A novel EEG channel selection and classification methodology for multi‐class motor imagery‐based BCI system design

Abstract

Multi-class MI EEG analysis is an extensively used paradigm in BCI. However, multiple EEG channels lead to redundant information extraction and would reduce the distinction among various MI tasks. Therefore, optimal channel selection from multi-channel EEG activity still remains a challenging task. In this study, to enhance the multi-class BCI system's performance, a novel channel selection, and features optimization methodology have been proposed.First, multi-channel EEG dataset is reduced to an optimum no. of channels subset using developed MDA-SOGWO based EEG channel selection criterion.After that, the discriminable feature-set is generated from time, frequency, and FAWT based time-frequency domain of EEG dataset. Then, an informative feature-set is constructed from extracted features through CCA-RFE based feature selection criterion. Finally, training and validation of selected feature-set are carried out using ELM, LDA, RF, and MLP classifiers. The proposed methodology is evaluated on multi-class MI datasets of BCI Competition IV 2a and BCI Competition III 3a, yielding classification accuracy of 85.65% and 95.15%, respectively. The classification results pose the proposed methodology as an efficient method for future BCI system design.

Optimal Channel Selection of Multiclass Motor Imagery Classification Based on Fusion Convolutional Neural Network with Attention Blocks

“…Further clarification can be found in Figure 8, which illustrates the distribution of these optimal channels, which were selected from 22 main electrodes following the International 10-20 system. Subject 4 in Testing [2,6,7,10,12,15,17,18,19,20,21,22] 12 61.22% Subject 5 in Testing [1,2,3,5,8,9,13,14,15,16,17,20,21] 13 57.81% Subject 6 in Testing [3,4,5,7,8,10,12,15,17 While no identical electrodes were selected, the results indicate that some electrodes were used more frequently than others. Figure 9 shows the average number of electrodes chosen for all subjects according to the different classification processes applied in Table 10.…”

“…Different methods for selecting channels have been used with the BCI IV 2a dataset, specifically when dealing with four-class classification. Researchers in [18][19][20][21][22] employ three main classification techniques: one-vs-one, one-vs-rest, and multiclass classification. In both one-vs-one and one-vs-rest, the means are derived from multiple binary classifications.…”

“…The application of the Firefly algorithm aimed at reducing the number of channels involved in calculating weighted scores for each channel near a candidate solution. Using both one-vsone and one-vs-rest approaches, the researchers in [20] utilized various ML techniques to compute the means of several binary classifications. They demonstrated the advantages of incorporating feature extraction, feature selection, and MDA-SOGWO channel selection to enhance classification accuracy, elevating it from 67.04% to 80.82%.…”

“…Based on the channel selection process, when selecting the optimal channels, it is possible to improve the classification results while ensuring that no significant channels are removed [3]. Due to this, the majority of researchers [18][19][20][21][22] selected different channels for each subject individually. Selecting a fixed channel for all subjects is a problem that has not previously been addressed.…”

Energy-Efficient Deep Neural Networks for EEG Signal Noise Reduction in Next-Generation Green Wireless Networks and Industrial IoT Applications

“…Two primary modes, bipolar and unipolar, exist for the recording of scalp EEG signals. Filtering approach: Known for its speed and classifier independence, it often requires additional refinement for accuracy [37].…”