Who are the best researchers in BCI

Brain-computer interface translates brain signals into handwriting

"I'm fine so far, and how about you?" A standard phrase that can be quickly typed on the mobile phone or the keyboard. People who are not able to write independently, this sentence costs significantly more time: they would need around a minute to access it via a special human-machine interface, a brain-computer interface (BCI) Bring screen. Now a new type of BCI enables paralyzed people to write on the computer more than twice as fast as before. To do this, it uses implanted electrodes to measure the brain activity that occurs when the user imagines writing a letter by hand. A decoder then translates the brain signals into individual letters and characters. The research team led by Francis Willett from Stanford University reports on his work in "Nature". With the help of this BCI, the test person could write up to 90 characters per minute. Over 94 percent of the signs were correct. The BCI thus almost reached the pace of people of the same age group who type an average of 115 characters per minute on their smartphones.

The 65-year-old test person had been paraplegic from the neck down since an accident on the spinal cord. Only minimal movements and twitching of the hand were possible for him. In order to be able to use the BCI, brain electrodes were implanted into the area of ​​the cerebrum, which is responsible for the movement of the hands. The research team instructed him to imagine handwriting words and sentences. In the meantime, the electrodes measured the resulting electrical signals. First, the BCI learned the alphabet. In the next step, the test person imagined writing down words and sentences.

The more he practiced with the BCI, the better the software got: Using the previous sentences, the program calculated the probability of a certain letter appearing next, which kept its error rate low. Even when the test person thought up their own sentences, the BCI reached almost 74 characters per minute with an error rate of 8.5 percent. During the test series, which lasted several days, the researchers had to repeatedly feed the algorithm with modified exercise sets in order to recalibrate it. This was necessary because the measured nerve signals changed slightly over time, for example due to minimal movements of the electrodes or remodeling processes in the brain.

Are the benefits worth the delicate intervention?

BCIs offer people who no longer speak or can no longer move new ways of communicating. So far, the scientists have used the signals that arise when someone imagines gross motor movements. Research has shown that the patterns of activity that the brain uses to control such movements can still be induced in the brain many years after paralysis. However, common BCIs do not use implanted electrodes, but conventional EEG. The electrical brain activity is measured by electrodes on the outside of the skull, which only gives a much more coarse picture of the brain activity; but that is enough to let test subjects control a cursor on the computer screen, for example. One advantage is that it does not require surgery.

In previous experiments with implanted electrodes, the Stanford University team managed to get their patients to write at a maximum of 40 characters per minute. Here, too, users should imagine movements - in this case, as if they were typing on a computer keyboard with hand and arm. Small and more complex movements, such as writing by hand, could be used for faster communication. Willett and his team suspect that their BCI can also be used to write so much faster because the signal patterns in the brain that trigger the letters are very different. This makes it easier for the program to distinguish between characters than, for example, moving a cursor in a straight line.

In the future, the BCI could offer patients with locked-in syndrome a new way of communicating. Those affected are no longer able to control most of their muscles at will. This can occur, for example, after a stroke of the brain stem or in the context of advanced amyotrophic lateral sclerosis (ALS).