Pocket-Sized AI Brain Inspired by Monkey Neurons

Researchers develop a highly efficient AI model that mimics the human visual system

Apr. 11, 2026 at 10:49pm

Got story updates? Submit your updates here. ›

An abstract, highly structured painting in earthy tones of green, brown, and blue, featuring sweeping geometric arcs, concentric circular forms, and precise botanical spirals, conceptually representing the efficient neural architecture and visual processing capabilities of a new compact AI model.A compact, energy-efficient AI model inspired by the visual processing of primate brains could revolutionize artificial intelligence and our understanding of the human mind.Cold Spring Harbor Today

Scientists have created a groundbreaking AI model that is inspired by monkey neurons and is just a fraction of the size of traditional AI systems, yet retains nearly the same performance. This compact model not only saves energy but also functions more like a living brain, offering insights into how the human visual system works and the potential for more human-like artificial intelligence.

Why it matters

This discovery could revolutionize our understanding of the brain and lead to the development of more efficient and human-like AI systems. The ability to create powerful AI models that are small and energy-efficient has major implications for applications like self-driving cars, which could run on less powerful computers while still accurately distinguishing pedestrians from objects.

The details

Researchers, led by Ben Cowley from Cold Spring Harbor Laboratory, have developed an AI model that mimics a part of the brain's visual system. Starting with 60 million variables, they managed to compress it into a version with just 10,000 variables, retaining nearly the same performance. The team utilized data from macaque monkeys to understand the human visual system and created an AI model that they could comprehend, simulating the visual system's V4 neurons. These neurons encode colors, textures, and curves, forming proto-objects. By training their model on macaque data and employing compression techniques, they achieved a remarkably small model that could be observed to reveal insights into the specialized nature of V4 neurons.

  • The study was published in Nature in April 2026.

The players

Ben Cowley

A researcher at Cold Spring Harbor Laboratory who led the team that developed the pocket-sized AI brain.

Mitya Chklovskii

A group leader at the Simons Foundation's Flatiron Institute who agrees that the compact AI model could lead to a deeper understanding of human brains and more human-like artificial intelligence.

Got photos? Submit your photos here. ›

What they’re saying

“This is incredibly small. It's something we could send in a tweet or an email.”

— Ben Cowley, Researcher, Cold Spring Harbor Laboratory

“Maybe we should update the foundations of the artificial networks.”

— Mitya Chklovskii, Group Leader, Simons Foundation's Flatiron Institute

What’s next

The researchers plan to continue studying the behavior of the artificial neurons in their compact AI model to gain further insights into the human visual system and explore the potential for more efficient and human-like artificial intelligence.

The takeaway

This groundbreaking discovery of a pocket-sized AI brain inspired by monkey neurons could lead to a paradigm shift in the development of artificial intelligence. By creating a highly efficient model that functions more like a living brain, researchers are poised to unlock new possibilities for energy-saving, human-like AI systems with wide-ranging applications.