Published in the journal “Neuron” Friday, the study reported 800,000 mouse and human embryo brain cells hooked up to a computer and virtual paddle learned to detect where the game’s electronic ball was, then control and volley the ball back with the paddle.
Because neurons are wired to receive external sensory input, as the study explained, scientists were able to inform the cells whether the paddle was connecting with the ball over time.
Al Jazeera, which posted a video of the brain cells playing the game, reported that the cells not only learned how to play the game, but improved and played longer rallies over time.
“We’ve made huge strides with silicon computing, but they’re still rigid and inflexible,” said Brett Kagan, an author of the study and chief scientific officer at Cortical Labs in Melbourne. “That’s something we don’t see with biology.”
Using software to study times when the neurons missed, the study’s researchers said they were able to demonstrate that “neurons could adapt activity to a changing environment, in a goal-oriented way, in real time.”
Conducted in part to understand how the brain works, the study’s significance, according to Kagan, lies in the cells’ ability to generalize what they have learned, unlike artificial intelligence.
“From worms to flies to humans, neurons are the starting block for generalized intelligence,” Kagan said in a statement to CNN. “So, the question was, can we interact with neurons in a way to harness that inherent intelligence?”
Kagan told CNN the researchers chose Pong because it was one of the first games used in machine learning and that the technology discovered could be used for “better drug discovery, disease modeling, and understanding how intelligence arises — which in turn could be used to develop new algorithms for machine learning.”
“It touches on the fundamental aspects of not only what it means to be human but what it means to be alive and intelligent at all, to process information and be sentient in an ever-changing, dynamic world,” Kagan said.