Human Brain Cells Trained to Play Doom in Lab

200,000 human neurons learn Doom via electrical rewards in Cortical Labs experiment, showing beginner skill.

  • Cortical Labs has successfully trained 200,000 living human neurons to play the 1993 video game Doom.
  • The neurons learn through electrical stimulation and feedback rewards, but currently play at a beginner’s skill level.
  • This experiment continues the long-running tradition of porting Doom to unconventional platforms, from blockchain networks to pregnancy tests.
  • Researchers built a custom platform to interact with the neurons using simple Python commands, accelerating development.
  • The system demonstrates neural adaptability, but the cells do not possess human cognition or understand the game.

Melbourne-based startup Cortical Labs has pushed the “Can it run Doom?” challenge to a new biological frontier by training clusters of 200,000 living human brain cells to navigate and shoot in the classic 1993 video game. The team connected the neurons to software that converts gameplay into electrical signals, as shown in a video posted last week.

- Advertisement -

For decades, Doom has served as an unofficial benchmark for engineers testing novel systems on everything from gut bacteria to blockchain networks. Consequently, the researchers initially relied on improvised low-level code to make their biological system functional.

However, development accelerated dramatically after they built a dedicated platform for high-level interaction. “It took the collaborator, Sean, who wrote Doom, the code for us a few days instead of 18 months,” application scientist Alon Loeffler told Decrypt.

The neurons learn through a reward system, receiving feedback when they correctly aim at or eliminate an enemy. Meanwhile, an AI refines how the game’s information is encoded into the electrical signals sent to the cell cluster.

Loeffler emphasized that the cells are reacting to inputs rather than comprehending the game. “The system doesn’t actually know it’s playing Doom,” he said, noting it requires a completely different programming mentality than traditional computing.

- Advertisement -

Despite using human-derived neurons, the system lacks structures for higher-order functionality or human cognition. Nevertheless, researchers observe clear evidence of learning and neural adaptability outside the brain.

✅ Follow BITNEWSBOT on Telegram, Facebook, LinkedIn, X.com, and Google News for instant updates.

Previous Articles:

- Advertisement -

Latest News

US Seizes $327K in USDT in Romance Scam Crackdown

The U.S. Justice Department seeks to forfeit nearly $328,000 in USDt linked to an...

Aave DAO Passes Contentious Proposal, But Changes Ahead

Aave Labs' controversial governance proposal passed an initial snapshot vote on Saturday with a...

Tether’s USAT Gets First Deloitte Attestation Report

Deloitte has issued the first attestation report for Tether USAT's reserves, revealing they slightly...

U.S. May Cap Critical AI Chip Sales to Chinese Firms

The Trump administration is reportedly discussing new export controls on advanced AI chips to...

Kalshi cancels $54 million in bets on Khamenei after killing.

Prediction platform Kalshi voided all wagers on a contract betting on Iranian Supreme Leader...

Must Read

The Best Bitcoin Casinos of 2025: An Expert’s Data-Driven Guide

Key TakeawaysA Deep Dive into the Top Bitcoin Casinos of 2025Bitcoin Casino Comparison Table1. Stake.com: Best for Variety & Integrated Sports Betting2. BC.Game: Best...
🔥 #AD Get 20% OFF any new 12 month hosting plan from Hostinger. Click here!