Error Potentials for Brain-Computer Interfaces

The idea to use EEG correlates of errors to correct or reinforce BCI operation has been proposed over a decade ago. Since then a body of evidence has corroborated this approach. In this paper we give an overview of our recent work exploring the possibilities of error-potential applications, involving removing constraints of laboratory paradigms to increase “real-life” validity, and investigating EEG feature-spaces to increase detection robustness.