Electroencephalography Signals Research Articles

A Data Augmentation Method for Motor Imagery EEG Signals Based on DCGAN-GP Network.

Motor imagery electroencephalography (EEG) signals have garnered attention in brain-computer interface (BCI) research due to their potential in promoting motor rehabilitation and control. However, the limited availability of labeled data poses challenges for training robust classifiers. In this study, we propose a novel data augmentation method utilizing an improved Deep Convolutional Generative Adversarial Network with Gradient Penalty (DCGAN-GP) to address this issue. We transformed raw EEG signals into two-dimensional time-frequency maps and employed a DCGAN-GP network to generate synthetic time-frequency representations resembling real data. Validation experiments were conducted on the BCI IV 2b dataset, comparing the performance of classifiers trained with augmented and unaugmented data. Results demonstrated that classifiers trained with synthetic data exhibit enhanced robustness across multiple subjects and achieve higher classification accuracy. Our findings highlight the effectiveness of utilizing a DCGAN-GP-generated synthetic EEG data to improve classifier performance in distinguishing different motor imagery tasks. Thus, the proposed data augmentation method based on a DCGAN-GP offers a promising avenue for enhancing BCI system performance, overcoming data scarcity challenges, and bolstering classifier robustness, thereby providing substantial support for the broader adoption of BCI technology in real-world applications.

Seizure detection via deterministic learning feature extraction

Epileptic seizures have a significant impact on the well-being of a large number of individuals worldwide. Utilizing electroencephalographic (EEG) signals for automatic seizure detection proves to be a valuable solution. However, dealing with raw EEG signals is inherently complex, necessitating a preliminary step of feature extraction prior to detection. Traditional feature extraction methods often amalgamate various types of features for seizure detection, as each type typically captures specific properties. In contrast, this paper focuses on detecting seizures by analyzing the system dynamics. The proposed Deterministic Learning Feature Extraction (DLFE) method extracts a single type of nonlinear dynamical feature rooted in the EEG system dynamics. DLFE employs deterministic learning to discern the inherent system dynamics of the EEG under both seizure and normal conditions. Through the feature extraction process, the infinite-dimensional system dynamics are transformed into feature vectors, exhibiting distinct distributions in seizure and normal states. This disparity can be effectively utilized for classification using standard classifiers. The performance of the proposed seizure detection method was assessed using the CHB-MIT and Bonn datasets. The average classification accuracy was found to be 98.63% with a specificity of 99.19% and a sensitivity of 98.06% on CHB-MIT dataset. Compared with the latest similar methods, the accuracy, specificity and sensitivity are improved by 0.31%, 0.21% and 0.05% respectively. Moreover, the performance was achieved with the short-time interval EEG signals within a few channels. The average classification accuracy was found to be 99.90% with a 0.22% improvement on Bonn dataset, which indicates the good generalization performance.

Judgments of Learning Reactively Improve Memory by Enhancing Learning Engagement and Inducing Elaborative Processing: Evidence from an EEG Study.

Making judgments of learning (JOLs) can reactively alter memory itself, a phenomenon termed the reactivity effect. The current study recorded electroencephalography (EEG) signals during the encoding phase of a word list learning task to explore the neurocognitive features associated with JOL reactivity. The behavioral results show that making JOLs reactively enhances recognition performance. The EEG results reveal that, compared with not making JOLs, making JOLs increases P200 and LPC amplitudes and decreases alpha and beta power. Additionally, the signals of event-related potentials (ERPs) and event-related desynchronizations (ERDs) partially mediate the reactivity effect. These findings support the enhanced learning engagement theory and the elaborative processing explanation to account for the JOL reactivity effect.

Electroencephalographic Signal Data Augmentation Based on Improved Generative Adversarial Network.

EEG signals combined with deep learning play an important role in the study of human-computer interaction. However, the limited dataset makes it challenging to study EEG signals using deep learning methods. Inspired by the GAN network in image generation, this paper presents an improved generative adversarial network model L-C-WGAN-GP to generate artificial EEG data to augment training sets and improve the application of BCI in various fields. The generator consists of a long short-term memory (LSTM) network and the discriminator consists of a convolutional neural network (CNN) which uses the gradient penalty-based Wasserstein distance as the loss function in model training. The model can learn the statistical features of EEG signals and generate EEG data that approximate real samples. In addition, the performance of the compressed sensing reconstruction model can be improved by using augmented datasets. Experiments show that, compared with the existing advanced data amplification techniques, the proposed model produces EEG signals closer to the real EEG signals as measured by RMSE, FD and WTD indicators. In addition, in the compressed reconstruction of EEG signals, adding the new data reduces the loss by about 15% compared with the original data, which greatly improves the reconstruction accuracy of the EEG signals' compressed sensing.

VLSI design for adjustable compression rate in lossless/lossy compression of EEG signal

In the realm of remote health monitoring, the compression of electroencephalography (EEG) signals at the edge holds significant importance. This paper presents a VLSI structure capable of adjustable compression rates to achieve both lossless and lossy compression of EEG data, catering to environments requiring flexible adjustment of compression types and rates. The architecture integrates two-dimensional discrete wavelet transform (2D-DWT) and the Set Partitioning in Hierarchical Trees (SPIHT). Notably, the Haar wavelet lifting structure is employed in 2D-DWT, and the SPIHT scanning sequence is optimized for enhanced resource efficiency. The proposed algorithm is validated using the CHB-MIT Scalp EEG Database, demonstrating compression ratios of ≥1.95 for normal EEG and ≥1.69 for epileptic EEG. Through synthesis and layout design in the 130 nm CMOS technology, the resulting circuit achieves a maximum clock frequency of 57 MHz and occupies an area of 81 kμm2, with an average power consumption of less than 0.71nj per data for 16-bit data. These findings affirm the efficacy of the proposed VLSI structure for achieving efficient lossless/lossy compression of edge EEG data.

Dynamic Cortical Connectivity During Propofol Sedation in Glioma Patients.

The behavioral manifestations and neurophysiological responses to sedation can assist in understanding brain function after neurological damage, and can be described by cortical functional connectivity. Glioma patients may experience neurological deficits that are not clinically detectable before sedation. We hypothesized that patients with gliomas exhibit distinct cortical connectivity patterns compared to non-neurosurgical patients during sedation. This is a secondary analysis of a previously published prospective observational study. Patients scheduled for resection of supratentorial glioma (n=21) or a non-neurosurgical procedure (n=21) under general anesthesia were included in this study. Frontal electroencephalography (EEG) signals were recorded at different sedation levels as assessed by the Observer Assessment of Alertness/Sedation (OAA/S) score. Kernel principal component analysis and k-means clustering were used to determine possible temporal dynamics from the weighted phase lag index characteristics. Ten EEG connectivity states were identified by clustering (76% consistency), each with unique properties. At OAA/S 3, the median (Q1, Q3) occurrence rates of state 6 (glioma group, 0.110 [0.083, 0.155] vs. control group, 0.070 [0.030, 0.110]; P=0.008) and state 7 (glioma group, 0.105 [0.083, 0.148] vs. control group: 0.065 [0.038, 0.090]; P=0.001), which are dominated by beta connectivity, were significantly different between the 2 groups, reflecting differential conversion of the beta band between the left and right brain regions. In addition, the temporal dynamics of the brain's functional connectivity was also reflected in the transition relationships between metastable states. There were differences in EEG functional connectivity, which is dynamic, between the glioma and nonglioma groups during sedation.

SK-MMFMNet: A multi-dimensional fusion network of remote sensing images and EEG signals for multi-scale marine target recognition

Intelligent recognition of multi-scale marine targets remains pivotal in studying marine resources and transportation. Multi-scale marine target recognition faces challenges such as blurred image, noise interference, varied target sizes, and random target positions. However, these hardly affect the judgment of human brain which could adeptly capture multi-scale targets and disregard noise interference. Therefore, this study proposes an innovative approach to recognize multi-scale marine targets through taking full advantages of the texture, color and structural information provided by remote sensing images and the quick classification ability of human brains, called Selective Kernel & Multi-dimensional Multimodal Data Fusion Module Network (SK-MMFMNet), which fuses remote sensing images and electroencephalography (EEG) signals to improve the accuracy of classifying multi-scale marine targets. In this study, we construct a multi-scale marine target dataset, which includes both remote sensing images of islands, wind turbines, and ships and their corresponding EEG signals from subjects while viewing remote sensing images. Then, the proposed approach extends the Multimodal Transfer Module (MMTM) based on attention mechanism to a dual fusion module across channel and spatial dimensions to fusing MobileNetV3 and EEGNet. Also, we embed the Selective Kernel Module into MobileNetV3 for addressing multi-scale features. The average experimental results across the three multi-scale marine target sub-dataset show that SK-MMFMNet exhibited accuracy improvements of 2.88 %, 21.60 %, and 1.08 %, moreover, F1-Score increments of 24.60 %, 162.22 %, and 14.32 % compared to MobileNetV3, EEGNet, and MMTMNet (MMTM-based fusion network). Visual analysis via Grad-CAM demonstrates that benefiting from EEG signals and Selective Kernel Module, our proposed SK-MMFMNet adjusts the network attention to exactly focus on the multi-scale target area, and thus achieves the best performance. Meanwhile, T-SNE visualization also proves the effectiveness of the three fusion modules and EEG signals for feature extraction. This study offers a valuable and promising insight for intelligent recognition of multi-scale marine targets.

The Ultra-Long-Term Sleep study: Design, rationale, data stability and user perspective.

Sleep deprivation and poor sleep quality are significant societal challenges that negatively impact individuals' health. The interaction between subjective sleep quality, objective sleep measures, physical and cognitive performance, and their day-to-day variations remains poorly understood. Our year-long study of 20 healthy individuals, using subcutaneous electroencephalography, aimed to elucidate these interactions, assessing data stability and participant satisfaction, usability, well-being and adherence. In the study, 25 participants were fitted with a minimally invasive subcutaneous electroencephalography lead, with 20 completing the year of subcutaneous electroencephalography recording. Signal stability was measured using covariance of variation. Participant satisfaction, usability and well-being were measured with questionnaires: Perceived Ease of Use questionnaire, System Usability Scale, Headache questionnaire, Major Depression Inventory, World Health Organization 5-item Well-Being Index, and interviews. The subcutaneous electroencephalography signals remained stable for the entire year, with an average participant adherence rate of 91%. Participants rated their satisfaction with the subcutaneous electroencephalography device as easy to use with minimal or no discomfort. The System Usability Scale score was high at 86.3 ± 10.1, and interviews highlighted that participants understood how to use the subcutaneous electroencephalography device and described a period of acclimatization to sleeping with the device. This study provides compelling evidence for the feasibility of longitudinal sleep monitoring during everyday life utilizing subcutaneous electroencephalography in healthy subjects, showcasing excellent signal stability, adherence and user experience. The amassed subcutaneous electroencephalography data constitutes the largest dataset of its kind, and is poised to significantly advance our understanding of day-to-day variations in normal sleep and provide key insights into subjective and objective sleep quality.

Evaluation of the Impact of VR Rural Streetscape Enhancement on Relaxation–Arousal Responses Based on EEG

As the quality of life and the spiritual and cultural well-being of the inhabitants progress, the current rural infrastructure has challenges in adequately addressing the physical and psychological requirements of individuals. This work presents a method for evaluating rural habitats by utilizing electroencephalography (EEG) signals and virtual reality (VR) technology to address the existing gap in physiological data evaluation in rural areas. We choose as experimental images the current situation (C1–C5) scenes of five rural main street nodes as well as the comparative transformation scenes (T1–T5). It then assesses the subjects’ subjective cognition and level of relaxation–arousal responses using the α/β value in the EEG data and the three subjective scale indexes of “Interest”, “Comfort”, and “Vitality”. The study’s findings demonstrated the following: 1. All three scores increased in the transformation scenarios, and subjects’ subjective assessments varied significantly across all five sets of scenes. 2. In all α/β values where significant differences existed, every electrode demonstrated a relaxation response in the transformation scenes (T1–T5) compared to the current scenes (C1–C5), whereas the T8 electrode demonstrated the only arousal response. 3. The Pz electrode in the parietal lobe area was found to be the most sensitive to the visual response to the green landscape and the form of the building façade along the street, and the T8 electrode in the right temporal lobe area was the most sensitive to the response to the overall perception of the surrounding environment, according to a comparison of the longitudinal α/β value. More options for rural streetscape design as well as fresh insights and methodologies for assessing the rural human environment in the future are anticipated from this study.

Brain-computer interface-based hand exoskeleton with bidirectional long short-term memory methods

It takes at least 3 months to restore hand and arm function to 70% of its original value. This condition certainly reduces the quality of life for stroke survivors. The effectiveness in restoring the motor function of stroke survivors can be improved through rehabilitation. Currently, rehabilitation methods for post-stroke patients focus on repetitive movements of the affected hand, but it is often stalled due to the lack of professional rehabilitation personnel. This research aims to design a brain-computer interface (BCI)-based exoskeleton hand motion control for rehabilitation devices. The Bidirectional long short-term memory (Bi-LSTM) method performs motion classification for the ESP32 microcontroller to control the movement of the DC motor on the exoskeleton hand in real-time. The statistical features, such as mean and standard deviation from the sliding windows process of electroencephalograph (EEG) signals, are used as the input for Bi-LSTM. The highest accuracy at the validation stage was obtained in the combination of mean and standard deviation features, with the highest accuracy of 91% at the offline testing stage and reaching an average of 90% in real-time (80%-100%). Overall, the control system design that has been made runs well to perform movements on the hand exoskeleton based on the classification of opening and grasping movements.

Generative Listener EEG for Speech Emotion Recognition Using Generative Adversarial Networks with Compressed Sensing.

Currently, emotional features in speech emotion recognition are typically extracted from the speeches, However, recognition accuracy can be influenced by factors such as semantics, language, and cross-speech datasets. Achieving consistent emotional judgment with human listeners is a key challenge for AI to address. Electroencephalography (EEG) signals prove to be an effective means of capturing authentic and meaningful emotional information in humans. This positions EEG as a promising tool for detecting emotional cues conveyed in speech. In this study, we proposed a novel approach named CS-GAN that generates listener EEGs in response to a speaker's speech, specifically aimed at enhancing cross-subject emotion recognition. We utilized generative adversarial networks (GANs) to establish a mapping relationship between speech and EEGs to generate stimulus-induced EEGs. Furthermore, we integrated compressive sensing theory (CS) into the GAN-based EEG generation method, thereby enhancing the fidelity and diversity of the generated EEGs. The generated EEGs were then processed using a CNN-LSTM model to identify the emotional categories conveyed in the speech. By averaging these EEGs, we obtained the event-related potentials (ERPs) to improve the cross-subject capability of the method. The experimental results demonstrate that the generated EEGs by this method outperform real listener EEGs by 9.31% in cross-subject emotion recognition tasks. Furthermore, the ERPs show an improvement of 43.59%, providing evidence for the effectiveness of this method in cross-subject emotion recognition.

A Study of Major Depressive Disorder Based on Resting-State Multilayer EEG Function Network

Depression is a complex mental disease with its pathological mechanism unclear. To depict the complete picture of the abnormal information interaction in a depressed brain, this study is the first to apply fully connected multilayer brain functional (FCMBF) network framework and proposed composite FCMBF (CFCMBF) network framework, combined with graph theory to analyze the within-frequency coupling (WFC) and cross-frequency coupling (CFC) of sensor-layer and source-layer electroencephalography (EEG) signals in relevant subjects. Results showed that in the sensor-layer FCMBF network, depressive patients showed significantly reduced functional connectivity, as well as abnormal global and local information processing abilities of the network, and these network properties were significantly correlated with depressive symptoms. In addition, from the perspective of depression recognition, we found that the sensor-layer CFCMBF network could achieve better classification accuracy, especially when using the overlapping degree of node under the right center region, its accuracy could reach <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$86.88\% \pm 9.25 \%$</tex-math> </inline-formula> . More importantly, the construction of the CFCMBF network has higher time efficiency and less information loss, since it not only measures the WFC and CFC between brain region representative signals (BRRSs) extracted from different brain regions, but also measures these two couplings between all nodes within each brain region. Although the FCMBF network contains more complete information by calculating WFC and CFC between all nodes distributed in each region, it will result in an enormous computational cost. In summary, this study proved the utility of multilayer brain network in revealing the abnormal brain interaction patterns of depression, and our proposed method might provide methodological support for efficient depression recognition research based on multilayer brain networks.

A Comparative Study on Feature Extraction Techniques for the Discrimination of Frontotemporal Dementia and Alzheimer's Disease with Electroencephalography in Resting-State Adults.

Early-stage Alzheimer's disease (AD) and frontotemporal dementia (FTD) share similar symptoms, complicating their diagnosis and the development of specific treatment strategies. Our study evaluated multiple feature extraction techniques for identifying AD and FTD biomarkers from electroencephalographic (EEG) signals. We developed an optimised machine learning architecture that integrates sliding windowing, feature extraction, and supervised learning to distinguish between AD and FTD patients, as well as from healthy controls (HCs). Our model, with a 90% overlap for sliding windowing, SVD entropy for feature extraction, and K-Nearest Neighbors (KNN) for supervised learning, achieved a mean F1-score and accuracy of 93% and 91%, 92.5% and 93%, and 91.5% and 91% for discriminating AD and HC, FTD and HC, and AD and FTD, respectively. The feature importance array, an explainable AI feature, highlighted the brain lobes that contributed to identifying and distinguishing AD and FTD biomarkers. This research introduces a novel framework for detecting and discriminating AD and FTD using EEG signals, addressing the need for accurate early-stage diagnostics. Furthermore, a comparative evaluation of sliding windowing, multiple feature extraction, and machine learning methods on AD/FTD detection and discrimination is documented.

Gradient Boosting for Predicting the Relation Between Bio-medical Signals and Seizures Using LGBM and XGBoost

Background and aim: In recent years, research in the fields of brain-computer interfacing techniques and related areas are developing at a very rapid rate with the help of exploding of Artificial Intelligence, Machine Learning and Deep Learning. A new concept of Gradient Boosting has become popular research area among the researchers related to the field of automatic classification of Electroencephalograph (EEG) signals for predication of mental health issues like seizures. Methods: However effective feature extraction from EEG and accurately classify them with efficient classifiers is still an important task and attracted wide attention in this area. Therefore in this paper, we presented the detailed mathematical analysis of these methods and ensemble learnings based EEG signals classification method for seizures classification in EEG using Extreme Gradient Boosting Model such as Light Gradient Boosting Machine Learning (LGBM) and XGBoost. Results: Time-frequency domain based non-linear features are selected from preprocessed EEG Dataset, and PCA (Principal Component Analysis) is used for dimensionality reduction for features engineering, then optimized feature based training and testing is done for two class classification in ensemble learning method i.e. LGBM and XGBoost. Finally, both models are tested with dataset of University of Bonn, Germany to classify the signals. Conclusions: In addition this paper highlights the Correlation Analysis Methodology to Identify Strong Predictor and Attributes Correlation-based Attribute Ranking for the Feature Engineering which has proved to be more efficient in EEG signals Classification and provide comparative analysis with other existing models for performance evaluation in terms of accuracy which is 87.34 and 92.31 for LBGM and XGBoost, sensitivity of 85.21 and 90.18 and specificity of 83.0 and 90.04 for LBGM and XGBoost.

Processing alcoholism eeg signals using neural networks

Alcoholism is one of the most common diseases, which creates problems for contemporary society. To date, there are still no standardized methods and tests to detect this disease. This study presents a new methodology based on an integrated approach to the study of Electroencephalography (EEG) signals, using a combination of nonlinear dynamics and machine learning methods. We applied signal encoding in the form of a 16-channel image, which was fed into the ResNet20 and GoogLeNet models for training. Wavelet transforms and fast Fourier transforms were considered for data preprocessing. In addition to the application of neural networks, the study of EEG signals was carried out by computing Lyapunov exponents and multiscale entropy. The question of choosing the most effective maternal wavelet was previously investigated. The wavelet transformations of Morlet, Meyer, Mexican Hat, Daubechies, and Gauss were studied. A spectrum of five Lyapunov exponents was calculated using the Sano – Savada method. To confirm the reliability of the results, the senior Lyapunov exponent was previously calculated by three more methods: Wolf, Kantz, Rosenstein. Previously, the methods were tested on such classical problems as the hyperchaotic Hénon map, logistic map, Rössler attractor and Lorenz attractor. The use of several approaches for the analysis of EEG signals confirmed the reliability of the results. The study demonstrates that the GoogLeNet model is more efficient than ResNet50. The experiment results showed that, based on the developed methodology, the GoogLeNet model is more efficient than ResNet50. The results of the experiment show that the wavelet transform, compared to the Fourier transform, gives higher accuracy for the analysis of EEG signals of patients with alcoholism.

Detection of multitask mental workload using gamma band power features

Cognitive fatigue occurs in various situations and is an essential condition to detect. In this study, how single and multi-tasking tests affect cognitive workload was examined, and multi-tasking was detected using electroencephalography (EEG) signals. In the cognitive workload paradigm, single-task tests with blocks 1 and 2 and multi-tasking tests with block 3 were created. EEG signals obtained from these blocks were treated as different frequency bands and lengths, and binary classification was performed. Two binary classifications were made: block 1–block 3 and block 2–block 3. According to the results, the highest classification accuracy for block 1–block 3 was obtained as 97.11% using the gamma frequency band and 5-s EEG length. For block 2–block 3, the highest classification accuracy was obtained as 90.88% using the gamma frequency band and 5-s EEG length. As a result, this study distinguished multi-tasking and single task with high classification accuracy. The developed model can be used to detect attention deficit and inability to focus. In addressing the prevalent challenges of distinguishing cognitive fatigue in single—task versus multitasking scenarios, our study offers a new method, which achieve a remarkable accuracy rate, thereby illuminating a new path in the research of cognitive fatigue.

Schizophrenia Patient Classification with Long Short-Term Memory Analysis of Electroencephalography Signals

Schizophrenia Patient Classification with Long Short-Term Memory Analysis of Electroencephalography Signals

A machine learning based depression screening framework using temporal domain features of the electroencephalography signals.

Depression is a serious mental health disorder affecting millions of individuals worldwide. Timely and precise recognition of depression is vital for appropriate mediation and effective treatment. Electroencephalography (EEG) has surfaced as a promising tool for inspecting the neural correlates of depression and therefore, has the potential to contribute to the diagnosis of depression effectively. This study presents an EEG-based mental depressive disorder detection mechanism using a publicly available EEG dataset called Multi-modal Open Dataset for Mental-disorder Analysis (MODMA). This study uses EEG data acquired from 55 participants using 3 electrodes in the resting-state condition. Twelve temporal domain features are extracted from the EEG data by creating a non-overlapping window of 10 seconds, which is presented to a novel feature selection mechanism. The feature selection algorithm selects the optimum chunk of attributes with the highest discriminative power to classify the mental depressive disorders patients and healthy controls. The selected EEG attributes are classified using three different classification algorithms i.e., Best- First (BF) Tree, k-nearest neighbor (KNN), and AdaBoost. The highest classification accuracy of 96.36% is achieved using BF-Tree using a feature vector length of 12. The proposed mental depressive classification scheme outperforms the existing state-of-the-art depression classification schemes in terms of the number of electrodes used for EEG recording, feature vector length, and the achieved classification accuracy. The proposed framework could be used in psychiatric settings, providing valuable support to psychiatrists.

Tripolar concentric ring electrodes for capturing localised electroencephalography signals during sleep.

By design, tripolar concentric ring electrodes (TCRE) provide more focal brain activity signals than conventional electroencephalography (EEG) electrodes placed further apart. This study compared spectral characteristics and rates of data loss to noisy epochs with TCRE versus conventional EEG signals recorded during sleep. A total of 20 healthy sleepers (12 females; mean [standard deviation] age 27.8 [9.6] years) underwent a 9-h sleep study. Participants were set up for polysomnography recording with TCRE to assess brain activity from 18 sites and conventional electrodes for EEG, eyes, and muscle movement. A fast Fourier transform using multitaper-based estimation was applied in 5-s epochs to scored sleep. Odds ratios with Bonferroni-adjusted 95% confidence intervals were calculated to determine the proportional differences in the number of noisy epochs between electrode types. Relative power was compared in frequency bands throughout sleep. Linear mixed models showed significant main effects of signal type (p < 0.001) and sleep stage (p < 0.001) on relative spectral power in each power band, with lower relative spectral power across all stages in TCRE versus EEG in alpha, beta, sigma, and theta activity, and greater delta power in all stages. Scalp topography plots showed distinct beta activation in the right parietal lobe with TCRE versus EEG. EEG showed higher rates of noisy epochs compared to TCRE (1.3% versus 0.8%, p < 0.001). TCRE signals showed marked differences in brain activity compared to EEG, consistent with more focal measurements and region-specific differences during sleep. TCRE may be useful for evaluating regional differences in brain activity with reduced muscle artefact compared to conventional EEG.

A brain functional network feature extraction method based on directed transfer function and graph theory for MI-BCI decoding tasks.

The development of Brain-Computer Interface (BCI) technology has brought tremendous potential to various fields. In recent years, prominent research has focused on enhancing the accuracy of BCI decoding algorithms by effectively utilizing meaningful features extracted from electroencephalographic (EEG) signals. This paper proposes a method for extracting brain functional network features based on directed transfer function (DTF) and graph theory. The method incorporates the extracted brain network features with common spatial pattern (CSP) to enhance the performance of motor imagery (MI) classification task. The signals from each electrode of the EEG, utilizing a total of 32 channels, are used as input signals for the network nodes. In this study, 26 healthy participants were recruited to provide EEG data. The brain functional network is constructed in Alpha and Beta bands using the DTF method. The node degree (ND), clustering coefficient (CC), and global efficiency (GE) of the brain functional network are obtained using graph theory. The DTF network features and graph theory are combined with the traditional signal processing method, the CSP algorithm. The redundant network features are filtered out using the Lasso method, and finally, the fused features are classified using a support vector machine (SVM), culminating in a novel approach we have termed CDGL. For Beta frequency band, with 8 electrodes, the proposed CDGL method achieved an accuracy of 89.13%, a sensitivity of 90.15%, and a specificity of 88.10%, which are 14.10, 16.69, and 11.50% percentage higher than the traditional CSP method (75.03, 73.46, and 76.60%), respectively. Furthermore, the results obtained with 8 channels were superior to those with 4 channels (82.31, 83.35, and 81.74%), and the result for the Beta frequency band were better than those for the Alpha frequency band (87.42, 87.48, and 87.36%). Similar results were also obtained on two public datasets, where the CDGL algorithm's performance was found to be optimal. The feature fusion of DTF network and graph theory features enhanced CSP algorithm's performance in MI task classification. Increasing the number of channels allows for more EEG signal feature information, enhancing the model's sensitivity and discriminative ability toward specific activities in brain regions. It should be noted that the functional brain network features in the Beta band exhibit superior performance improvement for the algorithm compared to those in the Alpha band.