Detecting Words From Brain Signals, University of Utah Reveals

Medgadget -- Researchers from the University of Utah have shown that they can detect articulated words from signals received by electrodes on the brain’s surface. This might one day enable patients with locked-in syndrome to communicate with their surroundings. The researchers used grids of microelectrodes placed on the cortical surface over speech centers during craniotomy in a patient with severe epileptic seizures. The microelectrodes consisted of 16 nonpenetrating microwires at millimetre intervals in a 4x4 grid pattern. They recorded local field potentials from the surface of face motor cortex and Wernicke’s area. Brain signals corresponding to ten different words were analyzed: yes, no, hot, cold, hungry, thirsty, hello, goodbye, more and less. Accurately distinguishing between two different words, for example yes and no, was possible up to ninety percent of the time, while accuracy dropped to forty-eight percent when distinguishing between all ten words. These results were achieved without any patient training. There is still plenty of opportunity to improve on these results while reducing invasiveness and the researchers go as far as to speculate about wireless systems able to record high-resolution cortical surface potentials functioning as brain-computer interfaces. The results are published in the September issue of the Journal of Neural Engineering.

MORE ON THIS TOPIC