Joint hybrid recursive feature elimination based channel selection and ResGCN for cross session MI recognition.

In the field of brain-computer interface (BCI) based on motor imagery (MI), multi-channel electroencephalography (EEG) data is commonly utilized for MI task recognition to achieve sensory compensation or precise human-computer interaction. However, individual physiological differences, environmental variations, or redundant information and noise in certain channels can pose challenges and impact the performance of BCI systems. In this study, we introduce a channel selection method utilizing Hybrid-Recursive Feature Elimination (H-RFE) combined with residual graph neural networks for MI recognition. This channel selection method employs a recursive feature elimination strategy and integrates three classification methods, namely random forest, gradient boosting, and logistic regression, as evaluators for adaptive channel selection tailored to specific subjects. To fully exploit the spatiotemporal information of multi-channel EEG, this study employed a graph neural network embedded with residual blocks to achieve precise recognition of motor imagery. We conducted algorithm testing using the SHU dataset and the PhysioNet dataset. Experimental results show that on the SHU dataset, utilizing 73.44% of the total channels, the cross-session MI recognition accuracy is 90.03%. Similarly, in the PhysioNet dataset, using 72.5% of the channel data, the classification result also reaches 93.99%. Compared to traditional strategies such as selecting three specific channels, correlation-based channel selection, mutual information-based channel selection, and adaptive channel selection based on Pearson coefficients and spatial positions, the proposed method improved classification accuracy by 34.64%, 10.8%, 3.25% and 2.88% on the SHU dataset, and by 46.96%, 5.04%, 5.81% and 2.32% on the PhysioNet dataset, respectively.

Large scale investigation of the effect of gender on mu rhythm suppression in motor imagery brain-computer interfaces.

The utmost issue in Motor Imagery Brain-Computer Interfaces (MI-BCI) is the BCI poor performance known as 'BCI inefficiency'. Although past research has attempted to find a solution by investigating factors influencing users' MI-BCI performance, the issue persists. One of the factors that has been studied in relation to MI-BCI performance is gender. Research regarding the influence of gender on a user's ability to control MI-BCIs remains inconclusive, mainly due to the small sample size and unbalanced gender distribution in past studies. To address these issues and obtain reliable results, this study combined four MI-BCI datasets into one large dataset with 248 subjects and equal gender distribution. The datasets included EEG signals from healthy subjects from both gender groups who had executed a right- vs. left-hand motor imagery task following the Graz protocol. The analysis consisted of extracting the Mu Suppression Index from C3 and C4 electrodes and comparing the values between female and male participants. Unlike some of the previous findings which reported an advantage for female BCI users in modulating mu rhythm activity, our results did not show any significant difference between the Mu Suppression Index of both groups, indicating that gender may not be a predictive factor for BCI performance.

Electroencephalogram-based adaptive closed-loop brain-computer interface in neurorehabilitation: a review.

Brain-computer interfaces (BCIs) represent a groundbreaking approach to enabling direct communication for individuals with severe motor impairments, circumventing traditional neural and muscular pathways. Among the diverse array of BCI technologies, electroencephalogram (EEG)-based systems are particularly favored due to their non-invasive nature, user-friendly operation, and cost-effectiveness. Recent advancements have facilitated the development of adaptive bidirectional closed-loop BCIs, which dynamically adjust to users' brain activity, thereby enhancing responsiveness and efficacy in neurorehabilitation. These systems support real-time modulation and continuous feedback, fostering personalized therapeutic interventions that align with users' neural and behavioral responses. By incorporating machine learning algorithms, these BCIs optimize user interaction and promote recovery outcomes through mechanisms of activity-dependent neuroplasticity. This paper reviews the current landscape of EEG-based adaptive bidirectional closed-loop BCIs, examining their applications in the recovery of motor and sensory functions, as well as the challenges encountered in practical implementation. The findings underscore the potential of these technologies to significantly enhance patients' quality of life and social interaction, while also identifying critical areas for future research aimed at improving system adaptability and performance. As advancements in artificial intelligence continue, the evolution of sophisticated BCI systems holds promise for transforming neurorehabilitation and expanding applications across various domains.

E-SAT: An extreme learning machine based self attention approach for decoding motor imagery EEG in subject-specific tasks.

The advancements in Brain-Computer Interface (BCI) have substantially evolved people's lives by enabling direct communication between the human brain and external peripheral devices. In recent years, the integration of machine larning (ML) and deep learning (DL) models have considerably imrpoved the performances of BCIs for decoding the motor imagery (MI) tasks. However, there still exist several limitations, e.g., extensive training time and high sensitivity to noises or outliers with those existing models, which largely hinder the rapid developments of BCIs. To address such issues, this paper proposes a novel extreme learning machine (ELM) based self-attention (E-SAT) mechanism to enhance subject-specific classification performances. Specifically, for E-SAT, ELM is employed both to imrpove self-attention module generalization ability for feature extraction and to optimize the model's parameter initialization process. Meanwhile, the extracted features are also classified using ELM, and the end-to-end ELM based setup is used to evaluate E-SAT performances on different MI EEG signals. Extensive experiments with different datasets, such as BCI Competition III Dataset IV-a, IV-b and BCI Competition IV Datasets 1,2a,2b,3, are conducted to verify the effectiveness of proposed E-SAT strategy. Results show that E-SAT outperforms several state-of-the-art (SOTA) existing methods in subject-specific classification on all the datasets, with an average classification accuracy of 99.8%,99.1%,98.9%,75.8%, 90.8%, and 95.4%, being achieved for each datasets, respectively. The experimental results not only show outstanding performance of E-SAT in feature extractions, but also demonstrate that it helps achieves the best results among nine other robust ones. In addition, results in this study also demonstrate that E-SAT achieves exceptional performance in both binary and multi-class classification tasks, as well as for noisy and non-noisy datatsets. .

Motor imagery with cues in virtual reality, audio and screen.

OBJECTIVE: Training plays a significant role in motor imagery (MI), particularly in applications such as Motor Imagery-based Brain-Computer Interface (MIBCI) systems and rehabilitation systems. Previous studies have investigated the intricate relationship between cues and MI signals. However, the medium of presentation still remains an emerging area to be explored, as possible factors to enhance Motor Imagery signals.. Approach: We hypothesise that the medium used for cue presentation can significantly influence both performance and training outcomes in MI tasks. To test this hypothesis, we designed and executed an experiment implementing no- feedback MI. Our investigation focused on three distinct cue presentation mediums -audio, screen, and virtual reality(VR) headsets-all of which have potential implications for BCI use in the Activities of Daily Lives. Main Results: The results of our study uncovered notable variations in MI signals depending on the medium of cue presentation, where the analysis is based on 3 EEG channels. To substantiate our findings, we employed a comprehensive approach, utilizing various evaluation metrics including Event- Related Synchronisation(ERS)/Desynchronisation(ERD), Feature Extraction (using Recursive Feature Elimination (RFE)), Machine Learning methodologies (using Ensemble Learning), and participant Questionnaires. All the approaches signify that Motor Imagery signals are enhanced when presented in VR, followed by audio, and lastly screen. Applying a Machine Learning approach across all subjects, the mean cross-validation accuracy (Mean ± Std. Error) was 69.24 ± 3.12, 68.69 ± 3.3 and 66.1±2.59 when for the VR, audio-based, and screen-based instructions respectively. Significance: This multi-faceted exploration provides evidence to inform MI- based BCI design and advocates the incorporation of different mediums into the design of MIBCI systems, experimental setups, and user studies. The influence of the medium used for cue presentation may be applied to develop more effective and inclusive MI applications in the realm of human-computer interaction and rehabilitation.

I see artifacts: ICA-based EEG artifact removal does not improve deep network decoding across three BCI tasks.

\textit{Objective.} In this paper, we conduct a detailed investigation on the effect of IC-based noise rejection methods in neural network classifier-based decoding of electroencephalography (EEG) data in different task datasets. \textit{Approach.} We apply a pipeline matrix of two popular different Independent Component (IC) decomposition methods (Infomax, AMICA) with three different component rejection strategies (none, ICLabel, and MARA) on three different EEG datasets (Motor imagery, long-term memory formation, and visual memory). We cross-validate processed data from each pipeline with three architectures commonly used for EEG classification (two convolutional neural networks (CNN) and one long short term memory (LSTM) based model. We compare decoding performances on within-participant and within-dataset levels. \textit{Main Results.} Our results show that the benefit from using IC-based noise rejection for decoding analyses is at best minor, as component-rejected data did not show consistently better performance than data without rejections---especially given the significant computational resources required for ICA computations. \textit{Significance.} With ever growing emphasis on transparency and reproducibility, as well as the obvious benefits arising from streamlined processing of large-scale datasets, there has been an increased interest in automated methods for pre-processing EEG data. One prominent part of such pre-processing pipelines consists of identifying and potentially removing artifacts arising from extraneous sources. This is typically done via Independent Component (IC) based correction for which numerous methods have been proposed, differing not only in the decomposition of the raw data into ICs, but also in how they reject the computed ICs. While the benefits of these methods are well established in univariate statistical analyses, it is unclear whether they help in multivariate scenarios, and specifically in neural network based decoding studies. As computational costs for pre-processing large-scale datasets are considerable, it is important to consider whether the tradeoff between model performance and available resources is worth the effort.

Translational Bioethics in China: Brain-Computer Interface Research as a Case Study.

The research and development of emerging technologies has potential long-term and societal impacts that pose governance challenges. This essay summarizes the development of research ethics in China over the past few decades, as well as the measures taken by the Chinese government to build its ethical governance system of science and technology after the occurrence of the CRISPR-babies incident. The essay then elaborates on the current problems of this system through the case study of ethical governance of brain-computer interface research, and explores how the transition from research ethics to translational bioethics, which encourages interdisciplinary collaboration and focuses on societal implications, may respond to the challenges of ethical governance of science and technology.

Enhancing Brain-Computer Interfaces through Kriging-Based Fusion of Sparse Regression Partial Differential Equations to Counter Injection Molding View of Node Displacement Effects.

Injection molding is an efficient and precise manufacturing technology that is widely used in the production of plastic products. In recent years, injection molding technology has made significant progress, especially with the combination of in-mold electronics (IME) technology, which makes it possible to embed electronic components directly into the surface of a product. IME technology improves the integration and performance of a product by embedding conductive materials and functional components in the mold. Brain-computer interfaces (BCIs) are a rapidly growing field of research that aims to capture, analyze, and feedback brain signals by directly connecting the brain to external devices. The Utah array, a high-density microelectrode array, has been widely used for the recording and transmission of brain signals. However, the traditional fabrication method of the Utah array suffers from high cost and low integration, which limits its promotion in practical applications. The lines that receive EEG signals are one of the key parts of a brain-computer interface system. The optimization of injection molding parameters is particularly important in order to effectively embed these lines into thin films and to ensure the precise displacement of the line nodes and the stability of signal transmission during the injection molding process. In this study, a method based on the Kriging prediction model and sparse regression partial differential equations (PDEs) is proposed to optimize the key parameters in the injection molding process. This method can effectively predict and control the displacement of nodes in the film, ensure the stability and reliability of the line during the injection process, and improve the accuracy of EEG signal transmission and system performance. The optimal injection parameters were finally obtained: a holding pressure of 525 MPa, a holding time of 50 s, and a melting temperature of 285 °C. Under this condition, the average node displacement of UA was reduced from the initial 0.19 mm to 0.89 µm, with an optimization rate of 95.32%.

Improving classification performance of motor imagery BCI through EEG data augmentation with conditional generative adversarial networks.

In brain-computer interface (BCI), building accurate electroencephalogram (EEG) classifiers for specific mental tasks is critical for BCI performance. The classifiers are developed by machine learning (ML) and deep learning (DL) techniques, requiring a large dataset for training to build reliable and accurate models. However, collecting large enough EEG datasets is difficult due to intra-/inter-subject variabilities and experimental costs. This leads to the data scarcity problem, which causes overfitting issues to training samples, resulting in reducing generalization performance. To solve the EEG data scarcity problem and improve the performance of the EEG classifiers, we propose a novel EEG data augmentation (DA) framework using conditional generative adversarial networks (cGANs). An experimental study is implemented with two public EEG datasets, including motor imagery (MI) tasks (BCI competition IV IIa and III IVa), to validate the effectiveness of the proposed EEG DA method for the EEG classifiers. To evaluate the proposed cGAN-based DA method, we tested eight EEG classifiers for the experiment, including traditional MLs and state-of-the-art DLs with three existing EEG DA methods. Experimental results showed that most DA methods with proper DA proportion in the training dataset had higher classification performances than without DA. Moreover, applying the proposed DA method showed superior classification performance improvement than the other DA methods. This shows that the proposed method is a promising EEG DA method for enhancing the performances of the EEG classifiers in MI-based BCIs.

A multistrategy differential evolution algorithm combined with Latin hypercube sampling applied to a brain-computer interface to improve the effect of node displacement.

Injection molding is a common plastic processing technique that allows melted plastic to be injected into a mold through pressure to form differently shaped plastic parts. In injection molding, in-mold electronics (IME) can include various circuit components, such as sensors, amplifiers, and filters. These components can be injected into the mold to form a whole within the melted plastic and can therefore be very easily integrated into the molded part. The brain-computer interface (BCI) is a direct connection pathway between a human or animal brain and an external device. Through BCIs, individuals can use their own brain signals to control these components, enabling more natural and intuitive interactions. In addition, brain-computer interfaces can also be used to assist in medical treatments, such as controlling prosthetic limbs or helping paralyzed patients regain mobility. Brain-computer interfaces can be realized in two ways: invasively and noninvasively, and in this paper, we adopt a noninvasive approach. First, a helmet model is designed according to head shape, and second, a printed circuit film is made to receive EEG signals and an IME injection mold for the helmet plastic parts. In the electronic film, conductive ink is printed to connect each component. However, improper parameterization during the injection molding process can lead to node displacements and residual stress changes in the molded part, which can damage the circuits in the electronic film and affect its performance. Therefore, in this paper, the use of the BCI molding process to ensure that the node displacement reaches the optimal value is studied. Second, the multistrategy differential evolutionary algorithm is used to optimize the injection molding parameters in the process of brain-computer interface formation. The relationship between the injection molding parameters and the actual target value is investigated through Latin hypercubic sampling, and the optimized parameters are compared with the target parameters to obtain the optimal parameter combination. Under the optimal parameters, the node displacement can be optimized from 0.585 to 0.027 mm, and the optimization rate can reach 95.38%. Ultimately, by detecting whether the voltage difference between the output inputs is within the permissible range, the reliability of the brain-computer interface after node displacement optimization can be evaluated.

A MAUDE database analysis on the new generation of active bone conduction hearing implants.

Objective: To analyze medical device reports (MDR) submitted to the Food and Drug Administration's (FDA) Manufacturer and User Device Facility Experience (MAUDE) database to identify adverse events (AEs) in patients implanted with novel active bone conduction hearing implants (BCIs). Methods: We conducted a search of the FDA MAUDE database on the newest generation of BCIs. Data were collected concerning device malfunctions, patient injuries, factors triggering these incidents, and the subsequent actions taken. Results: In total, 93 (16.7%) device malfunctions and 465 (83.3%) patient injuries with 358 subsequent interventions were identified, resulting in 558 AEs. Although the absolute AE number per device cannot be identified, the following trends were detected: Among the 494 AEs associated with OSI200, 55 (11.1%) reported device malfunctions and 454 (88.9%) cited patient injuries. Out of the 64 AEs linked to BCI602, 28 (59.4%) were associated with malfunctions, whereas 26 (40.6%) involved patient injuries. The most frequently reported particular AEs for the OSI200 were infection (n = 171, 34.6%), extrusion of the device (n = 107, 21.7%), and pain (n = 51, 10.3%). Conversely, no device output (n = 20, 31.3%) and loss of osseointegration (n = 7, 10.9%) were the most reported AEs for the BCI602. Various AEs led to 214 explanations and 77 revision surgeries. Sixty-seven AEs reported conservative treatment. Conclusion: The current study provides an overview of the most commonly reported complications with new active BCIs. Although providing an overview, given the limitations of the FDA MAUDE database, our results have to be interpreted with caution. Level of Evidence: 4.

Automatic Recognition of Multiple Emotional Classes from EEG Signals through the Use of Graph Theory and Convolutional Neural Networks.

Emotion is a complex state caused by the functioning of the human brain in relation to various events, for which there is no scientific definition. Emotion recognition is traditionally conducted by psychologists and experts based on facial expressions-the traditional way to recognize something limited and is associated with errors. This study presents a new automatic method using electroencephalogram (EEG) signals based on combining graph theory with convolutional networks for emotion recognition. In the proposed model, firstly, a comprehensive database based on musical stimuli is provided to induce two and three emotional classes, including positive, negative, and neutral emotions. Generative adversarial networks (GANs) are used to supplement the recorded data, which are then input into the suggested deep network for feature extraction and classification. The suggested deep network can extract the dynamic information from the EEG data in an optimal manner and has 4 GConv layers. The accuracy of the categorization for two classes and three classes, respectively, is 99% and 98%, according to the suggested strategy. The suggested model has been compared with recent research and algorithms and has provided promising results. The proposed method can be used to complete the brain-computer-interface (BCI) systems puzzle.

A Spatio-Temporal Capsule Neural Network with Self-Correlation Routing for EEG Decoding of Semantic Concepts of Imagination and Perception Tasks.

Decoding semantic concepts for imagination and perception tasks (SCIP) is important for rehabilitation medicine as well as cognitive neuroscience. Electroencephalogram (EEG) is commonly used in the relevant fields, because it is a low-cost noninvasive technique with high temporal resolution. However, as EEG signals contain a high noise level resulting in a low signal-to-noise ratio, it makes decoding EEG-based semantic concepts for imagination and perception tasks (SCIP-EEG) challenging. Currently, neural network algorithms such as CNN, RNN, and LSTM have almost reached their limits in EEG signal decoding due to their own short-comings. The emergence of transformer methods has improved the classification performance of neural networks for EEG signals. However, the transformer model has a large parameter set and high complexity, which is not conducive to the application of BCI. EEG signals have high spatial correlation. The relationship between signals from different electrodes is more complex. Capsule neural networks can effectively model the spatial relationship between electrodes through vector representation and a dynamic routing mechanism. Therefore, it achieves more accurate feature extraction and classification. This paper proposes a spatio-temporal capsule network with a self-correlation routing mechaninsm for the classification of semantic conceptual EEG signals. By improving the feature extraction and routing mechanism, the model is able to more effectively capture the highly variable spatio-temporal features from EEG signals and establish connections between capsules, thereby enhancing classification accuracy and model efficiency. The performance of the proposed model was validated using the publicly accessible semantic concept dataset for imagined and perceived tasks from Bath University. Our model achieved average accuracies of 94.9%, 93.3%, and 78.4% in the three sensory modalities (pictorial, orthographic, and audio), respectively. The overall average accuracy across the three sensory modalities is 88.9%. Compared to existing advanced algorithms, the proposed model achieved state-of-the-art performance, significantly improving classification accuracy. Additionally, the proposed model is more stable and efficient, making it a better decoding solution for SCIP-EEG decoding.

Enhancing EEG-Based MI-BCIs with Class-Specific and Subject-Specific Features Detected by Neural Manifold Analysis.

This paper presents an innovative approach leveraging Neuronal Manifold Analysis of EEG data to identify specific time intervals for feature extraction, effectively capturing both class-specific and subject-specific characteristics. Different pipelines were constructed and employed to extract distinctive features within these intervals, specifically for motor imagery (MI) tasks. The methodology was validated using the Graz Competition IV datasets 2A (four-class) and 2B (two-class) motor imagery classification, demonstrating an improvement in classification accuracy that surpasses state-of-the-art algorithms designed for MI tasks. A multi-dimensional feature space, constructed using NMA, was built to detect intervals that capture these critical characteristics, which led to significantly enhanced classification accuracy, especially for individuals with initially poor classification performance. These findings highlight the robustness of this method and its potential to improve classification performance in EEG-based MI-BCI systems.

Optimizing Real-Time MI-BCI Performance in Post-Stroke Patients: Impact of Time Window Duration on Classification Accuracy and Responsiveness.

Brain-computer interfaces (BCIs) are promising tools for motor neurorehabilitation. Achieving a balance between classification accuracy and system responsiveness is crucial for real-time applications. This study aimed to assess how the duration of time windows affects performance, specifically classification accuracy and the false positive rate, to optimize the temporal parameters of MI-BCI systems. We investigated the impact of time window duration on classification accuracy and false positive rate, employing Linear Discriminant Analysis (LDA), Multilayer Perceptron (MLP), and Support Vector Machine (SVM) on data acquired from six post-stroke patients and on the external BCI IVa dataset. EEG signals were recorded and processed using the Common Spatial Patterns (CSP) algorithm for feature extraction. Our results indicate that longer time windows generally enhance classification accuracy and reduce false positives across all classifiers, with LDA performing the best. However, to maintain the real-time responsiveness, crucial for practical applications, a balance must be struck. The results suggest an optimal time window of 1-2 s, offering a trade-off between classification performance and excessive delay to guarantee the system responsiveness. These findings underscore the importance of temporal optimization in MI-BCI systems to improve usability in real rehabilitation scenarios.

Sensorimotor brain-computer interface performance depends on signal-to-noise ratio but not connectivity of the mu rhythm in a multiverse analysis of longitudinal data.

Objective.Serving as a channel for communication with locked-in patients or control of prostheses, sensorimotor brain-computer interfaces (BCIs) decode imaginary movements from the recorded activity of the user's brain. However, many individuals remain unable to control the BCI, and the underlying mechanisms are unclear. The user's BCI performance was previously shown to correlate with the resting-state signal-to-noise ratio (SNR) of the mu rhythm and the phase synchronization (PS) of the mu rhythm between sensorimotor areas. Yet, these predictors of performance were primarily evaluated in a single BCI session, while the longitudinal aspect remains rather uninvestigated. In addition, different analysis pipelines were used to estimate PS in source space, potentially hindering the reproducibility of the results.Approach.To systematically address these issues, we performed an extensive validation of the relationship between pre-stimulus SNR, PS, and session-wise BCI performance using a publicly available dataset of 62 human participants performing up to 11 sessions of BCI training. We performed the analysis in sensor space using the surface Laplacian and in source space by combining 24 processing pipelines in a multiverse analysis. This way, we could investigate how robust the observed effects were to the selection of the pipeline.Main results.Our results show that SNR had both between- and within-subject effects on BCI performance for the majority of the pipelines. In contrast, the effect of PS on BCI performance was less robust to the selection of the pipeline and became non-significant after controlling for SNR.Significance.Taken together, our results demonstrate that changes in neuronal connectivity within the sensorimotor system are not critical for learning to control a BCI, and interventions that increase the SNR of the mu rhythm might lead to improvements in the user's BCI performance.

Improving non-invasive trajectory decoding via neural correlates of continuous erroneous feedback processing.

Objective. Over the last decades, error-related potentials (ErrPs) have repeatedly proven especially useful as corrective mechanisms in invasive and non-invasive brain-computer interfaces (BCIs). However, research in this context exclusively investigated the distinction of discrete events intocorrectorerroneousto the present day. Due to this predominant formulation as a binary classification problem, classical ErrP-based BCIs fail to monitor tasks demanding quantitative information on error severity rather than mere qualitative decisions on error occurrence. As a result, fine-tuned and natural feedback control based on continuously perceived deviations from an intended target remains beyond the capabilities of previously used BCI setups.Approach.To address this issue for future BCI designs, we investigated the feasibility of regressing rather than classifying error-related activity non-invasively from the brain.Main results.Using pre-recorded data from ten able-bodied participants in three sessions each and a multi-output convolutional neural network, we demonstrated the above-chance regression of ongoing target-feedback discrepancies from brain signals in a pseudo-online fashion. In a second step, we used this inferred information about the target deviation to correct the initially displayed feedback accordingly, reporting significant improvements in correlations between corrected feedback and target trajectories across feedback conditions.Significance.Our results indicate that continuous information on target-feedback discrepancies can be successfully regressed from cortical activity, paving the way to increasingly naturalistic, fine-tuned correction mechanisms for future BCI applications.