Feature
When Robots See Red
January 23, 2007
An IU researcher is a leader in the field of neurorobotics, building brainy bots that are giving new insights into human diseases
Olaf Sporns sits among his "creatures." Although Strider (on the left) is resting on a pedestal, his favorite red ball is nearby.
Strider is a junkie—but you'd be surprised by his "drug" of choice. It's the color red. Red tennis balls, to be exact.
You see, Strider is a robot, one of several built and studied by IU psychology professor Olaf Sporns in his Computational Cognitive Neuroscience Laboratory. A world-renowned researcher in the emerging field of neurorobotics, Sporns is conducting innovative research that is shedding new light on human pathological states, such as addiction.
By devising computational models that simulate mammalian brain circuits and embedding them in his "creatures," as he affectionately calls them, Sporns is able to create robots that can sense their environments and share important data about their brains.
"The robot becomes a tool that helps us make predictions about how biological brains work," says Sporns. "In our lab, we're looking at the mechanisms of reward processing and how rewards affect our behavior. This is important because we know that if reward mechanisms in the brain are broken, serious things can result—like drug addiction."
In the case of Strider and his compulsion for the color red, Sporns programmed Strider's brain to receive a computational "reward" whenever his camera lens (or "eye") spied that color. The next step was to set Strider loose among a variety of different-colored tennis balls and see what happens. It's amazing to watch. When Strider is placed in a circular, white pen with blue, green, red, and orange tennis balls, he is drawn to the red ball. He appears to drink in the color (thus getting his reward), and his movements become more precise as he seeks it out—he won't let it out of his sight. It's eerily similar to the way an alcoholic might treat his bottle of Jack Daniels.
"It's been demonstrated that if you figure out that something is rewarding you will seek it out. Strider did that on his own without us interfering as experimenters," says Sporns. "We can give the robot a reward, and when that happens, there's a change in the way its neurons respond and its behavior—the robot learns."
Sporns found that Strider's interactions with the environment and the way he sought stimuli affected the way his brain developed and created patterns in synapses. In the case of drug and alcohol addiction, however, research shows that drugs actually change the way the brain responds to reward.
"This really interests us," says Sporns. "Drugs like cocaine have a way of overriding normal reward mechanisms because they act chemically on those systems. And because they're acting chemically, the addict can't unlearn them." Sporns soon plans to program Strider's brain to interpret his tennis balls as drug substances—and then look at the robot's brain circuits and see how they've changed.
This research promises to have important implications on the way scientists understand addiction. For that reason, Sporns' work has been funded by the National Institute on Drug Abuse, which is interested in discovering how certain addictions come about and why people experience relapses, even when there are negative consequences (job losses, family breakups).
Sporns couldn't be prouder of his bots. "I'm convinced that the convergence of robotics and neurosciences will be instrumental in learning how the brain works as a system. It's exciting to see how this field has developed."
Related story: Good information — it's not all about the brain
