Dragonflies are seasonal animals and normally fly around as winged adults only during the summer. Which is why Anthony Leonardo and his lab mates at Janelia Farm have been learning how to nurture nymphs in their laboratory: that way they can maintain a steadier year-round supply of some of the world’s most astounding insect hunters.
The Leonardo group aims to reveal and decipher the rules that guide the insects’ behavior. They’ve constructed a unique indoor flight arena, with walls covered in murals depicting bucolic wetland scenery, to explain the brain-cell-encoded protocols that orchestrate the dragonfly’s actions to pursue and capture prey—in this case, fruit flies.
A foundational component of the task is to precisely characterize the head, wing, and body motions of the hunting dragonflies as they take off from a perch in pursuit of a flying meal-to-be. The researchers use a pair of high-speed cameras that capture simultaneous recordings of dragonflies’ foraging flights at 1,000 frames per second within a specified cubic-meter volume of the arena.
By associating particular pixel addresses of the cameras’ two-dimensional light sensors with particular locations in the three-dimensional arena, the resulting video data enables the researchers to computationally reconstruct and analyze the insects’ trajectories.
A second, larger set of cameras mounted around the periphery of the flight arena captures finer body motions from hunting dragonflies. To facilitate this, the researchers glue reflective balls the size of sesame seeds to the insects’ heads, body, and wings. Leonardo describes the setup as a motion capture system, similar to the ones commercial moviemakers use to make animations, like Gollum in The Lord of the Rings.
“We are capturing behaviors that are over before you otherwise would see them,” says Leonardo, referring, for example, to movements of the dragonfly’s head and on-the-fly changes in the flapping of its wings. In time, with tiny neuron-surveilling backpacks, Leonardo hopes to capture the imagery while simultaneously recording from, for example, “steering neurons,” whose activity depends on input from the insects’ 50,000-facet compound eyes. That activity, in turn, guides the activity of neurons that control wing muscles.
“We want to ask deep computational questions about how organisms solve important problems like capturing prey,” says Leonardo. The camera-acquired data and computational analyses of behavior, he adds, will help explain how the underlying neuronal circuits are pulling the strings of perception and behavior.
-- Ivan Amato