Researchers at the University of California, San Francisco, have developed a device that allows paralyzed people to control a robotic arm by simply imagining the action.
The device worked for a record seven months without requiring any adjustments. Previously, such devices only worked for a day or two.
Researchers at the University of California, San Francisco, have developed a device that allows paralyzed people to control a robotic arm by simply imagining the action.
The device worked for a record seven months without requiring any adjustments. Previously, such devices only worked for a day or two.
The device, known as a brain-computer interface (BCI), is based on an artificial intelligence model that adapts to subtle changes in the brain as a person repeats a movement — or, in this case, an imagined movement — gradually improving their ability to perform it, writes SciTechDaily.
“This combination of human-AI learning is the next step for these brain-computer interfaces,” said neuroscientist Karunesh Ganguly, MD, professor of neurology and member of the UCSF Weill Institute for Neuroscience.
The key point was discovering how brain activity changed from day to day as the study participant repeatedly imagined making certain movements. After the AI was programmed to take these changes into account, it worked for several months.
The researchers worked with a man who had been paralyzed by a stroke several years earlier. He could neither speak nor move.
Tiny sensors were implanted on the surface of his brain that could record brain activity when he imagined movement.
To see if their brain patterns changed over time, Ganguly asked the participant to imagine moving different parts of their body, such as their arms, legs, or head.
Although he couldn't actually move, the participant's brain could still generate signals to move when he imagined doing so. The BCI recorded the brain's representation of these movements through sensors in his brain.
Ganguly then asked the study participant to imagine making simple movements with their fingers, hands, or thumbs for two weeks, while sensors recorded their brain activity to train the AI.
The participant then tried to control the robotic arm and hand. But the movements were still not very precise.
So Ganguly had the participant practice on a virtual robotic arm that gave him feedback on the accuracy of his visualizations. Eventually, he got the virtual arm to do what he wanted.
When the man began training with a real robot arm, it only took him a few training sessions to transfer his skills to the real world.
He could make the robotic arm pick up blocks, rotate them, and move them to new locations. He was even able to open a cupboard, take out a cube, and bring it to a water dispenser.
Several months later, the participant was still able to control the robot arm after a 15-minute “tuning-in” to adjust to how his perception of the movements had changed since he started using the device.
Researchers are now refining artificial intelligence models to make the robotic arm move faster and smoother, and plan to test BCIs in the home. For people with paralysis, being able to eat or drink water on their own could be life-changing.
