Mathematical Models Provide Explanations
Visual perceptions, smells or spoken words: all of our sensory impressions and cognitive and emotional experiences are processed in the brain. This is done by transmitting information, in the form of electric impulses, from one of the some 86 billion neurons to the others at lightning-fast speeds. In turn, each neuron has up to 10,000 junctions, which are known as synapses. Thus the brain is a gigantic control centre.
Scientists from Computational and Systems Neuroscience (INM-6) at the Institute of Neuroscience and Medicine at Forschungszentrum Jülich are developing, among other things, simplified models of the neurons in the brain, which depict examples of the vast number of neural connections. This will help to determine whether the differing activity of neurons is caused by cellular structure or is simply a result of the properties of the individual cell. The model reduction also makes it possible to observe artificially altered variations of brain activity – for instance, those that occur in the context of neurodegenerative diseases – and their effects on the entire neural network.
In order to simulate larger neuronal circuits on a computer, Jülich scientists, together with other researchers, are developing the Neural Simulation Technology (NEST) software. The global NEST Initiative was founded in 2001. Its members aim to continually improve simulation technology and make it available to the scientific community free of charge.
To improve access to simulation technology and supercomputers for neuroscientists, and to maintain simulation tools built on decades of experience on a long-term basis, Forschungszentrum Jülich has founded the Simulation Lab Neuroscience, which, in its function as the “Bernstein Facility Simulation and Database Technology”, will take on a special role in the national Bernstein Network Computational Neuroscience.
New knowledge about the processes in the brain is also incorporated into the BrainScaleS research project, which is funded by the European Union, and in which the research team from INM-6 also participates. BrainScaleS aims to analyse the functional principles and energy efficiency of the brain, and to use this data as a model for developing more powerful and energy-efficient computers.