Facing robot opponents puts table tennis players' brains on high

Wearing an electrode-studded cap, a table tennis player stares down an opponent. This is no flesh-and-blood adversary; the robotic metal barrel across the table fires a ball every few seconds. According to a study published today in eNeuro, the player's brain reacts differently when going up against a human opponent or the cold, calculating skill of a machine.

The study brings brain imaging into a real-world competitive setting, says Thorben Hülsdünker, neuroscientist and sport scientist at LUNEX University, who was not involved with the research. "It's always questionable if you can transfer the results we have in a standardized lab, where we have a task on a computer, to an on-court situation," he says.

Sport scientists like Hülsdünker have long used electroencephalogram (EEG) and other brain imaging techniques to study how the brain activity of elite athletes fluctuates while they perform in a lab. To track this brain activity in a more natural setting, they equipped athletes with EEG caps, which measure electrical pulses in the brain, and a portable backpack imaging device. In the past when researchers tried to use EEG caps to record activity during strenuous activity, the movement feeds a lot of noise into the data. To get around this, the University of Florida (UF) group developed an EEG cap with a higher density of electrodes, featuring 120 rather than the usual 16, 32, or 64. An additional 120 electrodes measure noise in the data so it can be subtracted out from the true brain signal, a bit like noise-canceling headphones do.

"This means an awful lot of electrodes on a small space in your head," says Daniel Ferris, a UF neuroscientist and co-author of the study.

The researchers first tested the cap on the tennis court, but the players’ abrupt neck and body movements made it too difficult to get accurate recordings. So, the researchers turned to the smaller version: table tennis. With the cap in place, players faced off against either a robot or a human serving the ball, says Amanda Studnicki, a co-author of the study and UF graduate student.

Either way, right before a player received the serve, brain activity spiked in a region involved in planning and integrating visual cues and movement. When facing a human opponent, these neurons typically fired in unison, a sign that the brain is in an idle state. But against a robot opponent, a player's brain activity looked different: It was less coordinated, with brain cells firing at different times. In essence, when faced with a more inscrutable opponent, the brain was busy making calculations and predictions, trying to figure out when the ball would arrive. This led to an enhanced state of expectation and attention.

"Your brain is in a very different sort of state when you're playing with the machines shooting balls at you," Ferris says.

Many elite athletes, including those in table tennis, already use machines to train. But this study suggests the practice may not perfectly mimic facing human opponents, the authors note. "You want to get as close to a representation of what you're going to have to perform," Ferris says. "Given how different the brain dynamics are in a lot of these areas [when facing a robot], I think it's not a great match."

Still, Studnicki says "Robots are convenient—you can get a lot of repetition, so I think they are still worthwhile." But training against people will provide "more variability that you aren't getting with a robot."

It's not yet clear why the brain responds differently to a robot. The authors speculate the machine's lack of body language may trigger the different response.

Beyond sports, the portable cap could be used to monitor brain activity during movement in the daily lives of people who are healthy or have movement disorders such as Parkinson's disease. In the meantime, Ferris is glad to have made this first step with table tennis. "[It's] a good testbed. … It's a way to test whether we can actually figure out what your brain is doing."