Dragonfly Designs Inspire Engineering
Dragonflies are among the best flyers in the insect world. Their twin pairs of paper-thin wings allow them to hover and move in all directions, even in mating. When the time comes to dart after prey at high speed, they rarely miss.
What's their secret? One is "selective attention" -- a trait previously known only in primates, according to new research from the University of Adelaide, Australia. Selective attention is the ability to focus on one object and exclude others. Just as a tennis player must focus on the ball and ignore the cheers of the crowd, a dragonfly must pick out one target from a swarm of insects and avoid being distracted by all the others.
"What's exciting for us is that this is the first direct demonstration of something akin to selective attention in humans shown at the single neuron level in an invertebrate," Associate Professor [David] O'Carroll says.The paper by Weiderman and O'Carroll in Current Biology explains what they observed at the single neuron level:
"Recent studies reveal similar mechanisms at work in the primate brain, but you might expect it there. We weren't expecting to find something so sophisticated in lowly insects from a group that's been around for 325 million years...." (Emphasis added.)
Responses to individual targets moving at different locations within the receptive field differ in both magnitude and time course. However, responses to two simultaneous targets exclusively track those for one target alone rather than any combination of the pair. Irrespective of target size, contrast, or separation, this neuron selects one target from the pair and perfectly preserves the response, regardless of whether the "winner" is the stronger stimulus if presented alone.The authors found a moment of decision before the dragonfly "locks in" on one target, just like a heat-seeking missile or anti-torpedo torpedo analyzing its inputs:
Our data make a compelling case that CSTMD1 reflects competitive selection of one target. We emphasize "competitive," because the attended target is not always the same between trials or even within a trial, as seen in strikingly perfect switches from one to the other.... Competition is further suggested by rare examples where the activity observed under Pair stimulation initially lags both T1 and T2 responses... suggesting initial conflict in the underlying neural network before resolution of competition by a "winning" target.This apptitude for "highly accurate encoding of single stimuli" was closer to that seen in primates than in birds. The authors did not delve into the question of how two groups of animals separated by vast distances on Darwin's tree of life might have converged on the same dynamic, complex skill. Instead of helping Darwinian evolutionists, their study is poised to inspire engineers:
We previously showed that CSTMD1 still responds robustly to a target even when it is embedded within a high-contrast natural scene containing numerous potential distracters. Taken together with recent evidence that the behavioral state of insects strongly modulates responses of neurons involved in visuomotor control, our new data thus suggest a hitherto unexpected sophistication in higher-order control of insect visual processing, akin to selective attention in primates. Perhaps the most remarkable feature of our data is that once the response "locks" onto a target (or following a switch), the second target exerts no influence on the neuron's response: the distracter is ignored completely.
Our finding of a process analogous to selective attention in primates is particularly exciting because insects have proved to be powerful tools for "circuit-busting" neuronal computations in biological motion vision, inspiring substantial breakthroughs in computational models with diverse applications.In a news release, O'Conner added,
"We believe our work will appeal to neuroscientists and engineers alike. For example, it could be used as a model system for robotic vision. Because the insect brain is simple and accessible, future work may allow us to fully understand the underlying network of neurons and copy it into intelligent robots," he says.Another paper on dragonflies shows that these marvels of the insect world are equipped with navigational equipment that can do vector calculus. In the Proceedings of the National Academy of Sciences, Gonzalez-Bellido and a team at the Howard Hughes Medical Institute discerned "Eight pairs of descending visual neurons in the dragonfly [that] give wing motor centers accurate population vector of prey direction."
Intercepting a moving object requires prediction of its future location. This complex task has been solved by dragonflies, who intercept their prey in midair with a 95% success rate. In this study, we show that a group of 16 neurons, called target-selective descending neurons (TSDNs), code a population vector that reflects the direction of the target with high accuracy and reliability across 360°. The TSDN spatial (receptive field) and temporal (latency) properties matched the area of the retina where the prey is focused and the reaction time, respectively, during predatory flights. The directional tuning curves and morphological traits (3D tracings) for each TSDN type were consistent among animals, but spike rates were not. Our results emphasize that a successful neural circuit for target tracking and interception can be achieved with few neurons and that in dragonflies this information is relayed from the brain to the wing motor centers in population vector form.The dragonfly's head gear not only includes neurons that map inputs into a dynamic vector, but transmitters that send the vector map to the wing motors. Obviously this mapping process is occurring simultaneously with the selective attention target lock-on mechanism. It happens so fast, the dragonfly has its meal in the blink of an eye.
These two papers illustrate a growing willingness in the science world to view organisms from an engineering perspective, seeking to understand how life works and to imitate it. The excitement of discovery in the researchers' descriptions is nearly palpable. No longer are living things merely products of happenstance and blind churning; they are designs worthy of being copied. The more this "reverse-engineering" approach advances, the more Darwinian evolution will retreat.
Biomimetics is de facto intelligent design science. May it thrive in 2013!
Photo credit: Nedster78/Flickr.