Search

The application of intracranial electroencephalogram (iEEG) to predict seizures remains challenging. Although channel selection has been utilized in seizure prediction and detection studies, most of them focus on the combination with conventional machine learning methods. Thus, channel selection combined with deep learning methods can be further analyzed in the field of seizure prediction. Given this, in this work, a novel iEEG-based deep learning method of One-Dimensional Convolutional Neural Networks (1D-CNN) combined with channel increment strategy was proposed for the effective seizure prediction. First, we used 4-sec sliding windows without overlap to segment iEEG signals. Then, 4-sec iEEG segments with an increasing number of channels (channel increment strategy, from one channel to all channels) were sequentially fed into the constructed 1D-CNN model. Next, the patient-specific model was trained for classification. Finally, according to the classification results in different channel cases, the channel case with the best classification rate was selected for each patient. Our method was tested on the Freiburg iEEG database, and the system performances were evaluated at two levels (segment- and event-based levels). Two model training strategies (Strategy-1 and Strategy-2) based on the K-fold cross validation (K-CV) were discussed in our work. (1) For the Strategy-1, a basic K-CV, a sensitivity of 90.18%, specificity of 94.81%, and accuracy of 94.42% were achieved at the segment-based level. At the event-based level, an event-based sensitivity of 100%, and false prediction rate (FPR) of 0.12/h were attained. (2) For the Strategy-2, the difference from the Strategy-1 is that a trained model selection step is added during model training. We obtained a sensitivity, specificity, and accuracy of 86.23%, 96.00% and 95.13% respectively at the segment-based level. At the event-based level, we achieved an event-based sensitivity of 98.65% with 0.08/h FPR. Our method also showed a better performance in seizure prediction compared to many previous studies and the random predictor using the same database. This may have reference value for the future clinical application of seizure prediction. peerReviewed

Dysconnectivity of large-scale brain networks has been linked to major depression disorder (MDD) during resting state. Recent researches show that the temporal evolution of brain networks regulated by oscillations reveals novel mechanisms and neural characteristics of MDD. Our study applied a novel coupled tensor decomposition model to investigate the dysconnectivity networks characterized by spatio-temporal-spectral modes of covariation in MDD using resting electroencephalography. The phase lag index is used to calculate the functional connectivity within each time window at each frequency bin. Then, two adjacency tensors with the dimension of time frequency connectivity subject are constructed for the healthy group and the major depression group. We assume that the two groups share the same features for group similarity and retain individual characteristics for group differences. Considering that the constructed tensors are nonnegative and the components in spectral and adjacency modes are partially consistent among the two groups, we formulate a double-coupled nonnegative tensor decomposition model. To reduce computational complexity, we introduce the lowrank approximation. Then, the fast hierarchical alternative least squares algorithm is applied for model optimization. After clustering analysis, we summarize four oscillatory networks characterizing the healthy group and four oscillatory networks characterizing the major depression group, respectively. The proposed model may reveal novel mechanisms of pathoconnectomics in MDD during rest, and it can be easily extended to other psychiatric disorders. peerReviewed

Averaging amplitudes over consecutive time samples (i.e., time window) is widely used to calculate the peak amplitude of event-related potentials (ERPs). Cluster analysis of the spatio-temporal ERP data is a promising tool to determine the time window of an ERP of interest. However, determining an appropriate number of clusters to optimally represent ERPs is still challenging. Here, we develop a new method to estimate the optimal number of clusters utilizing consensus clustering. Various polarity dependent clustering methods, namely, k-means, hierarchical clustering, fuzzy c-means, self-organizing map, spectral clustering, and Gaussian mixture model, are used to configure consensus clustering after assessing them individually. When a range of clusters is applied many times, the optimal number of clusters should correspond to the expectation, which is the average of the obtained mean inner-similarities of estimated time windows across all conditions and groups converge in the satisfactory thresholds. In order to assess our method, the proposed method has been applied to simulated data and prospective memory experiment ERP data aimed to qualify N2 and P3, and N300 and prospective positivity components, respectively. The results of determining the optimal number of clusters meet at six cluster maps for both ERP data. In addition, our results revealed that the proposed method could be reliably applied to ERP data to determine the appropriate time window for the ERP of interest when the measurement interval is not accurately defined.

Combining magnetic resonance imaging (MRI) data from multi-site studies is a popular approach for constructing larger datasets to greatly enhance the reliability and reproducibility of neuroscience research. However, the scanner/site variability is a significant confound that complicates the interpretation of the results, so effective and complete removal of the scanner/site variability is necessary to realize the full advantages of pooling multi-site datasets. Independent component analysis (ICA) and general linear model (GLM) based denoising methods are the two primary methods used to denoise scanner/site-related effects. Unfortunately, there are challenges with both ICA-based and GLM-based denoising methods to remove site effects completely when the signals of interest and scanner/site-related noises are correlated, which may occur in neuroscience studies. In this study, we propose an effective and powerful denoising strategy that implements dual-projection (DP) theory based on ICA to remove the scanner/site-related effects more completely. This method can separate the signal effects correlated with noise variables from the identified noise effects for removal without losing signals of interest. Both simulations and vivo structural MRI datasets, including a dataset from Autism Brain Imaging Data Exchange II and a traveling subject dataset from the Strategic Research Program for Brain Sciences, were used to test the proposed GLM- and ICA-based denoising methods and our DP-based ICA denoising method. Results show that DP-based ICA

For real-world automatic sleep-stage classification tasks, various existing deep learning-based models are biased toward the majority with a high proportion. Because of the unique sleep structure, most of the current polysomnography (PSG) datasets suffer an inherent class imbalance problem (CIP), in which the number of each sleep stage is severely unequal. In this study, we first define the class imbalance factor (CIF) to describe the level of CIP quantitatively. Afterward, we propose two balancing methods to alleviate this problem from the dataset quantity and the relationship between the class distribution and the applied model, respectively. The first one is to employ the data augmentation (DA) with the generative adversarial network (GAN) model and different intensities of Gaussian white noise (GWN) to balance samples, thereinto, GWN addition is specifically tailored to deep learning-based models, which can work on raw electroencephalogram (EEG) data while preserving their properties. In addition, we try to balance the relationship between the imbalanced class and biased network model to achieve a balanced state with the help of class distribution and neuroscience principles. We further propose an effective deep convolutional neural network (CNN) model utilizing bidirectional long short-term memory (Bi-LSTM) with single-channel EEG as the baseline. It is used for evaluating the efficiency of two balancing approaches on three imbalanced PSG datasets (CCSHS, Sleep-EDF, and Sleep-EDF-V1). The qualitative and quantitative evaluation of experimental results demonstrates that the proposed methods could not only show the superiority of class balancing through the confusion matrix and classwise metrics, but also get better N1 stage and whole stages classification accuracies compared to other state-of-the-art approaches. peerReviewed

Predicting adverse outcomes is essential for pregnant women with systemic lupus erythematosus (SLE) to minimize risks. Applying statistical analysis may be limited for the small sample size of childbearing patients, while the informative medical records could be provided. This study aimed to develop predictive models applying machine learning (ML) techniques to explore more information. We performed a retrospective analysis of 51 pregnant women exhibiting SLE, including 288 variables. After correlation analysis and feature selection, six ML models were applied to the filtered dataset. The efficiency of these overall models was evaluated by the Receiver Operating Characteristic Curve. Meanwhile, real-time models with different timespans based on gestation were also explored. Eighteen variables demonstrated statistical differences between the two groups; more than forty variables were screened out by ML variable selection strategies as contributing predictors, while the overlap of variables were the influential indicators testified by the two selection strategies. The Random Forest (RF) algorithm demonstrated the best discrimination ability under the current dataset for overall predictive models regardless of the data missing rate, while Multi-Layer Perceptron models ranked second. Meanwhile, RF achieved best performance when assessing the real-time predictive accuracy of models. ML models could compensate the limitation of statistical methods when the small sample size problem happens along with numerous variables acquired, while RF classifier performed relatively best when applied to such structured medical records.

Objective Cluster analysis of spatio-temporal event-related potential (ERP) data is a promising tool for exploring the measurement time window of ERPs. However, even after preprocessing, the remaining noise can result in uncertain cluster maps followed by unreliable time windows while clustering via conventional clustering methods. Methods We designed an ensemble deep clustering pipeline to determine a reliable time window for the ERP of interest from temporal concatenated grand average ERP data. The proposed pipeline includes semi-supervised deep clustering methods initialized by consensus clustering and unsupervised deep clustering methods with end-to-end architectures. Ensemble clustering from those deep clusterings was used by the designed adaptive time window determination to estimate the time window. Results After applying simulated and real ERP data, our method successfully obtained the time window for identifying the P3 components (as the interest of both ERP studies) while additional noise (e.g., adding 20 dB to −5 dB white Gaussian noise) was added to the prepared data. Conclusion Compared to the state-of-the-art clustering methods, a superior clustering performance was yielded from both ERP data. Furthermore, more stable and precise time windows were elicited as the noise increased. Significance Our study provides a complementary understanding of identifying the cognitive process using deep clustering analysis to the existing studies. Our finding suggests that deep clustering can be used to identify the ERP of interest when the data is imperfect after preprocessing. peerReviewed

Background Neural stem cells (NSCs) are believed to have the most therapeutic potential for neurological disorders because they can differentiate into various neurons and glial cells. This research evaluated the safety and efficacy of intranasal administration of NSCs in children with cerebral palsy (CP). The functional brain network (FBN) analysis based on electroencephalogram (EEG) and voxel-based morphometry (VBM) analysis based on T1-weighted images were performed to evaluate functional and structural changes in the brain. Methods A total of 25 CP patients aged 3–12 years were randomly assigned to the treatment group (n = 15), which received an intranasal infusion of NSCs loaded with nasal patches and rehabilitation therapy, or the control group (n = 10) received rehabilitation therapy only. The primary endpoints were the safety (assessed by the incidence of adverse events (AEs), laboratory and imaging examinations) and the changes in the Gross Motor Function Measure-88 (GMFM-88), the Activities of Daily Living (ADL) scale, the Sleep Disturbance Scale for Children (SDSC), and some adapted scales. The secondary endpoints were the FBN and VBM analysis. Results There were only four AEs happened during the 24-month follow-up period. There was no significant difference in the laboratory examinations before and after treatment, and the magnetic resonance imaging showed no abnormal nasal and intracranial masses. Compared to the control group, patients in the treatment group showed apparent improvements in GMFM-88 and ADL 24 months after treatment. Compared with the baseline, the scale scores of the Fine Motor Function, Sociability, Life Adaptability, Expressive Ability, GMFM-88, and ADL increased significantly in the treatment group 24 months after treatment, while the SDSC score decreased considerably. Compared with baseline, the FBN analysis showed a substantial decrease in brain network energy, and the VBM analysis showed a significant increase in gray matter volume in the treatment group after NSCs treatment. Conclusions Our results showed that intranasal administration of NSCs was well-tolerated and potentially beneficial in children with CP. Trial registration: The study was registered in ClinicalTrials.gov (NCT03005249, registered 29 December 2016, https://www.clinicaltrials.gov/ct2/show/NCT03005249) and the Medical Research Registration Information System (CMR-20161129-1003).

Publisher Copyright: Author The human brain can be described as a complex network of functional connections between distinct regions, referred to as the brain functional network. Recent studies show that the functional network is a dynamic process and its community structure evolves with time during continuous task performance. Consequently, it is important for the understanding of the human brain to develop dynamic community detection techniques for such time-varying functional networks. Here, we propose a temporal clustering framework based on a set of network generative models and surprisingly it can be linked to Block Component Analysis to detect and track the latent community structure in dynamic functional networks. Specifically, the temporal dynamic networks are represented within a unified three-way tensor framework for simultaneously capturing multiple types of relationships between a set of entities. The multi-linear rank-(Lr,Lr,1) block term decomposition (BTD) is adopted to fit the network generative model to directly recover underlying community structures with the specific evolution of time from the temporal networks. We apply the proposed method to the study of the reorganization of the dynamic brain networks from electroencephalography (EEG) data recorded during free music listening. We derive several network structures (Lr communities in each component) with specific temporal patterns (described by BTD components) significantly modulated by musical features, involving subnetworks of frontoparietal, default mode, and sensory-motor networks. The results show that the brain functional network structures are dynamically reorganized and the derived community structures are temporally modulated by the music features. The proposed generative modeling approach can be an effective tool for describing community structures in brain networks that go beyond static methods and detecting the dynamic reconfiguration of modular connectivity elicited by continuously naturalistic tasks.

Nonnegative tensor decomposition is a versatile tool for multiway data analysis, by which the extracted components are nonnegative and usually sparse. Nevertheless, the sparsity is only a side effect and cannot be explicitly controlled without additional regularization. In this paper, we investigated the nonnegative CANDECOMP/PARAFAC (NCP) decomposition with the sparse regularization item using $$l_1$$ l 1 -norm (sparse NCP). When high sparsity is imposed, the factor matrices will contain more zero components and will not be of full column rank. Thus, the sparse NCP is prone to rank deficiency, and the algorithms of sparse NCP may not converge. In this paper, we proposed a novel model of sparse NCP with the proximal algorithm. The subproblems in the new model are strongly convex in the block coordinate descent (BCD) framework. Therefore, the new sparse NCP provides a full column rank condition and guarantees to converge to a stationary point. In addition, we proposed an inexact BCD scheme for sparse NCP, where each subproblem is updated multiple times to speed up the computation. In order to prove the effectiveness and efficiency of the sparse NCP with the proximal algorithm, we employed two optimization algorithms to solve the model, including inexact alternating nonnegative quadratic programming and inexact hierarchical alternating least squares. We evaluated the proposed sparse NCP methods by experiments on synthetic, real-world, small-scale, and large-scale tensor data. The experimental results demonstrate that our proposed algorithms can efficiently impose sparsity on factor matrices, extract meaningful sparse components, and outperform state-of-the-art methods.

Epileptic seizure detection using scalp electroencephalogram (sEEG) and intracranial electroencephalogram (iEEG) has attracted widespread attention in recent two decades. The accurate and rapid detection of seizures not only reflects the efficiency of the algorithm, but also greatly reduces the burden of manual detection during long-term electroencephalogram (EEG) recording. In this work, a stacked one-dimensional convolutional neural network (1D-CNN) model combined with a random selection and data augmentation (RS-DA) strategy is proposed for seizure onset detection. Firstly, we segmented the long-term EEG signals using 2-s sliding windows. Then, the 2-s interictal and ictal segments were classified by the stacked 1D-CNN model. During model training, a RS-DA strategy was applied to solve the problem of sample imbalance, and the patient-specific model was trained with event-based K-fold (K is the number of seizures per patient) cross validation for detecting all seizures of each patient. Finally, we evaluated the performances of the proposed approach in the two levels: the segment-based level and the event-based level. The proposed method was tested on two long-term EEG datasets: the CHB-MIT sEEG dataset and the SWEC-ETHZ iEEG dataset. For the CHB-MIT sEEG dataset, we achieved 88.14% sensitivity, 99.62% specificity and 99.54% accuracy in the segment-based level. From the perspective of the event-based level, 99.31% sensitivity, 0.2/h false detection rate (FDR) and mean 8.1-s latency were achieved. For the SWEC-ETHZ iEEG dataset, in the segment-based level, 90.09% sensitivity, 99.81% specificity and 99.73% accuracy were obtained. In the event-based level, 97.52% sensitivity, 0.07/h FDR and mean 13.2-s latency were attained. From these results, we can see that our method can effectively use both sEEG and iEEG data to detect epileptic seizures, and this may provide a reference for the clinical application of seizure onset detection.

Mental fatigue is a common phenomenon with implicit and multidimensional properties. It brings dynamic changes in functional brain networks. However, the challenging problem of false positives appears when the connectivity is estimated by electroencephalography (EEG). In this article, we propose a novel framework based on spatial clustering to explore the sources of mental fatigue and functional activity changes caused by them. To suppress false positive observations, spatial clustering is implemented in brain networks. The nodes extracted by spatial clustering are registered back to the functional magnetic resonance imaging (fMRI) source space to determine the sources of mental fatigue. The wavelet entropy of EEG in a sliding window is calculated to find the temporal features of mental fatigue. Our experimental results show that the extracted nodes correspond to the fMRI sources across different subjects and different tasks. The entropy values on the extracted nodes demonstrate clearer staged decreasing changes (deactivation). Additionally, the synchronization among the extracted nodes is stronger than that among all the nodes in the deactivation stage. The initial time of the strong synchronized deactivation is consistent with the subjective fatigue time reported by the subjects themselves. It means the synchronization and deactivation correspond to the subjective feelings of fatigue. Therefore, this functional activity pattern may be caused by sources of mental fatigue. The proposed framework is useful for a wide range of prolonged functional imaging and fatigue detection studies.

Repeating graphics are common research objects in modern design education. However, we do not exactly know the attentional processes underlying graphic artifacts consisting of repeating rhythms. In this experiment, the event-related potential, a neuroscientific measure, was used to study the neural correlates of repeating graphics within graded orderliness. We simulated the competitive identification process of people recognizing artifacts with graded repeating rhythms from a scattered natural environment with the oddball paradigm. In the earlier attentional processing related to the P2 component around the Fz electrode within the 150−250 ms range, a middle-grade repeating rhythm (Target 1) did not show a difference from a high-grade repeating rhythm (Target 2). However, in the later cognitive processes related to the P3b component around the Pz electrode within the 300−450 ms range, Target 1 had longer peak latency than Target 2, based on similar waveforms. Thus, we may suppose that the arrangement of the repeating graphics did not influence the earlier attentional processing but affected the later cognitive part, such as the categorization task in the oddball paradigm. Furthermore, as evidenced by the standard deviation wave across the trials, we suggest that the growing standard deviation value might represent the gradual loss of attentional focus to the task after the stimulus onset and that the zero-growth level may represent similar brain activity between trials.

Nonnegative tensor decomposition has become increasingly important for multiway data analysis in recent years. The alternating proximal gradient (APG) is a popular optimization method for nonnegative tensor decomposition in the block coordinate descent framework. In this study, we propose an inexact version of the APG algorithm for nonnegative CANDECOMP/PARAFAC decomposition, wherein each factor matrix is updated by only finite inner iterations. We also propose a parameter warm-start method that can avoid the frequent parameter resetting of conventional APG methods and improve convergence performance. By experimental tests, we find that when the number of inner iterations is limited to around 10 to 20, the convergence speed is accelerated significantly without losing its low relative error. We evaluate our method on both synthetic and real-world tensors. The results demonstrate that the proposed inexact APG algorithm exhibits outstanding performance on both convergence speed and computational precision compared with existing popular algorithms.

Accurate sleep stage classification is vital to assess sleep quality and diagnose sleep disorders. Numerous deep learning based models have been designed for accomplishing this labor automatically. However, the class imbalance problem existing in polysomnography (PSG) datasets has been barely investigated in previous studies, which is one of the most challenging obstacles for the real-world sleep staging application. To address this issue, this paper proposes novel methods with signal-driven and image-driven ways of noise addition to balance the imbalanced relationship in the training dataset samples. We evaluate the effectiveness of the proposed methods which are integrated into a convolutional neural network (CNN) based model. Experimental results evaluated on Sleep-EDF-V1, Sleep-EDF and CCSHS databases demonstrate that the proposed balancing approaches with specific tensity Gaussian white noise could enhance the overall or stage N1 recognition to some degree, especially the combination of two types of Data augmentation (DA) strategies shows the superiority of overall accuracy improvement. peerReviewed

Emotion recognition based on electroencephalography (EEG) has attracted high interest in fields such as health care, user experience evaluation, and human–computer interaction (HCI), as it plays an important role in human daily life. Although various approaches have been proposed to detect emotion states in previous studies, there is still a need to further study the dynamic changes of EEG in different emotions to detect emotion states accurately. Entropy-based features have been proved to be effective in mining the complexity information in EEG in many areas. However, different entropy features vary in revealing the implicit information of EEG. To improve system reliability, in this paper, we propose a framework for EEG-based cross-subject emotion recognition using fused entropy features and a Bidirectional Long Short-term Memory (BiLSTM) network. Features including approximate entropy (AE), fuzzy entropy (FE), Rényi entropy (RE), differential entropy (DE), and multi-scale entropy (MSE) are first calculated to study dynamic emotional information. Then, we train a BiLSTM classifier with the inputs of entropy features to identify different emotions. Our results show that MSE of EEG is more efficient than other single-entropy features in recognizing emotions. The performance of BiLSTM is further improved with an accuracy of 70.05% using fused entropy features compared with that of single-type feature.

Background: Neural stem cells (NSCs) are believed to have the most therapeutic potential for neurological disorders because they can differentiate into various neurons and glial cells. This research evaluated the safety and efficacy of intranasal administration of NSCs in children with cerebral palsy (CP). The functional brain network (FBN) analysis based on electroencephalogram (EEG) and voxel-based morphometry (VBM) analysis based on T1-weighted images were performed to evaluate functional and structural changes in the brain.Methods: 25 CP patients aged 3-12 years old were randomly assigned to the treatment group (n=15), which received intranasal infusion of NSCs loaded by nasal patches and rehabilitation therapy, or the control group (n=10), which received rehabilitation therapy only. The primary endpoints were the safety (assessed by the incidence of adverse events (AEs), laboratory and imaging examinations) and the changes in the Gross Motor Function Measure-88 (GMFM-88), the Activities of Daily Living (ADL) scale, the Sleep Disturbance Scale for Children (SDSC), and some adapted scales. The secondary endpoints were the FBN and VBM analysis.Results: There were only 4 AEs happened during the 24 months follow-up period. No significant difference in the laboratory examinations before and after treatment, and the magnetic resonance imaging showed no abnormal nasal and intracranial masses. Compared to the control group, patients in the treatment group showed apparent improvements in GMFM-88 and ADL 24 months after treatment. Compared with the baseline, the scale scores of the Fine Motor Function, Sociability, Life Adaptability, Expressive Ability, GMFM-88, and ADL increased significantly in the treatment group 24 months after treatment, while the SDSC score decreased considerably. Compared with baseline, the FBN analysis showed a substantial decrease in brain network energy, and the VBM analysis showed a significant increase in gray matter volume in the treatment group after NSCs treatment.Conclusions: Our results showed that intranasal administration of NSCs was well tolerated and potentially beneficial in children with CP.Trial registration: The study was registered in ClinicalTrials.gov (NCT03005249, registered 29 December 2016, https://www.clinicaltrials.gov/ct2/show/NCT03005249) and the Medical Research Registration Information System (CMR-20161129-1003).

Novel metaphors in literary texts (hereinafter referred to as literary metaphors) seem to be more creative and open-ended in meaning than metaphors in non-literary texts (non-literary metaphors). However, some disagreement still exists on how literary metaphors differ from non-literary metaphors. Therefore, this study explored the neural mechanisms of literary metaphors extracted from modern Chinese poetry by using the methods of Event-Related Potentials (ERPs) and Event-Related Spectral Perturbations (ERSPs), as compared with non-literary conventional metaphors and literal expressions outside literary texts. Forty-eight subjects were recruited to make the semantic relatedness judgment after reading the prime-target pairs in three linguistic conditions. According to the ERPs results, the earliest differences were presented during the time window of P200 component (170-260 ms) in the frontal and central areas, with the amplitude of P200 for literary metaphors more positive than the other two conditions, reflecting the early allocation of attention and the early conscious experience of the experimental stimuli. Meanwhile, significant differences were presented during the time window of N400 effect (430-530 ms), with the waveform of literary metaphors more negative than others in the frontal and central topography of scalp distributions, suggesting more efforts in retrieving conceptual knowledge for literary metaphors. The ERSPs analysis revealed that the frequency bands of delta and theta were both involved in the cognitive process of literary metaphor comprehension, with delta band distributed in the frontal and central scalp and theta band in parietal and occipital electrodes. Increases in the two power bands during different time windows provided extra evidences that the processing of literary metaphors required more attention and effort than non-literary metaphors and literal expressions in the semantic related tasks, suggesting that the cognitive process of literary metaphors was distinguished by different EEG spectral patterns.

Background: Temporal principal component analysis (tPCA) has been widely used to extract event-related potentials (ERPs) at group level of multiple subjects ERP data and it assumes that the underlying factor loading is fixed across participants. However, such assumption may fail to work if latency and phase for one ERP vary considerably across participants. Furthermore, effect of number of trials on tPCA decomposition has not been systematically examined as well, especially for within-subject PCA. New method: We reanalyzed a real ERP data of an emotional experiment using tPCA to extract N2 and P2 from single-trial EEG of an individual. We also explored influence of the number of trials (consecutively increased from 10 to 42 trials) on PCA decomposition by comparing temporal correlation, the statistical result, Cronbach’s alpha, spatial correlation of both N2 and P2 for the proposed method with the conventional time-domain analysis, trial-averaged group PCA, and single-trial-based group PCA. Results: The results of the proposed method can enhance spatial and temporal consistency. We could obtain stable N2 with few trials (about 20) for the proposed method, but, for P2, approximately 30 trials were needed for all methods. Comparison with Existing Method(s): About 30 trials for N2 were required and the reconstructed P2 and N2 were poor correlated across participants for the other three methods. Conclusion: The proposed approach may efficiently capture variability of one ERP from an individual that cannot be extracted by group PCA analysis. peerReviewed

Temperature-related reflex epilepsy most often takes the form of hot water epilepsy, but very rarely, reflex epilepsy is related to cold temperature. We report a 70-year-old male who had seizures triggered by cold sensations in the body. Four antiepileptic drugs were taken during the drug treatment, and oxcarbazepine was the most effective at stopping the seizures. We implemented clinical seizure induction and obtained EEG data from an interictal period and two complete ictal periods. Source estimation was performed to identify and map the primary sources involved in the seizures on the cortical level. We found that β rhythm appeared on the prefrontal lobes during the whole ictal period. The low-frequency slow δ and θ rhythms, especially the δ rhythm, appeared in the occipital lobe in the early ictal stage and propagated to the right temporal lobe in the mid-late ictal stage. The prefrontal lobe and right temporal lobe were mainly involved in the generation and propagation of the epileptic activities. This study provides a valuable reference for clinical drug therapy and provides insights into the characteristics of the brain activities involved in cold-induced reflex epilepsy. [Published with video sequences].

Human behavior in health and disease is associated with dynamic functional networks, which are characterized as hierarchical ensembles with segregated spatial architecture and corresponding temporal architecture. How best to extract dynamical functional network structure and coefficient time series from neural data remains an open problem. We developed a new method, which called stability-driven non-negative matrix factorization (staNMF), that combines non-negative matrix factorization with a novel data-driven model selection criterion to decompose dynamic functional networks into sets of superposed subnetworks. To evaluate our method, we first use simulated noisy connectivity networks to demonstrate that staNMF is able to find the correct number of subnetworks, obtain the connectivity matrix of each subnetwork, and recover the corresponding activation coefficient time series. We then apply staNMF to a resting state EEG dataset recorded from 41 five-year-old children with autism spectrum disorder (ASD) and 44 age-matched typically-developing (TD) children. StaNMF extracted several robust and biologically interpretable subnetworks for which significant differences between ASD and TD could be found. Specifically, the strength of the activation coefficient time series of one identified local connections subnetwork was higher in ASD children than in TD. Several subnetworks that had interhemispheric, long-range connections had lower energy of activation coefficient time series in ASD compared to TD. In conclusion, the analyses on simulated and real EEG data demonstrate that staNMF can be used to characterize dynamic functional networks, making it a promising tool for investigating disorders using resting-state EEG.

Recently, exploring brain activity based on functional networks during naturalistic stimuli especially music and video represents an attractive challenge because of the low signal-to-noise ratio in collected brain data. Although most efforts focusing on exploring the listening brain have been made through functional magnetic resonance imaging (fMRI), sensor-level electro- or magnetoencephalography (EEG/MEG) technique, little is known about how neural rhythms are involved in the brain network activity under naturalistic stimuli. This study exploited cortical oscillations through analysis of ongoing EEG and musical feature during freely listening to music. We used a data-driven method that combined music information retrieval with spatial Fourier Independent Components Analysis (spatial Fourier–ICA) to probe the interplay between the spatial profiles and the spectral patterns of the brain network emerging from music listening. Correlation analysis was performed between time courses of brain networks extracted from EEG data and musical feature time series extracted from music stimuli to derive the musical feature related oscillatory patterns in the listening brain. We found brain networks of musical feature processing were frequency-dependent. Musical feature time series, especially fluctuation centroid and key feature, were associated with an increased beta activation in the bilateral superior temporal gyrus. An increased alpha oscillation in the bilateral occipital cortex emerged during music listening, which was consistent with alpha functional suppression hypothesis in task-irrelevant regions. We also observed an increased delta–beta oscillatory activity in the prefrontal cortex associated with musical feature processing. In addition to these findings, the proposed method seems valuable for characterizing the large-scale frequency-dependent brain activity engaged in musical feature processing., Brain Topography, 33, ISSN:0896-0267, ISSN:1573-6792

In diagnosing sleep disorders, sleep stage classification is a very essential yet time-consuming process. Various existing state-of-the-art approaches rely on hand-crafted features and multi-modality polysomnography (PSG) data, where prior knowledge is compulsory and high computation cost can be expected. Besides, it is a big challenge to handle the task with raw single-channel electroencephalogram (EEG). To overcome these shortcomings, this paper proposes an end-to-end framework with a deep neural network, namely SingleChannelNet, for automatic sleep stage classification based on raw single-channel EEG. The proposed model utilizes a 90s epoch as the textual input and employs two multi-convolution (MC) blocks and several max-average pooling (M-Apooling) layers to learn different scales of feature representations. To demonstrate the efficiency of the proposed model, we evaluate our model using different raw single-channel EEGs (C4/A1 and Fpz-Cz) on two public PSG datasets (Cleveland children’s sleep and health study: CCSHS and Sleep-EDF database expanded: Sleep-EDF). Experimental results show that the proposed architecture can achieve better overall accuracy and Cohen’s kappa (CCSHS: 90.2%–86.5%, Sleep-EDF: 86.1%–80.5%) compared with state-of-the-art approaches. Additionally, the proposed model can learn features automatically for sleep stage classification using different single-channel EEGs with distinct sampling rates and without using any hand-engineered features. peerReviewed

Tucker decomposition can provide an intuitive summary to understand brain function by decomposing multi-subject fMRI data into a core tensor and multiple factor matrices, and was mostly used to extract functional connectivity patterns across time/subjects using orthogonality constraints. However, these algorithms are unsuitable for extracting common spatial and temporal patterns across subjects due to distinct characteristics such as high-level noise. Motivated by a successful application of Tucker decomposition to image denoising and the intrinsic sparsity of spatial activations in fMRI, we propose a low-rank Tucker-2 model with spatial sparsity constraint to analyze multi-subject fMRI data. More precisely, we propose to impose a sparsity constraint on spatial maps by using an lp norm (0

Driver distraction diverting drivers' attention to unrelated tasks and decreasing the ability to control vehicles, has aroused widespread concern about driving safety. Previous studies have found that driving performance decreases after distraction and have used vehicle behavioral features to detect distraction. But how brain activity changes while distraction remains unknown. Electroencephalography (EEG), a reliable indicator of brain activities has been widely employed in many fields. However, challenges still exist in mining the distraction information of EEG in realistic driving scenarios with uncertain information. In this paper, we propose a novel framework based on Multi-scale entropy (MSE) in a sliding window and Bidirectional Long Short-term Memory Network (BiLSTM) to explore the distraction information of EEG to detect driver distraction based on multi-modality signals in real traffic. Firstly, MSE with sliding window is implemented to extract the EEG features to determine the distraction position. Statistical analysis of vehicle behavioral data is then performed to validate driving performance indeed changes around distraction position. Finally, we use BiLSTM to detect driver distraction with MSE and other traditional features. Our results show that MSE notably decreases after distraction. Consistent with the result of MSE, driving performance significantly deviates from the normal state after distraction. Besides, BiLSTM performance of MSE outperforms other entropy-based methods and is better than behavioral features. Additionally, the accuracy is improved again after adding MSE feature to behavioral features with a 3% increasement. The proposed framework is useful for mining brain activity information and driver distraction detection applications in realistic driving scenarios. peerReviewed

Seizure prediction using intracranial electroencephalogram (iEEG) is still challenging because of complicated signals in spatial and time domains. Feature selection in the spatial domain (i.e., channel selection) has been largely ignored in this field. Hence, in this paper, a novel approach of iEEG channel selection strategy combined with one-dimensional convolutional neural networks (1D-CNN) was presented for seizure prediction. First, 15-sec and 30-sec iEEG segments with an increasing number of channels (from one channel to all channels) were sequentially fed into 1D-CNN models for training and testing. Then, the channel case with the best classification rate was selected for each participant. We tested our method on the Freiburg iEEG dataset. A sensitivity of 89.03-90.84%, specificity of 98.99-99.73%, and accuracy of 98.07-98.99% were achieved at the segment-based level. At the event-based level, we attained a sensitivity of 98.48-98.85% and a false prediction rate (FPR) of 0-0.02/h. peerReviewed

High dimensionality data have become common in neuroimaging fields, especially group-level functional magnetic resonance imaging (fMRI) datasets. fMRI connectivity analysis is a widely used, powerful technique for studying functional brain networks to probe underlying mechanisms of brain function and neuropsychological disorders. However, data-driven technique like independent components analysis (ICA), can yield unstable and inconsistent results, confounding the true effects of interest and hindering the understanding of brain functionality and connectivity. A key contributing factor to this instability is the information loss that occurs during fMRI data reduction. Data reduction of high dimensionality fMRI data in the temporal domain to identify the important information within group datasets is necessary for such analyses and is crucial to ensure the accuracy and stability of the outputs. In this study, we describe an fMRI data reduction strategy based on an adapted neighborhood preserving embedding (NPE) algorithm. Both simulated and real data results indicate that, compared with the widely used data reduction method, principal component analysis, the NPE-based data reduction method (a) shows superior performance on efficient data reduction, while enhancing group-level information, (b) develops a unique stratagem for selecting components based on an adjacency graph of eigenvectors, (c) generates more reliable and reproducible brain networks under different model orders when the outputs of NPE are used for ICA, (d) is more sensitive to revealing task-evoked activation for task fMRI, and (e) is extremely attractive and powerful for the increasingly popular fast fMRI and very large datasets. peerReviewed

The waveform in the time domain, spectrum in the frequency domain, and topography in the space domain of component(s) of interest are the fundamental indices in neuroscience research. Despite the application of time–frequency analysis (TFA) to extract the temporal and spectral characteristics of non-phase-locked component (NPLC) of interest simultaneously, the statistical results are not always expectedly satisfying, in that the spatial information is not considered. Complex Morlet wavelet transform is widely applied to TFA of event-related-potential (ERP) data, and mother wavelet (which should be firstly defined by center frequency and bandwidth (CFBW) before using the method to TFA of ERP data) influences the time–frequency results. In this study, an optimal set of CFBW was firstly selected from the number sets of CFBW, to further analyze for TFA of the ERP data in a cognitive experiment paradigm of emotion (Anger and Neutral) and task (Go and Nogo). Then tensor decomposition algorithm was introduced to investigate the NPLC of interest from the fourth-order tensor. Compared with the TFA results which only revealed a significant difference between Go and Nogo task condition, the tensor-based analysis showed significant interaction effect between emotion and task. Moreover, significant differences were found in both emotion and task conditions through tensor decomposition. In addition, the statistical results of TFA would be affected by the selected region of interest (ROI), whereas those of the proposed method were not subject to ROI. Hence, this study demonstrated that tensor decomposition method was effective in extracting NPLC, by considering spatial information simultaneously as the potential to explore the brain mechanisms related to experimental design.

Ongoing electroencephalography (EEG) signals are recorded as a mixture of stimulus-elicited EEG, spontaneous EEG and noises, which poses a huge challenge to current data analyzing techniques, especially when different groups of participants are expected to have common or highly correlated brain activities and some individual dynamics. In this study, we proposed a data-driven shared and unshared feature extraction framework based on nonnegative and coupled tensor factorization, which aims to conduct group-level analysis for the EEG signals from major depression disorder (MDD) patients and healthy controls (HC) when freely listening to music. Constrained tensor factorization not only preserves the multilinear structure of the data, but also considers the common and individual components between the data. The proposed framework, combined with music information retrieval, correlation analysis, and hierarchical clustering, facilitated the simultaneous extraction of shared and unshared spatio-temporal-spectral feature patterns between/in MDD and HC groups. Finally, we obtained two shared feature patterns between MDD and HC groups, and obtained totally three individual feature patterns from HC and MDD groups. The results showed that the MDD and HC groups triggered similar brain dynamics when listening to music, but at the same time, MDD patients also brought some changes in brain oscillatory network characteristics along with music perception. These changes may provide some basis for the clinical diagnosis and the treatment of MDD patients.

Deep learning has achieved unprecedented success in sleep stage classification tasks, which starts to pave the way for potential real-world applications. However, due to its enormous size, deployment of deep neural networks is hindered by high cost at various aspects, such as computation power, storage, network bandwidth, power consumption, and hardware complexity. For further practical applications (e.g., wearable sleep monitoring devices), there is a need for simple and compact models. In this paper, we propose a lightweight model, namely LightSleepNet, for rapid sleep stage classification based on spectrograms. Our model is assembled by a much fewer number of model parameters compared to existing ones. Furthermore, we convert the raw EEG data into spectrograms to speed up the training process. We evaluate the model performance on several public sleep datasets with different characteristics. Experimental results show that our lightweight model using spectrogram as input can achieve comparable overall accuracy and Cohen’s kappa (SHHS100: 86.7%-81.3%, Sleep-EDF: 83.7%-77.5%, Sleep-EDF-v1: 88.3%-84.5%) compared to the state-of-the-art methods on experimental datasets. peerReviewed

Seizure prediction using intracranial electroencephalogram (iEEG) has attracted an increasing attention during recent years. iEEG signals are commonly recorded in the form of multiple channels. Many previous studies generally used the iEEG signals of all channels to predict seizures, ignoring the consideration of channel selection. In this study, a method of one-dimensional convolutional neural networks (1D-CNN) combined with channel selection strategy was proposed for seizure prediction. First, we used 30-s sliding windows to segment the raw iEEG signals. Then, the 30-s iEEG segments, which were in three channel forms (single channel, channels only from seizure onset or free zone and all channels from seizure onset and free zones), were used as the inputs of 1D-CNN for classification, and the patient-specific model was trained. Finally, the channel form with the best classification was selected for each patient. The proposed method was evaluated on the Freiburg Hospital iEEG dataset. In the situation of seizure occurrence period (SOP) of 30[Formula: see text]min and seizure prediction horizon (SPH) of 5[Formula: see text]min, 98.60[Formula: see text] accuracy, 98.85[Formula: see text] sensitivity and 0.01/h false prediction rate (FPR) were achieved. In the situation of SOP of 60[Formula: see text]min and SPH of 5[Formula: see text]min, 98.32[Formula: see text] accuracy, 98.48[Formula: see text] sensitivity and 0.01/h FPR were attained. Compared with the many existing methods using the same iEEG dataset, our method showed a better performance.

The countermeasure of driver fatigue is valuable for reducing the risk of accidents caused by vigilance failure during prolonged driving. Listening to the radio (RADIO) has been proven to be a relatively effective “in-car” countermeasure. However, the connectivity analysis, which can be used to investigate its alerting effect, is subject to the issue of signal mixing. In this study, we propose a novel framework based on clustering and entropy to improve the performance of the connectivity analysis to reveal the effect of RADIO to maintain driver alertness. Instead of reducing signal mixing, we introduce clustering algorithm to classify the functional connections with their nodes into different categories to mine the effective information of the alerting effect. Differential entropy (DE) is employed to measure the information content in different brain regions after clustering. Compared with the Louvain-based community detection method, the proposed method shows superior ability to present RADIO effect in the confused functional connection matrices. Our experimental results reveal that the active connection clusters distinguished by the proposed method gradually move from the frontal region to the parieto-occipital region with the progress of fatigue, consistent with the alpha energy changes in the two brain areas. The active clusters in the parieto-occipital region significantly decrease and the most active clusters remain in the frontal region when RADIO is taken. The estimation results of DE confirm the significant change (p

A better understanding of the tumor immune microenvironment (TIME) could lead to accurate diagnosis, prognosis, and treatment stratification. Although molecular analyses at the tissue and/or single cell level could reveal the cellular status of the tumor microenvironment, these approaches lack information related to spatial-level cellular distribution, co-organization, and cell-cell interaction in the TIME. With the emergence of computational pathology coupled with machine learning (ML) and artificial intelligence (AI), ML- and AI-driven spatial TIME analyses of pathology images could revolutionize our understanding of the highly heterogeneous and complex molecular architecture of the TIME. In this review we highlight recent studies on spatial TIME analysis of pathology slides using state-of-the-art ML and AI algorithms. PATIENT SUMMARY: This mini-review reports recent advances in machine learning and artificial intelligence for spatial analysis of the tumor immune microenvironment in pathology slides. This information can help in understanding the spatial heterogeneity and organization of cells in patient tumors.

Tucker decomposition (TKD) has been utilized to identify functional connectivity patterns using processed fMRI data, but seldom focuses on originally acquired fMRI data. This study proposes to decompose multi-subject fMRI data in a natural three-way of voxel × time × subject via TKD. Different from existing tensor decomposition algorithms such as canonical polyadic decomposition (CPD) for extracting shared spatial maps (SMs), we propose to extract both shared and individual SMs by exploring spatial-temporal-subject relationship contained in the core tensor. We test the proposed method using multi-subject resting-state fMRI data with comparison to CPD for evaluating shared SMs and independent vector analysis (IVA) for assessing individual SMs under different model orders. The results show that the proposed method yields better and more robust shared SMs than CPD and more consistent individual SMs than IVA, indicating the potential of TKD in providing group and individual brain networks in a high-dimensional coupling way.

© 2018 Elsevier B.V. Double coupled canonical polyadic decomposition (DC-CPD) decomposes multiple tensors with coupling in the first two modes, into minimal number of rank-1 tensors that also admit the double coupling structure. It has a particular interest in joint blind source separation (J-BSS) applications. In a preceding paper, we proposed an algebraic algorithm for underdetermined DC-CPD, of which the factor matrices in the first two modes of each tensor may have more columns than rows. It uses a pairwise coupled rank-1 detection mapping to transform a possibly underdetermined DC-CPD into an overdetermined DC-CPD, which can be solved algebraically via generalized eigenvalue decomposition (GEVD). In this paper, we generalize the pairwise or second-order coupled rank-1 detection mapping to an arbitrary order K≥2. Based on this generalized coupled rank-1 detection mapping, we propose a broad framework for the algebraic computation of DC-CPD, which consists of a series of algorithms with more relaxed working assumptions, each corresponding to a fixed order K≥2. Deterministic and generic uniqueness conditions are provided. We will show through analysis and numerical results that our new uniqueness conditions for DC-CPD are more relaxed than the existing results for DC-CPD and CPD. We will further show, through simulation results, the performance of the proposed algebraic DC-CPD framework in approximate DC-CPD and a J-BSS application, in comparison with existing DC-CPD and CPD algorithms. ispartof: SIGNAL PROCESSING-IMAGE COMMUNICATION vol:73 pages:22-36 status: published

Objective The study aims to develop an automatic sleep scoring method by fusing different polysomnography (PSG) signals and further to investigate PSG signals’ contribution to the scoring result. Methods Eight combinations of four modalities of PSG signals, namely electroencephalogram (EEG), electrooculogram (EOG), electromyogram (EMG), and electrocardiogram (ECG) were considered to find the optimal fusion of PSG signals. A total of 232 features, covering statistical characters, frequency characters, time-frequency characters, fractal characters, entropy characters and nonlinear characters, were derived from these PSG signals. To select the optimal features for each signal fusion, four widely used feature selection methods were compared. At the classification stage, five different classifiers were employed to evaluate the validity of the features and to classify sleep stages. Results For the database in the present study, the best classifier, random forest, realized the optimal consistency of 86.24% with the sleep macrostructures scored by the technologists trained at the Sleep Center. The optimal accuracy was achieved by fusing four modalities of PSG signals. Specifically, the top twelve features in the optimal feature set were respectively EEG features named zero-crossings, spectral edge, relative power spectral of theta, Petrosian fractal dimension, approximate entropy, permutation entropy and spectral entropy, and EOG features named spectral edge, approximate entropy, permutation entropy and spectral entropy, and the mutual information between EEG and submental EMG. In addition, ECG features (e.g. Petrosian fractal dimension, zero-crossings, mean value of R amplitude and permutation entropy) were useful for the discrimination among W, S1 and R. Conclusions Through exploring the different fusions of multi-modality signals, the present study concluded that the multi-modality of PSG signals’ fusion contributed to higher accuracy, and the optimal feature set was a fusion of multiple types of features. Besides, compared with manual scoring, the proposed automatic scoring methods were cost-effective, which would alleviate the burden of the physicians, speed up sleep scoring, and expedite sleep research.

Objective: Abnormal and dynamic epileptogenic networks cause difficulties for clinical epileptologists in the localization of the seizure onset zone (SOZ) and the epileptogenic zone (EZ) in preoperative assessments of patients with refractory epilepsy. The aim of this study is to investigate the characteristics of time-varying effective connectivity networks in various non-seizure and seizure periods, and to propose a quantitative approach for accurate localization of SOZ and EZ. Methods: We used electrocorticogram recordings in the temporal lobe and hippocampus from seven patients with temporal lobe epilepsy to characterize the effective connectivity dynamics at a high temporal resolution using the full-frequency adaptive directed transfer function (ffADTF) measure and five graph metrics, i.e., the out-degree (OD), closeness centrality (CC), betweenness centrality (BC), clustering coefficient (C), and local efficiency (LE). The ffADTF effective connectivity network was calculated and described in five frequency bands (δ, θ, α, β, and γ) and five seizure periods (pre-seizure, early seizure, mid-seizure, late seizure, and post-seizure). The cortical areas with high values of graph metrics in the transient seizure onset network were compared with the SOZ and EZ identified by clinical epileptologists and the results of epilepsy resection surgeries. Results: Origination and propagation of epileptic activity were observed in the high time resolution ffADTF effective connectivity network throughout the entire seizure period. The seizure-specific transient seizure onset ffADTF network that emerged at seizure onset time remained for approximately 20–50 ms with strong connections generated from both SOZ and EZ. The values of graph metrics in the SOZ and EZ were significantly larger than that in the other cortical areas. More cortical areas with the highest mean of graph metrics were the same as the clinically determined SOZ in the low-frequency δ and θ bands and in Engel Class I patients than in higher frequency α, β, and γ bands and in Engel Class II and III patients. The OD and C were more likely to localize the SOZ and EZ than CC, BC, and LE in the transient seizure onset network. Conclusion: The high temporal resolution ffADTF effective connectivity analysis combined with the graph theoretical analysis helps us to understand how epileptic activity is generated and propagated during the seizure period. The newly discovered seizure-specific transient seizure onset network could be an important biomarker and a promising tool for more precise localization of the SOZ and EZ in preoperative evaluations. peerReviewed

Spatial source phase, the phase information of spatial maps extracted from functional magnetic resonance imaging (fMRI) data by data-driven methods such as independent component analysis (ICA), has rarely been studied. While the observed phase has been shown to convey unique brain information, the role of spatial source phase in representing the intrinsic activity of the brain is yet not clear. This study explores the spatial source phase for identifying spatial differences between patients with schizophrenia (SZs) and healthy controls (HCs) using complex-valued resting-state fMRI data from 82 individuals. ICA is first applied to preprocess fMRI data, and post-ICA phase de-ambiguity and denoising are then performed. The ability of spatial source phase to characterize spatial differences is examined by the homogeneity of variance test (voxel-wise F-test) with false discovery rate correction. Resampling techniques are performed to ensure that the observations are significant and reliable. We focus on two components of interest widely used in analyzing SZs, including the default mode network (DMN) and auditory cortex. Results show that the spatial source phase exhibits more significant variance changes and higher sensitivity to the spatial differences between SZs and HCs in the anterior areas of DMN and the left auditory cortex, compared to the magnitude of spatial activations. Our findings show that the spatial source phase can potentially serve as a new brain imaging biomarker and provide a novel perspective on differences in SZs compared to HCs, consistent with but extending previous work showing increased variability in patient data.