Scientists Ready to Simulate 20 Billion Brain Neurons

Scientists just proved they can simulate 20 billion neurons firing at once, bringing us closer than ever to understanding the human brain.

Researchers at Germany's Jülich Research Centre successfully demonstrated last month that their supercomputer can handle a network matching the size of the human cerebral cortex. That means simulating 20 billion neurons and 100 trillion connections, all working together like the real thing.

The breakthrough builds on an exciting 2024 win when scientists completed the first complete map of a fruit fly's brain. That tiny organ, smaller than a grain of sand, contains almost 500 feet of wiring and 54.5 million synapses, showing just how complex even the simplest brains can be.

Now, Professor Markus Diesmann and his team are scaling up dramatically using JUPITER, currently the world's fourth most powerful supercomputer. The machine runs on thousands of graphics processing units, giving it the muscle needed to track billions of neurons firing simultaneously.

This attempt looks more promising than past efforts. The Human Brain Project from a decade ago fell short despite significant government funding. But technology has caught up to ambition, and the Jülich team has already proven their approach works.

"We know now that large networks can do qualitatively different things than small ones," Diesmann explained. Understanding how billions of neurons interact could reveal patterns and processes that smaller simulations simply can't show.

The team created what they call a "spiking neural network" that mimics how real neurons communicate. When scaled up to brain size on JUPITER, it maintains the complexity scientists need to study how signals travel and transform throughout the organ.

Why This Inspires

The human brain remains one of science's greatest mysteries, and every step toward understanding it opens new possibilities. These simulations could help researchers discover how memories form, how we learn, and potentially how to treat brain disorders affecting millions of people.

While scientists acknowledge they can't perfectly replicate an actual brain yet, each advance brings new insights. University of Sussex professor Thomas Nowotny noted that even brain-sized simulations teach us different lessons than the real thing, but those lessons still matter.

The progress from mapping a fly brain to simulating human-scale neural networks in just one year shows how fast this field is moving. What seemed impossible a decade ago is now happening in German research labs, powered by machines that can process trillions of connections.

We're watching scientists unlock the secrets of the most complex object in the known universe, one neuron at a time.