Everybody with a computer knows Pong. After all, this very first computer game has been around since 1972—over fifty years! Pong is essentially table tennis made virtual. Your “paddle”—a vertical white bar—moves up and down on one side of the black screen. The goal is to for your paddle to intercept the “ball”—a moving white dot—when it bounces your way.
Rendered in simple, two-dimensional graphics, Pong is so easy you might say anyone could master it. And Pong’s newest players? A group of lab-grown brain cells.
Ok, so brain cells stretch the concept of “anyone.” But these were neurons, powerful cells that detect information from the world around them. They interpret that info, then send out commands to the rest of the body.
The experiment was straightforward. A computer sent electrical signals to neurons grown on a silicon chip. These signals told the neurons where the ball was. Meanwhile, the neurons controlled the paddle by sending electrical pulses back to the computer. If the neurons moved the paddle to intercept the ball, they received an organized blast of electrical activity. If the neurons missed the ball, they received a blast of messy white noise. Brain cells like the predictability of organized activity, and they dislike white noise’s chaos.
It’s a basic system of incentives, like training your dog to sit. Over time, the neurons learned that moving the paddle to hit the ball was a good thing. Game on!
Despite the cells’ success, true human-machine hybrids are still a long way away. You could get a higher Pong score than the neurons did. For now.