Maxwell Turner, a fly neurobiologist and assistant professor in UAlbany's Department of Biological Sciences, explains why we've all been underestimating fruit flies — and how his research to understand how vision works in their brains may one day help humans with visual deficits see the world more clearly.
Max Turner is among many UAlbany faculty members whose work blends artificial intelligence tools with other scientific disciplines — in his case, fly neurobiology focused on understanding how vision works in the brain.
Max was among more than two-dozen AI-focused faculty members recruited to UAlbany with the help of new state funding several years ago as part of the largest cluster hire in University history to support UAlbany's Al Plus Initiative.
His work uses artificial intelligence to make sense of the enormous amount of complex, multidimensional data collected as his lab watches how individual neurons in fly brains fluoresce in response to visual stimuli in the lab's mini-movie theater — and how the flies react physically by moving across a fly-sized treadmill.
It's not actually a treadmill so much as it's a tiny foam ball suspended by a stream of air and marked in such a way that a camera can record the ball's movement and correlate those movements with what the fly saw. You can watch it for yourself in this brief clip.
For all this data to make sense, Max and his students need to know what the fly was reacting to. What did it see? How do we know what flies see? Are they shown big, color photos of over-ripe bananas? (No.)
We asked Max this in a portion of our conversation that didn't make the final cut.
How do you decide what those visual inputs look like? It seems to me have you have to know, sort of, what a predator looks like to a fruit fly in order to simulate that.
MT: "I think it starts by thinking about what natural fly behavior looks like. What do flies do with their visual systems? We don’t show flies human-inspired visual stimuli. They don’t watch movies or read books. We don’t show them those things. We show them things that we think flies use their vision for, and practically speaking it helps in designing stimuli to know that flies have pretty bad spatial resolution, meaning they have kind of chunky, pixelated vision. They’re really fast. They have really fast vision but very low spatial acuity. So we don’t need super high-resolution images of a very realistic-looking predator. You could just show a big dark spot that gets bigger at the right trajectory, at the right speed, and they’ll think, ‘Oh God, that’s a dragonfly or whatever, let’s get out of here.’”
As Max noted in our conversation, other common model species for scientists studying vision are mice, monkeys and Zebrafish. But what sets Drosophila melanogaster apart, he said, is that we have a full connectome — that is, a complete wiring diagram of the brain.
Why are you a fly guy and not a Zebrafish guy?
MT: “I’m a fly guy because I like getting into the biological mechanisms — the genes, the neuron types. We have a connectome. For a neuron that I’m interested in, I can find all that neuron’s inputs, all of its outputs. I have the genetic tools to record activity in that neuron, activate that neuron, silence that neuron in a way that’s just not possible in any other model species."
Not even Zebrafish?
MT: Not even Zebrafish.
Learn more about Turner Vision Lab.
If you want a deep dive into why Drosophila melanogaster became such a widely used model species in science, this article in the International Journal of Molecular Sciences is a great place to start.
Research by Erin Frick
Audio editing and production by Scott Freedman
Photos by Patrick Dodson
Written and hosted by Jordan Carleo-Evangelist
[0:01] Host: Welcome to The Short Version, the UAlbany podcast that tackles big ideas, big questions, and big news in less time than it takes to cross the Academic Podium. I'm Jordan Carleo-Evangelist in UAlbany's Office of Communications and Marketing.
What if — hear me out — we've been wrong about fruit flies all this time. Maybe not wrong — just thinking about them in the wrong way. To most of us, fruit flies or Drosophila melanogaster, if you want to be formal about it, are a pest.
They swarm our kitchens, they destroy perfectly good bananas, and my goodness they are prolific. But they do all these things with about a million times fewer neurons in their brains than humans. When you think about it like that, it all starts to look kind of impressive. They're good at flying around complex environments, evading predators and finding food and mates. What more in life do you need if you're a fruit fly?
These insights come from Max Turner, an assistant professor and neurobiologist in UAlbany's Department of Biological Sciences. Max isn't a fly guy so much as he is an eye guy and a brain guy. He studies how fruit fly brains process vision and what that can tell us about how humans see the world. The answer is: more than you might think given our 100-billion neuron advantage.
The process of seeing occurs largely and somewhat mysteriously in the brain. Until we can untangle the vast complexity of the human brain, Drosophila melanogaster is a pretty good proxy. Max's lab puts the fruit flies into a fly-sized movie theater of sorts — a flymax if you will — and then watches how neurons in their brains light up in response to visual stimuli. They also record how the flies move their bodies in response to what they see while walking on a treadmill. Yes, it’s very small.
With the help of AI, his lab then analyzes all this data to better understand how visual processing works in a healthy brain. The hope is that that understanding will shed light on how vision is impacted by disease and ultimately how we might treat it. It's the kind of work that requires a long view of the value of scientific research, knowing that a humble fruit fly may one day be the reason a visually impaired human, with all our billions of neurons, can see the world a little more clearly.
[2:29] Host: You did at one point study eyes, now you study brains. Did you know as an 8-year-old kid: I'm going to study vision? Or what was your sort of trajectory of interest? How did you end up specializing in this? Was it something you had an interest in early on or that your undergraduate work and graduate work or postdoc work sort of directed you toward?
[2:50] Max Turner: I always was interested in science broadly, and so it was really in graduate school that I got interested in vision specifically where I started studying the retina because it's a pretty simple little neural tissue. It is part of the brain. It's just out in your eye. It's really thin. We know all the cell types. It's really accessible to experiment, and when I finished my graduate work, I wanted to go downstream and look at more complicated computations. What is the brain proper doing with visual information? But I didn't want to lose that connection to the biological mechanisms that I loved about the retina, and that's why I went to fly. It might seem like a weird jump to go from mammalian retina to fly brain, but both give you that ability to connect biology to computation, and that's why I'm a fly neurobiologist today.
[3:39] Host: I was reading that one of the things about fruit flies is that we know, I think the word is the entire connectome. We have a full map of their brains, which is astonishing to me because even in a fruit fly that seems enormously complicated. But does the fact that we sort of have a complete map of what their brains look like, does that make it easier to study what's happening when a fruit fly sees an apple, or whatever a fruit fly sees?
[4:10] Max Turner: The connectome is sort of the wiring diagram of the brain. It's a really high-resolution picture of every neuron in the brain. There's about a hundred thousand neurons in the fly brain. We have closer to a hundred billion neurons in our brain. Every neuron in the fly brain and every connection between every neuron in the brain. And it does make it easier to study vision because we can map out pathways from the first cells that sense the light all the way to the muscles that produce some visually guided behavior in the animal.
[4:39] Host: You have a complete wiring diagram of a Drosophila brain. So we sort of know where things go. And I also understand from the bit that I've read that, well, they're not like us, but their genomes are like enough to us that you can say, “Alright, well it works this way in a fruit fly, and humans are different and enormously complex but they're close enough that we can understand some things about humans from what we can learn from fruit flies.” I think that's hard for a lot of people to understand.
[5:10] Max Turner: We are very different species, but our visual systems have to solve similar kinds of problems. So when you go and you look at how does a fruit fly brain process visual information of this type or in this particular behavioral context, you find surprising commonalities between how their brain does it and how our brain does it. But in flies, we have a situation where we can go and look at the cells and the genes and the molecules in a way that you couldn't in humans. So we don't know their perceptual experience obviously, but we have ways of looking at activity in their brain so we can see how their neurons respond to visual stimuli. And that tells us something about the kinds of information from the visual world that their brain is extracting. And we can compare that to what we know other species’ visual systems are doing as well.
[5:57] Host: And so take me through that process. There's a fruit fly, and then there's a hole in the back of its head. And then some stuff goes in. And then you put it in an IMAX, I think is how you described it. So take me through that — from fruit fly comes out of the fly room into that other room, what's happening there and what are we learning from that process?
[6:18] Max Turner: Yes, so in my lab we use imaging methods, microscopy methods. So we use genetic methods to direct the expression of these fluorescent kind of glowing molecules that we engineer in specific cell types in the fly brain. So we can say we want this particular fluorescent molecule to only show up in these visual neurons that we like that we're interested in studying. Then we can open up the back of the fly's head, put it under the microscope and watch those cells glow and blink away. And the brightness that we're measuring in those neurons is related to their electrical activity. So we can show visual stimuli to the fly and watch the neurons respond in images. We show all kinds of things so we can show stimuli that maybe looks sort of like an approaching predator to the fly. There are certain neurons that seem to be tuned to detect those kinds of visual stimuli and kick off escape responses in the fly.
We can show visual stimuli that might look like another fly. So we can record brain activity in a male fly and show something that looks like a female fly walking by, and there are neurons in the male fly's brain that seem to be tuned to detecting the visual signatures of a female fly and gets him in a courtship state, gets him hot and bothered by looking at this picture that he thinks is a female. I want to resist the urge to anthropomorphize the animal too much and think about their visual perception as being like ours because I'm sure it's very different. They do not use their visual systems in exactly the same way that we do, so they perceive reality in a different way. So as much as I try to stay in the mind of a fly when thinking about vision for my particular work —
[7:29] Host: I can just imagine you in the lab being like, “Be the fly. Be the fly.”
[8:03] Max Turner: [Laughter] Yeah.
[8:04] Host: So if you were doing this work with a different kind of animal that doesn't rely as heavily on vision — so say you're using bats, which are amazing but not known for their visual acuity or even using it really to navigate the world, do you expect that you could do the same experiment with a bat looking into a bat's brain while it is sitting in your little IMAX theater there? Would it look different?
[8:31] Max Turner: I think it would look different. You see specializations for species specific kinds of behaviors and stimuli, but I would also say even more strongly than that you see commonalities across different species that are imposed by the physical world that we live in. So even though flies have fly-specific visual stimuli that they are sort of tuned to, we still see commonalities in how visual processing happens between them and other species that maybe don't have those behaviors. And one of the things that we found in my lab is that flies, when they walk or fly, they'll make very rapid turns a couple times a second all day long. If you look at parts of the visual system of the fly during those movements, there are neurons that, just like ours, also kind of shut down and stop signaling during that very rapid movement. So it's kind of remarkable that flies and people have the same visual sampling strategy looking around the world — and also a similar strategy in the brain to deal with that really dynamic motion that maybe is kind of confusing.
[9:32] Host: This may seem like an absurd question to you, or a very basic question, but why does it matter if we understand what's happening in the brain when we see or don't see things? What is it helping us accomplish?
[9:46] Max Turner: There's a connection to human health and disease here that requires you to have something of a long-term view of biology research. Ultimately, psychiatric disorders, neurological disorders are rooted in something going wrong with neurons, so these are cells in the brain or genes, and it's really difficult to understand how you go from kind of small biological mechanisms to wide changes in brain activity or perception without knowing how that works normally in the non-diseased brain. And it takes a long-term view of biomedical research. And if we stop doing basic research today, I don't think it would prevent us from coming up with cures next year or even five years from now, but 30 years down the line, biomedical research would be at a complete standstill because we haven't built those kind of basic building blocks.
[10:37] Host: So you came to UAlbany when we were hiring a number of faculty members whose work was in all different kinds of disciplines but overlapped with computational applications and artificial intelligence. How does that work factor into the work that you're doing? What are you learning from that and what are you using it for?
[10:53] Max Turner: There's a few ways we use computational tools and AI tools. One is in the data analysis itself. So we collect pretty high-dimensional, complicated data, and part of what we do is we develop methods to analyze these data and make sense of them. So from the imaging experiments, for example, we also use AI tools to track animal behavior. So for example, if we want to measure visual behaviors and flies, we can train machine learning models to automatically classify the behavior for us in a very high-throughput way. The last way that we use it, we can build models of the visual system of the brain. So these are models that are meant to predict neural activity. If you give it a visual stimulus, it should predict the activity of neurons that we can then check against what actually happens.
[11:39] Host: So the other connection maybe to the computational side is what you're learning about the way fruit flies’ brains react to visual input. Is that helpful in building or improving computer vision and the way that non-biological systems perceive the world?
[12:02] Max Turner: There is some hope that we could use lessons from biology, broadly speaking, to improve computer algorithms or machines that do visual processing. And I mean, I think the reason to believe that is that brains are super-efficient computers and every brain, including a fly’s brain, has to deal with really complex visual information. And it has to do that with a limited number of neurons, right? I said that flies have a hundred thousand neurons in the whole brain, but with those hundred thousand neurons, they can fly around the room, find mates, avoid predators, find food, navigate. So if we can figure out the algorithms that the fly brain is using to process that information so efficiently, can we take some of those lessons and port them into the way we design computers or algorithms that we put on board computers to do computer vision? That's the hope. Real brains learn very efficiently. You don't need giant nuclear reactors to teach a brain how to recognize objects. Just a little bit of experience is enough. So transferring knowledge about how the brain does these tasks to machine learning models to make them more efficient, there's clearly room for improvement there.
[13:14] Host: Another thing, this is particular to your line of work, but you've chosen an animal that is most people really hate. Do you find have a greater appreciation for drosophila than you think most people do because you know a little bit more about its world?
[13:34] Max Turner: Yes, definitely. Before I started working in flies, I didn't frankly think bugs were that interesting. And now since working in flies in particular, they're just really amazing creatures. They can do such amazing behaviors with such small, simple brains. And it's easy to think that they are somehow a lower form of life or something like that. There are no lower or higher forms of life. No species occupies a special place in biology. And when you start to think about the way that the brain and the body are really exquisitely adapted for exactly what that animal needs to do to survive and thrive, it's amazing. So if you can resist the urge to compare yourself to a fly in terms of human-level tasks, you're always going to be able to do math better than a fly. You're always going to have better book recommendations than a fly. But you can't fly around the way a fly can. You can't do many of the things that flies can do with such a tiny brain. So we have to appreciate them on their own terms.
[14:30] Host: That was Max Turner, a fly neurobiologist and assistant professor in UAlbany's Department of Biological Sciences.
Max and his students literally watch fly neurons light up in real time to better understand how sight works in their brains with the hope of one day helping humans with visual deficits see better.
If you want to see Turner Lab’s fly treadmill in action — and trust me, I think you do — be sure to check out the longer version in our show notes.
The Short Version would not be possible without contributions from many people. Scott Freedman provided audio production and editing from the UAlbany Digital Media studio deep inside the Podium tunnels. My colleague Erin Frick contributed research.
We'll be back next week with another short conversation about something interesting. I'm Jordan Carleo-Evangelist here at the University at Albany, and this has been The Short Version.