Simulated brain seen for the first time

Simulated brain seen for the first time
Written by admin
Simulated brain seen for the first time

TU Graz has achieved a breakthrough in research
Simulated brain seen for the first time

providers on a given topic

For the first time, scientists from the Technical University of Graz have succeeded in simulating visual functions on a detailed mouse brain model. Until now, it was possible to model brain structures, but not to perform specific functions.

Scientists from TU Granz succeeded in replicating visual functions on a simulated mouse brain.
Scientists from TU Granz succeeded in replicating visual functions on a simulated mouse brain.

“The ground-breaking thing about our latest model is that for the first time we can let the simulation of the brain perform certain functions – in our case vision,” says neuroscientist Wolfgang Maass, who is responsible for the corresponding scientific research. paper along with his PostDocs Guozhang Chen and Franz Scherr. With the results in mind, the researchers hope for a new scientific method to be used in future research.

Neural network as a basis

The researchers chose to focus on the visual function because it is one of the central functions of artificial intelligence. An example is autonomous driving or image processing. Algorithms must interpret and learn from environmental data recorded by sensors. Another basis was decades of studies by the Allen Institute for Brain Science in Seattle, which was dedicated to decoding the visual cortex of mice.

“We converted this data into a simulated network of biological neurons – that is, into a computer model of a part of the brain – and were able to reproduce the visual function using this biological model,” explains Maass. The model can perform the most important visual tasks and is resistant to interference. In the next step, the differences between the biological visual function of the simulation and the visual function of artificial neural networks are to be investigated.

Better understanding of the brain

Neural networks the brain is not only extremely powerful, but also very energy efficient: neurons “fire” only when necessary for a certain task. Artificial neural networks replicate this behavior—however, their neurons and network architecture are significantly different from those in the brain. Biological simulation models allow researchers to better understand these processes.

Such findings can also be used in computer technology, as Maass notes: “We are currently starting a pilot project with processor manufacturer Intel, embedding our biological models into their neuromorphic chips to see if this actually makes them more energy efficient.”

(ID: 48761083)

About the author


Leave a Comment