EEG-based communication: presence of an error potential

Background: EEG-based communication could be a valuable new augmentative communication technology for those with severe motor disabilities. Like all communication methods, it faces the problem of errors in transmission. In the Wadsworth EEG-based brain-computer interface (BCI) system, subjects learn to use mu or beta rhythm amplitude to move a cursor to targets on a computer screen. While cursor movement is highly accurate in trained subjects, it is not perfect. Methods: In an effort to develop a method for detecting errors, this study compared the EEG immediately after correct target selection to that after incorrect selection. Results: The data showed that a mistake is followed by a positive potential centered at the vertex that peaks about 180 ms after the incorrect selection. Conclusion: The results suggest that this error potential might provide a method for detecting and voiding errors that requires no additional time and could thereby improve the speed and accuracy of EEG-based communication. (C) 2000 Elsevier Science Ireland Ltd. All rights reserved.