Search

Steady-state visually evoked potential (SSVEP) refers to the brain’s response to visual stimuli at different frequencies and is widely used in brain-computer interfaces (BCIs). Despite their potential, SSVEP-based BCIs face significant challenges in real-world applications, particularly in controlling assistive devices, prosthetics, and communication systems for individuals with disabilities. The challenges include suboptimal frequency detection accuracy and long calibration periods, which limit the effectiveness of SSVEPs and contribute to increased visual fatigue during extended sessions. This review addresses these challenges by offering an overview of feature extraction methods for SSVEP recognition. It includes mathematical explanations of the processes, highlights their strengths and limitations, compares them, and discusses future directions. Feature extraction techniques can be categorized into three groups: calibration-free, calibration-based, and deep learning. While calibration-free methods require minimal data, they typically achieve lower accuracy than calibration-based methods, which rely on training datasets to provide better accuracy and information transfer rates; however, the lengthy training sessions often make these algorithms unsuitable for everyday use. On the other hand, deep learning approaches have improved accuracy and adaptability by automatically extracting complex features from data and accommodating varying conditions, even with shorter time windows. However, they require large amounts of data for training to improve accuracy. To address this issue, both calibration-based and deep learning methods can benefit from transfer learning, which alleviates the need for extensive training data by sharing knowledge across subjects. This approach enhances recognition accuracy and reduces reliance on subject-specific training, ultimately making these methods more practical for real-world applications.

Brain-computer interface (BCI) technology can be an alternative solution for people with severe disabilities who have lost their movement and communication abilities. This study proposes a self-paced paradigm using visual flicker imagination as an alternative modality for steady-state evoked potential (SSVEP)-based BCI systems during visual-fatigue periods. This study comprises two main parts. First, we propose a visual-stimulus pattern with sequential numbers to exhibit the responses and observe the electroencephalogram (EEG) features of visual flicker imagination in offline testing. We confirmed that EEG signals in the occipital and frontal areas responded to visual flicker imagination. Second, an EEG-based BCI was demonstrated for assistive-communication system control using visual flicker imagination. Two experiments were conducted to verify the efficiency of the proposed system. Twelve volunteers participated in the experiment. The results revealed that the proposed SSVEP method and algorithm yielded average classification accuracies of approximately 93.8% and 77.1% for SSVEP and visual flicker imagination in real-time processing, respectively. The visual flicker-imagination approach can be utilized for real-time EEG-based BCI systems and can further enhance the efficiency of conventional SSVEP-based BCI systems during visual fatigue. We plan to improve and verify the proposed BCI system for practical use to control augmentative and alternative communication for people with severe disabilities.

With the development of the brain-computer interface, it is possible to assess the image quality based on an electroencephalogram (EEG) signal, which is essential to construct a quality assessment system that accords with the characteristics of the human visual system. However, as the complexity of image degradation levels rises, the analysis of brain features becomes progressively more challenging, leading the neural mechanisms still unclear. In this paper, an image quality assessment experiment with three levels of degradation to explore the neural mechanisms of subjects when confronted with images degraded at multiple levels. Subsequently, we proposed a dynamic model combined with Topological Data Analysis (TDA) and Grey Theory to extract the feature of brain response to different distortion-level images, which is called grey-topological data analysis (Grey-TDA) in this paper. The results indicate that the proposed method is effective in identifying brain responses under stimuli of varying image qualities, which contributes to the research of image quality assessment from the perspective of brain cognition.

Emotion recognition based on electroencephalography (EEG) has become an important topic in affective computing. However, current research does not utilize the relationship between channel frequency bands, and the trainable parameters and FLOPs of the models of mainstream methods from the past two years are very large. To solve this problem, this paper proposes a new emotion recognition framework to learn the frequency band relationship of EEG and the temporal-spatial representation of the frequency domain. First, we extract the differential entropy features of the frequency band at a frequency resolution of 2 Hz to retain rich frequency band information. Second, we design a temporal convolution and 3D convolutional neural network (TC3DNet) that is a fusion of Temporal Convolutional Network (TCN) and 3D convolutional neural network (3DCNN) in series. TCN is used to learn frequency band relationships, and 3DCNN is further used to learn temporal-spatial representation. The average recognition accuracy of TC3DNet in the benchmark EEG emotion recognition SEED dataset and SEED-IV dataset is 98.48% and 95.30%, which is better than current state-of-the-art methods. The impact of different frequency band divisions on the proposed framework is also explored. Furthermore, this study visualizes the extracted features, thereby enhancing the interpretability of our work.

Clinical methods for dementia detection are expensive and prone to human errors. Despite various computer-aided methods using electroencephalography (EEG) signals and artificial intelligence, a reliable detection of Alzheimer’s disease (AD) remains a challenge. The existing EEG-based machine learning models have limited performance or high computation complexity. Hence, there is a need for an optimal deep learning model for the detection of AD. This paper proposes a low-complexity EEG-based AD detection CNN called LEADNet to generate disease-specific features. LEADNet employs spatiotemporal EEG signals as input, two convolution layers for feature generation, a max-pooling layer for asymmetric spatiotemporal redundancy reduction, two fully-connected layers for nonlinear feature transformation and selection, and a softmax layer for disease probability prediction. Different quantitative measures are calculated using an open-source AD dataset to compare LEADNet and four pre-trained CNN models. The results show that the lightweight architecture of LEADNet has at least a 150-fold reduction in network parameters and the highest testing accuracy of 99.24% compared to pre-trained models. The investigation of individual layers of LEADNet showed successive improvements in feature transformation and selection for detecting AD subjects. A comparison with the state-of-the-art AD detection models showed that the highest accuracy, sensitivity, and specificity were achieved by the LEADNet model.

The main goal of this paper is to introduce a Motor Imagery (MI) classification system for electroencephalography (EEG) that is extremely precise. To achieve this goal, we propose using a feature-extracted deep one-dimension (1D) convolutional neural network (CNN) which provides a model that can be further improved through hyperheuristic multi-objective evolutionary search. We can improve the classification performance by training this deep CNN model with feature-extracted data from the Physionet MI dataset. We also present a semi-deep fine-tuning approach that can yield improvements with just four epochs. Our findings using the Physionet MI dataset illustrate that the approach we suggest surpasses most contemporary techniques used for classifying EEG signals. Our system is computationally efficient and can be trained using reliable EEG data for individual patients, allowing for accurate classification of their EEG records. Because of its straightforward and parameter-independent characteristics, our system is versatile and can be utilized with any EEG dataset.

Educational institutions have been adopting business simulation games as a resource for developing skills and competencies. However, the real-time study of aspects related to learning during their use still needs to further investigation. This work aims to present contributions of a new methodology that considers the analysis of the player’s experience from the physiological, behavioural, and psychological perspectives, supported by biometric devices. Its validation was based on data collection conducted at a university in Brazil, with ten participants during a game monitored by Electroencephalogram and Eye Tracking technologies, relating the behaviour of the cerebral cortex with areas of greater interest and gaze fixation of the players. Additionally, pre-and post-test questionnaires highlighted personal behaviours and collective patterns influenced by aspects such as tutorial and game design elements, prior knowledge, ergonomic issues, cognitive strategy definition, organization of executive functions, and memory formation. The results show that the proposed model can be beneficial in assessing the potential of this type of organizational environment simulation, both for those who use it in the teaching process and for serious game developers.

This study explores the convergence of biometric analytics and machine learning in online education, where the level of student participation directly impacts academic achievement. In this study, various machine learning models were employed to identify cognitive states using eye-tracking and electroencephalogram data, which can provide quantitative indicators of cognitive activity. A comparative analysis of convolutional neural network (CNN), logistic regression, decision tree, random forest, gradient boosting, support vector machine, and K-nearest neighbors (KNN) models was conducted. The receiver operating characteristic curve was used as a benchmark to assess the effectiveness of the models in interpreting the subtle patterns in biometric feedback. The results show that the CNN model exhibited outstanding performance, achieving an area under the curve value of 0.98 in interpreting intricate data that reflect the levels of student involvement. In addition, the gradient boosting and KNN models exhibited noteworthy accuracy, and the logistic regression and random forest models realized a crucial equilibrium by providing a high level of interpretability and strong performance. This is particularly important for educational applications where the model’s rationale must be transparent. The findings of this study support the idea that machine learning can considerably improve remote learning platforms, leading to more customized and interactive educational experiences. However, the findings also acknowledge relevant ethical concerns and the need to understand these advanced technologies comprehensively.

The electroencephalogram (EEG) is a non-invasive technique with high temporal resolution that has become the research frontier of brain-computer interface (BCI) systems. It is widely used in medical rehabilitation, gaming, and other industries. However, decoding EEG signals remains a challenging task. A network called MSCARNet, which combines multi-scale convolution and Riemannian geometry, was proposed for classifying motor imagery based on EEG. The network is supplemented by an attention mechanism and sliding window technique. The MSCARNet utilizes sliding windows to expand data dimensions and multiple convolution kernels to obtain spatial and temporal features. These features are then mapped to Riemannian space and undergo bilinear mapping and logarithmic operations for dimensionality reduction. This approach is beneficial in reducing the impact of noise and outliers and provides convenience for classification. Subject-dependent and subject-independent experiments were conducted using the BCI-IV-2a dataset to validate the effectiveness of the MSCARNet. The results show that the accuracy improved by approximately 4% compared to existing state-of-the-art methods. The hybrid network based on Riemannian space can effectively improve the accuracy of EEG motor imagery classification tasks without excessive preprocessing.

High-performance automated detection methods for epilepsy play a crucial role in clinical diagnostic support. To address the challenge of effectively extracting features from epileptic EEG signals, characterized by strong spontaneity and complexity, a novel feature extraction approach based on Window Kullback-Leibler Divergence (WKLD) is proposed, coupled with discrete wavelet analysis for EEG signal feature extraction. Then, a Residual Multidimensional Taylor Network (ResMTN) classifier is applied for epilepsy state classification. Experimental results demonstrate an accuracy of 98% in classifying EEG signals during seizure and interictal periods, with both specificity and sensitivity reaching 98.18%, outperforming existing widely-used feature extraction and classification methods.

In the rapidly evolving domain of brain-to-brain communication, safeguarding the transmission of information against adversarial threats is paramount. This study introduces an advanced approach to enhance the resilience and security of brain-to-brain communication systems utilizing electroencephalogram data against such threats through adversarial neural network training. Concentrating on event-related potentials and employing a diverse collection of eight datasets, our research rigorously evaluates and optimizes the system’s defense mechanisms against adversarial manipulations. We specifically target the optimization of trial durations and sampling rates to bolster system security. Our findings reveal a marked improvement in the system’s defensive capabilities, demonstrated by a significant increase in adversarial accuracy by 17% and enhancement in the area under the receiver operating characteristic curve by 0.12 points. These results underscore the efficacy of our approach in fortifying brain-to-brain communication systems against sophisticated cyber threats, marking a significant step forward in the secure and robust transmission of neural signals.

Accurate identification of electroencephalogram (EEG) signals forms the basis for the development and application of brain–computer interface (BCI) devices. Signal preprocessing is an essential part of most EEG classification and recognition systems.In this study, we proposed a method called Super-resolution level separation (SRLS) to add dimensions of information to the classification model. First, we calculated the correlation between the EEG signals acquired by each channel and divided these channels into multiple levels. Next, we used a super-resolution method to calculate common EEGs (C-EEGs) acquired by the channels. In addition, we designed an EEG–split-informer (ES-informer) model based on an informer model to enable small-sample users to obtain highly fitting C-EEGs. We then calculated the difference between the C-EEG and true EEG (T-EEG) to obtain the unique EEGs (U-EEGs) acquired by each channel, thus, adding dimensions to the data inputted to the classification model. Utilizing the 2008 2a motor imagery (MI) EEG dataset from the BCI Competition and the P300 paradigm data collected, EEGNet was employed as the classification model to validate the efficacy of the proposed SRLS method. The results of the experiments indicated that the SRLS method augmented the input dimension of the model and amplified the classification accuracy by over 7% for MI and 3.6% for P300. These findings demonstrate that SRLS is capable of enhancing the recognition accuracy of EEG.

In the field of neuroscience, the electroencephalogram (EEG) is a crucial indicator of emotion. The EEG emotion recognition method based on domain adaptation (DA) has good objectivity and high time resolution and is the preferred method to study the brain’s response to emotional stimuli. However, due to the obvious instability of EEG emotion characteristics, it is difficult to predict the emotion corresponding to EEG signals of cross-subjects by a model that combines all source domains into a single source. In order to solve the problem of cross-subject emotion analysis, we propose an EEG emotion recognition net with a cross-subject multi-source adaptive method (MASTF-net), where EEG features of different subjects are regarded as different domains. Through analyzing the invariance of the target domain and the uniqueness of the source domain, this method realizes the emotional analysis of different objects according to the spatio-temporal images and frequency domain information. First, features of EEG image are extracted from frequency and time dimensions. Secondly, combined with the serialized EEG frequency characteristics of local brain regions, independent classification module are established for different domains to recognize the emotion feature distribution of different subjects. In addition, a feature extraction method of differential entropy(DE) data of EEG is proposed based on frequency band division, which can provide stable feature input for our network structure. Finally, experiments are conducted on the SEED dataset. The experimental results show that our method has better classification accuracy in the experiment on the problem of cross multiple subjects. MASTF-net is superior to other relevant methods and models in multi-source domain. On the issue of cross subject emotion analysis, the highest accuracy of our method can reach $88.19\%$ .

Brain–computer interfaces (BCI) facilitate communication between the human brain and computational systems, additionally offering mechanisms for environmental control to enhance human life. The current study focused on the application of BCI for communication support, especially in detecting unspoken answers to questions. Utilizing a multistage neural network (MSNN) replete with convolutional and pooling layers, the proposed method comprises a threefold approach: electroencephalogram (EEG) measurements, EEG feature extraction, and answer classification. The EEG signals of the participants are captured as they mentally respond with “yes” or “no” to the posed questions. Feature extraction was achieved through an MSNN composed of three distinct convolutional neural network models. The first model discriminates between the EEG signals with and without discernible noise artifacts, whereas the subsequent two models are designated for feature extraction from EEG signals with or without such noise artifacts. Furthermore, a support vector machine is employed to classify the answers to the questions. The proposed method was validated via experiments using authentic EEG data. The mean and standard deviation values for sensitivity and precision of the proposed method were 99.6% and 0.2%, respectively. These findings demonstrate the viability of attaining high accuracy in a BCI by preliminarily segregating the EEG signals based on the presence or absence of artifact noise and underscore the stability of such classification. Thus, the proposed method manifests prospective advantages of separating EEG signals characterized by noise artifacts for enhanced BCI performance.

One approach to therapy and training for the restoration of damaged muscles and motor systems is rehabilitation. EEG-assisted Brain-Computer Interface (BCI) may assist in restoring or enhancing ‘lost motor abilities in the brain. Assisted by brain activity, BCI offers simple-to-use technology aids and robotic prosthetics. This systematic literature review aims to explore the latest developments in BCI and motor control for rehabilitation. Additionally, we have explored typical EEG apparatuses that are available for BCI-driven rehabilitative purposes. Furthermore, a comparison of significant studies in rehabilitation assessment using machine learning techniques has been summarized. The results of this study may influence policymakers’ decisions regarding the use of EEG equipment, particularly wireless devices, to implement BCI technology. Moreover, the literature review results offer suggestions for further study and new research areas. We plan to identify the additional characteristics of each EEG equipment and determine which one is most suited for each industry by measuring the user experience based on various devices in future research.

Brain Epilepsy seizure is a critical disorder, which is an uncontrolled burst of electrical activity of brain. The early detection of brain seizure can save the life of humans. The electroencephalogram (EEG) signals may be used to automatically identify brain seizures, which is one of the most prominent solutions for this issue. However, the conventional methods are failed to classify the brain seizure effectively. So, this work implemented the Brain Epilepsy Seizure-Detection-Network (BESD-Net) using deep learning, recurrent learning properties. Initially, the dataset pre-processing is performed, which eliminates the noise, unwanted data from EEG dataset. Then, the deep learning based customized convolution neural network (CCNN) is trained on the pre-processed EEG data for precise extraction of disease correlated features. The machine learning based exhaustive random forest (ERF) feature selection is used to optimize the features obtained from the CCNN, which are highly correlated with disease dependent properties. In conclusion, the recurrent neural network (RNN) based bi-directional long short-term memory (BLSTM) is used in order to detect brain seizures from the chosen ERF features. Training and testing of suggested methodology had made use of CHB-MIT Scalp EEG Database. The aforementioned model has achieved the values of 98.36%, 97.54%, 97.91%, 98% and 95.08% respectively for precision, sensitivity, F1-Score, accuracy and specificity. The findings of the simulations demonstrate that the suggested BESD-Net led to superior performance when compared to the technologies that are already in use.

The connectivity properties between EEG signaling channels were found to play an important role in emotion recognition. During the generation and change of emotions, the connectivity between EEG signal channels is not only manifested in the spatial domain at a specific time, but also interconnected between different time intervals, which is often overlooked. A novel approach is proposed to exploit the shapes of spatial-temporal connectivity for EEG-based emotion recognition. By quantifying the connectivity strength of EEG signal channels within and across time intervals, spatial-temporal connectivity features are extracted, and these features can be represented by shapes in the three-dimensional space of EEG signals. Through these shapes, a mapping of different emotional states and brain activity is constructed. Subsequently, a parallel multi-scale convolutional neural network is employed to extract discriminative features from these connectivity shapes, facilitating the classification and identification of emotional states. Experimental results on the DEAP dataset show that our method achieves excellent performance, with an average classification accuracy of 93.25% and 93.16% for the two emotion dimensions of valence and arousal, respectively.

In the case of paralysis with visual and tactile impairments, brain–computer interfaces (BCIs) based on auditory and mental imagery paradigms are alternative methods for controlling external devices. This study demonstrates the use of a hybrid BCI via auditory stimulation and speech imagination for assistive technology. The proposed auditory BCI using Thai vowel and numeral stimulus patterns as well as multi-loudspeaker position settings for multi-command BCI are investigated. To avoid auditory stimulation during resting periods, a speech imagery method is used to enable an audio stimulator. We observe the classification efficiency of speech imagery and auditory BCIs from selected electroencephalogram channels using the proposed algorithms. We examine the efficiency of using the proposed BCI in the presence of background noise (speech). One command is created using the proposed speech-imagination paradigm. Four commands are created using the proposed auditory stimulation paradigm. We design an experiment to verify the proposed BCI paradigm and classification algorithms for real-time processing. The results show that the average classification accuracy of the proposed auditory BCI using numeral stimuli and scatter patterns without speech noise is 72.2% to 83.3%, respectively. The efficiency under background noise is approximately two times lower than that without background noise.

Sleep is essential for physical and mental health. Polysomnography (PSG) procedures are labour-intensive and time-consuming, making diagnosing sleep disorders difficult. Automatic sleep staging using Machine Learning (ML) - based methods has been studied extensively, but frequently provides noisier predictions incompatible with typical manually annotated hypnograms. We propose an energy optimization method to improve the quality of hypnograms generated by automatic sleep staging procedures. The method evaluates the system’s total energy based on conditional probabilities for each epoch’s stage and employs an energy minimisation procedure. It can be used as a meta-optimisation layer over the sleep stage sequences generated by any classifier that generates prediction probabilities. The method improved the accuracy of state-of-the-art Deep Learning models in the Sleep EDFx dataset by 4.0% and in the DRM-SUB dataset by 2.8%.

Brain computer interface (BCI) systems for neurorehabilitation have received increasing attention over the past decade. These systems provide an alternative approach to restore lost motor functions in stroke patients by inducing brain plasticity. To utilize BCI systems for stroke rehabilitation, user movement intentions from electroencephalogram are used as triggers for rehabilitation tools such as electrical stimulators. A reliable movement intention detection plays a vital role in an effective rehabilitation to establish a relation between movement intentions and corresponding feedback from a rehabilitation tool. We therefore propose a novel movement intention detection algorithm based on time series shapelets. We collected a dataset of subjects performing a self-pace ankle dorsiflexion and tested the algorithm in both classification and pseudo-online detection. The classification results demonstrate that our proposed algorithm significantly outperforms six other algorithms and achieves the highest classification performance with F1-score of 0.82. For pseudo-online detection, our algorithm gains very high performance in all performance metrics with 69% True Positive Rate (TPR), 8 False Positives per minute (FPs/min), and 44 ms Detection Latency (DL). Our proposed algorithm does not only provide competitive performance to state-of-the-art algorithms in terms of TPR, but also maintain low FPs/min and DL. In addition, the DL of our proposed algorithm is low enough to induce effective brain plasticity for neurorehabilitation. These promising results enlighten the development of asynchronous BCI systems based on time series mining techniques to enhance stroke rehabilitation results.

Depression (DP) and schizophrenia (SCZ) are both highly prevalent psychiatric disorders, and their diagnosis depends on the examination of symptoms and clinical tests, which can be subjective. As a measure of real-time neural activity, Electroencephalographic (EEG) has shown its usability to classify people either as normal or as having DP or SCZ, but automatic classification between the three categories (DP, SCZ and the normal) was rarely reported. Here, we propose an automatic diagnostic framework based on a convolutional neural network called the Multi-Channel Frequency Network (MUCHf-Net), which automatically learns feature representations of EEGs that characterize them as normal, DP, or SCZ. Two EEG databases were used in this study, the first one contains EEGs from 300 individuals (DP: 100, SCZ: 100, normal: 100) collecting from our hospital, and the second contains EEGs from 30 individuals (DP: 10, SCZ: 10, normal: 10) from public available datasets, and the spectrum matrices from these multi-channel EEGs were feed into MUCHf-Net. The results showed that: (1) MUCHf-Net accurately distinguished normal EEGs from DP or SCZ EEGs (accuracy: 91.12%; F1 score: 0.8947); (2) low-frequency bands (delta, theta, alpha) contributed the most important information to the classification model; (3) features located in the frontal and parietal lobes contributed more than other regions did; (4) MUCHf-Net fine-tuned on public datasets also had high classification accuracy: 87.71% (triple: normal, SCZ or DP) and 79.27% (binary: psychiatric disorders (DP or SCZ) or normal). Our study shows that deep learning has the potential to become an important tool for assisting in the diagnosis of psychiatric disorders.

In this paper we deal with a simplified anisotropic rat head phantom development and the investigation of the influence of the anisotropic white matter on electroencephalogram source localization. The proposed phantom is based on the cubic cross cell composition combined with agar mixture to set desired electrical conductivity anisotropic ratio. For the fabrication of the phantom, the 3D printed technology is exploited. Starting from a real rat brain, we proposed a simplified brain model incorporating the actual dimensions, shape and conductivity parameters of both grey and white matter containing simultaneously relevant deep-brain electrical signal sources. Five testing dipoles were located in the areas corresponding to the active brain regions. A single dipole localization error was calculated by comparing an inverse solution with a dipole position obtained from a computer tomography image. Neglecting anisotropy had a rather weak effect on localization error of a single testing dipole in our model. The reliability map was computed and interpreted in terms of spatial similarity between distributed inverse solutions involving isotropic and anisotropic forward models. We found spatially specific error increases located close to the electrodes and in the vicinity of anisotropic compartment. Hence, areas to be most sensitive to neglecting anisotropy in our model were identified.

The most common neurological brain issue is Alzheimer’s disease, which can be diagnosed using a variety of clinical methods. However, the electroencephalogram (EEG) is shown to be effective in detecting Alzheimer’s disease. The purpose of this research is to develop a computer-aided diagnosis system that can diagnose Alzheimer’s disease using EEG data. In the present study, a band-pass elliptic digital filter was used to eliminate interference and disturbances from the EEG dataset. Next, the Discrete Wavelet Transform (DWT) technique has been employed to decompose the filtered signal into its frequency bands in order to extract the features of EEG signals. Then, different signal features such as logarithmic band power, standard deviation, variance, kurtosis, average energy, root mean square, and Norm have been integrated into the DWT technique to generate the feature vectors and improve the diagnosis performance. After that, nine machine learning approaches have been investigated to classify EEG features into their corresponding classes: linear discriminant analysis (LDA), quadratic discriminant analysis (QDA), support vector machine (SVM), Naïve bayes (NB), k-nearest neighbor (KNN), decision tree (DT), extreme learning machine (ELM), artificial neural network (ANN), and random forests (RF). Finally, the performance of the different proposed machine learning approaches have been compared and evaluated by computing the sensitivity, specificity, overall diagnosis accuracy, and area under the receiver operating characteristic (ROC) curves and plotting the ROC curves and confusion matrices for five classification problems. These investigations aim to compare the proposed approaches and recommend the best combination method for the diagnosis of Alzheimer’s disorders. According to the results, the KNN classifier achieved an average classification accuracy of 99.98% with an area under the ROC curve of 100%. Our findings show that the suggested methodologies are an appealing supplementary tool for identifying possible biomarkers to help in the clinical diagnosis of Alzheimer’s disease.

In recent years, deep learning has been increasingly used in the field of multimodal emotion recognition in conjunction with electroencephalogram. Considering the complexity of recording electroencephalogram signals, some researchers have applied deep learning to find new features for emotion recognition. In previous studies, convolutional neural network model was used to automatically extract features and complete emotion recognition, and certain results were obtained. However, the extraction of hierarchical features with convolutional neural network for multimodal emotion recognition remains unexplored. Therefore, this paper proposes a hierarchical fusion convolutional neural network model to mine the potential information in the data by constructing different network hierarchical structures, extracting multiscale features, and using feature-level fusion to fuse the global features formed by combining weights with manually extracted statistical features to form the final feature vector. This paper conducts binary classification experiments on the valence and arousal dimensions of the DEAP and MAHNOB-HCI data sets to evaluate the performance of the proposed model. The results show that the model proposed in this paper can achieve accuracies of 84.71% and 89.00% on the two corresponding data sets, indicating that the model proposed in this paper is superior to other deep learning emotion classification models in feature extraction and fusion.

Electroencephalogram (EEG) signals reflect the activities or electrical disturbances in neurons in the human brain. Therefore, these signals are vital for diagnosing certain brain disorders. This study mainly focused on the diagnosis of epilepsy and autism spectrum disorders (ASDs) through the analysis and processing of EEGs. In this study, artifacts were removed from the EEG datasets using Independent Component Analysis and were filtered using a fifth-order band-pass Butterworth filter to remove interference and noise. Next, new methods were used to extract the features of EEGs using common spatial pattern (CSP). It is known that conventional CSP uses variance. However, here the use of entropy, energy, and band power with CSP was proposed to extract features of EEGs. Then, in our investigation, four techniques were employed for classification, namely, linear discriminant analysis, support vector machine, k-nearest neighbor (KNN), and artificial neural network, with the aim of comparing the proposed methods and recommending the optimal combination for the diagnosis of epilepsy and ASDs. Finally, the effects of segment length, frequency band, and reduction number on the results were investigated. Two EEG datasets were employed to verify the proposed methods: the King Abdulaziz University dataset (for ASD) and the MIT dataset (for epilepsy). The results indicated that the extracted features based on CSP and band LBP produced the best performance and that the combination of CSP-LBP-KNN provided the best performance with average classification accuracy of approximately 98.46% and 98.62% for diagnosing ASDs and epilepsy, respectively.

Brain-computer interface (BCI) is a novel human-computer interaction model, which does not depend on the conventional output pathway (peripheral nerve and muscle tissue). In the past three decades, it has attracted the interest of researchers and gradually become a research hotspot. As a typical BCI application, the brain-controlled wheelchair (BCW) could provide a new communicating channel with the external environment for physically disabled people. However, the main challenge of BCW is how to decode multi-degree of freedom control instruction from electroencephalogram (EEG) as soon as possible. The research progress of BCW has been developed rapidly over the past fifteen years. In this review, we investigate the BCW from multiple perspectives, include the type of signal acquisition, the pattern of commands for the control system and the working mechanism of the control system. Furthermore, we summarize the development trend of BCW based on the previous investigation, and it is mainly manifested in three aspects: from a wet electrode to dry electrode, from single-mode to multi-mode, and from synchronous control to asynchronous control. With the continuous development of BCW, we also find new functions have been introduced into BCW to increase its stability and robustness. It is believed that BCW will be able to enter the real-life from the laboratory and will be widely used in rehabilitation medicine in the future.

Attentive learning is an important feature of the learning process. It provides a beneficial learning experience and plays a key role in generating positive learning outcomes. Most studies widely applied electroencephalogram (EEG) to measure human attention level. Although most studies use EEG handcrafted features and statistical methods to classify attention level, a more effective feature learning technique is still needed. In this paper, we aim to analyze participants’ EEG signals through a deep learning model and classify those signals as showing either attentive or inattentive behaviors. To carry out this research, we initially conducted a background study on attention and its detection in EEG. After that, we design a Troxler’s fading experiment and use an EEG device to collect data on participants’ attentive and inattentive behaviors during the test. The collected EEG data will be analyzed using a Convolution Attention Memory Neural Network (CAMNN) model to classify participants’ attention level. The proposed CAMNN model is optimized with Vector-to-Vector (Vec2Vec) modeling, where the model can be learned through deep neural networks in an end-to-end approach. The result shows that our model can achieve 92% accuracy and 0.92 F1 score which outperforms several existing neural network models such as Recurrent Neural Network (RNN), Long Short-Term Memory (LSTM), and Convolutional Neural Network (CNN), Deep Learning with Convolutional Neural Networks (deep ConvNets), and Compact Convolutional Network for EEG-based BCIs (EEGNet). This research can be useful for those who are interested in developing attention level monitoring or biofeedback system in areas such as educational classroom learning, medical research, and industrial operator.

Electroencephalography (EEG) is a non-invasive technology used for the human brain-computer interface. One of its important applications is the evaluation of the mental state of an individual, such as workload estimation. In previous works, common spatial pattern feature extraction methods have been proposed for the EEG-based workload detection. Recently, several novel methods were introduced to detect EEG pattern workloads. However, it is still unknown which one of these methods is the one that offers the best performance for the workload EEG pattern feature detection. In this article, four methods were used to extract workload EEG features: (a) common spatial pattern feature extraction; (b) temporally constrained sparse group spatial pattern feature extraction; (c) EEGnet; and (d) the new proposed shallow convolutional neural network for workload estimation (WLnet). The classification accuracy of these four methods was compared. Experimental results demonstrate that the proposed WLnet achieved the best detection accuracy in both stress and non-stress conditions. We believe that the proposed methods may be relevant to real-life applications of mental workload estimation.

Epilepsy is one of the world’s most common neurological diseases. Reliable early prediction and warning of seizures can provide timely treatment for patients with epilepsy, and improve their quality of life. Compared with most hand-designed prediction methods, an automatic prediction model that can process the original electroencephalogram (EEG) signals directly and take into account the leads optimization problem is needed. In this paper, we proposed an end-to-end automatic seizure prediction model based on the Batch Normalization Long Short Term Memory networks (BNLSTM) and Channel and Spatial attention (CASA). Firstly, raw EEG signals without any preprocessing are used as the input to the system, which can reduce the computation amount. Secondly, BNLSTM and CASA retained the time and spatial information of the raw EEG data respectively. Channel attention (CA) achieved the automatic optimization of EEG full-lead data and improved the prediction accuracy. Spatial attention (SA) achieved the adaptive learning of feature parameters. Finally, a fully connected layer is applied to predict the seizures. The performance of the seizure prediction model we proposed is evaluated on the data of 14 patients with Area Under the Curve (AUC) of 0.986, accuracy (Acc) of 0.956, specificity (Spe) of 0.968, and sensitivity (Sen) of 0.942. In addition, the proposed method provided an accurate prediction for all 50 seizures of the other 5 patients in the generalization dataset. Experimental results show that the proposed model has a certain generalization performance, which can provide a reliable basis for early warning of epileptic seizures.

Recently, electroencephalogram-based emotion recognition has become crucial in enabling the Human-Computer Interaction (HCI) system to become more intelligent. Due to the outstanding applications of emotion recognition, e.g., person-based decision making, mind-machine interfacing, cognitive interaction, affect detection, feeling detection, etc., emotion recognition has become successful in attracting the recent hype of AI-empowered research. Therefore, numerous studies have been conducted driven by a range of approaches, which demand a systematic review of methodologies used for this task with their feature sets and techniques. It will facilitate the beginners as guidance towards composing an effective emotion recognition system. In this article, we have conducted a rigorous review on the state-of-the-art emotion recognition systems, published in recent literature, and summarized some of the common emotion recognition steps with relevant definitions, theories, and analyses to provide key knowledge to develop a proper framework. Moreover, studies included here were dichotomized based on two categories: i) deep learning-based, and ii) shallow machine learning-based emotion recognition systems. The reviewed systems were compared based on methods, classifier, the number of classified emotions, accuracy, and dataset used. An informative comparison, recent research trends, and some recommendations are also provided for future research directions.

The educational concepts upholding the theory of brain dominance have been developed for more than 30 years. Some academicians developed a series of the syllabus to exploit the brain capability of students by training their weaker hemisphere of the brain. Prior to training the weaker side of the brain with the developed syllabus, the brain dominance of the student shall be determined. All the current methods used to determine brain dominance are questionnaire-based assessments. There is a possibility that questionnaire biases could exist and lead to inaccurate results. In this research, we introduce a deep-learning method to classify brain dominance based on the electroencephalogram (EEG) signal that reflects the bio-information of the brain. In this paper, we employ a series of EEG signal processing techniques and a state-of-the-art deep learning neural network namely Metric Learning Based Convolutional Neural Network (MLBCNN) to determine brain dominance. We prove that the brain dominance theory is valid and it can be determined by applying machine learning from the EEG signals. We also present the results that show the MLBCNN system can give the best performance as compared to the other benchmark neural network models of which its classification accuracy is 97.44%. Hence, this proposed method can contribute to the education field by providing a system to discover students’ brain dominance and keep track of their brain training progress. In this way, the potential and capability of their brain can be fully unleashed.

We present a transfer learning-based approach for decoding imagined speech from electroencephalogram (EEG). Features are extracted simultaneously from multiple EEG channels, rather than separately from individual channels. This helps in capturing the interrelationships between the cortical regions. To alleviate the problem of lack of enough data for training deep networks, sliding window-based data augmentation is performed. Mean phase coherence and magnitude-squared coherence, two popular measures used in EEG connectivity analysis, are used as features. These features are compactly arranged, exploiting their symmetry, to obtain a three dimensional “image-like” representation. The three dimensions of this matrix correspond to the alpha, beta and gamma EEG frequency bands. A deep network with ResNet50 as the base model is used for classifying the imagined prompts. The proposed method is tested on the publicly available ASU dataset of imagined speech EEG, comprising four different types of prompts. The accuracy of decoding the imagined prompt varies from a minimum of 79.7% for vowels to a maximum of 95.5% for short-long words across the various subjects. The accuracies obtained are better than the state-of-the-art methods, and the technique is good in decoding prompts of different complexities.

Scalp electroencephalogram (EEG) is a non-invasive measure of brain activity. It is widely used in several applications including cognitive tasks, sleep stage detection, and seizure prediction. When recorded over several hours, this signal is usually corrupted by noisy disturbances such as experimental errors, environmental interferences, and physiological artifacts. These may generate confounding factors and, therefore, lead to false results. Models able to minimise EEG artifacts are then necessary for improving further analysis and application. In this work, we developed an EEG artifact removal model based on deep convolutional neural networks. The proposed approach was applied on long-term EEGs, acquired from epileptic patients, available in the EPILEPSIAE database. The main goal of our work is to develop a model able to automatically and quickly remove artifacts from EEGs. To develop it, we used EEG segments, manually preprocessed by experts and named target EEG segments. Our approach was evaluated comparing denoised segments with the target segments. Furthermore, we compared our approach with other artifact removal models. Results show that the developed model was able to attenuate the influence of artifacts, present in long-term EEG signals, in a similar way to that performed by experts. Additionally, results evidence that our approach performs better than other artifact removal models, combining a minor reconstruction error with a fast processing. Being a fully automatic and fast model that does not require reference artifact templates, turns it suitable, for example, for continuous preprocessing of long-term electroencephalogram for sleep staging or seizure prediction.

This article mainly studies Electroencephalogram (EEG) mental recognition. Because the human brain is very complex and the EEG signal is greatly affected by the environment, EEG mental recognition can be attributed to domain adaptative problems. Our main work is as follows: (1) At present, most domain adaptation learning only learns the linear subspace of Reproducing Kernel Hilbert Space (RKHS), and RKHS itself does not. Given the complexity and nonlinearity of EEG mental recognition, we propose an EEG mental recognition algorithm based on two learning. The two learning is RKHS learning and RKHS subspace learning. The source dictionary regularized RKHS subspace learning we proposed applies to EEG mental recognition and is better than pure RKHS subspace learning, but not enough. To get satisfactory results, we learn RKHS before RKHS subspace. (2) According to Moore–Aronszajn theorem, RKHS can be uniquely generated by a kernel function. The existing RKHS is rarely learnable. It is difficult to find a kernel function that can be learned and optimized. In RKHS learning, this paper uses a learnable kernel function that we published, and the kernel function is easy to optimize. (3) In RKHS subspace learning, most of the existing methods adopt the Maximum Mean Discrepancy (MMD) criterion, but it cannot make the spatial distribution of the same category of source and target domain data overlap as much as possible, and the label of the target domain data is unknown. To solve this problem, this paper proposes a framework of RKHS subspace learning based on source domain dictionary regularization. The experimental results on the brain-computer interface international competition data set (BCI competition IV 2a) show that the effect of the algorithm proposed in this paper is better than that of the other five advanced domain adaptation learning algorithms.

In recent years, with the continuous development of artificial intelligence and brain-computer interface technology, emotion recognition based on physiological signals, especially electroencephalogram signals, has become a popular research topic and attracted wide attention. However, the imbalance of the data sets themselves, affective features’ extraction from electroencephalogram signals, and the design of classifiers with excellent performance, pose a great challenge to the subject. Motivated by the outstanding performance of deep learning approaches in pattern recognition tasks, we propose a method based on convolutional neural network with data augmentation method Borderline-synthetic minority oversampling technique. First, we obtain 32-channel electroencephalogram signals from DEAP data set, which is the standard data set of emotion recognition. Then, after data pre-processing, we extract features in frequency domain and data augmentation based on the data augmentation algorithm above for getting more balanced data. Finally, we train a one dimensional convolutional neural network for three classification on two emotional dimensions valence and arousal. Meanwhile, the proposed method is compared with some traditional machine learning methods and some existing methods by other researchers, which is proved to be effective in emotion recognition, and the average accuracy rate of 32 subjects on valence and arousal are 97.47% and 97.76% respectively. Compared with other existing methods, the performance of the proposed method with data augmentation algorithm Borderline-SMOTE shows its advantage in affective emotional recognition than that without Borderline-SMOTE.

This paper proposes two novel methods to classify semantic vigilance levels by utilizing EEG directed connectivity patterns with their corresponding graphical network measures. We estimate the directed connectivity using relative wavelet transform entropy (RWTE) and partial directed coherence (PDC) and the graphical network measures by graph theory analysis (GTA) at four frequency bands. The RWTE and PDC quantify the strength and directionality of information flow between EEG nodes. On the other hand, the GTA of the complex network measures summarizes the topological structure of the network. We then evaluate the proposed methods using machine learning classifiers. We carried out an experiment on nine subjects performing semantic vigilance task (Stroop color word test (SCWT)) for approximately 45 minutes. Behaviorally, all subjects demonstrated vigilance decrement as reflected by the significant increase in response time and reduced accuracy. The strength and directionality of information flow in the connectivity network by RWTE/PDC and the GTA measures significantly decrease with vigilance decrement, p

Discovering and exploiting shared, invariant neural activity in electroencephalogram (EEG) based classification tasks is of significant interest for generalizability of decoding models across subjects or EEG recording sessions. While deep neural networks are recently emerging as generic EEG feature extractors, this transfer learning aspect usually relies on the prior assumption that deep networks naturally behave as subject- (or session-) invariant EEG feature extractors. We propose a further step towards invariance of EEG deep learning frameworks in a systemic way during model training. We introduce an adversarial inference approach to learn representations that are invariant to inter-subject variabilities within a discriminative setting. We perform experimental studies using a publicly available motor imagery EEG dataset, and state-of-the-art convolutional neural network based EEG decoding models within the proposed adversarial learning framework. We present our results in cross-subject model transfer scenarios, demonstrate neurophysiological interpretations of the learned networks, and discuss potential insights offered by adversarial inference to the growing field of deep learning for EEG.

This paper presents a new stochastic adaptive estimation-identification technique for nonstationary systems. The proposed method enhances the initial results from an on average estimation, and its identification, through a generalized adaptable function based on the Exponential Forgetting Factor (EFF), and the Sliding Mode (SM) regarding the error identification. In this form, the presented process includes the function implementation in three stages-estimation, adaptive estimation, and adaptive estimation-identification, allowing us to observe the gradual convergence to a nonstationary reference signal. Simulations first introduce convergence level checks obtained from the estimation and identification of artificial signals. After that, the algorithm is applied for real references, considering the Electroencephalogram (EEG) signals taken from a public database, finding their internal nonstationary gains, indirectly. Finally, the results include a performance comparison between the proposed strategy concerning the Recursive Least Square (RLS), the Least Mean Square (LMS), and the Kalman Filter (KF).

Motor imagery (MI) is a typical BCI paradigm and has been widely applied into many aspects (e.g. brain-driven wheelchair and motor function rehabilitation training). Although significant achievements have been achieved, multiple motor imagery decoding is still unsatisfactory. To deal with this challenging issue, firstly, a segment of electroencephalogram was extracted and preprocessed. Secondly, we applied a filter bank common spatial pattern (FBCSP) with one-vs-rest (OVR) strategy to extract the spatio-temporal-frequency features of multiple MI. Thirdly, the F-score was employed to optimise and select these features. Finally, the optimized features were fed to the spiking neural networks (SNN) for classification. Evaluation was conducted on two public multiple MI datasets (Dataset IIIa of the BCI competition III and Dataset IIa of the BCI competition IV). Experimental results showed that the average accuracy of the proposed framework reached up to 90.09% (kappa: 0.868) and 81.33% (kappa: 0.751) on the two public datasets, respectively. The achieved performance (accuracy and kappa) was comparable to the best one of the compared methods. This study demonstrated that the proposed method can be used as an alternative approach for multiple MI decoding and it provided a potential solution for online multiple MI detection.

The proposed emotion recognition model is based on the hierarchical long-short term memory neural network (LSTM) for video-electroencephalogram (Video-EEG) signal interaction. The inputs are facial-video and EEG signals from the subjects when they are watching the emotion-stimulated video. The outputs are the corresponding emotion recognition results. Facial-video features and corresponding EEG features are extracted based on a fully connected neural network (FC) at each time point. These features are fused through hierarchical LSTM to predict the key emotional signal frames at the next time point until the emotion recognition result is calculated at the last time point. Specially, a self-attention mechanism is applied to show the correlation of the stacked LSTM at different hierarchies. In this process, the “selective focus” is used to analyze the human-emotional temporal sequences in each model, which improves the utilization of the key spatial EEG signals. Moreover, the process includes the temporal attention mechanism to predict the key signal frame at next time point, which utilizes the key emotion data in temporal domain. The experimental results prove that the classification rate (CR) and F1-score of the proposed emotion recognition model are significantly increased by at least 2% and 0.015, respectively, compared to other methods.

Multimodal emotion recognition has gained traction in affective computing research community to overcome the limitations posed by the processing a single form of data and to increase recognition robustness. In this study, a novel emotion recognition system is introduced, which is based on multiple modalities including facial expressions, galvanic skin response (GSR) and electroencephalogram (EEG). This method follows a hybrid fusion strategy and yields a maximum one-subject-out accuracy of 81.2% and a mean accuracy of 74.2% on our bespoke multimodal emotion dataset (LUMED-2) for 3 emotion classes: sad, neutral and happy. Similarly, our approach yields a maximum one-subject-out accuracy of 91.5% and a mean accuracy of 53.8% on the Database for Emotion Analysis using Physiological Signals (DEAP) for varying numbers of emotion classes, 4 in average, including angry, disgust, afraid, happy, neutral, sad and surprised. The presented model is particularly useful in determining the correct emotional state in the case of natural deceptive facial expressions. In terms of emotion recognition accuracy, this study is superior to, or on par with, the reference subject-independent multimodal emotion recognition studies introduced in the literature.

Along with extensive and successive applications, emotion recognition based on electroencephalogram has attracted more and more researchers. However, how to acquire sufficient and high-quality real emotion electroencephalogram to train emotion recognition model has always been a bottleneck issue in the electroencephalogram-based emotion recognition research field. Because the subject's emotion is easy to be affected by many factors and subjects do not always evoke emotion well, it's very hard to determine whether the real emotion electroencephalogram appears or not in the experiments. On the contrary to electroencephalogram, facial expression obtained without any intentional control is easy to be recognized by computers, and is also one of the primary and reliable cues for understanding emotions. Inspired by such a common sense, we proposed an approach to building up ground truth electroencephalogram dataset with visual indication. Firstly, the relationship between facial expression and electroencephalogram is analyzed in details from the viewpoints of biophysics and correlation. Secondly, based on the analysis result that when the subject's facial expression is evoked, it should be accompanied by the emotional electroencephalogram corresponding to it, the method of building up the ground truth electroencephalogram dataset with visual indication is put forward along with the computer automatic implementation. Thirdly, we have built up a ground truth electroencephalogram dataset which covers 3 kinds of emotions (i.e. joy, sadness, and neutral) of the subjects who are undergraduate and graduate students from Minzu University of China. Lastly, the validation of the established dataset is tested by the comparative experiments between the long short term memory emotion recognition models trained with two different datasets respectively. One dataset contains both truth and false electroencephalograms and another one only contains the truth data.

Objective: Classification of sleep-wake states using multichannel electroencephalography (EEG) data that reliably work for neonates. Methods: A deep multilayer perceptron (MLP) neural network is developed to classify sleep-wake states using multichannel bipolar EEG signals, which takes an input vector of size 108 containing the joint features of 9 channels. The network avoids any post-processing step in order to work as a full-fledged real-time application. For training and testing the model, EEG recordings of 3525 30-second segments from 19 neonates (postmenstrual age of 37 ± 05 weeks) are used. Results: For sleep-wake classification, mean Cohen's kappa between the network estimate and the ground truth annotation by human experts is 0.62. The maximum mean accuracy can reach up to 83% which, to date, is the highest accuracy for sleep-wake classification.

Lack of communication causes problems for patients with neurodegenerative diseases, so the need for alternative methods is required to convey their thoughts with caretakers, friends, and family members. Brain-computer interface (BCI) is a device to control external devices by using mental thoughts without any other muscle movements to improve the communication quality for the disabled individual without any other help. The techniques of measuring electrical signal around the scalp during some activities by using electrodes are called electroencephalogram (EEG). By combining these two technologies together to form a brain-computer interaction, this helps the paralyzed individual to communicate with others to share the thoughts. In this paper, we deal with the history of EEG, electrode placements with measurements, and signal ranges, and also further discussed some of the prominent studies completed in designing BCI using EEG. This helps the new researchers to know the EEG measurement and position completely and paved the new way to create EEG-based interface research.

Nowadays, artificial intelligence is widely used in many biomedical-oriented problems. Because of obtained effective and efficient results, the use of intelligent solution mechanisms by artificial intelligence techniques is mainly focused on healthcare applications. Moving from the explanations, the objective of this paper is to provide an adaptive neuro-fuzzy inference system (ANFIS) trained by a recent optimization algorithm: electro-search optimization (ESO) for predicting the electroencephalogram (EEG) time series. In detail, the research was directed to the EEG time series showing chaotic characteristics so that an effective hybrid system can be designed for having an idea about future states of human brain activity in the case of possible diseases. The developed ANFIS-ESO system was evaluated with five different EEG time series and the obtained findings were reported accordingly. In addition, the ANFIS-ESO system was compared with alternative techniques-systems in order to see the performances according to different systems. In the end, it is possible to mention that the ANFIS-ESO system provides well-enough results in terms of predicting EEG time series. As a result of encouraging results, ANFIS-ESO is currently used actively for real cases.

Multi-subject electroencephalogram (EEG) computing has attracted increasing interest in recent years. Commercial EEG headsets often lead to a space-required or cost-expensive setup for obtaining simultaneous recordings of multiple subjects. The sparsity or immobility of their electrode placements may compromise their realistic applicability. Thus, this work developed a cost-efficient, portable, and customizable system to mitigate these constraints. The developed system implemented two core hardware infrastructures, including an event broadcaster and a dry electrode-compatible EEG amplifier by assembling entirely low-cost, off-the-shelf electronic components. The broadcaster allowed distribution of event markers to multiple EEG amplifiers concurrently, whereas the amplifier transmitted the digitized, event-synchronized EEG signals wirelessly. By conducting an oddball event-related potential (ERP) experiment with simultaneous recordings of three subjects on 10 days, the system reliably captured the time- and phase-locked ERP components (e.g., N100 and P300 amplitudes) by single-subject, multi-subject, and multi-day analytical approaches. The practicality and stability of the proposed system was empirically demonstrated in terms of the signal quality, EEG-event synchronization, and inter-amplifier coordination for a multi-subject setup. This work sheds light on how to economically facilitate and scale up multi-subject EEG computing for fundamental research and for brain-computer interface (BCI) applications in a larger subject population.

The canonical correlation analysis (CCA)-based frequency recognition method is one of the widely used methods in steady-state visual evoked potential (SSVEP)-based brain-computer interface (BCI), and several extended version were proposed in the past decade. However, these methods only adopt the maximal correlation coefficient provided by the CCA and discard other coefficients. This operation may lead to the loss of discriminative information that exists in the discarded coefficients. In the current study, we proposed to fuse all the correlation coefficients of the CCA with a nonlinear weighting function when performing frequency recognition with CCA method, termed as FoCCA. Evaluated on a benchmark dataset of thirty-five subjects, the experimental results demonstrated that the classification accuracy and information transfer rate (ITR) of FoCCA are significantly higher than those of the standard CCA at various time windows. Fusing correlation coefficients could be a new strategy to improve the performance of other extended CCA methods, which holds the potential to implement high-performance SSVEP-based BCI systems in the future.

Depression has become a serious disease that affects people’s mental health. How to detect it promptly and accurately is a difficult task. Electroencephalogram can reflect the spontaneous biological potential signals in the cerebral cortex and is widely used in the prediction and diagnosis of depression. With electroencephalogram, the key and most difficult challenge is to find the brain regions and frequencies associated with depression, especially mild depression. At present, the most commonly used method is the combination of feature selection and classification algorithm for detection. However, the classification accuracy needs to be further improved. The differential evolution is a population-based adaptive global optimization algorithm. Due to its fast convergence and strong robustness, this paper uses it to optimize the extracted features to achieve better result. Then the k-nearest neighbor classification algorithm is used to classify patients with mild depression and normal people. The experiment is performed on a data set of 10 subjects with mild depression and 10 normal subjects. The results show that the method can find the relatively optimal features and distinguish the two groups of subjects better. It effectively improves the classification accuracy and efficiency, and is superior to other feature optimization methods.

This study presents an automatic sleep-stage classification system based on utilizing compressive sensing (CS) for data reduction. The amount of electroencephalogram (EEG) signal data required for sleep-stage classification can be significantly reduced by applying CS at the cost of distortion in the reconstructed EEG signal. A neural network trained on features extracted from the reconstructed EEG signal was implemented as a classifier to avoid compromising the accuracy of classification in the presence of distortion in the reconstructed EEG signal. A radial basis function (RBF) neural network that uses a simple Manhattan distance function instead of complicated computation, such as vector multiplication matrix (VMM), was selected to reduce the hardware complexity. A classification method that utilizes information from the previous sleep stage based on the unique nature of human sleep was also presented and implemented on the proposed RBF classifier for further improvement of the classification accuracy. The classification system has been evaluated using EEG data from eight subjects in the Cyclic Alternating Pattern (CAP) sleep dataset. The measurement results showed that the classification system achieved high classification accuracy comparable to the previously reported sleep-stage classifiers that do not utilize signal compression.

Fluctuation-based dispersion entropy (FDispEn) is a new approach to estimate the dynamical variability of the fluctuations of signals. It is based on Shannon entropy and fluctuation-based dispersion patterns. To quantify the physiological dynamics over multiple time scales, multiscale FDispEn (MFDE) is developed in this paper. MFDE is robust to the presence of baseline wanders or trends in the data. We evaluate MFDE, compared with popular multiscale sample entropy (MSE), multiscale fuzzy entropy (MFE), and the recently introduced multiscale dispersion entropy (MDE), on selected synthetic data and five neurological diseases' datasets: 1) focal and non-focal electroencephalograms (EEGs); 2) walking stride interval signals for young, elderly, and Parkinson's subjects; 3) stride interval fluctuations for Huntington's disease and amyotrophic lateral sclerosis; 4) EEGs for controls and Alzheimer's disease patients; and 5) eye movement data for Parkinson's disease and ataxia. The MFDE avoids the problem of the undefined MSE values and, compared with the MFE and MSE, leads to more stable entropy values over the scale factors for white and pink noises. Overall, the MFDE is the fastest and most consistent method for the discrimination of different states of neurological data, especially where the mean value of a time series considerably changes along with the signal (e.g., eye movement data). This paper shows that MFDE is a relevant new metric to gain further insights into the dynamics of neurological diseases' recordings. The MATLAB codes for the MFDE and its refined composite form are available in Xplore.