NEW ORLEANS—Although the technology is still in its infancy, brain–computer interfaces can enable patients with paralysis to communicate and perform simple movements, according to research presented at the 2013 Annual Meeting of the American Neurological Association. For example, the investigational technology has allowed patients with brainstem stroke to type messages on a computer and browse the Internet.

Researchers are working to improve the control that brain–computer interfaces provide over external devices, such as prosthetic or robotic limbs, said Leigh Hochberg, MD, PhD, a neurologist at Massachusetts General Hospital in Boston. Dr. Hochberg is also affiliated with Brown University and the Providence Veterans Affairs Medical Center. The technology could aid neurorehabilitation for people with paralysis and improve researchers’ understanding of diseases such as epilepsy, he added. Eventually, the interfaces could record hundreds of neurons in a patient with epilepsy and reveal their individual firing patterns, which may be heterogeneous.

How to Develop a Brain–Computer Interface
Although they can be designed in various ways, brain–computer interfaces all include the same basic components: a neural sensor that detects activity in the brain, a decoder that translates the activity into commands, and an assistive technology that executes the commands by moving prosthetic limbs or operating computers. The technology is being developed for patients who are unable to move because of a condition such as spinal cord injury, brain stem stroke, or amyotrophic lateral sclerosis (ALS). Researchers hope that, in the future, control will be provided over functional electrical stimulation devices that directly reanimate paralyzed limbs. These devices also could facilitate traditional neurorehabilitation techniques.

Neurologists creating a brain–computer interface must choose which type of signal (eg, electrical or chemical) the device will record in the brain, as well as which brain region the device will monitor. Neurologists also must choose an appropriate sensor for the interface, such as EEG or fMRI. Dr. Hochberg and his colleagues are studying an interface that monitors the action potentials and field potentials of individual neurons in the motor cortex.

Brain–computer interfaces have benefited from studies during which electrodes, which had been implanted into the brains of nonhuman primates, decoded the firing rates of individual neurons. Researchers at the University of Utah later created the Utah array, a 16-mm² platform containing 100 electrodes, each of which can record the activity of several neurons. Research during which Utah arrays were implanted in the cortex of nonhuman primates has recorded a much greater amount of data on neural activity than was possible previously.

Ongoing Trial Indicates the Technology’s Clinical Benefits
Dr. Hochberg and colleagues began an ongoing clinical trial in 2004 to study whether patients can use a brain–computer interface to control a device by thinking about moving their own hands. The researchers have enrolled nine people in the trial so far. The participants have conditions such as stroke, spinal cord injury, muscular dystrophy, or ALS. A Utah-type array is implanted into the motor cortex of each participant to decode the motor cortical activity associated with the intention to move.

The first participant in the trial was a 24-year-old man with spinal cord injury. After implantation into the right motor cortex, the participant successfully controlled a cursor on a computer screen by thinking about moving his dominant left hand. The patient also was able to open and close a prosthetic hand using the interface. He demonstrated his ability to draw by using the interface to operate drawing software.

Neurosurgeons placed the interface in the “knob” area of the patient’s motor cortex, which ordinarily controls the hand, although the traditional understanding is that the more laterally placed face area expands medially to take over the hand area after a high cervical spinal cord injury. When the researchers asked the participant to try to move his hand, some neurons immediately increased their firing rates in unique patterns. “This [result] has helped me to reconsider what I was taught about cortical plasticity … and about the somatotopic reorganization often reported after peripheral or distal central injury,” said Dr. Hochberg.

Another patient enrolled in the trial nine years after having a brain stem stroke that resulted in locked-in syndrome. After implantation, the woman was able to type on a virtual keyboard using the brain–computer interface. By imagining that she is squeezing her hand, the woman points to and clicks on letters on the keyboard to spell words. The woman used this technique to chat with BrainGate researchers and also joined “what we believe to be the world’s first brain–computer-interface Google chat,” said Dr. Hochberg.

In addition, the interface enabled the woman to control a robotic arm, which she used to pick up and drink from a thermos of coffee that had been placed in front of her. “Notably, the woman is using this device more than five years after it’s been implanted, which speaks for the potential of this technology to last long enough to be of clinical utility,” said Dr. Hochberg.

—Erik Greb

Brain–Computer Interfaces Could Aid Neurorehabilitation