BCI Switch Using Detection of Contingent Negative Variation and Its Evaluation Based on ROC Analysis

Abstract

Our aim of this study is to develop a brain-computer interface (BCI) for use in our day-to-day lives. However, most of conventional BCIs have not taken into account the user's intention, especially in a case that the BCIs are always-running. In such a situation, these BCIs must work only when a user desires to use, but these BCIs might also work by a user's unintentional thought. We, therefore, proposed a new BCI system called the BCI switch. The BCI switch was designed to be activated by detection of the contingent negative variation (CNV)-related potentials, depending on user intentions for startup or shutdown of a conventional BCI system. Detection of CNV-related potentials was chosen to improve the single-trial discrimination of user intentions to switch because CNV had a high signal-to-noise ratio and needed high concentration for its elicitation. In this study, we evaluated its efficacy by using a receiver operating characteristic analysis. The result suggests that our proposed system might work as a switch of startup or shutdown by single-trial detection of imaginary response to user's intention without physical reaction.