A Channel Attention Based MLP-Mixer Network for Motor Imagery Decoding With EEG

He, Yanbin; Lu, Zhiyang; Wang, Jun

doi:10.1109/icassp43922.2022.9747488

Cited by 11 publications

(4 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…With the rapid advancement of deep learning, methods including but not limited to convolution neural networks (CNNs) [6]- [8], recurrent neural networks (RNNs) [9]- [11], multilayer perceptrons (MLPs) [12], [13], and graph convolution networks (GCNs) [14]- [20] have demonstrated marked effectiveness in EEG classification tasks across Manuscript received September 12, 2023; This study was funded by the Key Program of National Natural Science Foundation of China (grant no. 61731003) and the Funds for National Natural Science Foundation of China (grant no.…”

Section: Introductionmentioning

confidence: 99%

3D Visual Discomfort Assessment With a Weakly Supervised Graph Convolution Neural Network Based on Inaccurately Labeled EEG

Lu,

Zhao,

Yao

2024

IEEE Trans. Neural Syst. Rehabil. Eng.

View full text Add to dashboard Cite

Visual discomfort significantly limits the broader application of stereoscopic display technology. Hence, the accurate assessment of stereoscopic visual discomfort is a crucial topic in this field. Electroencephalography (EEG) data, which can reflect changes in brain activity, have received increasing attention in objective assessment research. However, inaccurately labeled data, resulting from the presence of individual differences, restrict the effectiveness of the widely used supervised learning methods in visual discomfort assessment tasks. Simultaneously, visual discomfort assessment methods should pay greater attention to the information provided by the visual cortical areas of the brain. To tackle these challenges, we need to consider two key aspects: maximizing the utilization of inaccurately labeled data for enhanced learning and integrating information from the brain's visual cortex for feature representation purposes. Therefore, we propose the weakly supervised graph convolution neural network for visual discomfort (WSGCN-VD). In the classification part, a center correction loss serves as a weakly supervised loss, employing a progressive selection strategy to identify accurately labeled data while constraining the involvement of inaccurately labeled data that are influenced by individual differences during the model learning process. In the feature extraction part, a feature graph module pays particular attention to the construction of spatial connections among the channels in the visual regions of the brain and combines them with high-dimensional temporal features to obtain visually dependent spatio-temporal representations. Through extensive experiments conducted in various scenarios, we demonstrate the effectiveness of our proposed model. Further analysis reveals that the proposed model mitigates the impact of inaccurately labeled data on the accuracy of assessment.

show abstract

Section: Introductionmentioning

confidence: 99%

3D Visual Discomfort Assessment With a Weakly Supervised Graph Convolution Neural Network Based on Inaccurately Labeled EEG

Lu,

Zhao,

Yao

2024

IEEE Trans. Neural Syst. Rehabil. Eng.

View full text Add to dashboard Cite

show abstract

“…EEGNet and ShallowConvNet utilize convolutional layers to extract spatial and temporal patterns from EEG data. However, EEGNet may need help with capturing long-range temporal dependencies [43], while ShallowConvNet may not be as effective as deeper architectures in capturing complex patterns. DeepConvNet excels at capturing spatial and temporal patterns but requires much training data to avoid overfitting [44].…”

Section: Introductionmentioning

confidence: 99%

Kernel-Based Regularized EEGNet Using Centered Alignment and Gaussian Connectivity for Motor Imagery Discrimination

Tobón-Henao

Álvarez-Meza

Castellanos-Domínguez

2023

Computers

View full text Add to dashboard Cite

Brain–computer interfaces (BCIs) from electroencephalography (EEG) provide a practical approach to support human–technology interaction. In particular, motor imagery (MI) is a widely used BCI paradigm that guides the mental trial of motor tasks without physical movement. Here, we present a deep learning methodology, named kernel-based regularized EEGNet (KREEGNet), leveled on centered kernel alignment and Gaussian functional connectivity, explicitly designed for EEG-based MI classification. The approach proactively tackles the challenge of intrasubject variability brought on by noisy EEG records and the lack of spatial interpretability within end-to-end frameworks applied for MI classification. KREEGNet is a refinement of the widely accepted EEGNet architecture, featuring an additional kernel-based layer for regularized Gaussian functional connectivity estimation based on CKA. The superiority of KREEGNet is evidenced by our experimental results from binary and multiclass MI classification databases, outperforming the baseline EEGNet and other state-of-the-art methods. Further exploration of our model’s interpretability is conducted at individual and group levels, utilizing classification performance measures and pruned functional connectivities. Our approach is a suitable alternative for interpretable end-to-end EEG-BCI based on deep learning.

show abstract

“…Hou and Jia et al [6] applied long short-term memory (LSTM) to extract features and used graph convolution networks (GCN) to model topological structure. He et al [7] applied channel attention to multi-layer perceptron to adaptively learn the importance of each channel. Specific to the task of EEG emotion recognition, many researchers have also proposed some approaches.…”

Section: Introductionmentioning

confidence: 99%

GTSception: A Deep Learning EEG Emotion Recognition Model Based on Fusion of Global, Time Domain and Frequency Domain Feature Extraction

Jian

Zhang

Qiu

et al. 2021

Preprint

View full text Add to dashboard Cite

With the rapid development of deep learning in recent years, automatic electroencephalography (EEG) emotion recognition has been widely concerned. At present, most deep learning methods do not normalize EEG data properly and do not fully extract the features of time and frequency domain, which will affect the accuracy of EEG emotion recognition. To solve these problems, we propose GTScepeion, a deep learning EEG emotion recognition model. In pre-processing, the EEG time slicing data including channels were pre-processed. In our model, global convolution kernels are used to extract overall semantic features, followed by three kinds of temporal convolution kernels representing different emotional periods, followed by two kinds of spatial convolution kernels highlighting brain hemispheric differences to extract spatial features, and finally emotions are dichotomy classified by the full connected layer. The experiments is based on the DEAP dataset, and our model can effectively normalize the data and fully extract features. For Arousal, ours is 8.76% higher than the current optimal emotion recognition model based on Inception. For Valence, the best accuracy of our model reaches 91.51%.

show abstract

A Channel Attention Based MLP-Mixer Network for Motor Imagery Decoding With EEG

Cited by 11 publications

References 15 publications

3D Visual Discomfort Assessment With a Weakly Supervised Graph Convolution Neural Network Based on Inaccurately Labeled EEG

3D Visual Discomfort Assessment With a Weakly Supervised Graph Convolution Neural Network Based on Inaccurately Labeled EEG

Kernel-Based Regularized EEGNet Using Centered Alignment and Gaussian Connectivity for Motor Imagery Discrimination

GTSception: A Deep Learning EEG Emotion Recognition Model Based on Fusion of Global, Time Domain and Frequency Domain Feature Extraction

Contact Info

Product

Resources

About