基于模糊神经网络的自适应噪声消除方法在视觉诱发电位提取中的应用

Application of Adaptive Noise Cancellation with Adaptive-network-based Fuzzy Inference System for Esti- mation of VEP

目的本文旨在提取淹没在强背景噪声自发脑电信号(EEG)中的视觉诱发脑电信号。方法通过设计合适的自适应模糊神经网络(ANFIS),对视觉诱发脑电信号进行建模,从而采用自适应噪声消除方法滤除干扰信号,提取出视觉诱发脑电信号。结果经与目前临床通用的相干平均法比较,在波形整体和P100潜伏期的提取上,效果显著。结论基于ANFIS的自适应噪声消除方法可有效的用于诱发脑电信号VEP的检测。

Objective To estimate visual evoked potentials (VEP) from large background noise of e-lectroencephalogram (EEG). Method The adaptive-network-based fuzzy inference system (ANFIS) was carefully designed to model the VEP signal, the adaptive noise cancellation with ANFIS was applied for estimating VEPs. Result A series of computer experiments conducted on simulated and real-test responses have confirmed the superiority of the method developed in this paper. Conclusion The adaptive noise cancellation method with ANFIS can be efficiently utilized to estimate VEPs.