BACK_TO_FEEDAICRIER_2
Scientists simulate full 86B-neuron human brain
OPEN_SOURCE ↗
REDDIT · REDDIT// 5d agoRESEARCH PAPER

Scientists simulate full 86B-neuron human brain

Researchers have unveiled the Digital Twin Brain (DTB), a massive GPU-accelerated computing platform capable of simulating a human-scale spiking neuronal network with 86 billion neurons and 47.8 trillion synapses. By utilizing over 14,000 GPUs, the system achieves near real-time simulation, incorporating individual structural and functional imaging data to create personalized digital brain models for neuroscience and medical research.

// ANALYSIS

This breakthrough marks the transition from animal-scale simulations to full human-scale computational neuroscience, providing a "dry lab" for brain research that could revolutionize drug testing and AGI development.

  • The platform uses individual MRI and PET data to personalize the simulation, allowing researchers to study specific brain pathologies or lesion effects in a controlled digital environment.
  • Achieving real-time performance at this scale requires unprecedented computational infrastructure, demonstrating the massive GPU requirements for high-fidelity biological modeling.
  • While the simulation successfully reproduces resting-state activity and basic cognitive tasks, it remains a mathematical model of neuronal firing rather than a conscious entity, though it pushes the boundaries of how we define digital life.
  • The project provides a crucial feedback loop for AI architecture, suggesting that biological spiking patterns could inform the next generation of energy-efficient neural networks.
// TAGS
digital-twin-brainresearchgpureasoninginference

DISCOVERED

5d ago

2026-04-07

PUBLISHED

5d ago

2026-04-06

RELEVANCE

8/ 10

AUTHOR

Anen-o-me