illustration by Ping Zhu

Fly Walker

Researchers hope to learn how a fruit fly thinks by accompanying it on a stroll.

Imagine a day in the life of the common fruit fly. He wakes up after a few hours of sleep. Using light and smell to guide his movements, he zips around at 30 centimeters per second in search of decaying fruit, dodging obstacles along the way. If he runs into a lady friend, he quivers his wings and sings her a sweet serenade.

A simple life, perhaps, and yet extraordinary: Drosophila melanogaster carries out this daily choreography using a brain that’s less than a millimeter across and holds just 100,000 neurons—a tiny fraction of the 75 million in a mouse or 100 billion in a human.

Vivek Jayaraman wants to capture, in real time, how the fly’s brain responds to a changing environment. Ultimately, he hopes to uncover very basic patterns—“algorithms”—of fly brain activity that hold true in more complex brains including, presumably, ours.

As a first step toward that goal, Vivek’s team has designed a technique that monitors the activity of specific sets of neurons while a fly is walking. The experimental setup can be adapted to study flying behavior as well, the researchers say.

“Think of this preparation as a kind of virtual reality for the fly,” says Vivek, a group leader at HHMI’s Janelia Farm Research Campus. “You tightly control the sensory environment but allow the fly to control its own movements. By adjusting parameters in this world, you can get some sense of how the brain translates perception into action.”

WEB EXTRA:
Inside the Brain of a Fly on the Move

See what it takes to assemble a window into a fly’s brain.

Play Slideshow >>

But designing such virtual worlds is no easy feat. The most commonly used technique for measuring fly brain activity—calcium imaging—typically doesn’t allow the insects to be in a natural position, let alone move freely. In this kind of imaging, insects are genetically engineered so that certain sets of neurons carry calcium-sensitive proteins that glow green when activated. Specially designed microscopes can then pick up the light signals when those neurons are active, allowing scientists to study basic neural function in the flies.

For the past three years, Vivek and postdoctoral fellows Johannes Seelig and Eugenia Chiappe have been revamping their recording setup so that it can be used while a fly is walking.

With their technique, published in the July 2010 issue of Nature Methods, the fly is temporarily knocked out so that the skin on its head can be removed, exposing the brain. Researchers glue the neck and thorax of the insect to a chamber mounted on a two-photon microscope. The fly’s legs are left to walk freely atop a tiny foam ball that’s floating on a column of air.

The tethered fly watches visual stimuli on a special display designed by Janelia Farm fellow Michael Reiser. An optical tracker measures the fly’s precise leg movements and correlates them with the brain activity picked up by the microscope or an electrode.

In nature, of course, the fly does not find itself bound in these sorts of restraints. “This is a compromised solution,” Vivek admits. But he hopes that many interesting behaviors, such as those involving walking responses to visual input, are not so different in a tethered insect. In 5 to 10 years, he predicts, researchers will have figured out how to record from a fly while it’s walking freely across an open space. In the meantime, the tethered preparation will do just fine.

Fruit flies make ideal experimental subjects because researchers can easily tinker with their genes. Thanks to thousands of fly lines engineered over the years—many developed recently by Janelia Farm director Gerry Rubin—scientists can obtain flies that carry genetic tags in just about any type of brain cell.

“That’s the very attractive thing about working with Drosophila—you can get some type of fly, and it will always have the same cells labeled,” notes Seelig. “That’s very different from most other species, where you have to somewhat randomly pick a cell to record from.”

In the first official experiment performed with their new apparatus, for example, Vivek’s group focused on the fly’s optic lobes, brain regions that process vision.

As they describe in the July 22 issue of Current Biology, the researchers found that optic lobe neurons tune their responses according to the fly’s interaction with the environment. “The neuron changes completely when the fly walks—it becomes more sensitive to higher speeds of motion,” Chiappe explains.

In future experiments, the team will use the setup to record from different cell types. They’re also trying to figure out how to use different kinds of visual stimuli—such as pictures of objects with long edges—to prod the insects to walk in certain directions. “We’re basically watching flies walk on a ball and wondering how visual signals getting into their brain are transformed into actions,” Chiappe says. What they are finding out is anything but simple.

Slow and Steady Does It

If you’ve ever taken a close look at fruit flies hovering around your bananas, you can imagine the daily frustration of scientists who work with these tiny creatures. Manipulating Drosophila requires patience, persistence, and—above all—extremely steady hands. (Vivek, in fact, has given up coffee.)

“You can’t accidentally cut off a leg, or else you can’t measure their walking response. If you pull on a muscle the wrong way, the fly will get ticked off,” says Michael Reiser. Perturbed flies and their flailing limbs are all but impossible to mount on a tiny microscope platform.

One of the long-term goals of their collaboration is creating more automated devices to reduce this human error. “Right now, it’s just good old manual dexterity,” Reiser says. “We want to make it more like a surgery, where special tools can assist you.”

Unfortunately, the dearth of manual expertise has steered many other fly labs away from this kind of work. “A lot of people are turned off by the overwhelming technical challenge,” he says, “so it’s really important to demonstrate that these experiments are possible.”

Now that they’ve demonstrated that the new method can measure brain patterns during fly walking, the team hopes other labs will use it to probe other aspects of behavior. To that end, they have posted detailed instructions, photographs, and videos of their technique on an open-source website, www.flyfizz.org.
--- V.H.

Scientist Profile

Janelia Group Leader
Janelia Farm Research Campus
Neuroscience