Eon Systems Demonstrates the First Complete Fruit Fly Brain Simulation Driving a Virtual Body to Perform Natural Behaviors

Gate News: On March 9, Eon Systems, a full-brain simulation company, demonstrated the first closed-loop system driven by a complete brain simulation that controls a physically simulated body. The full-brain model, built from the connectome of a real fruit fly, drives a virtual fruit fly to perform natural behaviors such as walking and grooming within the MuJoCo physics engine. This is the first time a biological connectome-derived full brain simulation has been used to generate diverse natural behaviors through physical simulation.

The core of the system is the fruit fly whole-brain computational model published by Eon senior scientist Philip Shiu in Nature in October 2024. The team used electron microscopy data from the FlyWire connectome and machine learning-predicted neurotransmitter types to build a leak-integrate-and-fire model covering over 125,000 neurons and 50 million synaptic connections. The model relies solely on the connectome topology, synaptic weights, and excitation/inhibition mapping, without tuning parameters neuron-by-neuron. The model’s predictions matched 91% of 164 published experimental observations. When synaptic weights were randomly shuffled, only 1 out of 100 simulations reproduced correct behaviors, indicating the connectome structure strongly constrains brain function.

In this demonstration, the brain model was integrated into the NeuroMechFly v2 simulation framework, forming a complete closed loop: sensory input → whole-brain neural activity → motor output. Eon Systems is headquartered in San Francisco, with an advisory board including Harvard geneticist George Church and Mathematica founder Stephen Wolfram. The next goal is to simulate the mouse brain (about 70 million neurons, 560 times that of the fruit fly), with the ultimate aim of full human brain simulation.