Human brain cells on a chip learned to play Doom in a week

A clump of human brain cells can play classic computer games doom. Although its performance is not up to human standards, experts say it brings biological computers a step closer to useful real-world applications such as controlling robotic arms.

In 2021, Australian company Cortical Labs used computer chips containing neurons to Pon. The chip consisted of a cluster of more than 800,000 living brain cells grown on a microelectrode array that could both send and receive electrical signals. The researchers had to carefully train the chip to control the paddles on either side of the screen.

Now, Cortical Labs has developed an interface to easily program these chips using the popular programming language Python. Independent developer Sean Cole used Python to teach Chip how to play. doom, He did it in about a week.

“Unlike, Pon “While the work we did a few years ago represented years of painstaking scientific effort, this demonstration was performed in a matter of days by people with relatively little direct biological expertise,” said Brett Kagan of the Cortical Lab. “It’s this accessibility and flexibility that makes it really exciting.”

Neuron computer chips used about a quarter as many neurons as traditional computer chips. Pon demonstration, performance doom It’s better than players shooting randomly, but far behind the performance of the best human players. However, it can learn much faster than traditional silicon-based machine learning systems, and new learning algorithms should improve performance, Kagan said.

But comparing chips to the human brain is unhelpful, he says. “Yes, it’s living and certainly biological, but what it’s actually using is a material that can process information in a very specific way that silicon can’t replicate.”

“doom “It is much more complex than previous demonstrations, and our successful interaction with it highlights a real advance in the way we control and train living neural systems,” said Andrew Adamatski from the University of the West of England in Bristol, UK.

Steve Farber of the University of Manchester, UK, agrees: doom You can level up significantly by playing. PonBut he says there’s still much we don’t understand about how neurons play the game. For example, how neurons know what is expected of them, and how we can “see” a screen without eyes.

Still, this quantum leap in ability is exciting, and Yoshikatsu Hayashi of the University of Reading in the UK says it brings us much closer to useful real-world applications, such as controlling robotic arms with biological computers, a task Hayashi and colleagues are attempting with similar computers made of jelly-like hydrogels. “[Playing Doom] It’s like a simpler version that controls the whole arm,” Hayashi says.

“What’s interesting here is not only that biological systems can play a role.” Doom, “But what it does mean is that it can deal with complexity, uncertainty, and real-time decision-making. That’s much closer to the kinds of challenges that future biological or hybrid computers will need to deal with,” Adamatsky says.