Multi-class motor imagery EEG classification method with high accuracy and low individual differences based on hybrid neural network

Abstract

Objective. Most current methods of classifying different patterns for motor imagery EEG signals require complex pre-processing and feature extraction steps, which consume time and lack adaptability, ignoring individual differences in EEG signals. It is essential to improve algorithm performance with the increased classes and diversity of subjects. Approach. This study introduces deep learning method for end-to-end learning to complete the classification of four-class MI tasks, aiming to improve the recognition rate and balance the classification accuracy among different subjects. A new one-dimensional input data representation method is proposed. This representation method can increase the number of samples and ignore the influence of channel correlation. In addition, a cascade network of convolutional neural network and gated recurrent unit is designed to learn time-frequency information from EEG data without extracting features manually, this model can capture the hidden representations related to different MI mode of each people. Main results. Experiments on BCI Competition 2a dataset and actual collected dataset achieve high accuracy near 99.40% and 92.56%, and the standard deviation is 0.34 and 1.35 respectively. Results demonstrate that the proposed method outperforms the advanced methods and baseline models. Significance. Experimental results show that the proposed method improves the accuracy of multi-classification and overcomes the impact of individual differences on classification by training neural network subject-dependent, which promotes the development of actual brain-computer interface systems.

Classification and recognition of left and right hand motion imagination based on CNN-LSTM

“…The CNN-LSTM network model uses 1D-CNN to extract the features of motor imaging EEG signals, motor imagination EEG as input to one-dimensional CNNs, and the data input dimension needs to become one-dimensional [11]. EEG signals obtained by multichannel EEG acquisition devices can be expressed as:…”

EEG-ITNet: An Explainable Inception Temporal Convolutional Network for Motor Imagery Classification

“…Although MI-EEGNET performed better than previous architectures, the high number of trainable parameters made it challenging to interpret. Besides these CNN architectures investigated for EEG signal classification, several studies can be found in the literature that used recurrent neural network (RNN) and its variants [21][22][23], such as LSTM and gated recurrent units (GRU), for the classification of mental tasks based on EEG signals [24,25]. However, RNNs are less prevalent in this area due to their exploding/vanishing gradient or lack of memory problems [26].…”

Stability of mental motor-imagery classification in EEG depends on the choice of classifier model and experiment design, but not on signal preprocessing

“…We then focus on the machine-learning methods used to detect patterns in the extracted EEG signal features. There are recent approaches using deep learning algorithms (Craik et al, 2019 ; Liu et al, 2021 ; Zhang et al, 2021 ; Lomelin-Ibarra et al, 2022 ). However, in biomedical applications with many features to take into account, still linear classifiers (e.g., SVM) or KNN classifiers are widely used (Mahmoudi and Shamsi, 2018 ).…”