The human brain remains an enigma, but researchers at the Medical Center are probing its mysteries by connecting the brain's signals to external devices like prosthetic limbs.
Funded by a new $26 million grant from the Defense Advanced Research Projects Agency, the research team is developing human brain-machines that will enable a human to control a prosthetic arm in the same way one controls his or her own arm by thinking about it. The researchers believe the project has boundless potential to alter human capabilities.
"This project takes advantage of a long history of experience and work by many people... mostly on how individual neurons workâ??to investigate motor control," said Dr. Dennis Turner, a professor of neurosurgery and neurobiology, who is working on the project.
Previous work with monkeys has shown that brain signals can be interpreted by a machine to control a prosthetic device. Scientists in the lab of Dr. Miguel Nicolelis, professor of neurobiology, used the brain signals of a monkey at Duke to control a prosthetic arm at the Massachusetts Institute of Technology in Cambridge, Mass., over a network.
In the current project, Patrick Wolf, an associate professor of biomedical engineering, is working on using arrays of electrodes to transmit signals from the brain tissue to a usable device that will then interpret the signals. "There are many steps involved in this project," Wolf said. "Brain signals are obviously generated on the inside. The [interpreting] device is on the outside. Somehow we have to get the signals out to this device."
After the electrodes detect the brain signals, the signals are routed outside the brain and sent to an interpretive device, which then records and analyzes the signals, Turner said. The signals are then mapped out and searched for patterns.
During the next stage of analysis, a computer generates a model that interprets the original signal as a specific motion or sensation. In the final step, the computer transfers this information to the prosthetic arm.
"Different components of motion are encoded in brain signals--for instance, the position of your arm, or forces at the joint level," said Craig Henriquez, an associate professor of biomedical engineering who is working on the interpretive computational models. "The question is which parts of the brain control which features of motion."
Henriquez mentioned several obstacles the researchers would have to overcome. A major challenge is integrating the entire process into a single system, but before this can be done, Henriquez said the next big step lies in training the brain to become aware of its new arm, which requires constant visual and tactile feedback.
"There are regions of the brain that map the sensory world, and these have developed over your lifetime," Henriquez said. "This feedback, called proprioceptive feedback, is a relationship with the environment of how one thing in the world relates to another."
Henriquez illustrated this concept by referring to Duke basketball player Chris Duhon shooting a ball. Duhon constantly relies on proprioceptive feedback of where his arm is in space to know when he should release the ball.
The researchers are optimistic about the project and its potential. "We've made progress already getting signals out of the brain, and we're working very hard as a team," Henriquez said. They hope to see practical applications of the technology in five to 10 years.
Get The Chronicle straight to your inbox
Signup for our weekly newsletter. Cancel at any time.