The Challenges of Using Scalp-EEG Input Signals for Continuous Device Control

Whether aiming to control a computer cursor, a robotic arm, or a wheelchair, it remains a significant challenge to achieve responsive and reliable asynchronous control via EEG signals. The most promising scalp-recorded EEG signals for this task are sensorimotor rhythms and steady-state visual evoked potentials, which have both been demonstrated to be viable for continuous device operation in controlled laboratory settings. Several issues, such as handling signal nonstationarity and identifying reliable asynchronous modes of operation, must be addressed before these scalp-EEG signals can become practical for controlling devices outside of the laboratory.