Fly Flight Simulators

Though tethered in place, the fly still could beat its wings freely, and every wing stroke cast a shadow onto an infrared sensor positioned under each wing. By comparing the number of right wing strokes to left wing strokes, the researchers deduced the direction the fly intended to pilot itself.

The researchers used this panoramic display to show blobs and lines to flies.

“We found indeed that flies avoid little tiny, tiny objects. And we believe this is an important behavior because it enables them to avoid predators,” which appear as little blobs, says Dickinson.

While flies shun little blobs, they do advance toward long, skinny things, as previous research has shown. Dickinson explains, “We believe that the flies are basically attracted to vertical objects because it's a very simple algorithm that will lead them toward safe havens in their environment: plant stems, tree trunks, and the like, that they can perch on.”

Dickinson is a neuroethologist—a scientist who studies “how the brains of organisms work to make decisions that enable the organism to survive,” he says. Decision-making is a complex subject, even with flies, which have “a little tiny brain the size of a poppy seed,” he says.

The fly captivates Dickinson for precisely the reason it chafes the rest of us: its uncanny ability to dodge its archenemy, the swatter. He explains, “Almost anyone has suffered the frustration of trying to swat an annoying fly, but how many people have stopped to actually wonder how this little tiny machine actually works? How does its little tiny brain control its little tiny muscles to control its little tiny wings to produce the aerodynamic forces that can so effectively evade our fly swatters?” With the data from the flight simulator study, Dickinson and colleagues are getting closer to an answer.

For instance, the team tried tricking the flies by showing them long skinny stripes of light, and then shrinking down the stripes into predator-like blobs. As the stripes shrink, the flies start trying to buzz off. “So the fly's brain, at a certain point, begins to interpret this shorter stripe as a predator,” Dickinson explains. “We believe this is the very, very simple algorithm that flies use in order to steer toward safe havens and steer away from threatening stimuli,” he says.

Soft Spot for Insects?

To Dickinson, a fly is no mere pest; it’s an aerodynamic marvel. “It's hard to imagine a more sophisticated flying device,” he says. “This organism can land on the ceiling. It can fly briefly upside down. It can hover in place. It can make rapid turns to the left and the right. And all of this is done autonomously,” much like the tiny flying robots Dickinson says engineers are working to develop.

In fact, that’s one potential application of Dickinson’s research. "We're hoping to understand how a fly's brain works and apply that information to the design of small flying robots,” for planetary exploration or search-and-rescue, he explains.

So what’s behind the fly’s in-air virtuosity? Dickinson attributes the bug’s aerial acrobatics to its “very, very fast eyes” and speed at processing sensory information. “Flies also have tiny gyroscopes. They have little tiny wind detectors on their head,” he adds.

“Most people would swat flies without ever thinking about what they were destroying, but I must confess, I don't swat flies unless I absolutely have to. In fact, I'll usually try to induce them to land on my arm, and take a nice long look at them as long as they're happy to stay there.”

But there are limits even to Dickinson’s arthropod awe. He says, “If I notice that they're starting to bite me, well, then it's curtains for the poor fly. But until that point, I'm really more interested in how the creature works than in trying to swat it.”

Flies, of course, aren’t the only animals forced to make snap judgments. We humans also have to turn visual signals into safe movements. But while we have big brains to help us decide to stop at a red light, for instance, flies manage such decisions despite their budget-sized brain capacity. Dickinson says, “The human brain has about 300,000 times more neurons than a fly's brain, and therefore, all of these decisions the fly must be able to make with rather limited neural resources.”

Dickinson says the work won’t stop with flight simulation, so be ready for more buzz on fly cognition.

“This problem of determining what’s a safe perch and what’s a predator is just one of the many, many decisions that a fly has to make throughout the course of its life,” he says. “It has to decide when to take off. It has to determine in which direction to fly in order to find a nice piece of smelly fruit where it can find a good meal, a place to lay its eggs, and maybe some mates.”

To Dickinson, that’s all fodder for more research. “We hope to keep plugging away at trying to determine how the fly’s brain makes all of these very, very challenging decisions,” he says.

Authors: Gaby Maimon, Andrew D. Straw, and Michael H. Dickinson

Publication: Current Biology, online March 13, 2008

RESEARCH FUNDED BY: National Science Foundation, Office of Naval Research, the Air Force Office of Scientific Research, and Caltech Della Martin Fellowship