Live brain cell dish learned to play 1970s arcade game pong.
About 800,000 cells linked to a computer have gradually learned to detect the position of the game’s electronic ball and control a virtual racket, a team reports in the review neuron.
This new achievement is part of an effort to understand how the brain learns and how to make computers smarter.
“We’ve made huge strides with silicon computing, but they’re still rigid and inflexible,” says Brett Kagan, author of the study and scientific director of Cortical Labs in Melbourne, Australia. “That’s something we don’t see with biology.”
For example, computers and people can learn to make a cup of tea, Kagan says. But people are able to generalize what they’ve learned in ways that a computer can’t.
“You may never have been to someone else’s house, but with a little digging and research you can probably make a great cup of tea as long as I have the ingredients,” he says. But even a very powerful computer would struggle to perform this task in an unfamiliar environment.
Cortical Labs therefore tried to understand how living brain cells acquire this type of intelligence. And Kagan says the Pong experiment was a way for the company to answer a key question about how a network of brain cells learns to change its behavior:
“If we allow these cells to know the outcome of their actions, will they actually be able to change in a goal-oriented way,” says Kagan.
To find out, the scientists used a system they developed called DishBrain.
A layer of living neurons is grown on a special silicon chip at the bottom of a thumb-sized dish filled with nutrients. The chip, which is connected to a computer, can both detect the electrical signals produced by the neurons, and deliver them electrical signals.
To test the cells’ ability to learn, the computer generated a game of Pong, a two-dimensional version of table tennis that gained a cult following as one of the earliest and most basic video games.
Pong is played on a video screen. A black rectangle defines the table and a white slider represents each player’s paddle, which can be moved up or down to intercept a white ball.
In the simplified version used in the experiment, there was a single paddle on the left side of the virtual table, and the ball jumped to other sides until it escaped the paddle.
To get the brain cells to play the game, the computer sent them signals indicating where the bouncing ball was. At the same time, he began to monitor information coming from the cells in the form of electrical impulses.
“We took that information and allowed them to influence what game of Pong they were playing,” Kagan said. “So they can move the paddle.”
At first, the cells didn’t understand the signals coming from the computer, or didn’t know what signals to send the other way. They also had no reason to play the game.
So the scientists tried to motivate the cells using electrical stimulation: a well-organized burst of electrical activity if they were successful. When they were wrong, the result was a chaotic stream of white noise.
“If they hit the ball, we gave them something predictable,” Kagan says. “When they missed it, they got something that was totally unpredictable.”
The strategy was based on Principle of free energy, which states that brain cells want to be able to predict what is happening in their environment. They would therefore choose predictable stimulation rather than unpredictable stimulation.
The approach worked. The cells began to learn to generate patterns of electrical activity that would move the paddle past the ball, and gradually the rallies grew longer.
Brain cells have never been so good in Pong. But interestingly, human brain cells seemed to achieve a slightly higher level of play than mouse brain cells, Kagan says.
And the level of play was remarkable, given that each network contained fewer cells than a cockroach’s brain, Kagan says.
“If you could see a cockroach playing a game of Pong and it was able to hit the ball twice as often as it missed, you’d be pretty impressed with that cockroach,” he says.
The results point to a future in which biology helps computers become smarter by changing the way they learn, Kagan says.
But that future is probably still a long way off, says Steve M. Potteradjunct associate professor at Georgia Tech.
“The idea of a computer with living components is exciting and starting to become a reality,” he says. “However, the kinds of learning these things can accomplish are pretty rudimentary right now.”
Even so, Potter says the system that allowed cells to learn Pong could be a great tool for research.
“It’s kind of a semi-living animal model that you can use to study all kinds of mechanisms in the nervous system, not just to learn,” he says.