Human Brain Cells Play Doom: A Milestone in Biological Computing

In a groundbreaking demonstration of bio-engineering, a cluster of 200,000 living human neurons has successfully learned to play the iconic 1993 video game, Doom. Australian biotech firm Cortical Labs unveiled its CL1 “biological computer” on March 8, 2026, showcasing the neurons’ ability to navigate the complex 3D environment and even target enemies – a feat previously confined to silicon-based processors.

How Does a ‘Biological Computer’ Work?

The CL1 system works by translating the game’s visual data into electrical signals that the neurons can interpret. These neurons, grown on a microelectrode array and kept alive in a nutrient bath, respond to stimuli by firing their own electrical spikes. The system then decodes these spikes into commands, allowing the biological computer to move, turn, and fire Doom’s weapon. This process leverages reinforcement learning, where the neurons adapt their activity in response to feedback.

From Pong to Demons: The Evolution of Biological Computing

Cortical Labs’ journey from Pong to Doom highlights the increasing sophistication of biological computing. In 2022, the team made headlines by teaching a cluster of lab-grown neurons, known as DishBrain, to play Pong by sensing the ball position and moving a paddle accordingly [The Register]. Pong presented a relatively simple input-output relationship – a moving line and a bouncing square – whereas Doom introduces a far more complex challenge.

“Pong was much simpler. There was a direct relationship. The ball went up, the paddle went up. It was a direct input-output relationship. Doom was much more complex,” explained Alon Loeffler, a scientist at Cortical Labs [The Register].

Overcoming the Complexity of Doom

Getting neurons to cope with the chaos of Doom required building a more sophisticated interface between the digital game world and the biological language of neurons: electricity. According to Interesting Engineering, independent developer Sean Cole solved this challenge in just one week by creating a Python-based interface. This allowed the neurons to navigate the game’s 3D environment in real time, autonomously targeting enemies and firing weapons.

What’s Next for Biological Computing?

Even as the neurons’ performance resembles that of a complete beginner, this experiment represents a major milestone in “programmable biology” [Interesting Engineering]. The long-term goal is to understand how neurons learn and adapt, potentially aiding drug research or inspiring latest computing ideas. For now, a dish of human brain cells is discovering that Doom’s demons shoot back.

Key Takeaways

• A biological computer, the CL1, developed by Cortical Labs, has successfully run Doom using 200,000 human neurons.
• The system translates game data into electrical signals that neurons interpret and respond to.
• This achievement builds upon previous work where neurons were taught to play Pong.
• The research aims to further understanding of neuronal learning and adaptation.