MLP With Riemannian Covariance for Motor Imagery Based EEG Analysis

Abstract

Stroke is one of the leading causes of disability and incidence. For the treatment of prognosis of stroke patients, Motor imagery (MI) as a novel experimental paradigm, clinically it is effective because MI based Brain-Computer interface system can promote rehabilitation of stroke patients. There is being a hot and challenging topic to recognize multi-class motor imagery action classification accurately based on electroencephalograph (EEG) signals. In this work, we propose a novel framework named MRC-MLP. Multiple Riemannian covariance is used for EEG feature extraction. We make a multi-scale spectral division to filter EEG signals. They consist of different frequency bandwidths name sub-band. We concatenate and vectorize features extracted by Riemannian covariance on each sub-band. We design a fully connected MLP model with an improved loss function for motor imagery EEG classification. Furthermore, our proposed method MRC-MLP outperforms state-of-the-art methods and achieves approximately mean accuracy with 76%.

Classification of motor imagery brain wave for bionic hand movement using multilayer perceptron

“…The results of this study suggest a higher presentation than one obtained by P. Yang et al [22]. What distinguishes the two methods used is the choice of the activation function and the number of recognitions of classification patterns.…”

“…The more pattern recognition is classified, the more difficult and complex it is for the system to classify them. The studies carried out in [22] in which the EEG signal was classified for four motor imagery patterns with the activation function' softmax' showed an accuracy of 76%. According to [23,24,25], a hidden layer consisting of 2 or more can improve the accuracy of the classification system but slow down the training system.…”

Neural Decoding of EEG Signals with Machine Learning: A Systematic Review

“…Moreover, the performance of SVM was compared with that of Fisher linear discriminant, BPNN, and radial basis function neural network (RBF-NN) to achieve optimal performance of classification [ 260 ]. Yang et al [ 261 ] have proposed a new framework based on multiple Riemannian covariances and MLP for feature extraction and classification, respectively. Use of this framework to classify MI EEG signals achieved a mean accuracy of 76%.…”

“… 14 channels Own database PCA KNN SVM LR DT Accuracy = 70.6 [ 167 ] 2018 ER 32 subj. 32 channels DEAP database DWT KNN Accuracy = 87.1 [ 261 ] 2020 MI 9 subj. 22 channels BCI Competition IV MRC MLP Accuracy = 76 [ 277 ] 2011 SS 20 subj.…”

Mixed-Precision Quantization and Parallel Implementation of Multispectral Riemannian Classification for Brain-Machine Interfaces

“…Another effort for embedded BMI has been made by Belwafi et al [26] implementing a CSP-based classifier on a FPGA device. The multispectral and multiscale Riemannian classifiers proposed in [27], [28] outperform both EEGNET and CSP-based models by around 5% and 2% higher accuracy, respectively. However, their proposed models are still very challenging for embedded deployment on low-power resourceconstrained MCUs due to large memory footprint and high computational complexity.…”