The signals in the brain generated by thoughts and related to writing were translated into text in real time, allowing a paralyzed person to write 16 words per minute.
Using brain implants and a machine learning algorithm, the system decoded brain signals related to handwriting and introduced a paralyzed person to the digital reality of modern communication.
Also read: Neuralink and the brain-computer interface. What will Elon Musk show?
BrainGate on the way to breakthrough in the development of brain-computer interfaces
BrainGate is a consortium that has made significant progress in recent years in the development of brain-computer interfaces (BCIs). This includes a complex brain-controlled robotic arm, first unveiled in 2012, as well as the newer, high-capacity wireless BCI. The new project focused on developing a new brain-computer interface for handwriting that had been advanced under the direction of Frank Willett, a scientist at Stanford. This development was supervised by another neurobiologist Krishna Shenoy of the Howard Hughes Medical Institute, along with Henderson (a neurosurgeon from Stanford).
The new brain-computer interface is promising, but extremely radical
The subject of the experiment was a 65-year-old man who had suffered a severe and paralyzing spinal cord injury ten years earlier.
Over the course of the experiment, the man tried to move his paralyzed arm as he would when writing words physically. In his mind he imagined the act of ‘writing letters on top of each other with a pen on a yellow block’. As he did so, the decoder entered every letter the man imagined when he was “identified by the neural network,” Henderson added.
Also read: Valve is working on the brain-computer interface. Gabe Newell promises a future straight out of science fiction movies
Machine learning is not universal
While the system can distinguish letters with about 95% accuracy — allowing Henderson to achieve about three-quarters of the average typing speed of smartphone users over the age of 65 — it has some obvious limitations. Brain surgery is overtly invasive. Moreover, machine learning is not universal and must adapt to the cognitive nuances of each individual user, and involves a “highly compute-intensive” process that requires a “specialized, high-performance computer or computer cluster”. We’re far from proposing this new brain-computer interface to help paralyzed people, but the potential of the system could one day change the way we write forever.
“Total travelaholic. Subtly charming zombie geek. Friend of animals everywhere. Music buff. Explorer. Tv junkie.”