Innovative Digital Models Map Brain-Wide Neural Connections in Mice

Researchers at EPFL have developed a groundbreaking method to create highly realistic, comprehensive wiring maps of the mouse brain through advanced digital modeling. This approach combines experimental data with sophisticated mathematical and computational techniques, enabling the simulation of how neurons connect across the entire brain. By leveraging large datasets of biological axonal reconstructions, including collaborations with Southeast University in China, scientists employed machine learning to categorize neurons based on their wiring patterns. A key innovation was the creation of a computational method that produces synthetic axons mirroring the paths of real neurons, ensuring the resulting connectomes accurately emulate actual brain networks. These synthetic axons replicate essential features like morphology and connectivity, allowing the assembly of digital brain models that resemble biological networks, particularly long-range connections linking distant regions. This development facilitates filling gaps in existing connectome datasets, enabling large-scale brain simulations, hypothesis testing, and exploration of brain organization in ways previously impossible through traditional experimental techniques. While currently focused on the mouse brain, the same principles could eventually be applied to other species, including humans, providing valuable insights into brain structure, function, and neurological diseases.