Application of adaptive noise cancellation with neural-network-based fuzzy inference system for visual evoked potentials estimation.

This paper presents an application of adaptive noise cancellation with neural-network-based fuzzy inference system (NNFIS) for rapid estimation of visual evoked potentials (VEPs). Usually a recorded VEP is severely contaminated by background ongoing activities of the spontaneous EEG signal in the human brain. Many approaches have been adopted to enhance the signal-to-noise ratio (SNR) of the recorded signal. However, nonlinear dynamic methods are rarely investigated in view of their complexity, and the fact that the nonlinear characteristics of the signal are hard to determine in general. An adaptive noise cancellation method with NNFIS was carefully designed to estimate the VEP signal. NNFIS, based on Takagi and Sugeno's fuzzy model, has the advantage of being linear-in-parameter; thus the conventional adaptive methods can be efficiently utilized to estimate its parameters. Another advantage of NNFIS lies in that it can track the dynamic behavior of VEP in a real-time fashion because the VEP variation tracking is important for critical patient monitoring in the clinical situation. A series of computer experiments conducted on simulated and real-test responses have confirmed the superiority of the method developed in this paper.