An electrode selection approach in P300-based BCIs to address inter- and intra-subject variability

Abstract

Brain Computer Interface (BCI) technologies use neural activity to implement a direct communication channel for healthy and disable subjects. To achieve this, many investigations look to improve BCI precision by increasing the number of electrodes with standard configurations, ignoring inter- and intra-subject variability. To control this variability in event-related potential (ERP)-based BCIs we propose to investigate the cumulative peak difference, an intrinsic characteristic of ERP, as a measure for electrode selection. The results shown in this work indicate that the proposed method improved accuracy and bitrate in all analyzed electrode sets. Our work contributes to the management of inter- and intra-subject variability which helps to design accurate and low-cost BCIs.