Putting ideas into practice

Every movement that we execute - each grasp and every step - has its origin in the brain. The goal of Carsten Mehring and his research group at the Bernstein Center for Computational Neuroscience and the University of Freiburg is to use the brain signals for controlling prostheses or for using a computer, thus laying the foundations for developing a new prosthetic device for severely paralyzed patients. Together with colleagues from the University Hospital in Freiburg, the scientists showed that with the help of electrodes placed onto the brain, continuous arm movements can be predicted. Their work is published in the January edition of the 'Journal of Neuroscience Methods'.

Left panel: Example of movement trajectories fom an experiment. Subjects used a handle to follow several target points (yellow) with a cursor (green). The trajectory of the cursor and the target points already visited were invisible to the subject. Right panel: the cursor movement along the X-achsis (top) and the Y-achis (bottom) in such an experiment (green curve). For comparison, the reconstruction of the movement from the subjects brain activity is shown (red curve). (Modified reprint from: Tobias Pistohl, Tonio Ball, Andreas Schulze-Bonhage, Ad Aertsen, Carsten Mehring. Journal of Neuroscience Methods, 2008 Jan 15 167/1 pp. 105-114)

Scientists around Mehring used a so-called 'semi-invasive' approach known as the electro- corticogram (ECoG) for measuring electrical brain signals. 'We are looking for an optimal compromise between fully invasive and non-invasive methods,' says Mehring. In non-invasive methods, such as the EEG, electrodes are applied onto the scalp. The neural signal is measured through the skull and has a correspondingly low spatial resolution. In fully invasive methods, electrodes are inserted a few millimeters into the brain, so that the activity of individual neurons or groups of neurons can be registered. The signal is much more accurate and allows controlling complex movements. First clinical studies with this method have already been successfully carried out with severely paralyzed patients. However, it is still hard to judge whether the brain can be damaged by the implanted electrodes and how stable the measured signals are over longer periods of time.

With ECoG, the electrodes are placed directly onto the brain surface and do not penetrate the brain tissue. They measure voltage changes on the brain surface that arise from large groups of neurons. This method is less invasive, and the measured signals are presumably stable over a long time. 'We want to check whether this method is suitable for the control of movements, and thus constitutes a possible alternative to fully invasive methods,' says Mehring and continues: 'Our results give us hope that this could work.'

Mehring conducted his investigations in epilepsy patients who got electrodes implanted in preparation for a brain surgery. Their brain activity was recorded while they were using a handle to control a cursor on a monitor. Using mathematical algorithms, the scientists succeeded in extracting brain signals that correlated with the cursor movements and by which a continuous reconstruction of the movement was possible.

In a next step, Mehring and his colleagues want to investigate how well this strategy can be used to control a cursor on the screen without the subject moving its arm. 'Previous studies show that in this way, movement reconstruction from brain signals can still be improved, because the subject can learn to adapt its brain activity to the cursor's movement,' says Mehring. 'There is hope that, based on such methods, new prosthetic devices or communication tools for severely paralyzed patients can be developed in the future. For the practical application of such devices in patients, however, many scientific and technical problems still have to be solved'.