Visual Stimuli for P300-Based Brain-Computer Interfaces: Color, Shape, and Mobility

Abstract

The purpose of this study was to identify the impact of different discriminative features of stimuli in a P300 brain-computer interface paradigm on overall performance and evoked potentials. It has been shown that stimuli sets with a greater number of discriminative features yield better target selection accuracy. Target selection accuracy was significantly higher for the stimuli that differ from each other by color, shape, and semantics. Highest performance was achieved with the stimuli set containing the largest number of discriminative features, namely a set of nine different-colored letters. This result is mainly due to higher mean P300 peak amplitude for stimuli sets that contain more discriminative features. The results of the study can be used for designing a better user experience in brain-computer interfacing (BCI). Motion of the stimuli presentation point and characteristics of this motion (linear or pseudorandom) did not have any impact on BCI performance. This result is promising for future BCI designs with rapid serial visual presentation using mobile robots or augmented reality as stimuli presentation environment.