Citation:
A. Ravi, Pearce, S. , Zhang, X. , and Jiang, N. , “User-Specific Channel Selection Method to Improve SSVEP BCI Decoding Robustness Against Variable Inter-Stimulus Distance”, in 2019 IEEE EMBS Neural Engineering (NER) Conference, 2019.
Abstract:
Steady-state visual evoked potentials (SSVEP) are responses elicited when a user is presented with a repetitive visual stimulus. Change in stimuli proximity has been shown to have an influence on the performance of SSVEP-based BCI, where the inter-stimulus distance has a positive correlation with the overall performance. This limits the flexibility in stimulus design by imposing a constraint on the acceptable inter-stimulus distance, consequently limiting the range of applicability for SSVEP-based BCIs in real-world applications. Another limitation that needs to be addressed is the required number of EEG channels. In this study, we investigated these two challenges. A process of selecting optimal user-specific channel set was proposed. We demonstrated that the user-specific channel set is more robust against variable inter-stimulus distance. A significant improvement in accuracy (p=10 −3 ) of 5% and a reduction in variation (p=10 −3 ) of 55% was achieved on average when compared to the performance using the classic 3-channel set (O1, O2, Oz) and 6-channel set (O1, O2, Oz, PO3, PO4, POz).
