By R.J. Melby 

Writer/Reporter 

A new scientific breakthrough in the field of computer programming and engineering may have lasting implications. Cortical Labs released a video on YouTube on February the 25^th of this year showcasing their achievement as brain cells played Doom on their CL1. 

What may sound like a science fiction dream – or nightmare – the lab shows a new dimension for what can be done with computers. In just a week, after trial and error, the Cortical Labs team successfully taught a bundle of human brain cells to play the complicated shooting game, all through rigorous training via electric signals. Dr. Brett Kagan says, “Is it an esports champion? Absolutely not. Right now, the cells play a lot like a beginner who has never seen a computer… But they are learning.” 

Cortical Labs achieved previous success in 2022 by teaching “more than 800,000 living brain cells grown on top of microelectrode arrays” that they were able to have their computer play the more rudimentary game of Pong, a two-dimensional game from the seventies comprised of two paddles and a singular ball that bounces between them. This took 18 months. 

Kagan says, “We showed that biological neurons could play the game Pong… a massive milestone because it demonstrated adaptive real-time goal-directed learning… [But] Doom is chaos. It’s 3D. It has enemies. It needs to explore its environment. And it’s hard.” 

The team was inspired when they received a number of internet comments challenging them to do the same with Doom. In the programming world, Doom is the ultimate challenge, with programmers even getting the game to work on a Ti-84 calculator. 

What is notable is that, in contrast to the 800,000 cells needed for the previous Pong experiment, only 200,000 cells were used to play Doom. By translating commands into electrical signals, the team turned visuals into “electrical stimulation.” 

The main thing the company touts is the accessibility of their product. “We built the CL AP1 so that any user can interact with cells living in the CL1 with simple Python commands,” says David Hogan. 

While an impressive achievement, Kagan still says more needs to be done. “The cells need feedback. They need to know what the right actions are and what the wrong actions are and how they can be trained to improve… While there’s still a lot of work to do on this, the exciting thing is we’ve solved the interface problem. We have a way to interact with these cells in real time.” 

Users can learn more on the Cortical Labs website.