A Japanese-German team has performed the largest neural network simulation to date using a Japanese supercomputer

Largest neural network simulates 1 per cent of brain

A network consisting of 1.73 billion nerve cells connected by 10.4 trillion synapses has been created using the power of a Japanese supercomputer.

Despite being the biggest neuronal simulation to date, the process requiring 82,944 processors only represented about 1 per cent of the neuronal network of a human brain.

The supercomputer completed in 40 minutes the equivalent amount of activity that is observed to take place in the living brain within one second.

The teams from the Okinawa Institute of Technology Graduate University (OIST), Japan, and the Research Centre in Julich, Germany, ran the advanced simulation software NEST on the Japanese K supercomputer, capable of processing 8 petaflops of information per second.

Markus Diesmann, from the Institute of Neuroscience and Medicine at Jülich, and his colleagues started developing the NEST software in 2009. It is an open-source software freely available to researchers worldwide. 

“If peta-scale computers like the K Computer are capable of representing 1 per cent of the network of a human brain today, then we know that simulating the whole brain at the level of the individual nerve cell and its synapses will be possible with exa-scale computers hopefully available within the next decade,” said Diesmann.

The digital nerve cells during the simulation were randomly connected, while each synapsis was supplied with 24 bytes of memory, enabling the researchers to carry out detailed mathematical description of the computational capacity required. Overall, the experiment needed memory equivalent to that of 250,000 PCs.

The researchers believe that with the increasing computational capacity of personal computers, neuroscientists will be able in the future to investigate neuronal systems using either regular laptops or computer clusters.

In the future, the team hopes to investigate the neural control of movement and the mechanism of Parkinson’s disease using the K supercomputer and the NEST software.

“The new result paves the way for combined simulations of the brain and the musculoskeletal system using the K Computer. These results demonstrate that neuroscience can make full use of the existing peta-scale supercomputers,” said Kenji Doya from the OIST.

The technology will be used, for example, during the Human Brain Project (HBP), funded by the European Union that is scheduled to start in October this year.

Recent articles

Info Message

Our sites use cookies to support some functionality, and to collect anonymous user data.

Learn more about IET cookies and how to control them