A four-choice hybrid P300/SSVEP BCI for improved accuracy

Abstract

Brain-computer interfaces (BCIs) usually rely on one of three types of signals: P300, SSVEP, or ERD. Recent work has shown that ‘hybrid’ BCIs that combine different types of signals may be superior to their simpler counterparts. This paper introduces a new type of hybrid P300/SSVEP BCI with a four-choice system. We compared this new approach to conditions that used only a P300 or SSVEP BCI. In addition to collecting EEG and performance data based on offline and then online runs, we also used questionnaires to explore the subject’s experiences with the different conditions. Results showed that subjects could use the hybrid BCI, and did not consider it more annoying, difficult or fatiguing than the P300 or SSVEP BCI. The hybrid condition improved performance relative to the SSVEP condition, but not the P300 condition. The hybrid condition did not elicit significant differences in the EEG measures compared here. These results validate this new BCI approach and provide support for future hybrid efforts.