Non-invasive brain-computer interface offers more control than once thought

A brain-computer interface (BCI) that translates electrical signals detected from the scalp into a user's commands offers comparable precision, speed and accuracy to systems that rely on electrodes surgically implanted in the brain, researchers at the Department of Health's Wadsworth Center laboratories have shown.

It has been widely assumed that only invasive devices could control complex movements, such as operating a word processing program or a motorized wheelchair by thought alone.

Jonathan Wolpaw, M.D., and Dennis McFarland, Ph.D., published their findings online in the Proceedings of the National Academy of Sciences the week of December 6, 2004. The paper will appear in the journal's December 21, 2004, print edition. Dr. Wolpaw's laboratory pioneered BCI technology.

BCIs provide an alternative communication and control option for the severely disabled, such as individuals with Lou Gehrig's disease, brain and spinal injuries, cerebral palsy and other neurodegenerative diseases. The brain's electrical output is translated by a computer into physical outputs, such as moving a cursor on a computer screen.

In the Wadsworth system, users wear an electrode cap that detects electroencephalographic (EEG) activity from the scalp and records specific brain waves. An adaptive algorithm analyzes the output and focuses on the signals that provide people greatest control as they learn to use their thoughts to direct a cursor to a target on a computer screen. As the trainee improves, the algorithm adapts anew.

"Thanks to medical technology, people paralyzed by brain injuries or disorders are living longer. Brain-computer interface technology promises to improve their quality of life by offering a new chance to communicate and control their lives," said Commissioner of Health Antonia C. Novello, M.D., M.P.H., Dr.P.H.

In the newly published study, able-bodied and spinal cord-injured individuals who were trained in Dr. Wolpaw's lab achieved real-time, two-dimensional cursor control comparable to that reported in studies of non-human primates with implanted electrodes. Their marked improvement in the performance of a non-invasive BCI can be attributed to changes in the signal processing and to advances in the adaptive algorithm. The findings suggest that an EEG-based system, which does not require surgical implantation of electrodes in the brain, may be further improved and may eventually support such sophisticated tasks as operating a neuroprosthetic arm or having mouse-like control over a cursor.

Dr. Wolpaw's research is supported by the National Institutes of Health and the James S. McDonnell Foundation. He is chief of Wadsworth Center's Laboratory of Nervous System Disorders and a professor in the University at Albany's School of Public Health. He also is a member of the New York State Spinal Cord Research Board. This board dispenses $8.5 million annually to scientists in New York and elsewhere for research aimed at curing and restoring function after spinal cord injuries.