An auditory BCI using voluntary mental response

A novel brain computer interface (BCI) is implemented in this study, which only depends on auditory modality. The subjects voluntary recognition of the property (e.g. voice laterality) of a target human voice makes the discriminability between brain responses to target and non-target voices in a random sequence. EEG data from eight subjects showed that the amplitude of N2 and late positive component (LPC) elicited by target voice was significantly higher than that by non-target ones. The area of N2 and LPC component was used as features for identifying the target among digit numbers 1-8, achieving an average accuracy of 89%.