of cows and you select the best milk yielding individuals using their genotype as well as their phenotype. And these are the exact same concepts that we're using for our flies. So even though I'm not breeding farm animals or milk yielding cows, I am kind of breeding flies using the same concepts that come from agriculture and population genetics. So that's basically it. And so from that point on, I started working on population genetic systems.
Ana Marija Jaksic [00:49:53]:
And turns out flies are exceptionally good at modeling population dynamics in terms of genome dynamics or changes across the genome in large populations. And that's what brought me to the fly model and eventually to their brains, I guess. So now I'm the queen of flies.
Mizter Rad [00:50:20]:
But you also mentioned that flies have this numerical trackable brain, which is also important.
Ana Marija Jaksic [00:50:26]:
Yeah, exactly. I mean, the basic reason why I ended up studying brains is because I was interested in adaptation to temperatures. So I was studying how flies adapt to high and low temperatures because of the whole concern about the climate change. And it turned out that the way that flies adapt to high temperatures to modulate the function of their brain, so they downregulate the function of their brain as the temperatures goes up, and they upregulate the function of their brain if the temperatures go down. And that's what, what does that mean?
Mizter Rad [00:51:04]:
Up, regulate and downregulate? What do you mean with that?
Ana Marija Jaksic [00:51:07]:
Yeah. So a fly cannot regulate its own temperature. So whichever is the environmental temperature that it finds itself in, its whole body will experience this temperature. And there are some components within neurons. So there are some channels that move around different ions in and out of a neuronal, neuronal cell that regulate how a neuron functions, or how it sends signals from one neuron to another. And these ion channels are, turns out they are very much influenced by environmental temperatures. So the higher the temperature is, the more efficient these ion channels are. So your neurons tend to have a higher activity at higher temperatures.
Ana Marija Jaksic [00:51:59]:
You might think that higher activity is a good thing, but it turns out high activity can also lead to aberrant behaviors and aberrant brain function. So, for example, epilepsy is like high functioning, super high functioning neurons that fire in a non regulated way at really high frequency. So this is something that is really bad. So what evolution does for flies brains, it down regulates the function of these ion channels. It selects for less well performing ion channels to kind of buffer the effect of temperatures when it's hot, this is kind of. When it's hot, the opposite happens.
Mizter Rad [00:52:45]:
Hot is the other way around.
Ana Marija Jaksic [00:52:47]:
Yeah, exactly. Everything slows down. So your brain kind of slows down as well. So what evolution does, it kind of selects for ion channels that are more efficient than usual than in normal temperatures.
Mizter Rad [00:53:03]:
Interesting, interesting.
Ana Marija Jaksic [00:53:05]:
Yeah.
Mizter Rad [00:53:05]:
Tell me something, how do you see the world in 50 years from now?
Ana Marija Jaksic [00:53:10]:
Yeah, I'm a very positive and optimistic person, so I feel, and I would hope that in 50 years from now we would be closer to a Star Trek type of society where we rely heavily on. On research and discovery and development of technology that is there for the benefit of, you know, the entire societies and of the humankind, instead of, you know, benefiting for simple things like profits and individual gain. So I'm kind of hoping that as we go through the next couple of decades, we'll be more and more aware how important it is to go away from individual gains towards a more societally impactful gains. And I think nature will force us to do this. I think the environmental change is already forcing us to consider the future in a more inclusive way that relies on, and not just increasing profit, but increasing the advancement of the society as such. So as a whole. Yeah. And I think, and I think investing heavily in science and research and figuring out the truths of the universe is what is going to get us there.
Ana Marija Jaksic [00:54:50]:
I think it's kind of a win win situation. The society will get better and it will be easier to live. We would be a happier society in general, I think it's a very utopian view of the future, but I think that thinking anything else or having no hope about the progress of the humankind leaves you kind of paralyzed in the current situation. So I think we always have to strive for something better.
Mizter Rad [00:55:21]:
Absolutely. I think we need utopian minds like yours, because if you look at the news, if you get stuck on, on your feed, on Instagram or TikTok or whatever, you just see the negatives, and we need more positives.
Ana Marija Jaksic [00:55:37]:
I agree.
Mizter Rad [00:55:38]:
Otherwise things will not work out well. So I really appreciate you being here, Anna, with us today. It was a super interesting, out of the box conversation. I'm really grateful that we managed to meet, to talk, to set this time, and I wish you all the best.
Ana Marija Jaksic [00:55:58]:
Thank you. Thanks so much again for inviting me. I had a really great time discussing all of the ideas I have and all of the questions you have. It was a fantastic experience for me as well.
Mizter Rad [00:56:12]:
Thanks so much, Ana. Thank you so much. It was a pleasure to have you here in the Mizter Rad show. Hasta la vista. Ciao.
