Single-trial Electroencephalogram Research Articles

Subject-independent auditory spatial attention detection based on brain topology modeling and feature distribution alignment

Auditory spatial attention detection (ASAD) seeks to determine which speaker in a surround sound field a listener is focusing on based on the one’s brain biosignals. Although existing studies have achieved ASAD from a single-trial electroencephalogram (EEG), the huge inter-subject variability makes them generally perform poorly in cross-subject scenarios. Besides, most ASAD methods do not take full advantage of topological relationships between EEG channels, which are crucial for high-quality ASAD. Recently, some advanced studies have introduced graph-based brain topology modeling into ASAD, but how to calculate edge weights in a graph to better capture actual brain connectivity is worthy of further investigation. To address these issues, we propose a new ASAD method in this paper. First, we model a multi-channel EEG segment as a graph, where differential entropy serves as the node feature, and a static adjacency matrix is generated based on inter-channel mutual information to quantify brain functional connectivity. Then, different subjects’ EEG graphs are encoded into a shared embedding space through a total variation graph neural network. Meanwhile, feature distribution alignment based on multi-kernel maximum mean discrepancy is adopted to learn subject-invariant patterns. Note that we align EEG embeddings of different subjects to reference distributions rather than align them to each other for the purpose of privacy preservation. A series of experiments on open datasets demonstrate that the proposed model outperforms state-of-the-art ASAD models in cross-subject scenarios with relatively low computational complexity, and feature distribution alignment improves the generalizability of the proposed model to a new subject.

PSAEEGNet: pyramid squeeze attention mechanism-based CNN for single-trial EEG classification in RSVP task.

Accurate classification of single-trial electroencephalogram (EEG) is crucial for EEG-based target image recognition in rapid serial visual presentation (RSVP) tasks. P300 is an important component of a single-trial EEG for RSVP tasks. However, single-trial EEG are usually characterized by low signal-to-noise ratio and limited sample sizes. Given these challenges, it is necessary to optimize existing convolutional neural networks (CNNs) to improve the performance of P300 classification. The proposed CNN model called PSAEEGNet, integrates standard convolutional layers, pyramid squeeze attention (PSA) modules, and deep convolutional layers. This approach arises the extraction of temporal and spatial features of the P300 to a finer granularity level. Compared with several existing single-trial EEG classification methods for RSVP tasks, the proposed model shows significantly improved performance. The mean true positive rate for PSAEEGNet is 0.7949, and the mean area under the receiver operating characteristic curve (AUC) is 0.9341 (p < 0.05). These results suggest that the proposed model effectively extracts features from both temporal and spatial dimensions of P300, leading to a more accurate classification of single-trial EEG during RSVP tasks. Therefore, this model has the potential to significantly enhance the performance of target recognition systems based on EEG, contributing to the advancement and practical implementation of target recognition in this field.

AN EMD AND IMF ENERGY ENTROPY-BASED OPTIMIZED FEATURE EXTRACTION AND CLASSIFICATION SCHEME FOR SINGLE TRIAL EEG SIGNAL

For the purpose of improving the classification accuracy of single-trial electroencephalogram (EEG) signal during motor imagery (MI) process, this study proposed a classification method which combines intrinsic mode functions (IMFs) energy entropy and improved empirical mode decomposition (EMD) scheme. Singular value decomposition (SVD), Gaussian mixture model (GMM), EMD and IMF energy entropy were employed for the newly designed scheme. After removing noise and artifacts from acquired EEG signals in EEGLAB, SVD was applied, and the singular values were clustered by GMM. The insignificant characteristics indicated by the small SVD values were then removed, and the signals were reconstructed, feeding to EMD algorithm. Those IMFs mapping to [Formula: see text], [Formula: see text], [Formula: see text] and [Formula: see text] frequencies were selected as the major features of the Electroencephalogram signal. The SVM classifier with Radial Basis Function Neural Network, linear, and polynomial kernel functions and voting mechanism then kicked in for classification. The results were compared with that of the traditional EMD and EEMD through simulation, showing that the proposed scheme can eliminate mode mixing effectively and improve the single-trial EEG signal classification accuracy significantly, suggesting the probability of designing a more efficient EEG control system based on the proposed scheme.

Cortical entrainment to speech produced by cochlear implant talkers and normal-hearing talkers.

Cochlear implants (CIs) are commonly used to restore the ability to hear in those with severe or profound hearing loss. CIs provide the necessary auditory feedback for them to monitor and control speech production. However, the speech produced by CI users may not be fully restored to achieve similar perceived sound quality to that produced by normal-hearing talkers and this difference is easily noticeable in their daily conversation. In this study, we attempt to address this difference as perceived by normal-hearing listeners, when listening to continuous speech produced by CI talkers and normal-hearing talkers. We used a regenerative model to decode and reconstruct the speech envelope from the single-trial electroencephalogram (EEG) recorded on the scalp of the normal-hearing listeners. Bootstrap Spearman correlation between the actual speech envelope and the envelope reconstructed from the EEG was computed as a metric to quantify the difference in response to the speech produced by the two talker groups. The same listeners were asked to rate the perceived sound quality of the speech produced by the two talker groups as a behavioral sound quality assessment. The results show that both the perceived sound quality ratings and the computed metric, which can be seen as the degree of cortical entrainment to the actual speech envelope across the normal-hearing listeners, were higher in value for speech produced by normal hearing talkers than that for CI talkers. The first purpose of the study was to determine how well the envelope of speech is represented neurophysiologically via its similarity to the envelope reconstructed from EEG. The second purpose was to show how well this representation of speech for both CI and normal hearing talker groups differentiates in term of perceived sound quality.

Cortical Entrainment to Speech Produced by Cochlear Implant Users and Normal-Hearing Talkers.

The perceived sound quality of speech produced by hard-of-hearing individuals greatly depends on the degree and configuration of their hearing loss. A cochlear implant (CI) may provide some compensation and auditory feedback to monitor/control speech production. However, to date, the speech produced by CI users is still different in quality from that produced by normal-hearing (NH) talkers. In this study, we attempted to address this difference by examining the cortical activity of NH listeners when listening to continuous speech produced by 8 CI talkers and 8 NH talkers. We utilized a discriminative model to decode and reconstruct the speech envelope from the single-trial electroencephalogram (EEG) recorded from scalp electrode in NH listeners when listening to continuous speech. The correlation coefficient between the reconstructed envelope and original speech envelope was computed as a metric to quantify the difference in response to the speech produced by CI and NH talkers. The same listeners were asked to rate the perceived sound quality of the speech as a behavioral sound quality assessment. Both behavioral perceived sound quality ratings and the cortical entrainment to speech envelope were higher for the speech set produced by NH talkers than for the speech set produced by CI talkers.

Discriminability of single-trial EEG during decision-making of cooperation or aggression: a study based on machine learning

Decision-making is a very important cognitive process in our daily life. There has been increasing interest in the discriminability of single-trial electroencephalogram (EEG) during decision-making. In this study, we designed a machine learning based framework to explore the discriminability of single-trial EEG corresponding to different decisions. For each subject, the framework split the decision-making trials into two parts, trained a feature model and a classifier on the first part, and evaluated the discriminability on the second part using the feature model and classifier. A proposed algorithm and five existing algorithms were applied to fulfill the feature models, and the algorithm Linear Discriminative Analysis (LDA) was used to implement the classifiers. We recruited 21 subjects to participate in Chicken Game (CG) experiments. The results show that there exists the discriminability of single-trial EEG between the cooperation and aggression decisions during the CG experiments, with the classification accuray of 75% (±6%), and the discriminability is mainly from the EEG information below 40 Hz. The further analysis indicates that the contributions of different brain regions to the discriminability are consistent with the existing knowledge on the cognitive mechanism of decision-making, confirming the reliability of the conclusions. This study exhibits that it is feasible to apply machine learning methods to EEG analysis of decision-making cognitive process.

Identifying Fragmented Reading and Evaluating Its Influence on Cognition Based on Single Trial Electroencephalogram.

Background: The use of social media daily could nurture a fragmented reading habit. However, little is known whether fragmented reading (FR) affects cognition and what are the underlying electroencephalogram (EEG) alterations it may lead to.Purpose: This study aimed to identify whether individuals have FR habits based on the single-trial EEG spectral features using machine learning (ML), as well as to find out the potential cognitive impairment induced by FR.Methods: Subjects were recruited through a questionnaire and divided into FR and noFR groups according to the time they spent on FR per day. Moreover, 64-channel EEG was acquired in Continuous Performance Task (CPT) and segmented into 0.5–1.5 s post-stimulus epochs under cue and background conditions. The sample sizes were as follows: FR in cue condition, 692 trials; noFR in cue condition, 688 trials; FR in background condition, 561 trials; noFR in background condition, 585 trials. For these single-trials, the relative power (RP) of six frequency bands [delta (1–3 Hz), theta (4–7 Hz), alpha (8–13 Hz), beta1 (14–20 Hz), beta2 (21–29 Hz), lower gamma (30–40 Hz)] were extracted as features. After feature selection, the most important feature sets were fed into three ML models, namely Support-Vector Machine (SVM), K-Nearest Neighbor (KNN), and Naive Bayes to perform the identification of FR. RP of six frequency bands was also used as feature sets to conduct classification tasks.Results: The classification accuracy reached up to 96.52% in the SVM model under cue conditions. Specifically, among six frequency bands, the most important features were found in alpha and gamma bands. Gamma achieved the highest classification accuracy (86.69% for cue, 86.45% for background). In both conditions, alpha RP in central sites of FR was stronger than noFR (p < 0.001). Gamma RP in the frontal site of FR was weaker than noFR in the background condition (p < 0.001), while alpha RP in parieto-occipital sites of FR was stronger than noFR in the cue condition (p < 0.001).Conclusion: Fragmented reading can be identified based on single-trial EEG evoked by CPT using ML, and the RP of alpha and gamma may reflect the impairment on attention and working memory by FR. FR might lead to cognitive impairment and is worth further exploration.

P3-MSDA: Multi-Source Domain Adaptation Network for Dynamic Visual Target Detection.

Single-trial electroencephalogram detection has been widely applied in brain-computer interface (BCI) systems. Moreover, an individual generalized model is significant for applying the dynamic visual target detection BCI system in real life because of the time jitter of the detection latency, the dynamics and complexity of visual background. Hence, we developed an unsupervised multi-source domain adaptation network (P3-MSDA) for dynamic visual target detection. In this network, a P3 map-clustering method was proposed for source domain selection. The adversarial domain adaptation was conducted for domain alignment to eliminate individual differences, and prediction probabilities were ranked and returned to guide the input of target samples for imbalanced data classification. The results showed that individuals with a strong P3 map selected by the proposed P3 map-clustering method perform best on the source domain. Compared with existing schemes, the proposed P3-MSDA network achieved the highest classification accuracy and F1 score using five labeled individuals with a strong P3 map as the source domain. These findings can have a significant meaning in building an individual generalized model for dynamic visual target detection.

Sequential Attention-based Detection of Semantic Incongruities from EEG While Listening to Speech.

We propose a method with attention-based recurrent neural networks (ARNN) for detecting the semantic incongruities in spoken sentences using single-trial electroencephalogram (EEG) signals. 19 participants listened to sentences, some of which included semantically anomalous words. We recorded their EEG signals while they listened. Although previous detection approaches used a word's explicit onset, we used the EEG signals of the whole regions of each sentence, which made it possible to classify the correctness of the sentences without the onset information of the anomalous words. ARNN achieved 63.5% classification accuracy with a statistical significance above the chance level and also above the performances which includes onset information (50.9%). Our results also demonstrated that the attention weights of the model showed that the predictions depended on the feature vectors that are temporally close to the onsets of the anomalous words.

Lumos!: Electrophysiological tracking of (wizarding) world knowledge use during reading.

In Troyer and Kutas (2018), individual differences in knowledge of the world of Harry Potter (HP) rapidly modulated individuals' average electrical brain potentials to contextually supported words in sentence endings. Using advances in single-trial electroencephalogram analysis, we examined whether this relationship is strictly a result of domain knowledge mediating the proportion of facts each participant knew; we find it is not. Participants read sentences ending in a contextually supported word, reporting online whether they had known each fact. Participants' reports correlated with HP domain knowledge and reliably modulated event-related brain potentials to sentence-final words within 250 ms. Critically, domain knowledge had a dissociable influence in the same time window for endings that participants reported not having known and/or were less likely to be known/remembered across participants. We hypothesize that knowledge impacts written word processing primarily by affecting the neural processes of (implicit) retrieval from long-term memory (LTM): Greater knowledge eases otherwise difficult retrieval processes. (PsycINFO Database Record (c) 2020 APA, all rights reserved).

Bayesian Nonnegative Matrix Factorization With Dirichlet Process Mixtures

Nonnegative Matrix Factorization (NMF) is valuable in many applications of blind source separation, signal processing and machine learning. A number of algorithms that can infer nonnegative latent factors have been developed, but most of these assume a specific noise kernel. This is insufficient to deal with complex noise in real scenarios. In this paper, we present a hierarchical Dirichlet process nonnegative matrix factorization (DPNMF) model in which the Gaussian mixture model is used to approximate the complex noise distribution. Moreover, the model is cast in the nonparametric Bayesian framework by using Dirichlet process mixture to infer the necessary number of Gaussian components. We derive a mean-field variational inference algorithm for the proposed nonparametric Bayesian model. We first test the model on synthetic data sets contaminated by Gaussian, sparse and mixed noise. We then apply it to extract muscle synergies from the electromyographic (EMG) signal and to select discriminative features for motor imagery single-trial electroencephalogram (EEG) classification. Experimental results demonstrate that DPNMF performs better in extracting the latent nonnegative factors in comparison with state-of-the-art methods.

Classification of EEG-based single-trial motor imagery tasks using a B-CSP method for BCI

Classifying single-trial electroencephalogram (EEG) based motor imagery (MI) tasks is extensively used to control brain-computer interface (BCI) applications, as a communication bridge between humans and computers. However, the low signal-to-noise ratio and individual differences of EEG can affect the classification results negatively. In this paper, we propose an improved common spatial pattern (B-CSP) method to extract features for alleviating these adverse effects. First, for different subjects, the method of Bhattacharyya distance is used to select the optimal frequency band of each electrode including strong event-related desynchronization (ERD) and event-related synchronization (ERS) patterns; then the signals of the optimal frequency band are decomposed into spatial patterns, and the features that can describe the maximum differences of two classes of MI are extracted from the EEG data. The proposed method is applied to the public data set and experimental data set to extract features which are input into a back propagation neural network (BPNN) classifier to classify single-trial MI EEG. Another two conventional feature extraction methods, original common spatial pattern (CSP) and autoregressive (AR), are used for comparison. An improved classification performance for both data sets (public data set: 91.25%±1.77% for left hand vs. foot and 84.50%±5.42% for left hand vs. right hand; experimental data set: 90.43%±4.26% for left hand vs. foot) verifies the advantages of the B-CSP method over conventional methods. The results demonstrate that our proposed B-CSP method can classify EEG-based MI tasks effectively, and this study provides practical and theoretical approaches to BCI applications.

Identification of Auditory Object-Specific Attention from Single-Trial Electroencephalogram Signals via Entropy Measures and Machine Learning.

Existing research has revealed that auditory attention can be tracked from ongoing electroencephalography (EEG) signals. The aim of this novel study was to investigate the identification of peoples’ attention to a specific auditory object from single-trial EEG signals via entropy measures and machine learning. Approximate entropy (ApEn), sample entropy (SampEn), composite multiscale entropy (CmpMSE) and fuzzy entropy (FuzzyEn) were used to extract the informative features of EEG signals under three kinds of auditory object-specific attention (Rest, Auditory Object1 Attention (AOA1) and Auditory Object2 Attention (AOA2)). The linear discriminant analysis and support vector machine (SVM), were used to construct two auditory attention classifiers. The statistical results of entropy measures indicated that there were significant differences in the values of ApEn, SampEn, CmpMSE and FuzzyEn between Rest, AOA1 and AOA2. For the SVM-based auditory attention classifier, the auditory object-specific attention of Rest, AOA1 and AOA2 could be identified from EEG signals using ApEn, SampEn, CmpMSE and FuzzyEn as features and the identification rates were significantly different from chance level. The optimal identification was achieved by the SVM-based auditory attention classifier using CmpMSE with the scale factor τ = 10. This study demonstrated a novel solution to identify the auditory object-specific attention from single-trial EEG signals without the need to access the auditory stimulus.

Familiar/unfamiliar face classification from EEG signals by utilizing pairwise distant channels and distinctive time interval

The aim of the study is to classify single trial electroencephalogram and to estimate active regions/locations on skull in unfamiliar/familiar face recognition task. For this purpose, electroencephalographic signals were acquired from ten subjects in different sessions. Sixty-one familiar and fifty-nine unfamiliar face stimuli were shown to the subjects in the experiments. Since channel responses are different for familiar and unfamiliar classes, the channels discriminating the classes were investigated. To do so, three distances and four similarity measures were employed to assess the most distant channel pairs between familiar and unfamiliar classes for a 1-s time duration; 0.6 s from the stimulus to 1.6 s in a channel selection process. It is experimentally observed that this time interval is maintaining the greatest distance between two categories. The electroencephalographic signals were classified using the determined channels and time interval to measure accuracy. The best classification accuracy was 81.30% and was obtained with the Pearson correlation as channel selection method. The most discriminative channel pairs were selected from prefrontal regions.

Multiclass support matrix machine for single trial EEG classification

Abstract We propose a novel multiclass classifier for single trial electroencephalogram (EEG) data in matrix form, namely multiclass support matrix machine (MSMM), aiming at improving the classification accuracy of multiclass EEG signals, and hence enhancing the performance of EEG-based brain computer interfaces (BCIs) involving multiple mental activities. In order to construct the MSMM, we propose a novel objective function, which is composed of a multiclass hinge loss term and a combined regularization term. We first formulate the multiclass hinge loss by extending the margin rescaling loss to support matrix-form data. We then devise the regularization term by combining the squared Frobenius norm of tensor-form model parameter and the nuclear norm of matrix-form hyperplanes extracted from the model parameter. While the Frobenius norm prevents over-fitting when training the model, the nuclear norm captures the structural information within the matrix data. We further propose an efficient solver for MSMM based on the alternating direction method of multipliers (ADMM) framework. We conduct extensive experiments on two benchmark EEG datasets. Experimental results show that MSMM achieves much better performance than state-of-the-art classifiers and yields a mean kappa value of 0.880 and 0.648 for dataset IIIa of BCI III and dataset IIa of BCI IV, respectively. To our best knowledge, MSMM is the first classifier that supports multiclass classification for matrix-form EEG data. The proposed MSMM enables easier and more efficient implementation of robust multi-task BCIs, and therefore has potential to promote the wider use of BCI technology.

Noise-robust cortical tracking of attended speech in real-world acoustic scenes

Selectively attending to one speaker in a multi-speaker scenario is thought to synchronize low-frequency cortical activity to the attended speech signal. In recent studies, reconstruction of speech from single-trial electroencephalogram (EEG) data has been used to decode which talker a listener is attending to in a two-talker situation. It is currently unclear how this generalizes to more complex sound environments. Behaviorally, speech perception is robust to the acoustic distortions that listeners typically encounter in everyday life, but it is unknown whether this is mirrored by a noise-robust neural tracking of attended speech. Here we used advanced acoustic simulations to recreate real-world acoustic scenes in the laboratory. In virtual acoustic realities with varying amounts of reverberation and number of interfering talkers, listeners selectively attended to the speech stream of a particular talker. Across the different listening environments, we found that the attended talker could be accurately decoded from single-trial EEG data irrespective of the different distortions in the acoustic input. For highly reverberant environments, speech envelopes reconstructed from neural responses to the distorted stimuli resembled the original clean signal more than the distorted input. With reverberant speech, we observed a late cortical response to the attended speech stream that encoded temporal modulations in the speech signal without its reverberant distortion. Single-trial attention decoding accuracies based on 40–50s long blocks of data from 64 scalp electrodes were equally high (80–90% correct) in all considered listening environments and remained statistically significant using down to 10 scalp electrodes and short (<30-s) unaveraged EEG segments. In contrast to the robust decoding of the attended talker we found that decoding of the unattended talker deteriorated with the acoustic distortions. These results suggest that cortical activity tracks an attended speech signal in a way that is invariant to acoustic distortions encountered in real-life sound environments. Noise-robust attention decoding additionally suggests a potential utility of stimulus reconstruction techniques in attention-controlled brain-computer interfaces.

Single-trial EEG analysis using similaritymeasure.

Single-trial electroencephalogram (EEG) data are analyzed with similarity measure. Time-frequency representation is constructed from EEG signals. It is then weighted with t-statistics. Finally, the test data are discriminated with similarity measure. Compared with non-weighted version, the experimental results indicate that the proposed method obtains better results in classification accuracy.

EEG-Based Perceived Tactile Location Prediction

Previous studies have attempted to investigate the peripheral neural mechanisms implicated in tactile perception, but the neurophysiological data in humans involved in tactile spatial location perception to help the brain orient the body and interact with its surroundings are not well understood. In this paper, we use single-trial electroencephalogram (EEG) measurements to explore the perception of tactile stimuli located on participants’ right forearm, which were approximately equally spaced centered on the body midline, 2 leftward and 2 rightward of midline. An EEG-based signal analysis approach to predict the location of the tactile stimuli is proposed. Offline classification suggests that tactile location can be detected from EEG signals in single trial (four-class classifier for location discriminate can achieve up to 96.76%) with a short response time (600 milliseconds after stimulus presentation). From a human-machine-interaction (HMI) point of view, this could be used to design a real-time reactive control machine for patients, e.g., suffering from hypoesthesia.

The effects of facial color and inversion on the N170 event-related potential (ERP) component

Faces are important for social interaction because much can be perceived from facial details, including a person’s race, age, and mood. Recent studies have shown that both configural (e.g. face shape and inversion) and surface information (e.g. surface color and reflectance properties) are important for face perception. Therefore, the present study examined the effects of facial color and inverted face properties on event-related potential (ERP) responses, particularly the N170 component. Stimuli consisted of natural and bluish-colored faces. Faces were presented in both upright and upside down orientations. An ANOVA was used to analyze N170 amplitudes and verify the effects of the main independent variables. Analysis of N170 amplitude revealed the significant interactions between stimulus orientation and color. Subsequent analysis indicated that N170 was larger for bluish-colored faces than natural-colored faces, and N170 to natural-colored faces was larger in response to inverted stimulus as compared to upright stimulus. Additionally, a multivariate pattern analysis (MVPA) investigated face-processing dynamics without any prior assumptions. Results distinguished, above chance, both facial color and orientation from single-trial electroencephalogram (EEG) signals. Decoding performance for color classification of inverted faces was significantly diminished as compared to an upright orientation. This suggests that processing orientation is predominant over facial color. Taken together, the present findings elucidate the temporal and spatial distribution of orientation and color processing during face processing.

RSTFC: A Novel Algorithm for Spatio-Temporal Filtering and Classification of Single-Trial EEG.

Learning optimal spatio-temporal filters is a key to feature extraction for single-trial electroencephalogram (EEG) classification. The challenges are controlling the complexity of the learning algorithm so as to alleviate the curse of dimensionality and attaining computational efficiency to facilitate online applications, e.g., brain-computer interfaces (BCIs). To tackle these barriers, this paper presents a novel algorithm, termed regularized spatio-temporal filtering and classification (RSTFC), for single-trial EEG classification. RSTFC consists of two modules. In the feature extraction module, an l2 -regularized algorithm is developed for supervised spatio-temporal filtering of the EEG signals. Unlike the existing supervised spatio-temporal filter optimization algorithms, the developed algorithm can simultaneously optimize spatial and high-order temporal filters in an eigenvalue decomposition framework and thus be implemented highly efficiently. In the classification module, a convex optimization algorithm for sparse Fisher linear discriminant analysis is proposed for simultaneous feature selection and classification of the typically high-dimensional spatio-temporally filtered signals. The effectiveness of RSTFC is demonstrated by comparing it with several state-of-the-arts methods on three brain-computer interface (BCI) competition data sets collected from 17 subjects. Results indicate that RSTFC yields significantly higher classification accuracies than the competing methods. This paper also discusses the advantage of optimizing channel-specific temporal filters over optimizing a temporal filter common to all channels.