An Asynchronous Training-Free SSVEP-BCI Detection Algorithm for Non-Equal Prior Probability Scenarios

• Published in: IEEE transactions on neural systems and rehabilitation engineering Vol. 32; pp. 4120 - 4130
• Main Authors: Wang, Junsong, Cui, Yuntian, Zhang, Hongxin, Wu, Haolin, Yang, Chen
• Format: Journal Article
• Language: English
• Published: United States IEEE 2024
• Subjects: Adult; Algorithms; asynchronous; Background noise; Brain modeling; Brain-computer interface (BCI); Brain-Computer Interfaces; Detection Algorithms; Electroencephalography; Electroencephalography - methods; Evoked Potentials, Visual - physiology; Female; Humans; Male; Mathematical models; maximum a posteriori (MAP); mutual information rate (MIR); Photic Stimulation; Posterior probability; Probability; Probability distribution; Reproducibility of Results; Sensitivity and Specificity; Signal-To-Noise Ratio; steady-state visual evoked potential (SSVEP); Target recognition; Training; Visualization; Young Adult
• ISSN: 1534-4320; 1558-0210; 1558-0210
• DOI: 10.1109/TNSRE.2024.3496727

SSVEP-based brain-computer interface (BCI) systems have received a lot of attention due to their relatively high Signal to Noise Ratio (SNR) and less training requirements. Most of the existing steady-state visual evoked potential (SSVEP) detection algorithms treat the prior probability of each alternative target being selected as equal. In this study, the prior probability distribution of alternative targets was introduced into the SSVEP recognition algorithm, and an asynchronous training-free SSVEP-BCI detection algorithm for non-equal prior probability scenarios was proposed. This algorithm is based on the Spatio-temporal equalization multi-window technique (STE-MW) and introduces the Maximum A Posteriori criterion (MAP), which makes full use of prior information to improve the performance of the asynchronous training-free BCI system. In addition, we proposed a mutual information-based performance evaluation metric called Mutual information rate (MIR) specifically for non-equal prior probability scenarios. This evaluation framework is designed to provide a more accurate estimation of the information transmission performance of BCI systems in such scenarios. A 10-target simulated vehicle control offline experiment involving 17 subjects showed that the proposed method improved the average MIR by 6.48%. Online free control experiments involving 12 subjects showed that the proposed method improved the average MIR by 14.93%, and significantly reduced the average instruction time. The proposed algorithm is more suitable for practical engineering application scenarios that are asynchronous and training-free; the extremely high accuracy is guaranteed while maintaining a low false alarm rate, which can be applied to asynchronous BCI systems that require high stability.