June 17, 2013 | 2
Dragonflies are straight “A” hunters, capturing fruit flies in mid-air about 95 percent of the time, a grade that puts a head-of-the-class predator like a lion to shame.
The insect’s efficiency—combined with hackable biology (less moving parts—i.e., neurons) compared to any mammal big or small—makes the dragonfly an alluring organism to study the neural underpinnings of a basic but still complex behavior like prey capture.
Intrigued by the dragonfly, biologist Anthony Leonardo and colleagues from Intan Technologies and Duke University set about creating the instrumentation that will enable the researchers to monitor the activity of a group of neurons in the species Libellula lydia that appear to be essential for guiding the hunt. This summer, Leonardo’s group at the Howard Hughes Medical Institute’s Janelia Farm Research campus in Virginia wants to demonstrate what happens in the dragonfly’s nervous system during the course of carrying out a complex behavior—zooming in for the kill—over the second or so that it unfolds. “The dragonfly catches moving flies in the air,” Leonardo says. “In the process of doing that, it has to think of a moving fly or mosquito and think about where it’s going, where it is now, where it’s going to be in the future how its own body works and that kind of goal is constantly changing.”
Carrying out these experiments requires both tracking the dragonfly and devising the necessary instrumentation to monitor the 16 neurons hypothesized to steer the insect’s movements as it closes in on a Drosophila. The team was able to successfully outfit dragonflies with a set of small reflective balls on the head and wings to track them as they move through an insect version of the Roman Coliseum where they feast away on fruit flies.
The hard part is yet to come during coming months when the insects will be equipped with backpacks that can record signals from brain cells when going after a fly and then transmit them by a radio signal to a computer for analysis. A dragonfly weighs 400 milligrams, less than half the weight of a paper clip, so building a backpack that would not pin the insect to the ground or radically change its behavior is a major challenge.
The smallest practical battery for the telemetry in the backpack would have totaled about a third the weight of the dragonfly, and might have dampened the insect’s ardor for the hunt. So Leonardo and team designed a 40-milligran backpack that is powered by energy from radio waves. By doing so, they can record from the insect’s steering neurons that guide it during prey capture—the garb should enable monitoring other groups of neurons as well. The backpack has tiny wires, miniature sensors, that connect to the ventral nerve cord, the dragonfly equivalent of a spinal cord. The backpack should be able to transmit 5 megabits per second of information about what the insect’s neurons are doing as it descends upon its lunch.
If these tests go as planned, the work will provide new insights into how circuits operate during dynamic neural processes that take in sensory information and process it to make decisions about future actions. “Our hope is that what we learn about the dragonfly will be broadly applicable to how neurons solve problems in general, Leonardo says. “This is a broad class of computation problems that nervous systems have to solve—and in some what they’ve evolved to solve.”
DFlies at Work: Dinner Is Served