Please activate JavaScript!
Please install Adobe Flash Player, click here for download

uni'wissen 01-2015

Ayoung man sits in a black leather armchair. His hand rests on a keypad. On his head he wears a hood with an array of electrodes at- tached to it. He is participating in a neuroscience experiment as a test subject. His task is easy: All he has to do is move his finger at irregular inter- vals and push one of the buttons. No one tells him when to do it. He is free to decide when to move his hand – or so he thinks. In reality the electrodes on his head are not just measuring the activity in his brain: They are charged with a weak electric current the test subject is not con- sciously aware of. “This allows us to activate particular areas of the brain,” explains the Freiburg neurobiologist Prof. Dr. Carsten Mehring. “Different stimulation patterns raise or lower the probability that the test subject will move his finger. We can thus influence his decision.” The purpose of the experiment is to determine how the activity of particular nerve cell networks is connected with movement and behavior. The Complexity of Lifting a Coffee Cup Mehring’s team of brain researchers is studying what happens in the human brain when a person moves his or her hands or rides a bike. The researchers want to understand how the brain controls movements and learns new movement patterns. They are conducting behavioral experi- ments and applying methods like electroencepha- lography, in which electrodes measure electrical activity along the scalp. In addition, the scientists are using the findings from their fundamental research to develop brain-machine interfaces, for example to enable paralyzed persons to control a prosthetic arm via an implant in their brain. An ordinary everyday movement like lifting a coffee cup seems simple: One decides consciously to do it, and the rest happens automatically. How- ever, it also involves a multitude of unconscious processes. “Movement control is a difficult prob- lem for brain research,” says Mehring. “The human “When it comes to flexible movements, even a child can beat the best robots.” body has more than 600 muscles. We do not know precisely how our brain controls them.” This is why scientists have not yet succeeded in pro- gramming robots to move as skillfully as humans. “We can train a robot to pick up a particular chess piece, but the machine is limited in its capacity to generalize and transfer knowledge it has acquired to a new task,” explains Mehring. A chess computer can beat the best players, but “when it comes to flexible movements, even a child can beat the best robots.” Structural Learning The outer layer of neural tissue in the human brain is called the cerebral cortex. It can be divided into areas that serve different functions. Numerous neural networks are involved in con- trolling a movement. The motor area is responsible for deliberate movements: It passes signals on to the spinal cord, from where they reach the muscles. When the surroundings change, the motor system needs to adapt the movement commands. The visual area helps by processing visual impressions, and the auditory area sup- plies the necessary acoustic information. The somatosensory area provides information on the current state of the muscles, enabling the brain to correct a movement and make it as precise as possible. “These areas and the neural networks inside them work together to improve a move- ment pattern.” In cooperation with other scientists, Mehring developed a concept that explains how humans learn a new movement pattern and why they can generalize it. “If you’ve learned how to ride a bicycle you can generally ride any model, whether it’s a racing bike, a mountain bike, or a city bike – even though you need to activate different muscles for each of them.” Mehring’s team approached the problem of generalization from a mathematical perspective: A certain muscle activation pattern, such as that necessary to ride a mountain bike, 13uni wissen 01 2015 13 13uni wissen 01201513