Supercomputer Develops Most Detailed Virtual Mouse Brain Simulation

Scientists have achieved a significant milestone in neuroscience by creating one of the most comprehensive and biologically realistic animal brain simulations ever developed. Using the powerful supercomputer Fugaku, they have reconstructed the entire mouse cortex, which includes nearly ten million neurons and 26 billion synapses, capturing intricate details of both structure and activity.

This simulation enables researchers to explore brain function and neurological disorders like Alzheimer's and epilepsy within a virtual environment, allowing them to track how damage propagates through neural circuits.

The report states that this model can help investigate cognitive processes and consciousness, offering a new way to study how brain disorders develop and the mechanisms of brain waves in attention and seizure dynamics.

The collaborative effort involved scientists from the Allen Institute and Japan's University of Electro-Communications, among other organizations. Fugaku, one of the fastest supercomputers in the world, can perform over 400 quadrillion operations per second, making it a key tool for this groundbreaking work.

The simulation was made possible by converting biological data from the Allen Cell Types Database and the Allen Connectivity Atlas into a functioning digital representation of the cortex. Researchers utilized the Allen Institute's Brain Modeling ToolKit and a tool named Neulite to create virtual neurons that mimic real neuronal behavior, allowing for a vivid observation of brain activity.

The simulation not only captures the fine details of neuron structure and synapse activity but also reflects the electrical signaling that occurs in live brain cells. The researchers emphasize that this achievement marks the beginning of a new era in neuroscience, with aspirations to eventually model the entire human brain, leveraging all biological details uncovered by their research.

This work highlights the intersection of deep neuroscience knowledge and high-performance computing, paving the way for future studies and potential therapies for neurological diseases. The full findings will be presented at SC25, a leading global supercomputing conference in mid-November, showcasing the implications of this virtual cortex for understanding brain functions and disorders.