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

Joharah Khabti,Saad AlAhmadi,Adel Soudani

DOI: https://doi.org/10.3390/s24103168

IF: 3.9

2024-05-17

Sensors

Abstract:The widely adopted paradigm in brain–computer interfaces (BCIs) involves motor imagery (MI), enabling improved communication between humans and machines. EEG signals derived from MI present several challenges due to their inherent characteristics, which lead to a complex process of classifying and finding the potential tasks of a specific participant. Another issue is that BCI systems can result in noisy data and redundant channels, which in turn can lead to increased equipment and computational costs. To address these problems, the optimal channel selection of a multiclass MI classification based on a Fusion convolutional neural network with Attention blocks (FCNNA) is proposed. In this study, we developed a CNN model consisting of layers of convolutional blocks with multiple spatial and temporal filters. These filters are designed specifically to capture the distribution and relationships of signal features across different electrode locations, as well as to analyze the evolution of these features over time. Following these layers, a Convolutional Block Attention Module (CBAM) is used to, further, enhance EEG signal feature extraction. In the process of channel selection, the genetic algorithm is used to select the optimal set of channels using a new technique to deliver fixed as well as variable channels for all participants. The proposed methodology is validated showing 6.41% improvement in multiclass classification compared to most baseline models. Notably, we achieved the highest results of 93.09% for binary classes involving left-hand and right-hand movements. In addition, the cross-subject strategy for multiclass classification yielded an impressive accuracy of 68.87%. Following channel selection, multiclass classification accuracy was enhanced, reaching 84.53%. Overall, our experiments illustrated the efficiency of the proposed EEG MI model in both channel selection and classification, showing superior results with either a full channel set or a reduced number of channels.

engineering, electrical & electronic,instruments & instrumentation,chemistry, analytical

What problem does this paper attempt to address?

### What problem does this paper attempt to solve?

This paper primarily aims to address the multi-class classification problem based on motor imagery (MI) in brain-computer interface (BCI) systems. Specifically, the research team proposes a new method to improve the efficiency and accuracy of MI task detection by combining channel selection and classification techniques.

#### Main Issues:
1. **EEG Signal Complexity**: The inherent characteristics of EEG signals make the classification process complex, especially when dealing with potential tasks specific to individual participants.
2. **Noisy Data and Redundant Channels**: BCI systems may generate noisy data and redundant channels, which increase device and computational costs.
3. **Fixed Channel Selection**: Most existing studies select different channels for each participant individually, whereas this paper aims to find a fixed optimal set of channels for all participants.

#### Proposed Method:
- **Fusion of Convolutional Neural Network and Attention Module (FCNNA)**: A CNN model with multiple convolutional blocks is constructed, and an attention module (CBAM) is used to further enhance EEG signal feature extraction.
- **Genetic Algorithm for Channel Selection**: A genetic algorithm (GA) is used to select the optimal set of channels, introducing a new method that provides both fixed and variable channels suitable for all participants.
- **Validation Experiments**: Experiments conducted on public datasets validate the significant improvements of this method in both multi-class and binary classification, achieving a highest accuracy of 93.09% in the binary classification of left-hand and right-hand movements.

Through the above methods, this paper addresses the issues of signal complexity and redundant channels present in traditional methods and improves classification performance.