Optimization of stimulus color for peripheral SSVEP-based brain-computer interfaces

Background: Most existing steady-state visual evoked potential (SSVEP)-based brain-computer interfaces (BCIs) struggle to balance user experience with system performance. Although recent studies have shown that peripheral vision stimulation can evoke SSVEPs with high user comfort, the impact of stimulus color on peripheral SSVEP performance remains underexplored. Therefore, this study attempted to investigate the effect of stimulus color on peripheral SSVEPs. Methods: Four conventional stimulus colors (i.e., blue, green, red, and white) were evaluated using ultra-low frequency SSVEP stimuli, with the stimulation frequencies ranging from 2 Hz to 3.32 Hz. Based on the results, the optimized stimulus color was used to build a 12-target peripheral SSVEP-based BCI. Task-discriminant component analysis (TDCA) algorithm was adopted to detect SSVEPs. The feasibility of the proposed system was verified through offline experiments with 13 participants and online experiments with 11 participants. Results: The offline experiments with 13 participants showed no significant differences in classification accuracy and information transfer rates (ITRs) among the four-color paradigms. However, green stimulation received the highest subjective comfort ratings. Consequently, green stimulation was selected for building the 12-target peripheral SSVEP-based BCI. The online results achieved a mean classification accuracy of 89.93 ± 6.10% and an ITR of 47.96 ± 6.98 bits/min. Conclusion: The present findings support a comfort-driven color selection strategy for peripheral ultra-low-frequency SSVEP stimulation while maintaining comparable performance among the tested colors. These findings may provide practical guidance for more visually tolerable SSVEP-based BCI systems based on peripheral visual stimulation.