A research team of Korean scientists has invented a robot arm that moves based on imaginary brain signals. The technology is expected to be utilized in the development of medical aids.
KAIST said on Monday that a research team led by Professor Jeong Jae-seung of Brain Cognitive Science (photo) announced the research results in the Journal of Neural Engineering.
Brain signals that occur when people imagine are difficult to determine unless planting microelectrodes deep inside the brain accurately. This, however, risks injury to the brain. The research team focused on the cerebral cortex, where planted microelectrodes have less brain injury risk. “The cerebral cortex, where nerve cell bodies are concentrated, is the right part that can identify accurate imaginary brain signals. The signals are far more accurate than imaginary brain signals obtained from outside the scalp,” said Jang Sang-jin, a doctoral student. It is likened to a situation where the shouts of people outside a soccer stadium cannot be accurately identified. Still, inside the stadium, people can recognize noise from someone who sits beside them.
The research team added measurement accuracy by adding machine learning to the measured imaginary brain signals. As a result, the team delivered around 80% accuracy rate in the human direction of the arm. The research team also linked the measured imaginary brain signals to the arm of the robot to test whether the robot moved its arm in the intended direction. The robot successfully interpreted the intention for the four directions and accurately reached the target.
The technology is expected to help patients with movement disorders, including quadriplegia. However, there are tasks to overcome for commercialization, such as developing non-invasive technology. “We can analyze brain signals that differ by each person based on individual customization. The technology can contribute to commercializing robotic arms that will replace prosthetic arms,” said Professor Jeong.