Trial EEG Research Articles

A Lightweight GCT-EEGNet for EEG-Based Individual Recognition Under Diverse Brain Conditions

A robust biometric system is essential to mitigate various security threats. Electroencephalography (EEG) brain signals present a promising alternative to other biometric traits due to their sensitivity, non-duplicability, resistance to theft, and individual-specific dynamics. However, existing EEG-based biometric systems employ deep neural networks, such as convolutional neural networks (CNNs) and recurrent neural networks (RNNs), which face challenges such as high parameter complexity, limiting their practical application. Additionally, their ability to generalize across a large number of subjects remains unclear. Moreover, they have been validated on datasets collected in controlled environments, which do not accurately reflect real-world scenarios involving diverse brain conditions. To overcome these challenges, we propose a lightweight neural network model, GCT–EEGNet, which is based on the design ideas of a CNN model and incorporates an attention mechanism to pay attention to the appropriate frequency bands for extracting discriminative features relevant to the identity of a subject despite diverse brain conditions. First, a raw EEG signal is decomposed into frequency bands and then passed to GCT–EEGNet for feature extraction, which utilizes a gated channel transformation (GCT) layer to selectively emphasize informative features from the relevant frequency bands. The extracted features were used for subject recognition through a cosine similarity metric that measured the similarity between feature vectors of different EEG trials to identify individuals. The proposed method was evaluated on a large dataset comprising 263 subjects. The experimental results demonstrated that the method achieved a correct recognition rate (CRR) of 99.23% and an equal error rate (EER) of 0.0014, corroborating its robustness against different brain conditions. The proposed model maintains low parameter complexity while keeping the expressiveness of representations, even with unseen subjects.

A Within-Subject Multimodal NIRS-EEG Classifier for Infant Data.

Functional Near Infrared Spectroscopy (fNIRS) and Electroencephalography (EEG) are commonly employed neuroimaging methods in developmental neuroscience. Since they offer complementary strengths and their simultaneous recording is relatively easy, combining them is highly desirable. However, to date, very few infant studies have been conducted with NIRS-EEG, partly because analyzing and interpreting multimodal data is challenging. In this work, we propose a framework to carry out a multivariate pattern analysis that uses an NIRS-EEG feature matrix, obtained by selecting EEG trials presented within larger NIRS blocks, and combining the corresponding features. Importantly, this classifier is intended to be sensitive enough to apply to individual-level, and not group-level data. We tested the classifier on NIRS-EEG data acquired from five newborn infants who were listening to human speech and monkey vocalizations. We evaluated how accurately the model classified stimuli when applied to EEG data alone, NIRS data alone, or combined NIRS-EEG data. For three out of five infants, the classifier achieved high and statistically significant accuracy when using features from the NIRS data alone, but even higher accuracy when using combined EEG and NIRS data, particularly from both hemoglobin components. For the other two infants, accuracies were lower overall, but for one of them the highest accuracy was still achieved when using combined EEG and NIRS data with both hemoglobin components. We discuss how classification based on joint NIRS-EEG data could be modified to fit the needs of different experimental paradigms and needs.

Session-independent subject-adaptive mental imagery BCI using selective filter-bank adaptive Riemannian features.

The brain-computer interfaces (BCIs) facilitate the users to exploit information encoded in neural signals, specifically electroencephalogram (EEG), to control devices and for neural rehabilitation. Mental imagery (MI)-driven BCI predicts the user's pre-meditated mental objectives, which could be deployed as command signals. This paper presents a novel learning-based framework for classifying MI tasks using EEG-based BCI. In particular, our work focuses on the variation in inter-session data and the extraction of multi-spectral user-tailored features for robust performance. Thus, the goal is to create a calibration-free subject-adaptive learning framework for various mental imagery tasks not restricted to motor imagery alone. In this regard, critical spectral bands and the best temporal window are first selected from the EEG training trials of the subject based on the Riemannian user learning distance metric (Dscore) that checks for distinct and stable patterns. The filtered covariance matrices of the EEG trials in each spectral band are then transformed towards a reference covariance matrix using the Riemannian transfer learning, enabling the different sessions to be comparable. The evaluation of our proposed Selective Time-window and Multi-scale Filter-Bank with Adaptive Riemannian (STFB-AR) features on four public datasets, including disabled subjects, showed around 15% and 8% improvement in mean accuracy over baseline and fixed filter-bank models, respectively.

Multi-source transfer learning via optimal transport feature ranking for EEG classification

Motor imagery (MI) brain-computer interface (BCI) paradigms have been extensively used in neurological rehabilitation. However, due to the required long calibration time and non-stationary nature of electroencephalogram (EEG) signals, it is challenging to obtain a substantial sample size of EEG data. Transfer learning (TL), which leverages knowledge from existing subjects to enhance the learning performance in new subjects, has been employed to solve this problem. Previous studies often extract sample features directly from the two domains for transfer learning, leading to poor features and negative transfer. To overcome this limitation, an optimal transport feature selection method was developed in this study to select the most suitable features for migration to enhance the generalization capability of the TL. This was achieved by analyzing the diagonal value similarity of the optimal transport coupling matrix. First, the Riemannian tangent space mapping method was used to map the sample covariance matrix of EEG trials from the Riemannian manifold to tangent space. Secondly, the optimal subset of EEG features was selected via the optimal transport feature selection method. Subsequently, the separate distribution alignment method was employed to minimize dissimilarities between the source and target domains. Finally, the weighted voting mechanisms were integrated for decision fusion. The proposed multi-source transfer feature learning (MSTFL) method was validated based on two public BCI Competition IV datasets. Our results demonstrated superior classification accuracy of 85.93% and 79.48%, respectively, on the two public datasets, compared to state-of-the-art models. These findings indicate the effectiveness of our proposed MSTFL model for both feature extraction and classification of MI signals.

T-TIME: Test-Time Information Maximization Ensemble for Plug-and-Play BCIs.

An electroencephalogram (EEG)-based brain-computer interface (BCI) enables direct communication between the human brain and a computer. Due to individual differences and non-stationarity of EEG signals, such BCIs usually require a subject-specific calibration session before each use, which is time-consuming and user-unfriendly. Transfer learning (TL) has been proposed to shorten or eliminate this calibration, but existing TL approaches mainly consider offline settings, where all unlabeled EEG trials from the new user are available. This article proposes Test-Time Information Maximization Ensemble (T-TIME) to accommodate the most challenging online TL scenario, where unlabeled EEG data from the new user arrive in a stream, and immediate classification is performed. T-TIME initializes multiple classifiers from the aligned source data. When an unlabeled test EEG trial arrives, T-TIME first predicts its labels using ensemble learning, and then updates each classifier by conditional entropy minimization and adaptive marginal distribution regularization. Our code is publicized. Extensive experiments on three public motor imagery based BCI datasets demonstrated that T-TIME outperformed about 20 classical and state-of-the-art TL approaches. To our knowledge, this is the first work on test time adaptation for calibration-free EEG-based BCIs, making plug-and-play BCIs possible.

(Micro)saccade-related potentials during face recognition: A study combining EEG, eye-tracking, and deconvolution modeling.

Under natural viewing conditions, complex stimuli such as human faces are typically looked at several times in succession, implying that their recognition may unfold across multiple eye fixations. Although electrophysiological (EEG) experiments on face recognition typically prohibit eye movements, participants still execute frequent (micro)saccades on the face, each of which generates its own visuocortical response. This finding raises the question of whether the fixation-related potentials (FRPs) evoked by these tiny gaze shifts also contain psychologically valuable information about face processing. Here, we investigated this question by corecording EEG and eye movements in an experiment with emotional faces (happy, angry, neutral). Deconvolution modeling was used to separate the stimulus ERPs to face onset from the FRPs generated by subsequent microsaccades-induced refixations on the face. As expected, stimulus ERPs exhibited typical emotion effects, with a larger early posterior negativity (EPN) for happy/angry compared with neutral faces. Eye tracking confirmed that participants made small saccades in 98% of the trials, which were often aimed at the left eye of the stimulus face. However, while each saccade produced a strong response over visual areas, this response was unaffected by the face's emotional expression, both for the first and for subsequent (micro)saccades. This finding suggests that the face's affective content is rapidly evaluated after stimulus onset, leading to only a short-lived sensory enhancement by arousing stimuli that does not repeat itself during immediate refixations. Methodologically, our work demonstrates how eye tracking and deconvolution modeling can be used to extract several brain responses from each EEG trial, providing insights into neural processing at different latencies after stimulus onset.

Motor Imagery Classification for Asynchronous EEG-Based Brain-Computer Interfaces.

Motor imagery (MI) based brain-computer interfaces (BCIs) enable the direct control of external devices through the imagined movements of various body parts. Unlike previous systems that used fixed-length EEG trials for MI decoding, asynchronous BCIs aim to detect the user's MI without explicit triggers. They are challenging to implement, because the algorithm needs to first distinguish between resting-states and MI trials, and then classify the MI trials into the correct task, all without any triggers. This paper proposes a sliding window prescreening and classification (SWPC) approach for MI-based asynchronous BCIs, which consists of two modules: a prescreening module to screen MI trials out of the resting-state, and a classification module for MI classification. Both modules are trained with supervised learning followed by self-supervised learning, which refines the feature extractors. Within-subject and cross-subject asynchronous MI classifications on four different EEG datasets validated the effectiveness of SWPC, i.e., it always achieved the highest average classification accuracy, and outperformed the best state-of-the-art baseline on each dataset by about 2%.

Alignment-Based Adversarial Training (ABAT) for Improving the Robustness and Accuracy of EEG-Based BCIs.

Machine learning has achieved great success in electroencephalogram (EEG) based brain-computer interfaces (BCIs). Most existing BCI studies focused on improving the decoding accuracy, with only a few considering the adversarial security. Although many adversarial defense approaches have been proposed in other application domains such as computer vision, previous research showed that their direct extensions to BCIs degrade the classification accuracy on benign samples. This phenomenon greatly affects the applicability of adversarial defense approaches to EEG-based BCIs. To mitigate this problem, we propose alignment-based adversarial training (ABAT), which performs EEG data alignment before adversarial training. Data alignment aligns EEG trials from different domains to reduce their distribution discrepancies, and adversarial training further robustifies the classification boundary. The integration of data alignment and adversarial training can make the trained EEG classifiers simultaneously more accurate and more robust. Experiments on five EEG datasets from two different BCI paradigms (motor imagery classification, and event related potential recognition), three convolutional neural network classifiers (EEGNet, ShallowCNN and DeepCNN) and three different experimental settings (offline within-subject cross-block/-session classification, online cross-session classification, and pre-trained classifiers) demonstrated its effectiveness. It is very intriguing that adversarial attacks, which are usually used to damage BCI systems, can be used in ABAT to simultaneously improve the model accuracy and robustness.

The Neural Mechanisms of Group Membership Effect on Emotional Mimicry: A Multimodal Study Combining Electromyography and Electroencephalography.

Emotional mimicry plays a vital role in understanding others' emotions and has been found to be modulated by social contexts, especially group membership. However, the neural mechanisms underlying this modulation remain unclear. We explored whether and how group membership modulated emotional mimicry using a multimodal method combining facial electromyography (fEMG) and electroencephalography (EEG). We instructed participants to passively view dynamic emotional faces (happy vs. angry) of others (in-group vs. out-group) and simultaneously recorded their fEMG and EEG responses. Then, we conducted combined analyses of fEMG-EEG by splitting the EEG trials into two mimicry intensity categories (high-intensity mimicry vs. low-intensity mimicry) according to fEMG activity. The fEMG results confirmed the occurrence of emotional mimicry in the present study but failed to find a group membership effect. However, the EEG results showed that participants mimicked in-group happiness and anger more than out-group. Importantly, this in-group preference involved different neural mechanisms in happiness and anger mimicry. In-group preference for happiness mimicry occurred at multiple neural mechanisms such as N1 (at P7, Pz, and P8), P2 (at Pz and P8), N2 (at P8), and P3 (at P7, Pz, and P8); in-group preference for anger mimicry occurred at P1 (at P7) and P2 (at Pz). Our findings provide new neural evidence for the effect of group membership on emotional mimicry by uncovering the temporal dynamics of this effect.

Time–space–frequency feature Fusion for 3-channel motor imagery classification

Low-channel EEG devices are crucial for portable and entertainment applications. However, the low spatial resolution of EEG presents challenges in decoding low-channel motor imagery. This study introduces TSFF-Net, a novel network architecture that integrates time–space–frequency features, effectively compensating for the limitations of single-mode feature extraction networks based on time-series or time–frequency modalities. TSFF-Net comprises four main components: time–frequency representation, time–frequency feature extraction, time–space feature extraction, and feature fusion and classification. Time–frequency representation and feature extraction transform raw EEG signals into time–frequency spectrograms and extract relevant features. The time–space network processes time-series EEG trials as input and extracts temporal–spatial features. Feature fusion employs Maximum Mean Discrepancy (MMD) loss to constrain the distribution of time–frequency and time–space features in the Reproducing Kernel Hilbert Space, subsequently combining these features using a weighted fusion approach to obtain effective time–space–frequency features. Moreover, few studies have explored the decoding of three-channel motor imagery based on time–frequency spectrograms. This study proposes a shallow, lightweight decoding architecture (TSFF-img) based on time–frequency spectrograms and compares its classification performance in low-channel motor imagery with other methods using two publicly available datasets. Experimental results demonstrate that TSFF-Net not only compensates for the shortcomings of single-mode feature extraction networks in EEG decoding, but also outperforms other state-of-the-art methods. Overall, TSFF-Net offers considerable advantages in decoding low-channel motor imagery and provides valuable insights for algorithmically enhancing low-channel EEG decoding.

Deep Learning for EEG‐Based Alzheimer’s Disease Diagnosis

AbstractBackgroundElectroencephalography (EEG) could be a powerful tool to diagnose Alzheimer’s Disease (AD) because it is pervasive, non‐invasive, and cost‐effective [1]. However, EEG‐based AD detection has suffered from poor performance and low generalizability due to the noisy nature of the signal and high inter‐patient variability. Here, we show that emerging deep learning techniques can alleviate these challenges and pave the way for the clinical use of EEG in AD detection.MethodWe formulate the problem of EEG‐based AD detection as a binary classification task. We propose an end‐to‐end deep neural network that utilizes contrastive representation learning to automatically learn low‐dimensional features from EEG trials. These features are then fed into a non‐linear classifier that predicts the probability of AD.ResultWe evaluated the model on a public dataset comprising 23 subjects (12 AD; 11 control) that have 663 trials in total [2]. (1) Patient‐dependent setup. In line with existing work in EEG‐based AD detection, we randomly assigned 80% of trials for training and the remaining 20% for tests, where the same patient could have trials in both groups. Our model achieved an F1 score of 99.35% which is competitive with, if not superior to, the state‐of‐the‐art baselines. (2) Patient‐independent setup. We then investigate a more challenging setup, where trials from a particular patient were assigned either to training (19 subjects) or to testing (4 subjects). The deep learning system is evaluated on patients it has not encountered before. Our model attains an F1 score of 86.45%, outperforming baselines by a significant margin (>20%).ConclusionOur results demonstrate that our deep learning model significantly improves the accuracy of EEG‐based AD diagnosis. In particular, our model sets the state‐of‐the‐art performance in the patient‐independent evaluation. This suggests that our model can learn a distinctive pattern of AD from a small group of subjects and apply it to unseen individuals: which is indispensable for real‐world deployment.[1] Tait, X. et al. Eeg Microstate Complexity for Aiding Early Diagnosis of Alzheimer’s Disease. Scientific Reports, 2020.[2] K. Smith, et al. Accounting for the Complex Hierarchical Topology of Eeg Phase‐Based Functional Connectivity in Network Binarisation. PloS One, 2017.

Enhancing multi-class SSVEP classification performance using regularized covariance estimators, spatiotemporal beamforming, and short-time EEG trials

One of the main challenges in BCI technologies is the problem of poor information transfer rate (ITR) of BCI systems and its effect on reducing the commands users can give. At the state of the art, the Linearly Constrained Minimum Variance (LCMV) beamformer is one of the most effective strategies that can provide satisfactory performance in the analysis of steady-state visual evoked potentials (SSVEP). When only a short-period SSVEP signal is available, a poor estimate of the covariance matrix used in the original LCMV method is obtained. Therefore, the main drawback of the conventional LCMV beamformer and its spatiotemporal form (LCMVst) is that they require a long-period SSVEP signal to achieve a reliable estimate for the beamformer weights. To cope with the mentioned problems, this paper presents two types of regularized estimators to achieve a robust estimate for the covariance matrix when only a short-period SSVEP signal is available. These two types, which are user-independent, include the convex combination (CC) and general linear combination (GLC) methods. In the next step, in order to provide suitable beamformer weights, the CC and GLC covariance estimators are combined with the original LCMVst beamformer. Furthermore, two new subject-based electrode selection algorithms, i.e., sequential forward electrode selection (SFES) algorithm and common spatial pattern (CSP) based technique, are applied to pick out the best collection of electrodes and to maximize the classification accuracy. The introduced methods were assessed using a four-class SSVEP dataset recorded on 20 subjects. Experimental results indicate that the proposed estimators optimally classify SSVEPs by estimating proper beamformer weights. The proposed LCMVst-CC and LCMVst-GLC beamformers improved the average classification accuracy by about 27.75% in comparison to the original LCMV-based beamformer at the 0.25-second segment length. Using SSVEP stimulations, we show that our proposed beamformers offer a better classification performance than the conventional LCMV-based spatiotemporal beamformer for all stimulation times.

Unsupervised domain adaptation for cross-patient seizure classification

Objective. Epileptic seizure is a chronic neurological disease affecting millions of patients. Electroencephalogram (EEG) is the gold standard in epileptic seizure classification. However, its low signal-to-noise ratio, strong non-stationarity, and large individual difference nature make it difficult to directly extend the seizure classification model from one patient to another. This paper considers multi-source unsupervised domain adaptation for cross-patient EEG-based seizure classification, i.e. there are multiple source patients with labeled EEG data, which are used to label the EEG trials of a new patient. Approach. We propose an source domain selection (SDS)-global domain adaptation (GDA)-target agent subdomain adaptation (TASA) approach, which includes SDS to filter out dissimilar source domains, GDA to align the overall distributions of the selected source domains and the target domain, and TASA to identify the most similar source domain to the target domain so that its labels can be utilized. Main results. Experiments on two public seizure datasets demonstrated that SDS-GDA-TASA outperformed 13 existing approaches in unsupervised cross-patient seizure classification. Significance. Our approach could save clinicians plenty of time in labeling EEG data for epilepsy patients, greatly increasing the efficiency of seizure diagnostics.

WRA-MTSI: A Robust Extended Source Imaging Algorithm Based on Multi-Trial EEG.

Reconstructing brain activities from electroencephalography (EEG) signals is crucial for studying brain functions and their abnormalities. However, since EEG signals are nonstationary and vulnerable to noise, brain activities reconstructed from single-trial EEG data are often unstable, and significant variability may occur across different EEG trials even for the same cognitive task. In an effort to leverage the shared information across the EEG data of multiple trials, this paper proposes a multi-trial EEG source imaging method based on Wasserstein regularization, termed WRA-MTSI. In WRA-MTSI, Wasserstein regularization is employed to perform multi-trial source distribution similarity learning, and the structured sparsity constraint is enforced to enable accurate estimation of the source extents, locations and time series. The resulting optimization problem is solved by a computationally efficient algorithm based on the alternating direction method of multipliers (ADMM). Both numerical simulations and real EEG data analysis demonstrate that WRA-MTSI outperforms existing single-trial ESI methods (e.g., wMNE, LORETA, SISSY, and SBL) in mitigating the influence of artifacts in EEG data. Moreover, WRA-MTSI yields superior performance compared to other state-of-the-art multi-trial ESI methods (e.g., group lasso, the dirty model, and MTW) in estimating source extents. WRA-MTSI may serve as an effective robust EEG source imaging method in the presence of multi-trial noisy EEG data.

Classification of Motor Imagery Using Trial Extension in Spatial Domain with Rhythmic Components of EEG

Electrical activities of the human brain can be recorded with electroencephalography (EEG). To characterize motor imagery (MI) tasks for brain–computer interface (BCI) implementation is an easy and cost-effective tool. The MI task is represented by a short-time trial of multichannel EEG. In this paper, the signal of each channel of raw EEG is decomposed into a finite set of narrowband signals using a Fourier-transformation-based bandpass filter. Rhythmic components of EEG are represented by each of the narrowband signals that characterize the brain activities related to MI tasks. The subband signals are arranged to extend the dimension of the EEG trial in the spatial domain. The spatial features are extracted from the set of extended trials using a common spatial pattern (CSP). An optimum number of features are employed to classify the motor imagery tasks using an artificial neural network. An integrated approach with full-band and narrowband signals is implemented to derive discriminative features for MI classification. In addition, the subject-dependent parameter optimization scheme enhances the performance of the proposed method. The performance evaluation of the proposed method is obtained using two publicly available benchmark datasets (Dataset I and Dataset II). The experimental results in terms of classification accuracy (93.88% with Dataset I and 91.55% with Dataset II) show that it performs better than the recently developed algorithms. The enhanced MI classification accuracy is very much applicable in BCI implementation.

Classification of Epileptic Seizure Types Using Multiscale Convolutional Neural Network and Long Short-Term Memory

The accurate classification of seizure types using electroencephalography (EEG) signals plays a vital role in determining a precise treatment plan and therapy for epilepsy patients. Among the available deep network models, Convolutional Neural Networks (CNNs) are the most widely adopted models for learning and representing EEG signals. However, typical CNNs have high computational complexity, leading to overfitting problems. This paper proposes the design of two effective, lightweight deep network models; the 1D multiscale neural network (1D-MSCNet) model and the Long Short-term Memory (LSTM)-based compact CNN (EEG-LSTMNet) model. The 1D-MSCNet model comprises three modules: a spectral–temporal convolution module, a spatial convolution module, and a classification module. It extracts features from input EEG trials at multiple frequency/time ranges, identifying relationships between the spatial distribution of their channels. The EEG-LSTMNet model includes three convolutional layers, namely temporal, depthwise, and separable layers, a single LSTM layer, and two fully connected classification layers to extract discriminative EEG feature representations. Both models have been applied to the same EEG trials collected from the Temple University Hospital (TUH) database. Results revealed F1-score values of 96.9% and 98.4% for the 1D-MSCNet and EEG-LSTMNet, respectively. Based on the demonstrated outcomes, both models outperform related state-of-the-art methods due to their architectures’ adoption of 1D modules and layers that reduce the computational effort needed, solve the overfitting problem, and enhance classification efficiency. Hence, both models could be valuable additions for neurologists to help them decide upon precise treatments and drugs for patients depending on their type of seizure.

A shallow mirror transformer for subject-independent motor imagery BCI

Objective:Motor imagery BCI plays an increasingly important role in motor disorders rehabilitation. However, the position and duration of the discriminative segment in an EEG trial vary from subject to subject and even trial to trial, and this leads to poor performance of subject-independent motor imagery classification. Thus, determining how to detect and utilize the discriminative signal segments is crucial for improving the performance of subject-independent motor imagery BCI. Approach:In this paper, a shallow mirror transformer is proposed for subject-independent motor imagery EEG classification. Specifically, a multihead self-attention layer with a global receptive field is employed to detect and utilize the discriminative segment from the entire input EEG trial. Furthermore, the mirror EEG signal and the mirror network structure are constructed to improve the classification precision based on ensemble learning. Finally, the subject-independent setup was used to evaluate the shallow mirror transformer on motor imagery EEG signals from subjects existing in the training set and new subjects. Main results:The experiments results on BCI Competition IV datasets 2a and 2b and the OpenBMI dataset demonstrated the promising effectiveness of the proposed shallow mirror transformer. The shallow mirror transformer obtained average accuracies of 74.48% and 76.1% for new subjects and existing subjects, respectively, which were highest among the compared state-of-the-art methods. In addition, visualization of the attention score showed the ability of discriminative EEG segment detection. This paper demonstrated that multihead self-attention is effective in capturing global EEG signal information in motor imagery classification. Significance:This study provides an effective model based on a multihead self-attention layer for subject-independent motor imagery-based BCIs. To the best of our knowledge, this is the shallowest transformer model available, in which a small number of parameters promotes the performance in motor imagery EEG classification for such a small sample problem.

Lightweight Source-Free Transfer for Privacy-Preserving Motor Imagery Classification

Transfer learning utilizes data or knowledge in one problem to help solve a related problem. It is particularly useful in electroencephalogram (EEG) based motor imagery (MI) classification, to handle high intra-subject and/or cross-subject variations. This paper considers offline unsupervised cross-subject MI classification, i.e., we have labeled EEG trials from several source subjects, but only unlabeled EEG trials from the target subject. Existing transfer learning approaches usually make use of the source domain data directly in target model learning. To protect the privacy of the source subjects, we propose lightweight source-free transfer (LSFT), which first generates source models locally and encapsulates them as model application programming interfaces (APIs), then constructs a virtual intermediate domain to transfer the knowledge in the source domains to the target domain, and finally performs feature adaptation learning. Compared with existing deep transfer learning approaches, LSFT does not need to transfer from massive source data or models, is computationally efficient, and has a small number of parameters. Experiments on four benchmark MI datasets demonstrated that LSFT outperformed 13 different approaches, including several state-of-the-art transfer learning approaches that make use of the source domain samples or model parameters directly.

Data augmentation of SSVEPs using source aliasing matrix estimation for brain-computer interfaces.

Currently, ensemble task-related component analysis (eTRCA) and task discriminative component analysis (TDCA) are the state-of-the-art algorithms for steady-state visual evoked potential (SSVEP)-based brain-computer interfaces (BCIs). However, training the BCIs requires multiple calibration trials. With insufficient calibration data, the accuracy of the BCI will degrade, or even become invalid with only one calibration trial. However, collecting a large amount of electroencephalography (EEG) data for calibration is a time-consuming and laborious process, which hinders the practical use of eTRCA and TDCA. This study proposed a novel method, namely Source Aliasing Matrix Estimation (SAME), to augment the calibration data for SSVEP-BCIs. SAME could generate artificial EEG trials with the featured SSVEPs. Its effectiveness was evaluated using two public datasets (i.e., Benchmark, BETA). When combined with SAME, both eTRCA and TDCA had significantly improved performance with a limited number of calibration data. Specifically, SAME increased the average accuracy of eTRCA and TDCA by about 12% and 3%, respectively, with as few as two calibration trials. Notably, SAME enabled eTRCA and TDCA to work well with a single calibration trial, achieving an average accuracy >90% for the Benchmark dataset and >70% for the BETA dataset with 1-second EEG. SAME is an effective method for SSVEP-BCIs to augment the calibration data, thereby significantly enhancing the performance of eTRCA and TDCA. We propose a new data-augmentation method that is compatible with the state-of-the-art algorithms of SSVEP-based BCIs. It can significantly reduce the efforts required to calibrate SSVEP-BCIs, which is promising for the development of practical BCIs.

Decoding Multi-Class Motor Imagery and Motor Execution Tasks Using Riemannian Geometry Algorithms on Large EEG Datasets.

The use of Riemannian geometry decoding algorithms in classifying electroencephalography-based motor-imagery brain-computer interfaces (BCIs) trials is relatively new and promises to outperform the current state-of-the-art methods by overcoming the noise and nonstationarity of electroencephalography signals. However, the related literature shows high classification accuracy on only relatively small BCI datasets. The aim of this paper is to provide a study of the performance of a novel implementation of the Riemannian geometry decoding algorithm using large BCI datasets. In this study, we apply several Riemannian geometry decoding algorithms on a large offline dataset using four adaptation strategies: baseline, rebias, supervised, and unsupervised. Each of these adaptation strategies is applied in motor execution and motor imagery for both scenarios 64 electrodes and 29 electrodes. The dataset is composed of four-class bilateral and unilateral motor imagery and motor execution of 109 subjects. We run several classification experiments and the results show that the best classification accuracy is obtained for the scenario where the baseline minimum distance to Riemannian mean has been used. The mean accuracy values up to 81.5% for motor execution, and up to 76.4% for motor imagery. The accurate classification of EEG trials helps to realize successful BCI applications that allow effective control of devices.