Scientists from the Allen Institute and Japan’s University of Electro-Communications have built one of the most detailed virtual models of the mouse cortex to date, simulating roughly 9 million neurons and 26 billion synapses across 86 regions on the Fugaku supercomputer.
What they built
Researchers created a biophysically detailed, whole‑cortex simulation of the mouse brain that reproduces both structure and activity. The model comprises about 9 million neurons, 26 billion synapses, and 86 interconnected regions, offering a digital testbed to examine phenomena such as Alzheimer’s disease, epilepsy, attention, and other aspects of brain function. (alleninstitute.org)
How it works
The team integrated large biological datasets from the Allen Cell Types Database and the Allen Mouse Brain Connectivity Atlas, then used the Allen Institute’s Brain Modeling ToolKit (BMTK) together with a lightweight neuron simulator called Neulite to translate equations into spiking, communicating virtual neurons. (celltypes.brain-map.org)
The supercomputer behind it
Fugaku—developed by RIKEN and Fujitsu—can execute more than 400 quadrillion operations per second and consists of 158,976 compute nodes. That horsepower enabled the large‑scale, biophysically detailed simulation to run at whole‑cortex scale. (fujitsu.com)
What the researchers say
“This shows the door is open. We can run these kinds of brain simulations effectively with enough computing power,” said Anton Arkhipov, Ph.D., an investigator at the Allen Institute. Tadashi Yamazaki, Ph.D., of the University of Electro‑Communications added: “Fugaku is used for research in a wide range of computational science fields, such as astronomy, meteorology, and drug discovery… On this occasion, we utilized Fugaku for a neural circuit simulation.” (sciencedaily.com)
Where the work is being presented
According to the institutions, the full paper is scheduled for release at SC25, the International Conference for High Performance Computing, Networking, Storage, and Analysis, held November 16–21, 2025, in St. Louis, Missouri. (uec.ac.jp)
Who’s involved
The collaboration is led by the Allen Institute and the University of Electro‑Communications, with contributions from RIST, Yamaguchi University, and RIKEN’s Center for Computational Science. Contributors named in project materials include Laura Green, Ph.D.; Beatriz Herrera, Ph.D.; Kael Dai, B.Sc.; Rin Kuriyama, M.Sc.; and Kaaya Akira‑Tamura, Ph.D. (uec.ac.jp)
Why it matters
By uniting rich, publicly available brain data with high‑performance computing, the project provides a scalable way to probe how damage spreads through circuits and to explore hypotheses about cognition and disease in silico—potentially informing future therapeutic strategies. The researchers say this milestone advances their long‑term aim of building whole‑brain models, eventually even human models, grounded in biological detail. (sciencedaily.com)
