Deep temporal networks for eeg based motor imagery recognition UKF is applied to the common spatio-spectral pattern Brain–computer interface (BCI) technology converts electroencephalogram (EEG) signals into control commands to help patients with motor disorders, such as stroke and AbstractThe electroencephalogram (EEG) based motor imagery (MI) signal classification, also known as motion recognition, is a highly popular area of research due to its applications in robotics, gaming, and medical fields. ,According to the proposed algorithm, Abstract. Computers in Biology and Medicine, Vol. Human brains manifest diverse responses to visual stimuli, as recorded by EEG, and this has led to the development of applications in various fields like Neuroscience [1], Biometrics [2], neuromarketing [3], etc. Traditionally, for classification tasks based on EEG, researchers would extract features from raw signals manually which is often time consuming and requires adequate domain knowledge. 760979 Motor imagery (MI) is a commonly used brain–computer interface paradigm, and decoding the MI-EEG signals has been an active research area in recent years. A fully connected, layer-based classifier is used to derive classification results. Parikh, D. However, the classification accuracy of CNNs is compromised when target data are distorted. In: Journal of Neuroscience 24 (43), 9674–9680, (2004). Deep temporal features of the EEG signals are extracted using multiple layers of temporal convolutional modules, and the graph convolutional network module is employed to learn the topology of electrode connections, facilitating the decoding of motor imagery EEG signals. 265–272. most signi cant featur es in MI-EEG data, a temporal convolu tional network for high-level temporal fea ture extraction, and a convolu tional-based sliding window for e cient MI-EEG data augmen Introduction During electroencephalography (EEG)-based motor imagery-brain-computer interfaces (MI-BCIs) task, a large number of electrodes are commonly used, and consume much computational resources. León: I am pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. This paper proposes a novel architecture of a deep neural network for EEG-based motor imagery classification that allows us to analyze the learned network weights from a neurophysiological perspective, thus providing an insight into the underlying patterns inherent in motor imagery induced EEG signals. & Bi, X. The EEG signal detects brain activity in various patterns, one popular pattern being motor imagery (MI) (Autthasan et al. 期刊：Nature 主题：MI-EEG-Transformer. Yang et al. Batra, Grad-cam: Visual explanations from deep networks via gradient-based Deep Spatial-Temporal Neural Network for Classification of EEG-Based Motor Imagery Hinton G. 2. Advancements in deep learning (DL) have significantly increased the accuracy of decoding EEG signals for MI-based BCI applications [7, 8, 9], yet several issues still hinder DL models from reaching practical use []. Deep learning for EEG-based Motor Imagery classification: accuracy-cost trade-off . As a challenging topic in brain-computer interface (BCI) research, motor imagery classification In this work we proposed Sinc-EEGNet, a lightweight convolutional neural network for EEG-BCI-based motor imagery classification that learns optimal band decomposition and spatial filtering, mimicking the behavior of the well-known FBCSP but learning the filters directly from the raw EEG data. However, decoding EEG signals poses significant challenges due to their complexity, dynamic nature, and low signal-to A comprehensive review of Deep Learning-based Motor Imagery EEG classification from various perspectives. , Xiao W. Deep spatial-temporal neural network for classification of EEG-based motor imagery. Sun et al. Signal Process. Until recent years, numerous models had been proposed, ranging from classical algorithms like Common Spatial Pattern (CSP) to deep learning models such as convolutional neural networks (CNNs) and transformers. Author Jr. 15:760979. ,According to the proposed algorithm, a five-layer neural network Electroencephalography (EEG) is a non-invasive technique used for recording electrical brain activity with a wide range of applications in cognitive neuroscience morales_time-frequency_2022 , clinical diagnosis smith_eeg_2005 ; loo_clinical_2005 , and brain-computer interfaces mcfarland_eeg-based_2017 ; lotte_review_2018 . , 2008) is used in the data acquisition stage. Deep temporal features of the EEG This paper proposes a model based on deep learning network to decode EEG-based MI actions, combining deep separation convolution network (DSCNN) and bidirectional long short-term memory (BLSTM Electroencephalogram (EEG) [1, 2] is a comprehensive reflection of the physiological activity of cerebral cortex and scalp brain cells, which contains a large amount of physiological and disease information. However, numerous studies have demonstrated that the optimal convolution scale varies across subjects and even within different sessions for Motor imaging EEG signal recognition is an important and challenging research problem in human-computer interaction. Subsequently, the reconstructed EEG signal is used as the input of the proposed deep residual convolutional network. , et al. 2023, 13, 18813. 2017; Ortner et al. proposed a classification framework based on LSTM networks, which uses a one-dimensional aggregation The effect of motor imagery on spinal segmental excitability. Introducing Temporal-FocalNets, a new framework that enhances the decoding of Motor Imagery EEG signals using a temporal focal modulation system inspired by image recognition techniques. The research introduces the time distributed long short-term memory (TD Brain-Computer Interface (BCI) technology has garnered significant attention in recent years for its potential to facilitate direct communication between the brain and machines [1], [2]. Dawwd, J. : Classification of EEG motor imagery tasks utilizing 2D temporal patterns with deep learning. 2021. This dataset is intended for research on brain–computer interfaces (BCIs) based on motor imagery. MI refers to a pattern of brain activity in which the subject imagines moving a specific part of their body (such as the left or right hand, tongue, or foot) without physically moving it. EEG decoding plays a crucial role in the development of motor imagery brain-computer interface. Until recent years, numerous models had been proposed, ranging from classical algorithms like Common Spatial Pattern (CSP) to deep learning models such as Due to raw EEG signals being characterized by low spatial resolution and low signal-to-noise ratios [4], this poses a great challenge in developing effective BCIs. & Cui, W. Qiu S, Du C, et al. Moreover, they introduced two methods for automatically selecting the optimal number of channels. , Zhang S. and attention-based temporal convolutional networks (ATCNet) are characterized by the ability to achieve better classification results using fewer training parameters. , Wojcik G. Unlike the existing deep neural networks in the literature, the proposed network allows us to analyze the learned network weights from a neurophysiological perspective, thus providing an insight into the underlying patterns inherent in motor imagery induced EEG The multi-scale deep convolutional neural networks are introduced to deal with the representation for imagined motor Electroencephalography (EEG) signals through deep learning algorithm, referred to as Deep Motor Features (DeepMF), for brain EEG‑based motor imagery (MI) plays a pivotal role in BCI, enabling the translation of thought into actionable commands (RNN)33–35, temporal convolutional network (TCN)36,37, deep belief In this article, we provide a brief overview of the EEG-based classification of motor imagery activities using machine learning methods. Optimal Deep Learning-Based Recognition for EEG Signal Motor Imagery (ODLR-EEGSM) is a novel approach presented in EEG-Based Motor Imagery Classification with Deep 245 In terms of deep learning methods, Daoud et al. Recently, a lot of efforts have been made to improve MI signal classification using a Deep Learning has grasped great attention for recognition of Electroencephalography. Das, R. This hybrid approach uniquely combines the temporal feature extraction capabilities of Transformers with the spatial dependency modeling power of GCNs. It can be viewed as a psychological rehearsal of motor actions without any actual movement output, primarily associated with the brain’s motor cortex mu (8–12 Hz) and beta (18–26 Hz) EEG signal rhythms. In this paper, we propose a novel architecture of a deep neural network The electroencephalogram (EEG) based motor imagery (MI) signal classification, also known as motion recognition, is a highly popular area of research due to its applications in robotics, gaming, and medical fields. In such a way, users can further manipulate external devices or exchange information with the surroundings (Pereira et al. However, the problem is ill-posed as these signals are highly nonlinear, unpredictable, and noisy, hence A novel decoding framework for motor imagery EEG signals, known as the Spatial Filter Temporal Graph Convolutional Network (SF-TGCN), is proposed in this study. Improving EEG-Based Motor Imagery Classification via Spatial and Temporal Recurrent Neural Networks Annu Int Conf IEEE Eng Med Biol Soc. However, most current methods face two main issues: (1) They usually rely on convolutional neural networks to extract temporal features of MI signals without fully considering the brain’s functional connectivity during MI tasks. The electroencephalogram (EEG) based motor imagery (MI) signal Motor imagery (MI) classification is key for brain-computer interfaces (BCIs). However, traditional intention-recognition methods face several challenges, such as prolonged training times and limited cross-subject adaptability, which restrict their practical application. Unlike the existing deep neural networks in the literature, the proposed This paper presents the unscented Kalman filter UKF to the BCI signal processing to classify the EEG-based motor imagery signals. Neurocomputing, Vol. , 2019; In the past decade, Electroencephalogram (EEG) has been applied in many fields, such as Motor Imagery (MI) and Emotion Recognition. By leveraging the unique characteristics of task-related brain signals, this system facilitates enhanced communication with these devices. Specifically, researchers have focused on the recognition of motor imagery (MI) based on EEG and translating brain activities into specific motor intentions. The area of brain-computer interface research is widely spreading as it has a diverse array of potential applications. EEG-based BCI for real-world applications has recently attracted increasing interest, including robots [4] and mind-controlled wheelchairs [5]. e. However, the problem is ill-posed as these signals are non-stationary and noisy. , Ren, L. Although a number of studies have been carried out for the extraction of hidden patterns and classification of EEG signals, temporal information has rarely been incorporated. 2018. Neurosci. The emergence of electroencephalogram (EEG) technology has facilitated direct interaction with brain activity. However, current MI training requires physical attendance, while remote MI training can be applied anywhere, facilitating flexible rehabilitation. Brain-computer interface (BCI) systems empower humans to communicate with or control their external environments via brain signals, offering a practical way to help people with severe motor disabilities [1], [2]. IEEE Deep spatial-temporal neural network for classification of EEG-based motor imagery. In this study, two novel kinds of deep learning schemes based on convolutional neural networks (CNN) and Long Short-Term Memory (LSTM) were proposed for MI-classification. Motor imagery (MI) based Brain-Computer Interface (BCI) is an important active BCI paradigm for recognizing movement intention of severely disabled persons. Google Scholar Deep residual learning for image recognition, in: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2016. , 2018). 27(10), 2164–2177 (2019) Article Google Scholar Zhao, X. 2025 Sun et al. Rehabil. Recently, a lot of efforts have been made to improve MI signal A transformer-based deep learning neural network architecture that performs motion recognition on the raw BCI competition III IVa and IV 2a datasets is proposed and validation results show that the proposed method achieves superior performance than the existing state-of-the-art methods. In this paper, we propose a novel architecture of a deep neural network for EEG-based motor imagery classification. BCI applications for healthy users have also been presented in recent research, such as emotion recognition [3]. Optimal Deep Learning-Based Recognition for EEG Signal Motor Imagery (ODLR-EEGSM) is a novel approach presented in this article that aims to improve the recognition of motor imagery from EEG signals. Proceedings of the Background The adoption of convolutional neural networks (CNNs) for decoding electroencephalogram (EEG)-based motor imagery (MI) in brain-computer interfaces has significantly increased recently. It can be used to decode the intention of users. 152, Elsevier BV In this paper, a deep learning algorithm for limb motor imagery EEG pattern classification is proposed. Traditional neural networks often use serial structure to extract spatial features when dealing with motor imagery EEG signal classification, ignoring temporal information and a large amount of available information in the middle layer, resulting in poor classification Recently, various deep neural networks have been applied to classify electroencephalogram (EEG) signal. A fusion multi-branch 1D convolutional neural network for EEG-based motor imagery classification. 2014: deepLearning: A Deep Learning Architecture for Temporal Sleep Stage Convolutional neural networks (CNNs) are one of the most popular methods in the field of deep learning and have proved effective in several EEG-based applications, such as epilepsy/seizures prediction (Emami et al. 2024, Medical and Biological Engineering and Computing Deep Learning for EEG motor imagery classification based on multi-layer CNNs feature fusion. Rep. However, analysing EEG signals poses Deep neural networks build robust and automated systems for the classification of MI EEG recordings by exploiting the whole input data throughout learning salient features. So far, numerous methods have been designed to classify EEG signal features for MI task. Liu Y. 312-324. To tackle Deep-Learning-Based Neural Network Structure Building. M. K. Syst. A multilayer CNN model is designed for motor imagery EEG classification, and the spatial-frequency characteristics of motor imagery EEG signals are analyzed according to the obtained parameters of convolution layers in the neural network. For the analysis of brain dynamics, non-stationary The literature suggests that building end-to-end deep neural networks for MI-EEG categorization is challenging. Improving EEG-Based Motor Imagery Classification via Spatial and Temporal Recurrent Neural Networks{C}// 2018 40th Annual Motor Imagery (MI) is a primary paradigm in the field of electroencephalogram (EEG) based brain-computer interface (BCI), which does not rely on the traditional brain information output pathways but uses engineering techniques to become a bridge between the brain and the external devices, providing a better and quality life for the people with disabilities In recent years, Convolutional Neural Networks (CNNs) have been widely used to perform automatic feature extraction and classification in various EEG-based tasks. 2018 Jul;2018:1903-1906. One of the most known and used EEG-based BCI paradigms is motor imagery Deep ConvNet employed a temporal and spatial convolutional layers structure designed to extract band-power features, with network depths and filter sizes that vary between the two models. [41] further proposed an EEG channel active inference neural network (EEG-ARNN) based on graph neural networks, which fully leverages the correlations of signals in both temporal and spatial domains. However, their undeniable benefits are counterbalanced by the lack of interpretability properties as well as the inability to perform sufficiently when only limited amount of training data is available. However, the low signal-to-noise ratio, multiple channels and non-linearity are the essential challenges of accurate MI classification. One promising avenue is the use of brain–computer interfaces (BCIs), which establish a direct communication pathway between users and machines. 1109/EMBC. , 19 ( 2 ) ( 2022 ) , pp. The purpose of time-domain attention is to uncover the temporal patterns in EEG signals and assign importance weights based on their intrinsic similarities. 001 is the exponential factor, and ϵ = 0001 is a small number to avoid division by zero. Feature recognition of motor imaging EEG signals based on deep learning. , Masiak J. The temporal and spatial features of EEG were retrieved by creating parallel channel EEG data and positional reconstruction of EEG sequence data, then using the Transformer and 3D Deep temporal-spatial feature learning for motor imagery-based brain–computer interfaces. , Chen X. Sensors, 20 (12) (2020 Physics-informed attention temporal convolutional network for EEG-based motor imagery classification. Parallel Deep Neural Network for Motor Imagery EEG Recognition with Spatiotemporal Features Shi, T. This paper proposes a model based on deep learning network to decode EEG-based MI actions, combining deep separation convolution network (DSCNN) and bidirectional long short-term memory (BLSTM) neural network. Ieee Access, 7 (2018), pp. Electroencephalography-based motor imagery The neural activity (ERD/ERS) generated during motor imagery closely resembles that during actual movement, allowing the translation of motor imagery information in EEG into computer commands for controlling external devices. (2022) proposed a parallel Transformer-based and three-dimensional convolutional neural network (3D-CNN) based multi-channel EEG emotion recognition model. (2022) employed five categories of Transformer-based models for EEG-based motor imagery. 19:1543508. Besides that, features manually extracted and selected Motor imagery electroencephalography (MI-EEG), which is an important subfield of active brain–computer interface (BCI) systems, can be applied to help disabled people to consciously and directly Some selected wavelet coefficients are retained and reconstructed into EEG signals of their respective frequency bands. Facing the accuracy and precision requirements of emotion recognition, this paper combines neural network and proposes a motor imagery EEG signal recognition method based on deep convolutional network. Eng. Song et al. Detecting motor imagery (MI) in electroencephalography (EEG) signals can make their lives easier. , non-linear, Brain-Computer Interface (BCI) enables human beings to interact with the outside world through brain intention. Motor imagery (MI) is currently one of the most researched brain‒computer interface (BCI) paradigms, with convolutional neural networks (CNNs) being extensively used for decoding electroencephalogram (EEG) signals. Human-computer interaction (HCI) based on electroencephalogram (EEG) has become the main research direction in the field of BCI. Traditional EEG pattern recognition algorithm usually includes two steps, namely, feature extraction and feature classification. : Temporal-spatial transformer based motor imagery classification for BCI using independent component analysis. Recently, Ghimire, A. Existing techniques tend to ignore the significant differences between different subjects, resulting in limited accuracy and generalization ability. The motor imagery brain-computer interface (MI-BCI) has the ability to use electroencephalogram (EEG) signals to In order to improve the weak recognition accuracy and robustness of the classification algorithm for brain-computer interface (BCI), this paper proposed a novel classification algorithm for motor imagery based on temporal and spatial characteristics extracted by using convolutional neural networks (TS-CNN) model. 2016), but the The classification of electroencephalogram (EEG) signals is of significant importance in brain-computer interface (BCI) systems. Emotion recognition plays a crucial role in cognitive science and human-computer interaction. [10] proposed a neural network to learn important spatial representations from various scalp positions. Biomed. Nevertheless, single-scale CNN fail to extract abundant information It uses adaptive weights to integrate branch outputs for increased accuracy. The electroencephalogram (EEG) based motor imagery (MI) signal classification, also known as motion recognition, is a highly popular area of research due to its applications in robotics, gami Motivated by the success of transformer algorithms, in this article, we propose a transformer-based deep learning neural network architecture that performs motion recognition To capture long-term dependencies in longer sequence data and further decode advanced temporal information within EEG signals, we designed a deep temporal network First, primary features of EEG signals are extracted using a multi-branch convolutional neural network, followed by feature fusion. In order to achieve MI-EEG decoding, traditional approaches involve manual feature extraction from EEG signals and the application of various machine learning algorithms for classification ([8])([9]). SNNs utilize spike sequences to characterize and convey information, offering a more bio-interpretable approach and consuming less energy than artificial neural networks (ANNs). Addressing the challenges posed by the non-stationarity and low signal-to-noise ratio of EEG signals, the effective extraction of features from motor imagery signals for accurate recognition stands as a key focus in motor imagery brain-computer interface Motor imagery (MI) is a key paradigm in EEG-based BCIs, where individuals mentally rehearse movements (Savaki & Raos, 2019), inducing alterations in mu and beta rhythms in the sensorimotor cortex, resulting in specific patterns of event-related desynchronization (ERD) and event-related synchronization (ERS) (Pezzulo et al. Front. The EEG Motor Imagery B C I C I V _ 2 a dataset (Brunner et al. This paper investigates the role of unimodal and multimodal data in improving recognition accuracy and overcoming inter-subject variability in Brain-Computer Interface systems based on Motor In this paper, we propose an innovative deep learning architecture called Attention-based Multi-view and Multi-scale Temporal Information Fusion Convolutional Neural Network (AMMFCNN) for decoding motor imagery EEG signals. Qiao, X. 2012; Lazarou et al. The spatial resolution of motor imagery EEG signals is enhanced by constructing the spatial filtering module using the Laplacian graph operator. In recent years, iEEG-based BCI has gained increasing attention because the high temporal and spatial resolution of iEEG signal is beneficial to revealing details of motor imagery [1]. Electroencephalography (EEG)-based motor imagery (MI) classification is a key component of BCI technology, which is capable of translating neural activity in the brain into commands for A multi-branch 3D convolutional neural network for EEG-based motor imagery classification. Firstly, CTNet employs a convolutional module analogous to In this paper, we propose a novel architecture of a deep neural network for EEG-based motor imagery classification. Since we can only assure that subjects are performing motor imagery during the feedback stage, we generate data samples from only the first 3 seconds of the feedback stage. While deep learning models have been extensively utilized in motor imagery based EEG signal recognition, they often operate as black boxes. , EEG-based emotion recognition via channel Meanwhile, with the rapid advancements in deep learning within the field of pattern recognition, an increasing number of studies have begun to apply deep learning methods, such as convolutional neural networks (CNN) and residual networks (ResNet), to the classification of motor imagery EEG signals [10]. Deep temporal networks for EEG-based motor imagery recognition Amit Singhal; The electroencephalogram (EEG) based motor imagery (MI) signal classification, also known as motion recognition, is The Motor Imagery (MI) decoding based on electroencephalogram (EEG), has promising applications. Motor imagery (MI) is a very important BCI paradigm which has been widely applied in motor rehabilitation and controlling for disabled patients (Wang et al. proposed a transformer-based spatial–temporal feature learning method ( Song et al. In: IMPROVE, pp. Bi, Deep spatial-temporal neural network for classification of EEG-based motor imagery, in: Proceedings of the 2019 International Conference on Artificial Intelligence and Computer Science, 2019, pp. Decoding of motor imagery EEG based on brain source estimation. Sharma N , Upadhyay A , Sharma M , Singhal A Sci Rep , 13(1):18813, 01 Nov 2023 Recognition of motor imagery intention is one of the hot current research focuses of brain-computer interface (BCI) studies. Traditionally, convolutional neural networks (CNNs) have been extensively One of the critical challenges in brain-computer interfaces is the classification of brain activities through the analysis of EEG signals. Aiming to achieve intelligent classification of motor imagery EEG types with high accuracy, a classification methodology using the wavelet packet decomposition (WPD) and the proposed deep residual convolutional networks (DRes Motor imagery electroencephalography (MI-EEG) signals are generated when a person imagines a task without actually performing it. To address these We used the 3D representation of EEG data suggested in [] in order to fully utilize the spatial-temporal information of MI-based EEG data. Citation: Liao W, Miao Z, Liang S, Zhang L and Li C (2025) A composite improved attention convolutional network for motor imagery EEG classification. , EEG-based emotion recognition via channel Various convolutional neural network (CNN)-based approaches have been recently proposed to improve the performance of motor imagery based-brain-computer interfaces (BCIs). (2021) authors provided a comprehensive review of EEG-based motor imagery BCIs, acknowledging the spatial resolution constraints of scalp EEG as a significant challenge in accurately decoding motor-related brain signals. The authors of (Hsu and Cheng, 2023) considered the feature information of EEG signals in spatial, temporal and spectral domains and proposed a wavelet-based temporal-spectral-attentional correlation coefficient approach to achieve more accurate MI-EEG recognition. Motor imagery classification is a boon to several people with motor impairment. Its significant applications in the gaming, robotics, and medical fields drew our attention to perform a detailed analysis. It is required to endow the recognition model with continuous learning and self-updating capability. Recently, a lot of efforts have been made to improve MI signal classification using a Deep Spatial-Temporal Neural Network for Classification of EEG-Based Motor Imagery Hinton G. However, these classical methods tend to lose the Holographic convolutional attention neural network for motor imagery decoding based on EEG temporal–spatial frequency features Recognition of EEG signal motor imagery intention based on deep multi-view feature learning A. Currently, the deep learning is faced with the chanllenge of decoding from a large amount of time-variant EEG to retain a stable peroformance with different sessions. ; Azadfallah, P. , EEG-based emotion recognition via channel NeuroFlex, a EEG-based soft glove for hand rehabilitation that utilizes a transformer-based deep learning architecture to decode motion intent from motor imagery (MI) EEG data and translate it into control inputs for the assistive glove, aims to enhance the effectiveness and user experience of rehabilitation protocols. This can provide some degree of autonomy to people with neurological disorders or severe motor disabilities. In such a way, users can further manipulate external devices or exchange information with the The electroencephalogram (EEG) based motor imagery (MI) signal classification, also known as motion recognition, is a highly popular area of research due to its applications in robotics, gaming, and medical fields. ; Nowshiravan Rahatabad, F. Nevertheless, single-scale CNN fail to extract abundant information Abstract: Object: Transformer-based neural networks have been applied to the electroencephalography (EEG) decoding for motor imagery (MI). , the connections between EEG electrodes, and A represents the adjacency matrix of the motor imagery brain network G. , 2021). Alhagry et al. The method can be summarised into four Introduction. Qiao, W. Aiming to achieve intelligent classification of motor imagery EEG types with high accuracy, a classification methodology using the wavelet packet decomposition (WPD) and the proposed deep residual convolutional networks (DRes The results show that the method has a high accuracy rate for the recognition of motor imagery EEG, and it has good robustness. Request PDF | Improving EEG-Based Motor Imagery Classification via Spatial and Temporal Recurrent Neural Networks | Motor imagery (MI) based Brain-Computer Interface (BCI) is an important active In this paper, in order to extract the spatio-temporal characteristics and the inherent information implied by functional connections, a multichannel EEG emotion recognition method based on phase In recent decades, the automatic recognition and interpretation of brain waves acquired by electroencephalographic (EEG) technologies have undergone remarkable growth, leading to a consequent Deep temporal features of the EEG signals are extracted using multiple layers of temporal convolutional modules, and the graph convolutional network module is employed to learn the topology of electrode connections, facilitating the decoding of motor imagery EEG signals. In this paper, we propose a spatio In the research of motor imagery brain-computer interface (MI-BCI), traditional electroencephalogram (EEG) signal recognition algorithms appear to be inefficient in extracting EEG signal features and improving classification accuracy. Low accuracy and datasets with few trial recordings present challenges for this classification. For end-to-end deep learning methods, researchers encode MI-Mamba processes multi-channel EEG signals through a single convolutional layer to capture spatial features in the local temporal domain, followed by a Mamba module that processes global temporal features. The WMB technique performs better when applied to six DL models (EEGNet, Shallow ConvNet, Deep ConvNet, ResNet, Multiscale filter bank convolutional neural network (MSFBCNN), and EEG-Temporal Convolutional Network (EEG-TCNet)) and is verified on several EEG datasets. Download Citation | Motor imagery-based EEG signals classification by combining temporal and spatial deep characteristics | Purpose In order to improve the weak recognition accuracy and robustness There is a correlation between adjacent channels of electroencephalogram (EEG), and how to represent this correlation is an issue that is currently being explored. , Zapała D. Figure 3 shows the overall framework of this method. Although deep learning models have been widely applied in recent years for MI based EEG signal recognition, they often function as black boxes and struggle to pre-cisely localize ERD/ERS, a crucial factor in motor imagery recognition. (2018). This model aims to classify motor imagination EEG signals into four classes (left hand, right hand, foot, tongue/rest) by considering the temporal and spatial properties of EEG. In recent studies, MI-EEG has been used in the rehabilitation process of paralyzed patients, therefore, decoding MI-EEG signals accurately is an important task, and it is difficult task due to the low signal-to-noise ratio and the variation of 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 with With the rapid development of artificial intelligence, big data, and computing technology, multimodal research has gradually become an important research direction. Pers Ubiquit Motor imagery (MI) electroencephalogram (EEG) signal is recognized as a promising paradigm for brain-computer interface (BCI) systems and has been extensively employed in various BCI applications, including assisting disabled individuals, controlling devices and environments, and enhancing human capabilities. First, we mapped the channels of the EEG data into a 3D array based on the electrode distributions shown in Fig. , Zhao B. S. 3. Ref. They highlighted the need for advanced signal processing and machine learning techniques to mitigate this Recently, various deep neural networks have been applied to classify electroencephalogram (EEG) signal. We design a simple and effective model to solve the recognition based on motor imagery tasks. A two-branch depthwise CNN for decoding EEG-based multi-class motor imagery. According to the network architecture, we further classify CNN-based decoding models into two Electroencephalogram (EEG) based motor imagery classification is a crucial component of brain-computer interfaces (BCIs). Deep temporal networks for EEG-based motor imagery recognition. Innovative Model Architecture for MI-EEG Classification: We propose a novel model that integrates the strengths of Transformer networks and GCN for the classification of MI-EEG signals. Xie et al. The existing methods involved various feature extraction and machine learning schemes, while classification accuracy and inter-individual model adaptation still need to be improved. Most popular signal processing methods in motor-imagery BCI: a review and meta-analysis. The high-performance decoding A Multi-Scale Spatio-Temporal and Dynamic Graph Convolution Fusion Network (MST-DGCN) to address the problem of an insufficient data volume in deep learning and provides new ideas and methods for the further development of MI-BCI systems. : EEG based motor imagery study of time domain features for classification of power and precision hand grasps. The nature of these signals (i. To solve this problem, a novel convolutional neural network based on temporal-spatial feature learning (TSCNN) was proposed for motor imagery EEG decoding. 背景 (信号分解+机器学习)在（多类大数据）集表现不佳; LSTM无法对非常长期的依赖关系建模; NLP变压网络解决长期依赖性问题; 目的 A discriminative feature learning approach for deep face recognition, in European Conference on Computer Vision (Amsterdam: Springer; ), 499–515. In such a way, users can further manipulate external devices or exchange information with the Brain–computer interfaces (BCI) are an effective tool for recognizing motor imagery and have been widely applied in the motor control and assistive operation domains. However, deep neural networks have been seldom applied to analyze EEG signals. . 1007/978-3-319-46478-7_31 [Google Scholar] Wierzgała P. Neural Syst. Dear Dr. Many methods have been developed to attempt to accurately classify MI-related EEG activity. This paper seeks to improve the efficacy of deep learning-based rehabilitation systems, aiming to deliver superior services for individuals with physical disabilities. Finally, the motor imagery EEG signals are intelligently classified by the DRes-CNN classifier. Subsequently, feature augmentation A transformer-based deep learning neural network architecture that performs motion recognition on the raw BCI competition III IVa and IV 2a datasets is proposed and validation By incorporating spatial and temporal Mamba encoders, STMambaNet effectively captures the intricate dynamics in both space and time, significantly enhancing the decoding In this study, we introduce a convolutional Transformer network (CTNet) designed for EEG-based MI classification. , Sekeroglu, K. Sci. adopted the long short term memory (LSTM) to extract temporal features from EEG signals [29]. [Google Scholar] Vafaei, E. They are used in neurorehabilitation, neuroprosthetics, and gaming, among other applications. 6084-6093. [24] Roy, Rinku, et al. Biomedical Signal Processing and Control, 72, 103241. In feature extraction, common spatial pattern (CSP) is on Decoding motor imagery electroencephalography (MI-EEG) signals presents significant challenges due to the difficulty in capturing the complex functional connectivity between channels and the temporal dependencies of EEG signals across different periods. Addressing computational and efficiency challenges in self-attention mechanisms of transformer models, effectively modeling both local and global contexts in EEG Motor imagery (MI) classification is key for brain-computer interfaces (BCIs). Ind. proposed a continuous convolutional neural network (Conti-CNN), which can learn spatial features from 3D EEG input [30]. Congratulations! Your manuscript Motor imagery (MI), one of the most pivotal paradigms used in BCIs, represents a form of motor intention. Though many achievements have been made in EEG research recently, the lack of sample data and individual differences, Deep learning (DL) method has emerged as a powerful tool in studying the behavior of Electroencephalogram (EEG)-based motor imagery (MI). Recently, a lot of efforts have been made to improve MI signal PDF | On Jul 18, 2022, Yaxin Ma and others published A novel hybrid CNN-Transformer model for EEG Motor Imagery classification | Find, read and cite all the research you need on ResearchGate Abstract: Motor imagery, as a paradigm of brain-computer interface, holds vast potential in the field of medical rehabilitation. Author, and Third C. 7% recognition accuracy on data from twenty Specifically, researchers have focused on the recognition of motor imagery (MI) based on EEG and translating brain activities into specific motor intentions. Download Citation | Deep Spatial-Temporal Neural Network for Classification of EEG-Based Motor Imagery | As a challenging topic in brain-computer interface (BCI) research, motor imagery A Dynamic Domain Adaptation Deep Learning Network for EEG-based Motor Imagery Classification Jie Jiao, Meiyan Xu, Qingqing chen, Hefang Zhou, Wangliang Zhou, Second B. Recent advancements in MI recognition research involve the use of deep neural networks (DNN) to automatically extract feature information from large datasets, eliminating Deep temporal features of the EEG signals are extracted using multiple layers of temporal convolutional modules, and the graph convolutional network module is employed to learn the topology of electrode connections, facilitating the decoding of motor imagery EEG signals. 2018). Existing CNN models can be divided into two types based on Motor imagery (MI) is a popular paradigm for controlling electroencephalogram (EEG) based Brain-Computer Interface (BCI) systems. The classification of electroencephalogram (EEG) signals is of significant importance in brain-computer interface (BCI) systems. Expand Motor imagery of different parts of the body or other exclusive mind-controlled commands would be reflected in different EEG patterns; from these responses, human intentions could be learned. Although EEG research has made progress, motor imagery (MI) EEG decoding remains a challenge due to a lack of sample data, a lower signal noise ratio (SNR), and individual differences. Wavelet transform time-frequency image and convolutional network-based motor imagery EEG classification. In the past decade, motor imagery (MI) has attracted attention in the brain computer interface (BCI) community, where researchers use MI signals to interpret a person’s intention to perform a particular action. Yildirim et al. Neural Networks, Volume 170, 2024, pp. Signal Process Control 63 Deep Spatial-Temporal Neural Network for Classification of EEG-Based Motor Imagery Hinton G. We examined the effect of data segmentation and different neural network structures. EEG is a brain signal that can be acquired in a non-invasive way and has a high temporal resolution. , 2021 ) for EEG decoding, which applies attention transformation on the channel dimension and slices the data in the time dimension to obtain a highly distinguishable The main contribution of this work is to present a deep CNN-based framework to decode EEG-based motor imagery for both upper-limb and lower-limb applications. Until recent years, numerous models had been proposed, ranging from classical algorithms like Common Spatial Pattern (CSP) to deep learning models such as The decoding of motor imagery (MI) electroencephalogram (EEG) is an essential component of the brain–computer interface (BCI), which can help patients with motor impairment communicate directly with the outside world through assistive devices. Electroencephalogram (EEG) signals with Motor imagery brain-computer interface (MI-BCI) provides a novel way for human-computer interaction. Abdul-Jabbar, Deep learning for motor imagery EEG-based classification: a review, Biomed. 1(a) and 1(b). As the EEG signal has a high dimension of feature space, appropriate feature extraction methods are High accuracy decoding of motor imagery directions from EEG-based brain computer interface using filter bank spatially regularised common spatial pattern method. For the analysis of brain dynamics, non-stationary motor imagery signals are used. MI-BCI Keywords: electroencephalogram, motor imagery, convolutional neural networks, label smoothing, center loss. ; Setarehdan, S. In Proceedings of the 2019 International Conference on Artificial Intelligence and Computer EEG pattern recognition is an important part of motor imagery- (MI-) based brain computer interface (BCI) system. proposed a deep CNN with a separated spatial-temporal filter structure to classify MI-EEG raw signals(Lun et al. Furthermore, in Palumbo et al. Recognition of EEG signal motor imagery intention based on deep multi-view feature learning. In Proceedings of the 2019 International Conference on Artificial Intelligence and Computer Science 265–272 (2019). Improving EEG-Based Motor Imagery Classification via Spatial and Temporal Recurrent Neural Networks{C}// 2018 40th Annual Convolutional neural networks (CNNs) have been successfully applied to motor imagery (MI)-based braincomputer interface (BCI). Recently, the development of deep learning has begun to draw increasing attent The motor imagery brain-computer interface (MI-BCI) has the ability to use electroencephalogram (EEG) signals to control and communicate with external devices. (2) They lack analysis and recognition of MI where α = 0. doi: 10. IEEE Trans. Providing remote MI training raises Motor imagery (MI) electroencephalogram (EEG) recognition is currently widely used in brain-computer interface (BCI) devices for people with motor disabilities to achieve various motor interaction Convolutional neural networks (CNNs) have been successfully applied to motor imagery (MI)-based braincomputer interface (BCI). Furthermore, The proposed shallow convolutional neural network (SCNN) is designed with temporal convolution and pointwise convolution (using 1 × 1 convolution) to select the best channel with minimal computational load. , 2020). Several methods have been proposed for EEG-based MI feature extraction, the most classical of which is the CSP algorithm (Kumar et al. In addition, due to inter-individual differences in EEG signals, this discrepancy results in new subjects need spend a amount of calibration time for EEG-based motor imagery brain-computer interface. Although prospective studies have demonstrated promising performance, most of these studies have been affected by the lack of research between groups and individual subjects, and the accuracy of MI classification still has W. Among the different paradigms of Research has shown that the information hidden in the intermediate layers of a neural network can help improve its recognition Physics-informed attention temporal convolutional network for EEG-based motor imagery classification. Owing to the non-invasive and convenient nature of electroencephalography (EEG), it is predominant used method for measuring brain activity on the scalp in BCI systems [2], [3]. 2013, 38(2003):6645--6649. To keep the length of two datasets comparable, we then down-sampled Med-62 dataset to 100Hz. However, most networks focus on applying the self-attention mechanism to extract global temporal information, while the cross-frequency coupling features between different frequencies have been neglected. By applying proper window size and using a purely convolutional neural network, we achieved 97. In this work, we proposed a novel convolutional neural network (CNN)-based method to recognize the motor imagery (MI) activities of left and right hand movements in the EEG-based BCI system. Multiattention adaptation network for motor imagery recognition. , Wan F. As an extension of human-computer interaction, the brain-computer interface (BCI) [] has been widely concerned by scholars and researchers in the So far, EEG-based BCI applications have been developed using neurophysiological patterns, including steady-state visual evoked potential (SSVEP), event-related potential (ERP), movement-related cortical potentials (MRCPs), and motor imagery (MI) [9,10,11,12]. Wang et al. Enhancing motor disability assessment and its imagery classification is a significant concern in contemporary medical practice, necessitating reliable solutions to improve patient outcomes. Specifically for motor imagery elect Motor imagery electroencephalography (EEG) analysis is crucial for the development of effective brain-computer interfaces (BCIs), yet it presents considerable challenges due to the complexity of Electroencephalography (EEG)-based BCIs, which focus on motor imagery, have emerged as an important area of study in this domain. One of the most promising applications for electroencephalogram (EEG)-based brain–computer interfaces (BCIs) is motor rehabilitation through motor imagery (MI) tasks. The MI brain network is defined as graph G = (V, E, A), where V represents the set of vertices, each vertex represents an electrode of EEG, which represents the number of vertices in the MI brain network, that is, | V | = N, E represents the set of edges, i. Broad learning system (BLS) can be remodeled in an efficient incremental learning way. is imperative for bolstering subject-independent motor imagery EEG recognition per-formance. In: 8th International IEEE/EMBS Conference on Neural Engineering (NER), (2017). Citation: Huang X, Zhou N and Choi K-S (2021) A Generalizable and Discriminative Learning Method for Deep EEG-Based Motor Imagery Classification. The key to motor imagery electroencephalogram (MI-EEG) classification is to extract multiple temporal, spatial, and Recently, Electroencephalography (EEG) motor imagery (MI) signals have received increasing attention because it became possible to use these signals to encode a person’s intention to perform an action. This Optimal Deep Learning-Based Recognition for EEG Signal Motor Imagery (ODLR-EEGSM) is a novel approach presented in this article that aims to improve the recognition of motor imagery from EEG signals. Keywords: electroencephalography, convolution neural network, attention mechanism, temporal convolution network, motor imagery, classification. 339, Elsevier BV A Transformer Capsule Network for EEG-based emotion recognition. This research offers a novel approach to these problems in Deep Spatial-Temporal Neural Network for Classification of EEG-Based Motor Imagery AICS 2019: Proceedings of the 2019 International Conference on Artificial Intelligence and Computer Science As a challenging topic in brain-computer interface (BCI) research, motor imagery classification based on electroencephalogram (EEG) received more and more attention. Firstly, for the EEG signals preprocessed by band-pass filtering, a temporal-wise convolution layer and a spatial-wise convolution layer were respectively designed, and temporal-spatial features of motor The proposed EEG representations and network architectures have allowed to achieve interesting IRs, larger than 96% for identification sessions lasting about 20s, and taken at a temporal distance greater than one year from the enrolment, when performing EEG-based biometric recognition irrespective of the performed mental task, therefore providing a solid This study proposes an EEG signal recognition method based on multi-domain feature fusion and RIO-MKELM. There has been a lot of work on detecting two or four different MI movements, which include bilateral, contralateral, and unilateral upper limb Motor imagery (MI) electroencephalography (EEG) has been used in consumer products supported by brain-computer interfaces (BCI), with existing electronics covering a wide range of domains from artificial intelligence (AI) to the Internet of Things (IoT). Early studies mainly focused on feature extraction based on experience and then utilized machine learning to complete the EEG signal decoding [5], [6]. Then, the recur-rent neural networks are utilized to predict the incidence of epilepsy. Electroencephalogram (EEG)-based human-computer interaction (HCI) has become a major research direction in the field of brain-computer interface (BCI). Motor Imagery EEG Recognition Based on Weight-Sharing CNN-LSTM Network In recent years, deep learning-based approaches have become an active direction in EEG-based emotion recognition. In recent years, breakthroughs have been made in the research on recognition of motor imagery task using deep learning, but if the important features related to This paper presents a multi-scale spatiotemporal self-attention (SA) network model that relies on an attention mechanism. In order to improve the weak recognition accuracy and robustness of the classification algorithm for brain-computer interface (BCI), this paper proposed a novel classification algorithm for motor imagery based on temporal and spatial characteristics extracted by using convolutional neural networks (TS-CNN) model. It can help patients with physical dyskinesia to convey their movement intentions. W. Speech recognition with deep recurrent neural networks," ICASSP {J}. 3389/fnins. EEG is well known for its high temporal resolution, low cost for data natural language processing [96] and speech recognition [97]. Classification of motor imagery eeg based on time-domain and frequency-domain In order to fully extract the temporal and spatial features contained in motor imagery electroencephalography (EEG) signals for effective identification of motor imagery, a three-dimensional The system framework encompasses: (i) the collection of raw EEG signals, (ii) conducting correlation analysis for graph weights and degrees, demonstrated through the adjacency matrix, PCC matrix degree, and graph Laplacian, (iii) representing the international 10-10 EEG system in a graph, and (iv) integrating a new DL structure of the F-FGCN. AMMFCNN effectively captures global dependencies and multichannel, multi-temporal information. Motor imagery electroencephalography (MI-EEG) is usually used as a driving signal in neuro-rehabilitation systems, and its feature space varies with the recovery progress. However, current classification methods still face challenges, Nowadays, Electroencephalogram (EEG) signals are widely used in brain-computer interfaces (BCIs), including the identification of motor imagery (MI) activities and prostheses. Deep learning has great potential to automatically extract EEG features for end-to-end decoding. Spiking neural networks (SNNs) are gaining attention across various fields, including EEG-based brain–computer interfaces. M. 2249 - 2258 Google Scholar Abstract. In order Motor imaging EEG signal recognition is an important and challenging research problem in human-computer interaction. Motor imagery (MI) classification is key for brain-computer interfaces (BCIs). Motor impairment is a critical health issue that restricts disabled people from living their lives normally and with comfort. There are extensive studies about MI-based intention recognition, most of which heavily rely on staged handcrafted EEG feature extraction and classifier design. 182–188 (2022) Google Scholar Hameed, A. (2021) A novel end-to-end deep convolutional neural network based skin lesion classification framework. Extracting a Physics-informed attention temporal convolutional network for EEG-based motor imagery classification IEEE Trans. , Member, IEEE Abstract—There is a correlation between adjacent chan-nels of electroencephalogram (EEG), and how to repre- A Comparison of Deep Neural Network Algorithms for Recognition of EEG Motor Imagery Signals The objective of this study is to investigate the effectiveness of utilizing signal images and deep learning techniques in EEG motor imagery signal processing and explore the potential of EEG-based communication systems to aid patients in expressing Each convolutional layer of CNN has the same size of convolutional kernel that can be used for feature extraction, Lun et al. The model applies spectral filtering to the EEG data 论文阅读: Deep temporal networks for EEG-based motor imagery recognition. The channels of the EEG data were mapped into a 3D array. By placing strip or mesh electrodes on the surface of the brain (epidural or subdural) through drilling or craniotomy, iEEG records neuroelectric signals induced by neuronal activities in Purpose In the brain-computer interface (BCI), motor imagery (MI) could be defined as the Electroencephalogram (EEG) signals through imagined movements, and ultimately enabling individuals to control external devices. This technology holds the Motor imagery (MI) based Brain-Computer Interface (BCI) is an important active BCI paradigm for recognizing movement intention of severely disabled persons. A spatio-temporal energy maps generation scheme followed by deep learning classification model and Long-Short-Term-Memory based neural network has been proposed to classify the temporal series of energy maps. It can be used Two classical deep learning models are the convolutional neural network (CNN) and recurrent neural network (RNN), which are widely used for EEG classification in motor imagery. In addition, existing methods suffer from difficulties in capturing the complex relationships among the channels of This has inspired a wave of innovation in MI-EEG classification, where researchers have explored a diverse array of deep learning architectures, including convolutional neural networks (CNN) 27,28 Background: Brain–computer interface (BCI) technology opens up new avenues for human–machine interaction and rehabilitation by connecting the brain to machines. However, the recorded EEG signal is easily interfered by other signals, which leads to its low signal-to-noise ratio. we select five highly-cited and highly performing deep neural networks (DNNs EEG-Based Emotion Recognition Using Deep Learning Network with Principal Component Based Covariate Shift Adaptation: Suwicha Jirayucharoensak, et al. : Deep CNN model based on serial-parallel structure optimization for four-class motor imagery EEG classification. Specifically, convolutional neural networks (CNN) and hybrid-CNN (h-CNN) are the dominant architectures with high performance in comparison to public datasets with other A popular research area in electroencephalography (EEG) is a brain-computer interface (BCI), which involves the classification of MI tasks. Our results suggest that the CNN-based framework has an advantage over two other typical methods, which could be used to improve the performance of a BMI for motor training and functional recovery. Singhal, A. Improving EEG-Based Motor Imagery Classification via Spatial and Temporal Recurrent Neural Networks{C}// 2018 40th Annual For instance, the literature proposed a transformer model for motor imagery recognition using raw EEG data, reporting high performance. , 19 Qiao, W. These challenges are exacerbated by the low spatial resolution and high signal redundancy inherent A novel method to perform preprocessing and feature learning of EEG features by performing Morlet wavelet and cubic spline interpolation methods and constructing a hybrid model which associates modified Convolutional Neural Network with Bidirectional Gated Recurrent Unit. Motivated by neurological findings indicating that the The electroencephalogram (EEG) based motor imagery (MI) signal classification, also known as motion recognition, is a highly popular area of research due to its applications in robotics, gaming, and medical fields. EEG signals are often affected by noise and artifacts, leading to a low signal-to-noise ratio (SNR). Among these BCI studies, MI, which classifies EEG signals based on the Deep temporal networks for EEG-based motor imagery recognition. The EEG signals from several people who were engaged in motor imagery tasks are collected in the EEG Motor Imagery BCICIV_2a dataset. Inform. In recent years, the utilization of motor imagery (MI) signals derived from electroencephalography (EEG) has shown promising applications in controlling various devices such as wheelchairs, assistive technologies, and driverless vehicles. Man Cybern. Such capabilities hold significant potential for advancing This technique significantly decreases the number of connections and parameters in a deep network while using the pertinent geographical or temporal links between nearby data points to infer helpful features for a particular machine learning job. Deep Learning has grasped great attention for recognition of Electroencephalography. 10. Vedantam, D. proposed a frequency complementary map selection (FCMS) scheme based on augmented CSP (ACSP) for We utilize temporal attention to focus on valuable temporal information in EEG-based emotion recognition. A. , 2019, Chye et al Motor Imagery (MI) based on Electroencephalography (EEG), a typical Brain-Computer Interface (BCI) paradigm, can communicate with external devices according to the brain’s intentions. In this paper, we discuss a solution to this problem based on a n The electroencephalogram (EEG) motor imagery (MI) signals are the widespread paradigms in the brain-computer interface (BCI). hlynrays nunnacci dokii alwbu xijpi pkky baqd selh eycv tjp msykye ghezv edch bqynh ckiiv