Deep learning techniques for classification of electroencephalogram (EEG) motor imagery (MI) signals: a review

Neural Computing and Applications, Journal Year: 2021, Volume and Issue: 35(20), P. 14681 - 14722

Published: Aug. 25, 2021

Language: Английский

---

A comprehensive survey on multimodal medical signals fusion for smart healthcare systems

Information Fusion, Journal Year: 2021, Volume and Issue: 76, P. 355 - 375

Published: July 5, 2021

Language: Английский

---

Physics-Informed Attention Temporal Convolutional Network for EEG-Based Motor Imagery Classification

IEEE Transactions on Industrial Informatics, Journal Year: 2022, Volume and Issue: 19(2), P. 2249 - 2258

Published: Aug. 9, 2022

The brain-computer interface (BCI) is a cutting-edge technology that has the potential to change world. Electroencephalogram (EEG) motor imagery (MI) signal been used extensively in many BCI applications assist disabled people, control devices or environments, and even augment human capabilities. However, limited performance of brain decoding restricting broad growth industry. In this article, we propose an attention-based temporal convolutional network (ATCNet) for EEG-based classification. ATCNet model utilizes multiple techniques boost MI classification with relatively small number parameters. employs scientific machine learning design domain-specific deep interpretable explainable features, multihead self-attention highlight most valuable features MI-EEG data, extract high-level convolutional-based sliding window data efficiently. proposed outperforms current state-of-the-art Competition IV-2a dataset accuracy 85.38% 70.97% subject-dependent subject-independent modes, respectively.

Language: Английский

---

A Multi-Branch Convolutional Neural Network with Squeeze-and-Excitation Attention Blocks for EEG-Based Motor Imagery Signals Classification

Diagnostics, Journal Year: 2022, Volume and Issue: 12(4), P. 995 - 995

Published: April 15, 2022

Electroencephalography-based motor imagery (EEG-MI) classification is a critical component of the brain-computer interface (BCI), which enables people with physical limitations to communicate outside world via assistive technology. Regrettably, EEG decoding challenging because complexity, dynamic nature, and low signal-to-noise ratio signal. Developing an end-to-end architecture capable correctly extracting data's high-level features remains difficulty. This study introduces new model for MI known as Multi-Branch EEGNet squeeze-and-excitation blocks (MBEEGSE). By clearly specifying channel interdependencies, multi-branch CNN attention employed adaptively change channel-wise feature responses. When compared existing state-of-the-art models, suggested achieves good accuracy (82.87%) reduced parameters in BCI-IV2a dataset (96.15%) high gamma dataset.

Language: Английский

---

Explainable AI-driven IoMT fusion: Unravelling techniques, opportunities, and challenges with Explainable AI in healthcare

Information Fusion, Journal Year: 2024, Volume and Issue: 110, P. 102472 - 102472

Published: May 16, 2024

Language: Английский

---

Attention-Inception and Long- Short-Term Memory-Based Electroencephalography Classification for Motor Imagery Tasks in Rehabilitation

IEEE Transactions on Industrial Informatics, Journal Year: 2021, Volume and Issue: 18(8), P. 5412 - 5421

Published: Dec. 3, 2021

In recent years, the contributions of deep learning have had a phenomenal impact on electroencephalography-based brain-computer interfaces. While decoding accuracy electroencephalography signals has continued to increase, process caused models continuously expand in terms size and computational resource requirements. However, due their increased requirements, it become difficult embed, store, execute for artificial intelligence things, cloud-based, or edge devices used rehabilitation. Hence, this article proposes novel learning-based lightweight model based attention-inception convolutional neural network long- short-term memory. The proposed achieves excellent public competition datasets while requiring few parameters low time. Using BCI IV 2a dataset high gamma dataset, achieved 82.8% 97.1% accuracies, respectively.

Language: Английский

---

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

IEEE Access, Journal Year: 2022, Volume and Issue: 10, P. 36672 - 36685

Published: Jan. 1, 2022

In recent years, neural networks and especially deep architectures have received substantial attention for EEG signal analysis in the field of brain-computer interfaces (BCIs). this ongoing research area, end-to-end models are more favoured than traditional approaches requiring transformation pre-classification. They can eliminate need prior information from experts extraction handcrafted features. However, although several learning algorithms been already proposed literature, achieving high accuracies classifying motor movements or mental tasks, they often face a lack interpretability therefore not quite by neuroscience community. The reasons behind issue be number parameters sensitivity to capture tiny yet unrelated discriminative We propose an architecture called EEG-ITNet comprehensible method visualise network learned patterns. Using inception modules causal convolutions with dilation, our model extract rich spectral, spatial, temporal multi-channel signals less complexity (in terms trainable parameters) other existing architectures, such as EEG-Inception EEG-TCNet. By exhaustive evaluation on dataset 2a BCI competition IV OpenBMI imagery dataset, shows up 5.9\% improvement classification accuracy different scenarios statistical significance compared its competitors. also comprehensively explain support validity illustration neuroscientific perspective. made code open at https://github.com/AbbasSalami/EEG-ITNet

Language: Английский

---

A compact multi-branch 1D convolutional neural network for EEG-based motor imagery classification

Biomedical Signal Processing and Control, Journal Year: 2022, Volume and Issue: 81, P. 104456 - 104456

Published: Dec. 15, 2022

Language: Английский

---

A Multibranch of Convolutional Neural Network Models for Electroencephalogram-Based Motor Imagery Classification

Biosensors, Journal Year: 2022, Volume and Issue: 12(1), P. 22 - 22

Published: Jan. 3, 2022

Automatic high-level feature extraction has become a possibility with the advancement of deep learning, and it been used to optimize efficiency. Recently, classification methods for Convolutional Neural Network (CNN)-based electroencephalography (EEG) motor imagery have proposed, achieved reasonably high accuracy. These approaches, however, use CNN single convolution scale, whereas best scale varies from subject subject. This limits precision classification. paper proposes multibranch models address this issue by effectively extracting spatial temporal features raw EEG data, where branches correspond different filter kernel sizes. The proposed method’s promising performance is demonstrated experimental results on two public datasets, BCI Competition IV 2a dataset High Gamma Dataset (HGD). technique show 9.61% improvement in accuracy EEGNet (MBEEGNet) fixed one-branch model, 2.95% variable model. In addition, ShallowConvNet (MBShallowConvNet) improved single-scale network 6.84%. outperformed other state-of-the-art methods.

Language: Английский

---

Dynamic Convolution With Multilevel Attention for EEG-Based Motor Imagery Decoding

IEEE Internet of Things Journal, Journal Year: 2023, Volume and Issue: 10(21), P. 18579 - 18588

Published: June 1, 2023

Brain–computer interface (BCI) is an innovative technology that utilizes artificial intelligence (AI) and wearable electroencephalography (EEG) sensors to decode brain signals enhance the quality of life. EEG-based motor imagery (MI) signal used in many BCI applications, including smart healthcare, homes, robotics control. However, restricted ability a major factor preventing from expanding significantly. In this study, we introduce dynamic attention temporal convolutional network (D-ATCNet) for decoding MI signals. The D-ATCNet model uses convolution (Dy-conv) multilevel performance classification with relatively small number parameters. has two main blocks: 1) 2) convolution. Dy-conv encode low-level MI-EEG information shifted window self-attention extract high-level encoded signal. proposed performs better than existing methods accuracy 71.3% subject independent 87.08% dependent using competition IV-2a data set.

Language: Английский

---