One Step Closer To Efficient Robotic Limbs

//

A device that would allow paralyzed people to use their thoughts to move robotic limbs fluidly and realistically is now one step closer to reality.

A team of scientists from Harvard, MIT and Massachusetts General Hospital led by Ziv Williams have found two groups of cells in one area of the monkey brain that allow the animals to remember a sequence of two movements at once. The team was then able to program a computer to interpret those brain patterns, in turn moving a cursor on a screen in the planned sequence.

The development is an improvement over current brain-machine interfaces, which focus on translating a single thought into a single movement in an external device.

NEWS: Brain In A Dish Flies Plane

DNEWS VIDEO: Mind-Controlled Robot Uses Human Brainwaves

Most real-world actions are multi-faceted. When planning to take a sip from a cup or play a song on a piano, for example, people imagine the fluid behavior, not each individual movement required to get it done.

To bring technology closer to the goal of fluid and efficient movements, the researchers trained two male rhesus monkeys to move a cursor on a computer screen to two targets that had previously flashed in front of them, one after the other. During each round, the researchers recorded activity in 281 neurons in two areas of the prefontal cortex, the part of the brain responsible for planning complex actions.

BLOG: Mind-Controlled Drone Takes Off

Using the information collected, the team reported in the journal Nature Neuroscience, they were then able to program computers to turn the activity of just a small number of monkey neurons into a two-stage action on the screen with an accuracy rate of more than 70 percent.

The findings could eventually lead to robotic limbs that will move more quickly, flexibly or efficiently.

The development of [brain-machine interfaces] that can perform and potentially execute sequential motor function more effectively in this way will require substantial technological innovations,” the researchers wrote. “But as a key initial step, it is necessary to consider a concurrent BMI architecture in which the elements of a planned motor task are decoded in parallel (at once).”

Credit: Victor Habbick Visions/Science Photo Library/Corbis