Author(s): Liang, LY (Liang, Liyan); Lin, JJ (Lin, Jiajun); Yang, C (Yang, Chen); Wang, YJ (Wang, Yijun); Chen, XG (Chen, Xiaogang); Gao, SK (Gao, Shangkai); Gao, XR (Gao, Xiaorong)

Source: JOURNAL OF NEURAL ENGINEERING Volume: 17 Issue: 4 Article Number: 046026 DOI: 10.1088/1741-2552/abaa9b Published: AUG 2020

Abstract: Objective. The design of the stimulation paradigm plays an important role in steady-state visual evoked potential (SSVEP)-based brain-computer interface (BCI) studies. Among various stimulation designs, the dual-frequency paradigm in which two frequencies are used to encode one target is of importance and interest. However, because the number of possible frequency combinations is huge, the existing dual-frequency modulation paradigms failed to optimize the encoding towards the best combinations. Thus, this work aiming at designing a new dual-frequency and phase modulation paradigm with the best combinations stimuli. Approach. This study proposed a dual-frequency and phase modulation method, which can achieve a large number of targets by making different combinations of two frequencies and an initial phase. This study also designed a set of methods for quickly optimizing the stimulation codes for the dual-frequency and phase modulation method. Main results. An online 40-class BCI experiment with 12 subjects obtained an accuracy of 96.06 +/- 4.00% and an averaged information transfer rate (ITR) of 196.09 +/- 15.25 bits min(-1), which were much higher than the existing dual-frequency modulation paradigms. Moreover, an offline simulation with a public dataset showed that the optimization method was also effective for optimizing the single-frequency and phase modulation paradigm. Significance. These results demonstrate the high performance of the proposed dual-frequency and phase modulation method and the high efficiency of the optimization method for designing SSVEP stimulation paradigms. In addition, the coding efficiency of the optimized dual-frequency and phase modulation paradigm is higher than that of the single-frequency and phase modulation paradigm, and it is expected to further realize the BCI paradigm with a large amount of targets.

Accession Number: WOS:000561554100001

PubMed ID: 32726763

Author Identifiers:

Author        Web of Science ResearcherID        ORCID Number

Chen, Xiaogang                  0000-0002-5334-1728

ISSN: 1741-2560

eISSN: 1741-2552

Full Text: https://iopscience.iop.org/article/10.1088/1741-2552/abaa9b