Participants took part in 22 to 68 sessions, with each session lasting 24 minutes. At the end of their training, two participants were able to hit the target on a computer screen 92 percent of the time.
"[This is] compelling evidence that individuals can learn to control the spectral composition of their EEG, and that this allows them to exercise impressive control over the movement of a cursor displayed on a screen," said Emanuel Donchin, a psychology professor at the University of South Florida in Tampa.
The two study participants with spinal cord injuries performed better than the uninjured participants, possibly reflecting greater motivation or injury-associated brain changes.
The computer program selected the brain waves controlling cursor movement based on a person's past performance.
"The computer automatically adapts to the person using the system," Wolpaw said. "It is an interaction between two adaptive controllersthe system and the person using it."
Wolpaw predicts future improvements of the non-invasive brain-computer interface will focus on three-dimensional movement.
In the future, users may be able to operate a robotic arm that could pick things up, or they may be able to control a neural prosthesis in which electrodes implanted in a paralyzed limb may be stimulated to get the muscles to move.
There is a lively debate about to what extent it's necessary to implant electrodes in the brain to achieve complex control.
Invasive brain-computer interfaces, in which electrodes are surgically implanted into the brain, have so far mainly been tested on monkeys. However Cyberkinetics, a neurotechnology company based in Foxborough, Massachusetts, has just initiated a study in which a human has been implanted with 80 electrodes.
Wolpaw said his study offers a strong non-invasive alternative. "If you can do as well, or nearly as well, with electrodes on the scalp as [with electrodes implanted in the brain], you might very well elect to do that," he said.
But scientists seem to agree that both invasive and non-invasive means of acquiring brain signals should continue to be developed, because both have potential benefits for different applications.
"For example, implanted brain-recording electrodes could be integrated with implanted stimulation systems that activate paralyzed muscles and generate useful movements," said Dawn Taylor, a biomedical engineering professor at Case Western Reserve University in Cleveland, Ohio.
"In this way, a paralyzed person could once again move their arms and hands just by thinking of doing so," she said. "With implanted electrodes, the person would not need a caregiver to put on and maintain an external electrode cap in order to move their hands."
Taylor said Wolpaw's study is encouraging, because it shows that scientists can get a lot of useful information from relatively noisy, low-resolution, non-invasive brain readings.
"However, this suggests that we could get even better results if we apply similar adaptive training techniques in people implanted with higher-resolution invasive electrodes," she said. "The good news is that severely paralyzed people will have multiple options for effectively controlling assistive devices using their brain activity."
SOURCES AND RELATED WEB SITES