The logic of a fruit fly

When a fruit fly catches a whiff of rotting banana, it knows exactly how to home in on that appetizing smell: Even in perfectly still air, a fly can easily locate its next meal. This uncanny ability is more than a reflex; it’s part of an adept strategy to survive and proof that even a brain with just 100,000 neurons (vs. the human brain’s 80 billion) can chart a course through the world.

Most people see fruit flies as little more than kitchen nuisances that appear to buzz around aimlessly. But for Gaby Maimon, head of the Laboratory of Integrative Brain Function at Rockefeller and an investigator at the Howard Hughes Medical Institute, flies are
windows into one of the most enduring questions in science: How do brains give rise to intelligence?

While human brains contain massive networks of cells, fruit fly brains are small, accessible, and surprisingly powerful. Over the past decade, Maimon’s lab has helped reveal how flies build internal understandings of the world, carry out math as they navigate, and make decisions about where to move. By combining cutting-edge genetic tools and technology optimized to image minuscule brains, Maimon and his team are revealing, neuron by neuron, and circuit by circuit, how these brains work.

Ultimately, Maimon’s basic research could shed light on how mammals, including humans, navigate the world and make decisions, as well as how these processes can go wrong in diseases like Alzheimer’s.

We spoke with Maimon about why he turned to flies, what his lab has uncovered, and why understanding these insects could help us understand ourselves.

To the average person, fruit flies don’t seem very smart. What makes them a good model for studying intelligence?

We always like to say that fruit fly brains may be small, but they’re not simple. Per neuron, the fly brain does as much or more than a mammalian brain. For instance, if a fly is hungry and it smells something good, logically, it turns upwind—a behavior that helps it arrive at its favorite environment: a rotting fruit. But what’s even more impressive is that if the wind dies down, we recently found that flies will use a working memory of the direction from which the wind was last coming to keep heading along that angle for a few minutes. And if a group of flies is hungry and they have no cues at all—no wind, no smell—they each pick a random direction and travel in a long, straight path, sticking to that direction for hours, effectively organizing a search party, where the hope is that at least a few flies chance upon a morsel of food and survive.

Alongside these navigational capacities, flies can be genetically modified with incredible precision, reproduce quickly, and have been studied for over a century to reveal fundamental principles of genetics, development, and evolution. This unique combination of experimental tractability and deep biological knowledge makes them an ideal system for investigating the brain from multiple perspectives. This is why, in 2006, I switched from studying primates to flies.

These flies’ brains are the size of poppy seeds. How do you study them?

When I began working with flies, methods existed to secure them in place and electrically record their brain activity. It wasn’t possible, however, to record electrical activity during behavior. So, as postdocs, my colleagues and I came up with a method to tether a fly in place and record both its wingbeats and the electrical activity of its neurons. Now, we typically use microscopes to image neural activity as flies walk on tiny floating balls, which act like spherical treadmills that allow them to navigate around a visual virtual environment. These methods, alongside other developments related to manipulating neuronal activity and characterizing the anatomy of the fly brain in fine detail, have opened a new window into understanding how neuronal circuits regulate behavior.

What have you discovered about how they work?

In the middle of every fly brain is a structure called the ellipsoid body, which literally looks like a donut. This structure has a set of neurons that break it up into sectors, like pizza slices divvying up a pie. When you measure the activity of these neurons, only a few neighboring ones are active at a time, creating a localized activity “bump” in one section of the donut. When a fly turns, the position of this bump rotates correspondingly, like a compass needle tracking the fly’s orientation in the world.

Our first major finding, in 2017, described a dedicated neuronal circuit that functions to rotate the ellipsoid-body compass signal at just the right speed to match how fast flies are turning in the world. In 2021, we further found that the fly brain calculates not just angles (like the compass direction) but also vectors, which are mathematical quantities that combine an angle and a length. Specifically, we found that another structure in the fly brain—called the fan-shaped body—encodes vectors and can add or rotate them as the fly moves through space. The fan-shaped body can calculate the direction in which a fly is traveling, and it does so via vector arithmetic. This traveling-direction signal is particularly important when the ellipsoid body compass is indicating one direction (which the fly is facing), but she is walking or flying in another direction.

“If we can figure out how memory works in the fly, it could reshape how we think about memory in all brains.”

More recently, we discovered how the same system can turn an internally generated goal into a motor action. Here, the goal is the direction in which the fly wants to progress forward, and the motor action is that act of turning so as to align the body with the goal direction. The circuit we discovered provided the first biological example of how a brain can convert a navigational goal into a steering signal.

Do human brains work the same way?

Humans don’t have the same brain structures as flies, but some of the core computational principles might be similar across nervous systems. The ellipsoid-body compass in flies, for example, works nearly identically to “head direction cells” that have been discovered in mammals. Other mechanisms might fundamentally differ between the fly and human brain, and our work could help to reveal these differences too.

What do you want to find out next?

We want to understand the nature of spatial memory. To this end, we are focusing on how flies can remember a direction or location over seconds, minutes, or even days. As part of this research direction, we’re building apparatuses that allow flies to live for weeks in a virtual environment while we simultaneously image their brain activity. Overcoming the significant technical challenges involved in this project is worth it, I think, because if we can figure out how memory works in the fly, it could reshape how we think about memory in all brains.

Where do you hope this work will ultimately lead us?

Brains remain fundamentally mysterious. Rigorous explanations of how sets of neurons give rise to function are still rare. By gaining a detailed understanding of how a small region in the middle of the insect brain works, our research provides a road map for how to go about achieving a similar understanding of bigger brains, like our own. Also, the insect brain holds a beauty all its own. Most species that share our planet are insects, and to understand how they master the art of living, in my view, deepens our own humanity. This sense of wonder is also why we pursue this work, and where such curiosity-driven science takes us is not wholly predictable.