uni'wissen 01-2012_ENG

Kumar is observing groups of nerve cells in the basal ganglia in his research: “We have two net works that influence each other in their activity.” The first region, the subthalamic nucleus, stimu lates the second, the globus pallidus, which for its part inhibits the first. The electrical activity of these two networks is in a state of balance in healthy people, but in the sick it oscillates periodi cally. This prevents signals from being passed on – a consciously controlled command, like “reach for the cup,” doesn’t come through any more. In order to determine the cause of the oscillations, Kumar and his colleagues are simu lating the dynamics of these networks. “We can draw inferences about the behavior without having to work directly on the human brain.” Alleviating Symptoms with Half As Many Impulses What is new is the research group’s idea of including another brain structure, the striatum, in their simulations. This region receives its impulses directly from the cerebral cortex and is referred The basal ganglia (colored) in the brain are among other things responsible for selecting and control- ling movements. They include two networks of nerve cells that influence each other and whose activity periodically oscillates in Parkinson’s patients: the subthalamic nucleus and the globus pallidus. In order to study the emergence of the disease, Arvind Kumar simulated this activity in a model, also taking account of the influence of the striatum, another brain structure. Graphic: Wrobel “We can draw inferences about the behavior without having to work directly on the human brain” device. DBS does not arrest the progress of the disease, but it does allow many patients to reduce the amount of drugs they have to take. Their quality of life improves. Modeling Networks of Nerve Cells on the Computer Kumar uses computer simulations to work out what exactly happens in the networks of the roughly hundred billion nerve cells in the human brain during this process. “When something is so microscopically small that I don’t see it and at the same time so complex that I don’t under stand it, I have to simplify it.” This field of research is known as computational neuro science: Engineers, mathematicians, and bio physicists study the properties of the networks in the brain and develop models to gain new insight into the causes and therapies of neuronal dysfunctions. In order to accomplish this, they break down that which is known about nerve cells, their behavior, and the patterns of the con nections between them into the simplest possible mathematical equations. “A lot of the details are lost, but it is an initial step toward solving this enormous task.” The computer can then use the equations to reconstruct the neuronal networks believed to be harbored by the brain – from the molecular interaction of individual cells to the communi cation between complex networks. The computer thus serves as a virtual mini laboratory in which scientists simulate and test hypotheses and make predictions. A computer program helps them to develop the models. To take an example: 3,000 nerve cells for one network, 2,000 for another, and a five percent degree of connections between them results in a model of two networks with over half a million connec tions. The scientists need only specify the elec trical activity; the rest is done by the machine. The resulting raw data is evaluated and statis tically analyzed just like data obtained from an experiment. 18

uni'wissen 01-2012_ENG

Pages