An international team has reconstructed, with unprecedented precision, the architecture and functioning of tens of thousands of neurons from a mouse brain sample, along with their 500 million connections.
This monumental work, published in
Nature, marks a turning point in neuroscience. It provides an unprecedented view of the mechanisms underlying vision in mice, while opening new perspectives for understanding certain human neurological disorders.
An unprecedented technological feat
For nine years, researchers combined electron microscopy and artificial intelligence to reconstruct this neural network in 3D. The analyzed fragment, taken from the visual cortex, was sliced into 28,000 ultra-thin sections before being digitized.
Before analyzing this neural network, the team had recorded neuron activity while the mouse watched movie clips. This dual structural and functional approach allows direct correlation between brain anatomy and its functioning.
The generated data, equivalent to 22 years of HD video, is now freely accessible. This repository will serve as a foundation for future research on neurological disorders related to abnormal connectivity.
Findings that challenge existing models
Contrary to expectations, inhibitory neurons don't simply moderate brain activity randomly. Their action is highly selective, targeting specific circuits to coordinate overall activity.
Researchers also identified new cell types and previously unknown connection patterns. These observations suggest the brain is far less disordered than previously thought, following organizational rules that remain to be elucidated.
These results could inspire new therapeutic approaches, particularly for conditions like autism and schizophrenia where neural communication is impaired.
Article author: Cédric DEPOND