📄 Showing the "Simple" version — "Advanced" is not ready yet. Add it to favorites to help prioritize it. Ask for the full read

The fruit fly Drosophila is a tiny creature with a brain the size of a poppy seed. It contains about one hundred thousand nerve cells, and scientists have already mapped all their connections, like a city subway map where stations are neurons and lines are the links between them. This map is called a connectome.

Researchers asked: can we use this ready-made natural circuit as a 'brain' for a computer model of a fly? They took the connectome of an adult Drosophila and turned it into a control program — a Neural Network Oscillations with the same connection structure. Then they placed the virtual fly in a simulator where it had to learn to walk, turn, and perform other movements. Learning proceeded by trial and error: successful steps earned rewards, falls earned penalties.

The result surprised them: the fly with its 'native' connectome learned faster and moved more steadily than models with random or simplified structures. It needed fewer attempts to master a new movement — much like an experienced dancer finds it easier to learn a new routine than a beginner. Moreover, this model was easier to analyze: you can trace how signals travel from the 'eyes' to the 'muscles' and which connections matter most.

This work is a step toward machines that learn not from scratch but by building on evolution-tested circuits. Perhaps future robots will have 'brains' inspired not by abstract algorithms but by real biology — from flies to humans.

Original abstract

as written by the authors arXiv:2602.17997

Animals perform coordinated whole-body movements under the control of neural systems shaped by brain-wide connectivity. The mapping of the whole-brain neural connections, or the connectomes, provides a natural graph for modeling sensorimotor information flow, yet its potential as a neural controller for embodied agents remains largely unexplored. Here, we introduce the Fly-connectomic Graph Model, which directly instantiates the whole-brain connectome of an adult Drosophila as a graph-structured neural controller for movements of a simulated biomechanical fruit fly via deep reinforcement learning. We achieve stable performance across diverse locomotion tasks, as well as better sample efficiency compared to both graph and non-graph baselines. Our results demonstrate a biologically informed way towards effective control policy design by translating whole-brain wiring principles into actionable architectural priors, while also improving the interpretability through dynamic information flow. This work also highlights the potential to bridge neuromechanics with embodied intelligence by providing a computational platform for investigating the sensorimotor transformation underlying animal behavior and a paradigm to advance the development of more nature-aligned intelligent systems.