Search

Steady-state visual evoked potential (SSVEP) is one of the most commonly used control signals in the brain–computer interface (BCI) systems. However, the conventional spatial filtering methods for SSVEP classification highly depend on the subject-specific calibration data. The need for the methods that can alleviate the demand for the calibration data becomes urgent. In recent years, developing the methods that can work in inter-subject scenario has become a promising new direction. As a popular deep learning model nowadays, Transformer has been used in EEG signal classification tasks owing to its excellent performance. Therefore, in this study, we proposed a deep learning model for SSVEP classification based on Transformer architecture in inter-subject scenario, termed as SSVEPformer, which was the first application of Transformer on the SSVEP classification. Inspired by previous studies, we adopted the complex spectrum features of SSVEP data as the model input, which could enable the model to simultaneously explore the spectral and spatial information for classification. Furthermore, to fully utilize the harmonic information, an extended SSVEPformer based on the filter bank technology (FB-SSVEPformer) was proposed to improve the classification performance. Experiments were conducted using two open datasets (Dataset 1: 10 subjects, 12 targets; Dataset 2: 35 subjects, 40 targets). The experimental results show that the proposed models could achieve better results in terms of classification accuracy and information transfer rate than other baseline methods. The proposed models validate the feasibility of deep learning models based on Transformer architecture for SSVEP data classification, and could serve as potential models to alleviate the calibration procedure in the practical application of SSVEP-based BCI systems. [ABSTRACT FROM AUTHOR]

The detection of attentive mental state plays an essential role in the neurofeedback process and the treatment of Attention Deficit and Hyperactivity Disorder (ADHD). However, the performance of the detection methods is still not satisfactory. One of the challenges is to find a proper representation for the electroencephalogram (EEG) data, which could preserve the temporal information and maintain the spatial topological characteristics. Inspired by the deep learning (DL) methods in the research of brain–computer interface (BCI) field, a 3D representation of EEG signal was introduced into attention detection task, and a 3D convolutional neural network model with cascade and parallel convolution operations was proposed. The model utilized three cascade blocks, each consisting of two parallel 3D convolution branches, to simultaneously extract the multi-scale features. Evaluated on a public dataset containing twenty-six subjects, the proposed model achieved better performance compared with the baseline methods under the intra-subject, inter-subject and subject-adaptive classification scenarios. This study demonstrated the promising potential of the 3D CNN model for detecting attentive mental state. [ABSTRACT FROM AUTHOR]

Effective frequency recognition algorithms are critical in steady-state visual evoked potential (SSVEP) based brain–computer interfaces (BCIs). In this study, we present a hierarchical feature fusion framework which can be used to design high-performance frequency recognition methods. The proposed framework includes two primary techniques for fusing features: spatial dimension fusion (SD) and frequency dimension fusion (FD). Both SD and FD fusions are obtained using a weighted strategy with a nonlinear function. To assess our novel methods, we used the correlated component analysis (CORRCA) method to investigate the efficiency and effectiveness of the proposed framework. Experimental results were obtained from a benchmark dataset of thirty-five subjects and indicate that the extended CORRCA method used within the framework significantly outperforms the original CORCCA method. Accordingly, the proposed framework holds promise to enhance the performance of frequency recognition methods in SSVEP-based BCIs. [ABSTRACT FROM AUTHOR]

Touch is one of the important ways to perceive the surrounding environment. The surface texture is an essential character of an object. Here, we propose a dual-layer shielding triboelectric tactile sensor with patterned flower-shaped holes. The object surface to be measured is able to rub against the tribo-layer through the flower-shaped holes to produce a signal output. The patterned design of the tactile sensor is evolved from the contact state of the fingertips during human finger touch. In addition, two shielding layers were designed. The inner-shielding layer is used to shield the influence of the human body potential on the output signal, in order to achieve more accurate identification. The outer-shielding layer enables the flower-shaped holes to play their structural role. The proposed tactile sensor has the advantage of a single channel, resulting in a smaller amount of data collection and processing than the multi-channel scheme. With a designed convolutional neural network model, the recognition accuracy reaches 96.03% when recognizing 7 objects with designed different surface textures and 92.58% when recognizing 5 kinds of fruits and vegetables. The tactile sensing module can be easily integrated into intelligent haptic prosthesis and play an important role in future human-machine interfaces. ● A wearable TENG tactile sensor for texture recognition has a specific pattern evolved from the contact situation. ● Dual-shielding layer to reduce the influence of human potential can greatly improve the SNR. ● A deep learning model based on CNN to recognize 7-classes of textures obtains a high recognition accuracy of 96.03%. [ABSTRACT FROM AUTHOR]

Abstract Our mysterious brain is believed to operate near a non-equilibrium point and generate critical self-organized avalanches in neuronal activity. A central topic in neuroscience is to elucidate the underlying circuitry mechanisms of neuronal avalanches in the brain. Recent experimental evidence has revealed significant heterogeneity in both synaptic input and output connectivity, but whether the structural heterogeneity participates in the regulation of neuronal avalanches remains poorly understood. By computational modeling, we predict that different types of structural heterogeneity contribute distinct effects on avalanche neurodynamics. In particular, neuronal avalanches can be triggered at an intermediate level of input heterogeneity, but heterogeneous output connectivity cannot evoke avalanche dynamics. In the criticality region, the co-emergence of multi-scale cortical activities is observed, and both the avalanche dynamics and neuronal oscillations are modulated by the input heterogeneity. Remarkably, we show similar results can be reproduced in networks with various types of in- and out-degree distributions. Overall, these findings not only provide details on the underlying circuitry mechanisms of nonrandom synaptic connectivity in the regulation of neuronal avalanches, but also inspire testable hypotheses for future experimental studies. [ABSTRACT FROM AUTHOR]

Multichannel frequency detection methods for SSVEP-based BCI have received increasing interest in recent years. Among the alternative methods, multivariate synchronization index (MSI) is a potential one to achieve robust performance for SSVEP-based BCI. This study further presents an extension to MSI (termed as EMSI) for frequency recognition, which leverage the method of time delay embedding. The new method incorporates the first-order time delayed version of the EEG data during calculation the synchronization index. The effectiveness of the proposed method is validated by comparing it with the standard MSI on the actual SSVEP datasets collected from eleven subjects. The experimental results indicate that the EMSI significantly outperforms the MSI at four time windows. It suggests that the EMSI is a promising methodology for frequency recognition and could further improve the performance of SSVEP-based BCI system. [ABSTRACT FROM AUTHOR]

Photoplethysmography (PPG)-based heart rate (HR) monitoring is a promising feature in modern wearable devices. However, it is difficult to accurately track HR during physical exercise since PPG signals are vulnerable to motion artifacts (MA). In this paper, an algorithm is presented to combine ensemble empirical mode decomposition (EEMD) with spectrum subtraction (SS) to track HR changes during subjects’ physical activities. In this algorithm, EEMD decomposes a PPG signal and an acceleration signal into intrinsic mode functions (IMFs), respectively. Then noise related IMFs are removed. Next the correlation coefficient is computed between the spectrum of the acceleration signal and that of the PPG signal in the band of [0.4 Hz–5 Hz]. If the coefficient is above 0.5, SS is used to remove the spectrum of the acceleration signal from the PPG's spectrum. Finally, a spectral peak selection method is used to find the peak corresponding to HR. Experimental results on datasets recorded from 12 subjects during fast running showed the superior performance of the proposed algorithm compared with a benchmark method termed TROIKA. The average absolute error of HR estimation was 1.83 beats per minute (BPM), and the Pearson correlation was 0.989 between the ground-truth and the estimated HR. [ABSTRACT FROM AUTHOR]

Abstract: Many motor imagery based BCI systems will utilize the common spatial pattern (CSP) feature for task classification. However, the frequency band and time interval involved for CSP feature extraction will have large effect on the BCI performance. In this paper, with aim to find the optimal frequency band and time interval for effective CSP feature extraction, an approach based on particle swarm optimization is proposed. In this approach, the frequency band and time interval are coded with the particles and the optimal settings of them can be simultaneously detected by the evolution of particles for individual subject. The results from eighteen BCI participants confirmed that the individual frequency band and time interval provided by particles could actually improve the discriminative ability of CSP features, and it has potential application in online BCI system. [Copyright &y& Elsevier]

Highlights: [•] A novel frequency recognition method (MSI) was proposed. [•] MSI showed better performance than traditional method when using short length data. [•] MSI showed better performance when using small number of channels. [•] MSI was successfully used in online BCI experiment. [ABSTRACT FROM AUTHOR]

• A Bayesian analysis (BA) based on Granger causality estimation was developed. • The proposed model estimated parameters more robustly than the traditional Granger causality analysis for time series with different kinds of noise. • The network patterns estimated by our proposed method were more similar to the predefined networks than traditional Granger causality analysis under noise conditions. • Both the simulation study and real fMRI data applications revealed that our proposed method was less influenced by the data length compared with traditional Granger causality analysis. Cognitive processes involve information integration among multiple encephalic regions, which can be measured by causal networks. However, the estimation of causal networks by means of some traditional methods with the least square will lead to distorted networks because of the unexpected outlier noise and the small number of signal samples in real applications. In this work, we adopted Bayesian inference to estimate parameters in a multivariate autoregression model (MVAR), to restrain the influence of outliers. Through the simulation study, we observed that our proposed method can efficiently suppress outlier influence and shows stable performance when sample sizes become small. Application to real motor imagery functional magnetic resonance imaging (fMRI) also revealed that the proposed approach can capture the inherent hemispheric lateralization of motor imagery even with a small number of fMRI samples. We compared our proposed Bayesian-based Granger analysis with traditional Granger causality analysis. The analyses conducted in the current work demonstrate the robustness of Bayesian-based Granger analysis to outlier conditions or physiological signals with small sample sizes. [ABSTRACT FROM AUTHOR]

As one of the important psychological stress reactions, Micro-expressions (MEs) are spontaneous and subtle facial movements, which usually occur in a high-stake situation and can reveal genuine human feelings and cognition. ME, Recognition (MER) has essential applications in many fields such as lie detection, criminal investigation, and psychological healing. However, due to the challenges of learning discriminative ME features via fleeting facial subtle reactions as well as the shortage of available MEs data, this research topic is still far from well-studied. To this end, in this paper, we propose a deep prototypical learning framework, namely ME-PLAN, with a local attention mechanism for the MER problem. Specifically, ME-PLAN consists of two components, i.e., a 3D residual prototypical network and a local-wise attention module, where the former aims to learn the precise ME feature prototypes through expression-related knowledge transfer and episodic training, and the latter could facilitate the attention to the local facial movements. Furthermore, to alleviate the dilemma that most MER methods need to depend on manually annotated apex frames, we propose an apex frame spotting method with Unimodal Pattern Constrained (UPC) and further extract ME key-frames sequences based on the detected apex frames to train our proposed ME-PLAN in an end-to-end manner. Finally, through extensive experiments and interpretable analysis regarding the apex frame spotting and MER on composite-database, we demonstrate the superiority and effectiveness of the proposed methods. • We explore deep prototypical learning for the MER problem, and try to address three challenges in this topic. • We propose a novel deep prototypical learning framework, namely ME-PLAN, using a 3D residual prototypical network with episodic training and a local-wise attention module to learn precise ME representation. • A novel apex frame spotting method based on Unimodal Pattern Constraint is proposed to effectively eliminate noise interference and accurately locate apex frames. • Extensive experimental results with a comparison of state-of-the-art methods have demonstrated the effectiveness of our ME-PLAN on apex frame spotting and MER tasks. [ABSTRACT FROM AUTHOR]

Temporal lobe epilepsy (TLE) is considered a brain activity disorder that is likely related to allomnesia and conscious disturbance. In rodents, TLE epileptiform discharges can be triggered by systemic administration of a dose of pilocarpine. However, how these pilocarpine-induced epileptiform discharges generate and propagate through the whole brain has not been well studied. In this study, we sought to assess alterations of activity and information flow in the default mode network (DMN) during TLE epileptiform discharges in pilocarpine-induced TLE rats. We identified that in resting state, the rat DMN could be divided into three subnetworks that constituted a frequency-specific information flow loop. This frequency-specific loop converted into frequency-independent flow patterns during the generation and propagation of epileptiform discharges. Moreover, the activity of the theta (4–8 Hz) and alpha (8–13 Hz) bands in each DMN subnetwork exhibited completely different alterations during epileptiform discharges. Overall, our results demonstrated frequency-dependent alterations of both activity and information flow patterns within the DMN during epileptiform discharge. These results suggest that the independence and cooperation of different frequency bands may serve as an underlying mechanism for the generation and propagation of epileptiform discharges in the brain during SE in pilocarpine-treated rats. [ABSTRACT FROM AUTHOR]

L1-regularized multiway canonical correlation analysis (L1-MCCA) has been introduced to reference signal optimization in steady-state visual evoked potential (SSVEP)-based brain–computer interface (BCI). The effectiveness of L1-regularization on significant trial selection highly depends on the regularization parameter setting, which can be typically determined by cross-validation (CV). However, CV will substantially reduce the practicability of BCI system due to additional data requirement for the parameter validation and relatively high computational cost. To solve the problem, this study proposes a Bayesian version of L1-MCCA (called SBMCCA) by exploiting sparse Bayesian learning. The SBMCCA method avoids CV and can efficiently estimate the model parameters under the Bayesian evidence framework. Experimental results show that the SBMCCA method achieved comparable recognition accuracy but much higher computational efficiency in contrast to the L1-MCCA method. [ABSTRACT FROM AUTHOR]