They're calling it a neuroprosthesis, and while it's only one patient for now, the team at the University of California, San Francisco (UCSF) hopes their device may help other paralyzed people communicate.
"To our knowledge, this is the first successful demonstration of direct decoding of full words from the brain activity of someone who is paralyzed and cannot speak," said Dr. Edward Chang, a neurosurgeon at UCSF who led the research team.
"It shows strong promise to restore communication by tapping into the brain's natural speech machinery," Chang said in a statement.
The team implanted an array of electrodes over the area of the brain that controls speech in a man who suffered a stroke that left him paralyzed and unable to speak at age 20.
"Since his injury, he has had extremely limited head, neck, and limb movements, and communicates by using a pointer attached to a baseball cap to poke letters on a screen," UCSF said in a statement.
"His cognitive function was intact," the team wrote in their report,
published in the New England Journal of Medicine.
The man, now in his late 30s, was prompted to use a limited vocabulary while the device was tuned using computer algorithms to translate electrical activity from his brain. These words were then projected onto a computer screen.
UCSF has a video of the man using the device. "Good morning," he is prompted via a computer screen. "Hello," comes the answer, a few seconds later, also typed as text across the screen.
Asked, "How are you today?" the patient answers, haltingly, "I am very good."
"We decoded sentences from the participant's cortical activity in real time at a median rate of 15.2 words per minute, with a median word error rate of 25.6%," the team wrote.
"We detected 98% of the attempts by the participant to produce individual words, and we classified words with 47.1% accuracy using cortical signals that were stable throughout the 81-week study period," researchers said.
Eventually, the patient, who asked not to be identified, helped the team make up a 50-word vocabulary that included words such as "yes," "no," "family," "clean" and "nurse." These were expanded to full sentences such as "No, I am not thirsty."
It's not a permanent fix -- the electrode is a large device that sits on top of the skull and cannot be used continuously. But it's not a one-experiment wonder, either, the researchers said.
"In previously reported brain--computer interface applications, decoding models often require daily recalibration before deployment with a user," the researchers wrote. This device, they said, was more stable.
"This is an important technological milestone for a person who cannot communicate naturally, and it demonstrates the potential for this approach to give a voice to people with severe paralysis and speech loss," said David Moses, a postdoctoral engineer in Chang's lab who worked on the study.
"This trial is just the beginning. This is the very first participant that's been in the trial, and the first set of experiments that were part of this trial to show that this is possible," Chang said.
"On the hardware side, we need to build systems that have higher data resolution to record more information from the brain, and more quickly. On the algorithm side, we need to have systems that can translate these very complex signals from the brain into spoken words, not text but actually oral, audible spoken words," he added.
"Probably one of the most important priorities is to expand the vocabulary so that it's not limited to the 50 words that we started with, but something that is generalizable to all of the words in English, for example. We also need to make sure that what we see in this one participant can be seen with other people for a broader patient population."
Other teams have tried to help paralyzed people speak.
In 2017, a team at the University of Tübingen in Germany
used a cap studded with electroencephalogram sensors to help patients paralyzed by amyotrophic lateral sclerosis (ALS) to transmit some simple thoughts.