Elon Musk’s company “Neuralink” has been making headlines with its so-called “human-machine interface”. Musk has been testing chips on monkeys as well as humans. The US Food and Drug Administration approved human trials in May 2023.
These chips allegedly convert brain signals into electrical impulses – and vice versa. Musk has recently nevertheless had to face a few setbacks.
Now “Neuralink” also has competition. An international team led by the Swiss Federal Institute of Technology in Lausanne (ETHL) has developed an alternative miniature brain-computer interface under the acronym MiBMI that can convert thoughts into text with great precision. According to the researchers, who presented their results in a specialist study, the tiny device could fundamentally change communication for people with severe paralysis. The chip, which is only about the size of a fingertip, can also be implanted in healthy people.
The MiBMI chip interprets brain signals generated during visualized writing and translates them into text on a screen in real time. It processes signals from up to 512 channels simultaneously, which should enable a fairly precise recording of brain activity. In tests with a person suffering from tetraplegia, the chip was able to correctly identify 90,8 percent of the visualized letters. It can also distinguish between 31 different symbols, including commands such as “space” and “backspace”.
The study authors point to the extraordinary efficiency of the MiBMI, which uses an innovative method called “distinctive neural codes” (DNCs) that achieves high accuracy with minimal power consumption. In addition, the chip only activates when there is relevant neural activity, which saves energy and improves signal quality. The current writing speed is 90 characters per minute – a significant improvement over existing aids.
The MiBMI technology could soon enable fully implantable brain-computer interfaces that do not require external devices and allow wireless communication with other devices. Sadly, nothing is yet known about its long-term safety and stability and neither about the adaptation to individual differences in brain signals.