UC San Francisco researchers have allowed a paralyzed man to control a robotic arm through a device that transmits the signals of his brain to a computer.
He could understand, move and release objects simply imagining himself by performing the actions.
The device, known as a cerebral computer interface (BCI), worked during a 7 -month record without the need to be adjusted. Until now, such devices have only worked for one day or two.
The BCI is based on an AI model that can adjust to the small changes that take place in the brain as a person repeats a movement, or in this case, an imagined movement, and learn to do it in a more refined way.
“This combination of learning between humans and AI is the following phase for these cerebral computer interfaces,” said neurologist Karuneh Ganguly, MD, PHD, professor of neurology and member of the Institute of Neurosciences of UCSF Weill. “It is what we need to achieve a sophisticated and realistic function.”
The study, which was funded by the National Health Institutes, appears on March 6 in Cell.
The key was the discovery of how the activity changes in the brain day by day as a study participant imagined repeatedly making specific movements. Once the AI was scheduled to take into account those shifts, it worked for months at the same time.
Location, location, location
Ganguly studied how patterns of brain activity in animals represent specific movements and saw that these representations changed the day as the animal learned. He suspected that the same thing happened in humans, and that was the reason why his BCI lost so quickly the ability to recognize these patterns.
Ganguly and Nikhilesh Natraj neurology researcher, PHD, worked with a study participant who had been paralyzed by a stroke years before. I couldn’t talk or move.
He had small sensors implanted on the surface of his brain that could collect brain activity when he imagined he moved.
To see if his brain patterns changed over time, Ganguly asked the participant to imagine different parts of his body, such as his hands, feet or head.
Although he couldn’t actually move, the participant’s brain could still produce the signals for a movement when he imagined doing so. The BCI recorded the representations of the brain of these movements through the sensors in their brain.
The Ganguly team discovered that the shape of representations in the brain remained the same, but their locations changed slightly day by day.
Virtual to reality
Ganguly then asked the participant to imagine making simple movements with their fingers, hands or thumbs over the course of two weeks, while the sensors recorded their brain activity to train the AI.
Then, the participant tried to control a robotic arm and hand. But the movements were not yet very precise.
Then, Ganguly had the practice of the participants in a virtual robot arm who commented on the precision of their visualizations. Finally, he got the virtual arm to do what he wanted to do.
Once the participant began to practice with the Real Robot ARM, he only took some practice sessions transfer their skills to the real world.
He could make the robotic arm collect blocks, turn them and move them to new locations. He could even open a cabinet, take out a cup and hold it in a water dispenser.
Months later, the participant could still control the robotic arm after a 15 -minute “adjustment” to adjust how their movement representations had been diverted since the device had begun to use.
Ganguly is now refining AI models to make the robotic arm move faster and without problems, and planning to try the BCI in a domestic environment.
For people with paralysis, the ability to feed or take a drink of water would change life.
Ganguly thinks this is within reach.
“I am very sure that we have learned how to build the system now and that we can make this work,” he said.