Brain-computer interfaces: signals, methods, and goals

A brain-computer interface (BCI) changes signals from the brain, which normally merely reflect brain activity into the end product of that activity, outputs that communicate the user's intent to the outside world (Wolpaw et al., 2002). BCI operation requires the effective interaction of two adaptive controllers: user and system. The user generates signals that encode intent; and the system translates these signals into commands that control a neuroprosthesis, an environmental control system, a word processor, or another external device. Thus, BCI use is essentially a new motor skill that user and system master and maintain. A variety of different brain signals could conceivably be used for BCI communication and control. These signals fall into two classes: spikes that reflect the action potentials of individual neurons, and field potentials that reflect the combined synaptic, neuronal, and axonal activity of groups of neurons (Niedermeyer and Lopes da Silva, 1999).