Electroencephalogram-Based Motor Imagery Signals Classification Using a Multi-Branch Convolutional Neural Network Model with Attention Blocks

Altuwaijri, Ghadir Ali; Muhammad, Ghulam

doi:10.3390/bioengineering9070323

Cited by 16 publications

(4 citation statements)

References 47 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Jiang et al improved the performance of their CNN model in the EEG-based emotion recognition task by incorporating the temporal-channel attention mechanism into the designed deep model [41]. Altuwaijri and Muhammad improved the performance of their CNN model by adding CBAM structure to multi-branch EEGNet through attention mechanism and fusion methods for EEG-based motor imagery classification [42]. Notably, the proposed attention-enhanced models demonstrated versatility in leveraging different EEG descriptions that consider time, frequency, and spatial information (sensor locations) interchangeably or in conjunction.…”

Section: Discussionmentioning

confidence: 99%

Assessment of Attention-based Deep Learning Architectures for Classifying EEG in ADHD and Typical Children

Han,

Jin,

2024

IJACSA

View full text Add to dashboard Cite

Although limited research has explored the integration of electroencephalography (EEG) and deep learning approaches for attention deficit hyperactivity disorder (ADHD) detection, using deep learning models for actual data, including EEGs, remains a difficult endeavour. The purpose of this work was to evaluate how different attention processes affected the performance of well-established deep-learning models for the identification of ADHD. Two specific architectures, namely long short-term memory (LSTM)+ attention (Att) and convolutional neural network (CNN)s+Att, were compared. The CNN+Att model consists of a dropout, an LSTM layer, a dense layer, and a CNN layer merged with the convolutional block attention module (CBAM) structure. On top of the first LSTM layer, an extra LSTM layer, including T LSTM cells, was added for the LSTM+Att model. The information from this stacked LSTM structure was then passed to a dense layer, which, in turn, was connected to the classification layer, which comprised two neurons. Experimental results showed that the best classification result was achieved using the LSTM+Att model with 98.91% accuracy, 99.87% accuracy, 97.79% specificity and 98.87% F1score. After that, the LSTM, CNN+Att, and CNN models succeeded in classifying ADHD and Normal EEG signals with 98.45%, 97.74% and 97.16% accuracy, respectively. The information in the data was successfully utilized by investigating the application of attention mechanisms and the precise position of the attention layer inside the deep learning model. This fascinating finding creates opportunities for more study on largescale EEG datasets and more reliable information extraction from massive data sets, ultimately allowing links to be made between brain activity and specific behaviours or task execution.

show abstract

Section: Discussionmentioning

confidence: 99%

Assessment of Attention-based Deep Learning Architectures for Classifying EEG in ADHD and Typical Children

Han,

Jin,

2024

IJACSA

View full text Add to dashboard Cite

show abstract

“…In an alternative scenario, the BCI IV 2a and HGD1 datasets were employed without the inclusion of pre-processing techniques. However, the classifier MPEEGCBAM 2 demonstrated accuracies of 82.85% and 95.45% for the respective datasets [24]. In the BCI competition IV dataset, the pre-processing steps consisted of applying the CAR technique and a bandpass filter.…”

Section: Literature Reviewmentioning

confidence: 99%

Robust CNN architecture for classification of reach and grasp actions from neural correlates: an edge device perspective

Sultan,

Ijaz,

Waris

et al. 2023

Meas. Sci. Technol.

View full text Add to dashboard Cite

Brain–computer interfaces (BCIs) systems traditionally use machine learning (ML) algorithms that require extensive signal processing and feature extraction. Deep learning (DL)-based convolutional neural networks (CNNs) recently achieved state-of-the-art electroencephalogram (EEG) signal classification accuracy. CNN models are complex and computationally intensive, making them difficult to port to edge devices for mobile and efficient BCI systems. For addressing the problem, a lightweight CNN architecture for efficient EEG signal classification is proposed. In the proposed model, a combination of a convolution layer for spatial feature extraction from the signal and a separable convolution layer to extract spatial features from each channel. For evaluation, the performance of the proposed model along with the other three models from the literature referred to as EEGNet, DeepConvNet, and EffNet on two different embedded devices, the Nvidia Jetson Xavier NX and Jetson Nano. The results of the Multivariant 2-way ANOVA (MANOVA) show a significant difference between the accuracies of ML and the proposed model. In a comparison of DL models, the proposed models, EEGNet, DeepConvNet, and EffNet, achieved 92.44 ± 4.30, 90.76 ± 4.06, 92.89 ± 4.23, and 81.69 ± 4.22 average accuracy with standard deviation, respectively. In terms of inference time, the proposed model performs better as compared to other models on both the Nvidia Jetson Xavier NX and Jetson Nano, achieving 1.9 sec and 16.1 sec, respectively. In the case of power consumption, the proposed model shows significant values on MANOVA (p < 0.05) on Jetson Nano and Xavier. Results show that the proposed model provides improved classification results with less power consumption and inference time on embedded platforms.

show abstract

“…For example, Xie et al ( 2022 ) have proposed a novel approach that utilizes multi-head self-attention combined with position embedding to enhance the classification performance of EEG on the Physionet dataset, achieving an accuracy of 68.54%. Furthermore, Altuwaijri and Muhammad ( 2022 ) have employed channel attention and spatial attention mechanisms to capture temporal and spatial features from EEG signals on the BCI-2a dataset, resulting in an accuracy of 83.63%. However, these methods lack comprehensive integration of multi-scale spatiotemporal features and also neglect adaptive attention selection for global features (Al-Saegh et al, 2021 ; Altaheri et al, 2021 ).…”

Section: Introductionmentioning

confidence: 99%

An improved model using convolutional sliding window-attention network for motor imagery EEG classification

Huang,

Zheng,

et al. 2023

Front. Neurosci.

View full text Add to dashboard Cite

IntroductionThe classification model of motor imagery-based electroencephalogram (MI-EEG) is a new human-computer interface pattern and a new neural rehabilitation assessment method for diseases such as Parkinson's and stroke. However, existing MI-EEG models often suffer from insufficient richness of spatiotemporal feature extraction, learning ability, and dynamic selection ability.MethodsTo solve these problems, this work proposed a convolutional sliding window-attention network (CSANet) model composed of novel spatiotemporal convolution, sliding window, and two-stage attention blocks.ResultsThe model outperformed existing state-of-the-art (SOTA) models in within- and between-individual classification tasks on commonly used MI-EEG datasets BCI-2a and Physionet MI-EEG, with classification accuracies improved by 4.22 and 2.02%, respectively.DiscussionThe experimental results also demonstrated that the proposed type token, sliding window, and local and global multi-head self-attention mechanisms can significantly improve the model's ability to construct, learn, and adaptively select multi-scale spatiotemporal features in MI-EEG signals, and accurately identify electroencephalogram signals in the unilateral motor area. This work provided a novel and accurate classification model for MI-EEG brain-computer interface tasks and proposed a feasible neural rehabilitation assessment scheme based on the model, which could promote the further development and application of MI-EEG methods in neural rehabilitation.

show abstract

Electroencephalogram-Based Motor Imagery Signals Classification Using a Multi-Branch Convolutional Neural Network Model with Attention Blocks

Cited by 16 publications

References 47 publications

Assessment of Attention-based Deep Learning Architectures for Classifying EEG in ADHD and Typical Children

Assessment of Attention-based Deep Learning Architectures for Classifying EEG in ADHD and Typical Children

Robust CNN architecture for classification of reach and grasp actions from neural correlates: an edge device perspective

An improved model using convolutional sliding window-attention network for motor imagery EEG classification

Contact Info

Product

Resources

About