Feature Extraction and Analysis of EEG Signals based Motor-Movement Imagery using Multiscale Wavelet transform and Adaptive Neuro Fuzzy Inference System (ANFIS) Algorithms

Abstract

The purpose of this paper using two algorithms for analyzing EEG signals for Motor Movement / Imagery of the BCI the placement 10-10 international system is used to recorded data which is related with both hand and feet and was adopted with imagination status, equally important and according to the electrical activity of the brain signals it consider as non-stationary, furthermore in this research I propose to apply Multiscale Wavelet Transform MSWT with four level of the decomposition of the EEG signals analysis as well as the Debauches of the wavelet families was used to compute the coefficients of the Two Dimensional 2D-DWT techniques which applied in this study besides it was superior for a features extraction secondly the ANFIS was applied as classification algorithms to classify the input of the data with split it into training 80% and testing 20% additionally I used 100% of the features for train, however I used five membership with gaussian function with three input is applied finally I conclude the accuracy of the training features was 100% while the performance was 100% for both testing and training.