Error-related EEG patterns during tactile human-machine interaction

Abstract

Recently, the use of brain-computer interfaces (BCIs) has been extended from active control to passive detection of cognitive user states. These passive BCI systems can be especially useful for automatic error detection in human-machine systems by recording EEG potentials related to human error processing. Up to now, these so-called error potentials have only been observed in the visual and auditory modality. However, new interfaces making use of the tactile sensory modality for conveying information to the user are on the rise. The present study aims at investigating the feasibility of BCI error detection during tactile human-machine interaction. Therefore, an experiment was conducted where EEG was measured while participants interacted with a tactile interface. During this interaction, errors of the user as well as of the interface were induced. It was shown that EEG patterns after erroneous behavior — either of the user or of the interface — significantly differed from patterns after correct responses.