Online Privacy-Preserving EEG Classification by Source-Free Transfer Learning.

Electroencephalogram (EEG) signals play an important role in brain-computer interface (BCI) applications. Recent studies have utilized transfer learning to assist the learning task in the new subject, i.e., target domain, by leveraging beneficial information from previous subjects, i.e., source domains. Nevertheless, EEG signals involve sensitive personal mental and health information. Thus, privacy concern becomes a critical issue. In addition, existing methods mostly assume that a portion of the new subject's data is available and perform alignment or adaptation between the source and target domains. However, in some practical scenarios, new subjects prefer prompt BCI utilization over the time-consuming process of collecting data for calibration and adaptation, which makes the above assumption difficult to hold. To address the above challenges, we propose Online Source-Free Transfer Learning (OSFTL) for privacy-preserving EEG classification. Specifically, the learning procedure contains offline and online stages. At the offline stage, multiple model parameters are obtained based on the EEG samples from multiple source subjects. OSFTL only needs access to these source model parameters to preserve the privacy of the source subjects. At the online stage, a target classifier is trained based on the online sequence of EEG instances. Subsequently, OSFTL learns a weighted combination of the source and target classifiers to obtain the final prediction for each target instance. Moreover, to ensure good transferability, OSFTL dynamically updates the transferred weight of each source domain based on the similarity between each source classifier and the target classifier. Comprehensive experiments on both simulated and real-world applications demonstrate the effectiveness of the proposed method, indicating the potential of OSFTL to facilitate the deployment of BCI applications outside of controlled laboratory settings.

Wasserstein generative adversarial network with gradient penalty and convolutional neural network based motor imagery EEG classification.

Objective.Due to the difficulty in acquiring motor imagery electroencephalography (MI-EEG) data and ensuring its quality, insufficient training data often leads to overfitting and inadequate generalization capabilities of deep learning-based classification networks. Therefore, we propose a novel data augmentation method and deep learning classification model to enhance the decoding performance of MI-EEG further.Approach.The raw EEG signals were transformed into the time-frequency maps as the input to the model by continuous wavelet transform. An improved Wasserstein generative adversarial network with gradient penalty data augmentation method was proposed, effectively expanding the dataset used for model training. Additionally, a concise and efficient deep learning model was designed to improve decoding performance further.Main results.It has been demonstrated through validation by multiple data evaluation methods that the proposed generative network can generate more realistic data. Experimental results on the BCI Competition IV 2a and 2b datasets and the actual collected dataset show that classification accuracies are 83.4%, 89.1% and 73.3%, and Kappa values are 0.779, 0.782 and 0.644, respectively. The results indicate that the proposed model outperforms state-of-the-art methods.Significance.Experimental results demonstrate that this method effectively enhances MI-EEG data, mitigates overfitting in classification networks, improves MI classification accuracy, and holds positive implications for MI tasks.

A Learnable and Explainable Wavelet Neural Network for EEG Artifacts Detection and Classification.

Electroencephalography (EEG) artifacts are very common in clinical diagnosis and can heavily impact diagnosis. Manual screening of artifact events is labor-intensive with little benefit. Therefore, exploring algorithms for automatic detection and classification of EEG artifacts can significantly assist clinical diagnosis. In this paper, we propose a learnable and explainable wavelet neural network (WaveNet) for EEG artifact detection and classification. The model is powered by the wavelet decomposition block based on invertible neural network, which can extract signal features without information loss, and a tree generator for building wavelet tree structure automatically. They provide the model with good feature extraction capabilities and explainability. To evaluate the model's performance more fairly, we introduce the base point level matching score (BASE) and the Event-Aligned Compensation Scoring (EACS) at the event level as two metrics for model performance evaluation. On the challenging Temple University EEG Artifact (TUAR) dataset, our model outperforms other baselines in terms of F1-score for both artifact detection and classification tasks. The case study also validates the model's ability to offer explainability for predictions based on frequency band energy, suggesting potential applications in clinical diagnosis.

Unsupervised heterogeneous domain adaptation for EEG classification.

Objective.Domain adaptation has been recognized as a potent solution to the challenge of limited training data for electroencephalography (EEG) classification tasks. Existing studies primarily focus on homogeneous environments, however, the heterogeneous properties of EEG data arising from device diversity cannot be overlooked. This motivates the development of heterogeneous domain adaptation methods that can fully exploit the knowledge from an auxiliary heterogeneous domain for EEG classification.Approach.In this article, we propose a novel model named informative representation fusion (IRF) to tackle the problem of unsupervised heterogeneous domain adaptation in the context of EEG data. In IRF, we consider different perspectives of data, i.e. independent identically distributed (iid) and non-iid, to learn different representations. Specifically, from the non-iid perspective, IRF models high-order correlations among data by hypergraphs and develops hypergraph encoders to obtain data representations of each domain. From the non-iid perspective, by applying multi-layer perceptron networks to the source and target domain data, we achieve another type of representation for both domains. Subsequently, an attention mechanism is used to fuse these two types of representations to yield informative features. To learn transferable representations, the maximum mean discrepancy is utilized to align the distributions of the source and target domains based on the fused features.Main results.Experimental results on several real-world datasets demonstrate the effectiveness of the proposed model.Significance.This article handles an EEG classification situation where the source and target EEG data lie in different spaces, and what's more, under an unsupervised learning setting. This situation is practical in the real world but barely studied in the literature. The proposed model achieves high classification accuracy, and this study is important for the commercial applications of EEG-based BCIs.

Beta-band power classification of go/no-go arm-reaching responses in the human hippocampus.

Objective.Can we classify movement execution and inhibition from hippocampal oscillations during arm-reaching tasks? Traditionally associated with memory encoding, spatial navigation, and motor sequence consolidation, the hippocampus has come under scrutiny for its potential role in movement processing. Stereotactic electroencephalography (SEEG) has provided a unique opportunity to study the neurophysiology of the human hippocampus during motor tasks. In this study, we assess the accuracy of discriminant functions, in combination with principal component analysis (PCA), in classifying between 'Go' and 'No-go' trials in a Go/No-go arm-reaching task.Approach.Our approach centers on capturing the modulation of beta-band (13-30 Hz) power from multiple SEEG contacts in the hippocampus and minimizing the dimensional complexity of channels and frequency bins. This study utilizes SEEG data from the human hippocampus of 10 participants diagnosed with epilepsy. Spectral power was computed during a 'center-out' Go/No-go arm-reaching task, where participants reached or withheld their hand based on a colored cue. PCA was used to reduce data dimension and isolate the highest-variance components within the beta band. The Silhouette score was employed to measure the quality of clustering between 'Go' and 'No-go' trials. The accuracy of five different discriminant functions was evaluated using cross-validation.Main results.The Diagonal-Quadratic model performed best of the 5 classification models, exhibiting the lowest error rate in all participants (median: 9.91%, average: 14.67%). PCA showed that the first two principal components collectively accounted for 54.83% of the total variance explained on average across all participants, ranging from 36.92% to 81.25% among participants.Significance.This study shows that PCA paired with a Diagonal-Quadratic model can be an effective method for classifying between Go/No-go trials from beta-band power in the hippocampus during arm-reaching responses. This emphasizes the significance of hippocampal beta-power modulation in motor control, unveiling its potential implications for brain-computer interface applications.

A diagonal masking self-attention-based multi-scale network for motor imagery classification.

Objective. Electroencephalography (EEG)-based motor imagery (MI) is a promising paradigm for brain-computer interface (BCI), but the non-stationarity and low signal-to-noise ratio of EEG signals make it a challenging task.Approach. To achieve high-precision MI classification, we propose a Diagonal Masking Self-Attention-based Multi-Scale Network (DMSA-MSNet) to fully develop, extract, and emphasize features from different scales. First, for local features, a multi-scale temporal-spatial block is proposed to extract features from different receptive fields. Second, an adaptive branch fusion block is specifically designed to bridge the semantic gap between these coded features from different scales. Finally, in order to analyze global information over long ranges, a diagonal masking self-attention block is introduced, which highlights the most valuable features in the data.Main results. The proposed DMSA-MSNet outperforms state-of-the-art models on the BCI Competition IV 2a and the BCI Competition IV 2b datasets.Significance. Our study achieves rich information extraction from EEG signals and provides an effective solution for MI classification.

Model and Data Dual-Driven Double-Point Observation Network for Ultra-Short MI EEG Classification.

Although deep networks have succeeded in various signal classification tasks, the time sequence samples used to train the deep models are usually required to reach a certain length. Especially, in brain computer interface (BCI) research, around 3.5s-long motor imagery (MI) Electroencephalography (EEG) samples are needed to obtain satisfactory classification performance. This time-span requirement of the training samples makes real-time MI BCI systems impossible to implement based on deep networks, which restricts the related researches within laboratory and makes practical application hard to accomplish. To address this issue, a double-point observation deep network (DoNet) is developed to classify ultra-short samples buried in noise. First, an analytical solution is developed theoretically to perform ultra-short signal classification based on double-point couples. Then, a signal-noise model is constructed to study the interference of noise on classification based on double-point couples. Based on which, an independent identical distribution condition is utilized to improve the classification accuracy in a data-driven manner. Combining the theoretical model and data-driven mechanism, DoNet can construct a steady data-distribution for the double-point couples of the samples with the same label. Therefore, the conditional probability of each double-point couple of a test sample can be obtained. With a voting strategy, the samples can be accurately classified by fusing these conditional probabilities. Meanwhile, the noise interference can be suppressed. DoNet has been evaluated on two public EEG datasets. Compared to most state-of-the-art methods, the 1s-long EEG signal classification accuracy has been improved by more than 3%.

Exploring inter-trial coherence for inner speech classification in EEG-based brain-computer interface.

Objective. In recent years, electroencephalogram (EEG)-based brain-computer interfaces (BCIs) applied to inner speech classification have gathered attention for their potential to provide a communication channel for individuals with speech disabilities. However, existing methodologies for this task fall short in achieving acceptable accuracy for real-life implementation. This paper concentrated on exploring the possibility of using inter-trial coherence (ITC) as a feature extraction technique to enhance inner speech classification accuracy in EEG-based BCIs.Approach. To address the objective, this work presents a novel methodology that employs ITC for feature extraction within a complex Morlet time-frequency representation. The study involves a dataset comprising EEG recordings of four different words for ten subjects, with three recording sessions per subject. The extracted features are then classified using k-nearest-neighbors (kNNs) and support vector machine (SVM).Main results. The average classification accuracy achieved using the proposed methodology is 56.08% for kNN and 59.55% for SVM. These results demonstrate comparable or superior performance in comparison to previous works. The exploration of inter-trial phase coherence as a feature extraction technique proves promising for enhancing accuracy in inner speech classification within EEG-based BCIs.Significance. This study contributes to the advancement of EEG-based BCIs for inner speech classification by introducing a feature extraction methodology using ITC. The obtained results, on par or superior to previous works, highlight the potential significance of this approach in improving the accuracy of BCI systems. The exploration of this technique lays the groundwork for further research toward inner speech decoding.

DCNet: A Self-Supervised EEG Classification Framework for Improving Cognitive Computing-Enabled Smart Healthcare.

Cognitive computing endeavors to construct models that emulate brain functions, which can be explored through electroencephalography (EEG). Developing precise and robust EEG classification models is crucial for advancing cognitive computing. Despite the high accuracy of supervised EEG classification models, they are constrained by labor-intensive annotations and poor generalization. Self-supervised models address these issues but encounter difficulties in matching the accuracy of supervised learning. Three challenges persist: 1) capturing temporal dependencies in EEG; 2) adapting loss functions to describe feature similarities in self-supervised models; and 3) addressing the prevalent issue of data imbalance in EEG. This study introduces the DreamCatcher Network (DCNet), a self-supervised EEG classification framework with a two-stage training strategy. The first stage extracts robust representations through contrastive learning, and the second stage transfers the representation encoder to a supervised EEG classification task. DCNet utilizes time-series contrastive learning to autonomously construct representations that comprehensively capture temporal correlations. A novel loss function, SelfDreamCatcherLoss, is proposed to evaluate the similarities between these representations and enhance the performance of DCNet. Additionally, two data augmentation methods are integrated to alleviate class imbalances. Extensive experiments show the superiority of DCNet over the current state-of-the-art models, achieving high accuracy on both the Sleep-EDF and HAR datasets. It holds substantial promise for revolutionizing sleep disorder detection and expediting the development of advanced healthcare systems driven by cognitive computing.

Alteration of cortical functional networks in mood disorders with resting-state electroencephalography.

Studies comparing bipolar disorder (BD) and major depressive disorder (MDD) are scarce, and the neuropathology of these disorders is poorly understood. This study investigated source-level cortical functional networks using resting-state electroencephalography (EEG) in patients with BD and MDD. EEG was recorded in 35 patients with BD, 39 patients with MDD, and 42 healthy controls (HCs). Graph theory-based source-level weighted functional networks were assessed via strength, clustering coefficient (CC), and path length (PL) in six frequency bands. At the global level, patients with BD and MDD showed higher strength and CC, and lower PL in the high beta band, compared to HCs. At the nodal level, compared to HCs, patients with BD showed higher high beta band nodal CCs in the right precuneus, left isthmus cingulate, bilateral paracentral, and left superior frontal; however, patients with MDD showed higher nodal CC only in the right precuneus compared to HCs. Although both MDD and BD patients had similar global level network changes, they had different nodal level network changes compared to HCs. Our findings might suggest more altered cortical functional network in patients with BD than in those with MDD.

An EEG Classification-Based Method for Single-Trial N170 Latency Detection and Estimation.

Event-related potentials (ERPs) can reflect the high-level thinking activities of the brain. In ERP analysis, the superposition and averaging method is often used to estimate ERPs. However, the single-trial ERP estimation can provide researchers with more information on cognitive activities. In recent years, more and more researchers try to find an effective method to extract single-trial ERPs, because most of the existing methods have poor generalization ability or suffer from strong assumptions about the characteristics of ERPs, resulting in unsatisfactory results under the condition of a very low signal-to-noise ratio. In this paper, an EEG classification-based method for single-trial ERP detection and estimation was proposed. This study used a linear generated EEG model containing templates of ERP local descriptors which include amplitude and latency, and this model can avoid the invalid assumption about ERPs taken by other methods. The purpose of this method is not to recover the whole ERP waveform but to model the amplitude and latency of ERP components. This method afterwards examined the three machine learning models including logistic regression, neural network, and support vector machine in the EEG signal classification for ERP detection and selected the best performed MLPNN model for detection. To get the utmost out of information produced in the classification process, this study also used extra information to propose a new optimization model, with which outperformed detection results were obtained. Performance of the proposed method is evaluated on simulated N170 and real P50 data sets, and the results show that the model is more effective than the Woody filter and the SingleTrialEM algorithm. These results are also consistent with the conclusion of sensory gating, which demonstrated good generalization ability.

A Deep Learning-Based Classification Method for Different Frequency EEG Data.

In recent years, the research on electroencephalography (EEG) has focused on the feature extraction of EEG signals. The development of convenient and simple EEG acquisition devices has produced a variety of EEG signal sources and the diversity of the EEG data. Thus, the adaptability of EEG classification methods has become significant. This study proposed a deep network model for autonomous learning and classification of EEG signals, which could self-adaptively classify EEG signals with different sampling frequencies and lengths. The artificial design feature extraction methods could not obtain stable classification results when analyzing EEG data with different sampling frequencies. However, the proposed depth network model showed considerably better universality and classification accuracy, particularly for EEG signals with short length, which was validated by two datasets.

The influence of the abundance and morphology of epileptiform discharges on diagnostic accuracy: How many spikes you need to spot in an EEG.

OBJECTIVE: The operational definition of interictal epileptiform discharges (IEDs) of the International Federation of Clinical Neurophysiology (IFCN) described six morphological criteria. Our objective was to assess the impact of pattern-repetition in the EEG-recording, on the diagnostic accuracy of using the IFCN criteria. For clinical implementation, specificity over 95% was set as target. METHODS: Interictal EEG-recordings of 20-minutes, containing sharp-transients, from 60 patients (30 with epilepsy and 30 with non-epileptic paroxysmal events) were evaluated by three experts, who first marked IEDs solely based on expert opinion, and then, independently from the first session evaluated the presence of the IFCN criteria for each sharp-transient. The gold standard was derived from long-term video-EEG recordings of the patients habitual paroxysmal episodes. RESULTS: Presence of at least one discharge fulfilling five criteria provided a specificity of 100% (sensitivity: 70%). For discharges fulfilling fewer criteria, a higher number of discharges was needed to keep the specificity over 95% (5 discharges, when only 3 criteria were fulfilled). A sequential combination of these sets of criteria and thresholds provided a specificity of 97% and sensitivity of 80%. CONCLUSIONS: Pattern-repetition and IED morphology influence diagnostic accuracy. SIGNIFICANCE: Systematic application of these criteria will improve quality of clinical EEG interpretation.

Diagnosis of Alzheimer's Disease by Time-Dependent Power Spectrum Descriptors and Convolutional Neural Network Using EEG Signal.

Using strategies that obtain biomarkers where early symptoms coincide, the early detection of Alzheimer's disease and its complications is essential. Electroencephalogram is a technology that allows thousands of neurons with equal spatial orientation of the duration of cerebral cortex electrical activity to be registered by postsynaptic potential. Therefore, in this paper, the time-dependent power spectrum descriptors are used to diagnose the electroencephalogram signal function from three groups: mild cognitive impairment, Alzheimer's disease, and healthy control test samples. The final feature used in three modes of traditional classification methods is recorded: k-nearest neighbors, support vector machine, linear discriminant analysis approaches, and documented results. Finally, for Alzheimer's disease patient classification, the convolutional neural network architecture is presented. The results are indicated using output assessment. For the convolutional neural network approach, the accurate meaning of accuracy is 82.3%. 85% of mild cognitive impairment cases are accurately detected in-depth, but 89.1% of the Alzheimer's disease and 75% of the healthy population are correctly diagnosed. The presented convolutional neural network outperforms other approaches because performance and the k-nearest neighbors' approach is the next target. The linear discriminant analysis and support vector machine were at the low area under the curve values.

Ambulatory EEG to Classify the Epilepsy Syndrome.

SUMMARY: Recording of interictal epileptiform discharges to classify the epilepsy syndrome is one of the most common indications for ambulatory EEG. Ambulatory EEG has superior sampling compared with standard EEG recordings and advantages in terms of cost-effectiveness and convenience compared with a prolonged inpatient EEG study. Ambulatory EEG allows for EEG recording in all sleep stages and transitional states, which can be very helpful in capturing interictal epileptiform discharges. In the absence of interictal epileptiform discharges or in patients with atypical events, the characterization of an epilepsy syndrome may require recording of the habitual events. Diagnostic ambulatory EEG can be a useful alternative to inpatient video-EEG monitoring in a selected number of patients with frequent events who do not require medication taper or seizure testing for surgical localization.

A Review of Automated Techniques for Assisting the Early Detection of Alzheimer's Disease with a Focus on EEG.

In this paper, we review state-of-the-art approaches that apply signal processing (SP) and machine learning (ML) to automate the detection of Alzheimer's disease (AD) and its prodromal stages. In the first part of the document, we describe the economic and social implications of the disease, traditional diagnosis techniques, and the fundaments of automated AD detection. Then, we present electroencephalography (EEG) as an appropriate alternative for the early detection of AD, owing to its reduced cost, portability, and non-invasiveness. We also describe the main time and frequency domain EEG features that are employed in AD detection. Subsequently, we examine some of the main studies of the last decade that aim to provide an automatic detection of AD and its previous stages by means of SP and ML. In these studies, brain data was acquired using multiple medical techniques such as magnetic resonance imaging, positron emission tomography, and EEG. The main aspects of each approach, namely feature extraction, classification model, validation approach, and performance metrics, are compiled and discussed. Lastly, a set of conclusions and recommendations for future research on AD automatic detection are drawn in the final section of the paper.

Adaptive transfer learning for EEG motor imagery classification with deep Convolutional Neural Network.

In recent years, deep learning has emerged as a powerful tool for developing Brain-Computer Interface (BCI) systems. However, for deep learning models trained entirely on the data from a specific individual, the performance increase has only been marginal owing to the limited availability of subject-specific data. To overcome this, many transfer-based approaches have been proposed, in which deep networks are trained using pre-existing data from other subjects and evaluated on new target subjects. This mode of transfer learning however faces the challenge of substantial inter-subject variability in brain data. Addressing this, in this paper, we propose 5 schemes for adaptation of a deep convolutional neural network (CNN) based electroencephalography (EEG)-BCI system for decoding hand motor imagery (MI). Each scheme fine-tunes an extensively trained, pre-trained model and adapt it to enhance the evaluation performance on a target subject. We report the highest subject-independent performance with an average (N=54) accuracy of 84.19% (±9.98%) for two-class motor imagery, while the best accuracy on this dataset is 74.15% (±15.83%) in the literature. Further, we obtain a statistically significant improvement (p=0.005) in classification using the proposed adaptation schemes compared to the baseline subject-independent model.

Active probing to highlight approaching transitions to ictal states in coupled neural mass models.

The extraction of electrophysiological features that reliably forecast the occurrence of seizures is one of the most challenging goals in epilepsy research. Among possible approaches to tackle this problem is the use of active probing paradigms in which responses to stimuli are used to detect underlying system changes leading up to seizures. This work evaluates the theoretical and mechanistic underpinnings of this strategy using two coupled populations of the well-studied Wendling neural mass model. Different model settings are evaluated, shifting parameters (excitability, slow inhibition, or inter-population coupling gains) from normal towards ictal states while probing stimuli are applied every 2 seconds to the input of either one or both populations. The correlation between the extracted features and the ictogenic parameter shifting indicates if the impending transition to the ictal state may be identified in advance. Results show that not only can the response to the probing stimuli forecast seizures but this is true regardless of the altered ictogenic parameter. That is, similar feature changes are highlighted by probing stimuli responses in advance of the seizure including: increased response variance and lag-1 autocorrelation, decreased skewness, and increased mutual information between the outputs of both model subsets. These changes were mostly restricted to the stimulated population, showing a local effect of this perturbational approach. The transition latencies from normal activity to sustained discharges of spikes were not affected, suggesting that stimuli had no pro-ictal effects. However, stimuli were found to elicit interictal-like spikes just before the transition to the ictal state. Furthermore, the observed feature changes highlighted by probing the neuronal populations may reflect the phenomenon of critical slowing down, where increased recovery times from perturbations may signal the loss of a systems' resilience and are common hallmarks of an impending critical transition. These results provide more evidence that active probing approaches highlight information about underlying system changes involved in ictogenesis and may be able to play a role in assisting seizure forecasting methods which can be incorporated into early-warning systems that ultimately enable closing the loop for targeted seizure-controlling interventions.

Cross-Domain Classification Model With Knowledge Utilization Maximization for Recognition of Epileptic EEG Signals.

Conventional classification models for epileptic EEG signal recognition need sufficient labeled samples as training dataset. In addition, when training and testing EEG signal samples are collected from different distributions, for example, due to differences in patient groups or acquisition devices, such methods generally cannot perform well. In this paper, a cross-domain classification model with knowledge utilization maximization called CDC-KUM is presented, which takes advantage of the data global structure provided by the labeled samples in the related domain and unlabeled samples in the current domain. Through mapping the data into kernel space, the pairwise constraint regularization term is combined together the predictive differences of the labeled data in the source domain. Meanwhile, the soft clustering regularization term using quadratic weights and Gini-Simpson diversity is applied to exploit the distribution information of unlabeled data in the target domain. Experimental results show that CDC-KUM model outperformed several traditional non-transfer and transfer classification methods for recognition of epileptic EEG signals.

Generalized Learning Riemannian Space Quantization: A Case Study on Riemannian Manifold of SPD Matrices.

Learning vector quantization (LVQ) is a simple and efficient classification method, enjoying great popularity. However, in many classification scenarios, such as electroencephalogram (EEG) classification, the input features are represented by symmetric positive-definite (SPD) matrices that live in a curved manifold rather than vectors that live in the flat Euclidean space. In this article, we propose a new classification method for data points that live in the curved Riemannian manifolds in the framework of LVQ. The proposed method alters generalized LVQ (GLVQ) with the Euclidean distance to the one operating under the appropriate Riemannian metric. We instantiate the proposed method for the Riemannian manifold of SPD matrices equipped with the Riemannian natural metric. Empirical investigations on synthetic data and real-world motor imagery EEG data demonstrate that the performance of the proposed generalized learning Riemannian space quantization can significantly outperform the Euclidean GLVQ, generalized relevance LVQ (GRLVQ), and generalized matrix LVQ (GMLVQ). The proposed method also shows competitive performance to the state-of-the-art methods on the EEG classification of motor imagery tasks.