Removal of EOG and EMG artifacts from EEG using combination of functional link neural network and adaptive neural fuzzy inference system

Abstract

A new way of removing electrooculogram (EOG) and electromyogram (EMG) artifacts from an electroencephalogram (EEG) was developed that involves combining an adaptive neural fuzzy inference system (ANFIS) and a functional link neural network (FLNN) to construct a filter. The ANFIS is divided into a nonlinear antecedent part and a linear consequent part; and the consequent part is replaced with the FLNN to enhance the nonlinear approximation ability of the method. An adaptive filtering algorithm adjusts the parameters of the fuzzy inference and neural network. Verification of the method showed that it was effective in removing EOG and EMG artifacts from an EEG. A comparison showed that this method provided better performance than an ANFIS, an RBF-ANFIS, or an adaptive FL-BPNN.