Real-time functional magnetic resonance imaging neurofeedback (rt-fMRI-NF) is a promising noninvasive brain computer interface (BCI) technique for enhancing self-regulation of affective brain states. However, conventional univariate rt-fMRI-NF approaches struggle to discriminate distributed neural patterns underlying distinct emotions. This study implemented an rt-fMRI semantic neurofeedback (rt-fMRI-sNF) paradigm incorporating real-time representational similarity analysis (rt-RSA) to enable navigation among emotional states. Four emotion-specific base patterns were first derived from functional localizer runs and then used as target patterns during neurofeedback. Using an RSA-informed circular semantic map (CSM), participants received real-time visual feedback indicating both the similarity and intensity of their current brain activity relative to target patterns. Participants were instructed to use mental imagery to shift their brain activity toward the specific target pattern and enhance its intensity. Analyses of localizer data revealed overlapping regional activations across emotions and demonstrated that RSA reliably distinguished between emotional states. Group-level mixed-effects modeling of neurofeedback performance indicated significant within-run improvements and higher initial performance in the second run. Together, these results demonstrate the methodological feasibility of an RSA-informed rt-fMRI-NF framework for multivariate brain-state modulation and establish a foundation for future studies examining its transferability and clinical relevance.

Continual learning aims to enable models to acquire new knowledge while retaining previously learned information. Prompt-based methods have shown remarkable performance in this domain; however, they typically rely on key-value pairing, which can introduce inter-task interference and hinder scalability. To overcome these limitations, we propose a novel approach employing task-specific Prompt-Prototype (ProP), thereby eliminating the need for key-value pairs. In our method, task-specific prompts facilitate more effective feature learning for the current task, while corresponding prototypes capture the representative features of the input. During inference, predictions are generated by binding each task-specific prompt with its associated prototype. Additionally, we introduce regularization constraints during prompt initialization to penalize excessively large values, thereby enhancing stability. Experiments on several widely used datasets demonstrate the effectiveness of the proposed method. In contrast to mainstream prompt-based approaches, our framework removes the dependency on key-value pairs, offering a fresh perspective for future continual learning research.

Immunohistochemical (IHC) biomarker prediction greatly benefits from multimodal data fusion. However, the simultaneous acquisition of genomic and pathological data is often constrained by cost or technical limitations. To address this, we propose a novel Genomics-guided Multimodal Knowledge Decomposition Network (GMKDN), a framework that effectively integrates genomics and pathology data during training while dynamically adapting to available data during inference. GMKDN introduces two key innovations: 1) the Batch-Sample Multimodal Knowledge Decomposition (BMKD) module, which decomposes input features into pathology-specific, modality-general, and genomics-specific components to reduce redundancy and enhance knowledge transferability, and 2) the Online Similarity-Preserving Knowledge Distillation (OSKD) module, which optimizes activation similarity matrices to facilitate robust knowledge transfer between teacher and student models. The BMKD module improves generalization across modalities, while the OSKD module enhances model robustness, particularly when certain modalities are unavailable during inference. Extensive evaluations conducted on the TCGA-BRCA dataset and an external test cohort (QHSU) demonstrate that GMKDN consistently outperforms state-of-the-art (SOTA) slide-based multiple instance learning (MIL) approaches as well as existing multimodal learning models, establishing a new benchmark for breast cancer biomarker prediction. Our code is available at https://github.com/qiyuanzz/GMKDN.

In recent years, with the advancements in brain science, spiking neural networks (SNNs) have garnered significant attention. SNNs can generate spikes that mimic the function of neurons transmission in humans brain, thereby significantly reducing computational costs by the event-driven nature during training. While deep SNNs have shown impressive performance on classification tasks, they still face challenges in more complex tasks such as object detection. In this paper, we propose SpikingYOLOX, extending the structure of the original YOLOX by introducing signed spiking neurons and fast Fourier convolution (FFC). The designed ternary signed spiking neurons could generate three kinds of spikes to obtain more robust features in the deep layer of the backbone. Meanwhile, we integrate FFC with SNN modules to enhance object detection performance, because its global receptive field is beneficial to the object detection task. Extensive experiments demonstrate that the proposed SpikingYOLOX achieves state-of-the-art performance among other SNN-based object detection methods.

The molecular classification of adult-type diffuse gliomas is essential for determining appropriate therapeutic strategies, but genomic sequencing remains costly. Recent advances in digital pathology and deep learning have led to several studies exploring molecular classification using multiple instance learning (MIL) on whole slide images (WSIs). However, achieving optimal classification performance using only histological slides is challenging due to the lack of guidance from genomic data. In this study, we propose a teacher-student distillation framework for glioma molecular classification using WSIs. Our method leverages a pretrained self-normalizing neural network (SNN) as the genomic teacher model, which selects genes based on survival analysis-driven criteria to guide the MIL-based student model in learning effective histological representations. During training, both genomic and pathological data are utilized, while inference relies solely on WSIs. Experimental validation on the TCGA GBM-LGG datasets shows that our approach outperforms state-of-the-art (SOTA) MIL models, highlighting its effectiveness in glioma diagnostic subtyping using WSIs.

Recently, virtual staining has emerged as a promising alternative to revolutionize histological staining by digitally generating stains. However, most existing methods suffer from the curse of staining unreality and unreliability. In this paper, we propose the Orthogonal Decoupling Alignment Generative Adversarial Network (ODA-GAN) for unpaired virtual immunohistochemistry (IHC) staining. Our approach is based on the assumption that an image consists of IHC staining-related features, which influence staining distribution and intensity, and staining-unrelated features, such as tissue morphology. Leveraging a pathology foundation model, we first develop a weakly-supervised segmentation pipeline as an alternative to expert annotations. We introduce an Orthogonal MLP (O-MLP) module to project image features into an orthogonal space, decoupling them into staining-related and unrelated components. Additionally, we propose a Dual-stream PatchNCE (DPNCE) loss to resolve contrastive learning contradictions in the staining-related space, thereby enhancing staining accuracy. To further improve realism, we introduce a Multi-layer Domain Alignment (MDA) module to bridge the domain gap between generated and real IHC images. Evaluations on three benchmark datasets show that our ODA-GAN reaches state-of-the-art (SOTA) performance. Our source code is available at https://github.com/ittong/ODA-GAN.

Brain-inspired Spiking Neural Networks (SNNs) have garnered significant attention due to their bio-plausibility and low power consumption advantages compared to Artificial Neural Networks (ANNs). However, the application of SNN in computer vision remains limited, primarily due to their inferior performance. In this work, we aim to bridge the performance gap between ANNs and SNNs in object detection by our Advanced SpikingYOLOX. The proposed approach extends the SpikingYOLOX with two key innovations: PSA-SNN and 2D-Spiking Transformer, both designed to enhance object detection performance. PSA-SNN extends spike-based self-attention by incorporating high-speed partial self-attention with an SNN-based 2D-Spiking Transformer in the deepest layer of the backbone, significantly improving feature extraction. The 2D-Spiking Transformer redefines the role of spiking neurons in Transformer sequences (Key, Query, Value), demonstrating that applying an additional spiking layer solely to the Value sequence yields the best performance while maintaining computational efficiency in spike-driven Transformers. We conduct extensive experiments on static images and the Advanced SpikingYOLOX achieves state-of-the-art performance among other SNN-based object detection methods. This work paves the way for more advanced SNN applications in object detection and broader computer vision tasks.

The increasing scale and complexity of modern IT systems necessitate advanced solutions for monitoring and managing performance anomalies. Artificial intelligence for IT operations (AIOps) has emerged as a promising approach to enhance the efficiency and effectiveness of IT operations. However, existing methods struggle with effectively detecting anomalies in multi-dimensional performance data and accurately identifying their root causes in complex interdependent systems. This paper proposes a novel framework, PA-Rank, that combines generative adversarial networks (GANs), reinforcement learning, and graph-based methods to address these challenges comprehensively. For anomaly detection, an unsupervised GAN-based model is developed to identify anomalous time periods and assign weighted scores to metrics, facilitating precise anomaly identification. For root cause localization, a Causal Graph Construction Model (CGCM) has been developed, utilizing a reinforcement learning-based causal discovery method that is integrated with graph attention networks (GAT) to construct a causal graph representing the relationships between metrics. A random walk algorithm further ranks metric importance during anomalies, enabling effective root cause localization. Extensive experiments on real-world datasets, including SMD, ASD, and DAMADICS, demonstrate the superiority of PA-Rank over traditional statistical and state-of-the-art machine learning methods. On the SMD dataset, the proposed framework achieved an F1 score of 0.9542 for anomaly detection and consistently identified root causes among top-ranked candidates on the Pymicro and RMS datasets with the highest PR@Avg scores. These results underscore PA-Rank’s efficacy in diagnosing performance anomalies and supporting efficient system maintenance.

Cognitive neuroscience is currently a field of research highly valued by many countries worldwide, and fostering corresponding international collaboration can accelerate the development of cognitive neuroscience in our country. However, challenges related to industry competition, privacy, and regulatory policies hinder international collaboration that relies on cross-server data sharing. Considering the current limitations of tensor decomposition methods in establishing constraints between cross-server data, this study leverages the advantages of federated learning to develop a federated non-negative coupled tensor decomposition framework (FCNCP), aimed at establishing coupling constraints across different servers while preserving privacy. In experiments validating the effectiveness of the algorithm, we conducted 50 decompositions on synthetic tensor data, achieving an average tensor fit coefficient of 0.996, and the results demonstrated successful establishment of the coupling constraint. In real ERP data decomposition experiments, we applied the FCNCP algorithm to decompose ERP tensor data collected during proprioceptive stimulation applied to the left and right hands. The decomposition results revealed symmetrical activation areas in the left and right hemispheres induced by contralateral stimulation, with components in the beta and gamma frequency bands. These components are consistent with findings from related studies in cognitive neuroscience, confirming that this method can effectively handle high-dimensional EEG data across servers. This study not only provides new tools and approaches for processing and analyzing high-dimensional EEG data across servers but also promotes the advancement of coupled tensor decomposition techniques and their integration with emerging federated learning frameworks, offering significant theoretical and practical value.

Automated detection of lymph node metastasis (LNM) holds great potential to alleviate the workload of doctors and reduce misinterpretations. Despite the practical successes achieved, effectively addressing the highly complex and heterogeneous tumor microenvironment remains an open and challenging problem, especially when tumor subtypes intermingle and are difficult to delineate. In this paper, we propose a multi-task adaptive resolution network, named MAR-Net, for LNM detection and subtyping in complex mixed-type cancers. Specifically, we construct a resolution-aware module to mine heterogeneous diagnostic information, which exploits the multi-scale pyramid information and adaptively combines multi-resolution structured features for comprehensive representation. Additionally, we adopt a multi-task learning approach that simultaneously addresses LNM detection and subtyping, reducing model instability during optimization and improving performance across both tasks. More importantly, to rectify the potential misclassification of tumor subtypes, we elaborately design a hierarchical subtying refinement (HSR) algorithm that leverages a generic segmentation model informed by pathologists' prior knowledge. Evaluations have been conducted on three private and one public cancer datasets (554 WSIs, 4.8 million patches). Our experimental results demonstrate that the proposed method consistently achieves superior performance compared to the state-of-the-art methods, achieving 0.5% to 3.2% higher AUC in LNM detection and 3.8% to 4.4% higher AUC in LNM subtyping.

Sleep stage classification plays a crucial role in sleep quality assessment and sleep disorder prevention. Nowadays, many studies have developed algorithms for this purpose, but they still face two challenges. The first is noise in physiological signals from various devices. The second challenge is that most studies simply concatenate multi-modal features without considering their correlations. To this end, we propose a framework, namely Diff-SleepNet, to efficiently classify sleep stages from multi-modal input. This framework begins with a diffusion model with peak signal-to-noise ratio (PNSR) loss function that adaptively filters noise. The filtered signals are then transformed into a multi-view spectrum through data pre-processing. These spectra are processed by a transformer-based backbone to extract multi-modal features. The production is fed into the following multi-scale attention module for robust feature fusion. The sleep stage category is finally determined by a fully connected layer. Our framework is trained and validated on three typical datasets, i.e., SHHS, Sleep-EDF-SC, and Sleep-EDF-X. Experimental results demonstrate that it is effective and has advantages over other peer methods.

OBJECTIVE: Intracranial electroencephalogram (iEEG) signals are generally recorded using multiple channels, and channel selection is therefore a significant means in studying iEEG-based seizure prediction. For n channels, [Formula: see text] channel cases can be generated for selection. However, by this means, an increase in n can cause an exponential increase in computational consumption, which may result in a failure of channel selection when n is too large. Hence, it is necessary to explore reasonable channel selection strategies under the premise of controlling computational consumption and ensuring high classification accuracy. Given this, we propose a novel method of channel reordering strategy combined with dual CNN-LSTM for effectively predicting seizures. METHOD: First, for each patient with n channels, interictal and preictal iEEG samples from each single channel are input into the CNN-LSTM model for classification. Then, the F1-score of each single channel is calculated, and the channels are reordered in descending order according to the size of F1-scores (channel reordering strategy). Next, iEEG signals with an increasing number of channels are successively fed into the CNN-LSTM model for classification again. Finally, according to the classification results from n channel cases, the channel case with the highest classification rate is selected. RESULTS: Our method is evaluated on the three iEEG datasets: the Freiburg, the SWEC-ETHZ and the American Epilepsy Society Seizure Prediction Challenge (AES-SPC). At the event-based level, the sensitivities of 100%, 100% and 90.5%, and the false prediction rates (FPRs) of 0.10/h, 0/h and 0.47/h, are achieved for the three datasets, respectively. Moreover, compared to an unspecific random predictor, our method also shows a better performance for all patients and dogs from the three datasets. At the segment-based level, the sensitivities-specificities-accuracies-AUCs of 88.1%-94.0%-93.5%-0.9101, 99.1%-99.7%-99.6%-0.9935, and 69.2%-79.9%-78.2%-0.7373, are attained for the three datasets, respectively. CONCLUSION: Our method can effectively predict seizures and address the challenge of an excessive number of channels during channel selection.

Driver distraction, a major cause of traffic crashes, is reported to reduce driving performance and be detected with vehicle behavioral features. It also induces physiological responses. Time and frequency-domain features of physiological signals have been used to study distraction, but they are susceptible to residual noise and tend to overlook complexity. Moreover, the resampling problem arises while analyzing physiological signals at multiple time scales. This paper proposes a novel framework based on multiscale entropy on absolute time scales (MSaE) and bidirectional long short-term memory (BiLSTM) network to mine the distraction information in multi-modality physiological signals and detect distraction automatically. Firstly, an entropy-based resampling method is adopted to find the suitable downsampling rates of electroencephalography (EEG), electrocardiogram (ECG), and electromyography (EMG). Then, calculating entropy with absolute time scales instead of relative time scales in a sliding window is utilized to explore the fluctuations of each signal while distraction. Afterward, ReliefF is selected from conventional feature selectors to identify the optimal feature set for each signal. Finally, BiLSTM with time dependency is designed to detect driver distraction with the selected feature set. The results illustrate significant distinctions in the MSaE of multiple physiological signals between normal and distracted driving. Additionally, MSaE, superior to traditional features, is selected as the most discriminative feature for each signal in distraction mining. Furthermore, the accuracy is further improved by about 8%, incorporating multi-modality features rather than vehicle behavioral features. This study indicates the potential of employing various signals to understand and detect driver distraction effectively.

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.

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.

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.

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$ \ell _{p} $norm (${0}< {p}\le {1}$), in addition to adding low-rank constraints on factor matrices via the Frobenius norm. We solve the constrained Tucker-2 model using alternating direction method of multipliers, and propose to update both sparsity and low-rank constrained spatial maps using half quadratic splitting. Moreover, we extract new spatial and temporal features in addition to subject-specific intensities from the core tensor, and use these features to classify multiple subjects. The results from both simulated and experimental fMRI data verify the improvement of the proposed method, compared with four related algorithms including robust Kronecker component analysis, Tucker decomposition with orthogonality constraints, canonical polyadic decomposition, and block term decomposition in extracting common spatial and temporal components across subjects. The spatial and temporal features extracted from the core tensor show promise for characterizing subjects within the same group of patients or healthy controls as well.

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.

To examine the electrophysiological underpinnings of the functional networks involved in music listening, previous approaches based on spatial independent component analysis (ICA) have recently been used to ongoing electroencephalography (EEG) and magnetoencephalography (MEG). However, those studies focused on healthy subjects, and failed to examine the group-level comparisons during music listening. Here, we combined group-level spatial Fourier ICA with acoustic feature extraction, to enable group comparisons in frequency-specific brain networks of musical feature processing. It was then applied to healthy subjects and subjects with major depressive disorder (MDD). The music-induced oscillatory brain patterns were determined by permutation correlation analysis between individual time courses of Fourier-ICA components and musical features. We found that (1) three components, including a beta sensorimotor network, a beta auditory network and an alpha medial visual network, were involved in music processing among most healthy subjects; and that (2) one alpha lateral component located in the left angular gyrus was engaged in music perception in most individuals with MDD. The proposed method allowed the statistical group comparison, and we found that: (1) the alpha lateral component was activated more strongly in healthy subjects than in the MDD individuals, and that (2) the derived frequency-dependent networks of musical feature processing seemed to be altered in MDD participants compared to healthy subjects. The proposed pipeline appears to be valuable for studying disrupted brain oscillations in psychiatric disorders during naturalistic paradigms.

Brain-computer interfaces (BCIs) can be used in real-time fMRI neurofeedback (rtfMRI NF) investigations to provide feedback on brain activity to enable voluntary regulation of the blood-oxygen-level dependent (BOLD) signal from localized brain regions. However, the temporal pattern of successful self-regulation is dynamic and complex. In particular, the general linear model (GLM) assumes fixed temporal model functions and misses other dynamics. We propose a novel data-driven analyses approach for rtfMRI NF using intersubject covariance (ISC) analysis. The potential of ISC was examined in a reanalysis of data from 21 healthy individuals and nine patients with post-traumatic stress-disorder (PTSD) performing up-regulation of the anterior cingulate cortex (ACC). ISC in the PTSD group differed from healthy controls in a network including the right inferior frontal gyrus (IFG). In both cohorts, ISC decreased throughout the experiment indicating the development of individual regulation strategies. ISC analyses are a promising approach to reveal novel information on the mechanisms involved in voluntary self-regulation of brain signals and thus extend the results from GLM-based methods. ISC enables a novel set of research questions that can guide future neurofeedback and neuroimaging investigations.

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.

Abstract 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.

Recent continuous task studies, such as narrative speech comprehension, show that fluctuations in brain functional connectivity (FC) are altered and enhanced compared to the resting state. Here, we characterized the fluctuations in FC during comprehension of speech and time-reversed speech conditions. The correlations of Hilbert envelope of source-level EEG data were used to quantify FC between spatially separate brain regions. A symmetric multivariate leakage correction was applied to address the signal leakage issue before calculating FC. The dynamic FC was estimated based on a sliding time window. Then, principal component analysis (PCA) was performed on individually concatenated and temporally concatenated FC matrices to identify FC patterns. We observed that the mode of FC induced by speech comprehension can be characterized with a single principal component. The condition-specific FC demonstrated decreased correlations between frontal and parietal brain regions and increased correlations between frontal and temporal brain regions. The fluctuations of the condition-specific FC characterized by a shorter time demonstrated that dynamic FC also exhibited condition specificity over time. The FC is dynamically reorganized and FC dynamic pattern varies along a single mode of variation during speech comprehension. The proposed analysis framework seems valuable for studying the reorganization of brain networks during continuous task experiments.

Canonical polyadic decomposition (CPD) of multi-subject complex-valued fMRI data can be used to provide spatially and temporally shared components among groups with both magnitude and phase information. However, the CPD model is not well formulated due to the large subject variability in the spatial and temporal modalities, as well as the high noise level in complexvalued fMRI data. Considering that the shift-invariant CPD can model temporal variability across subjects, we propose to further impose a phase sparsity constraint on the shared spatial maps to denoise the complex-valued components and to model the inter-subject spatial variability as well. More precisely, subject-specific time delays are first estimated for the complex-valued shared time courses in the framework of real-valued shift-invariant CPD. Source phase sparsity is then imposed on the complex-valued shared spatial maps. A smoothed ℓ0norm is specifically used to reduce voxels with large phase values after phase de-ambiguity based on the small phase characteristic of BOLD-related voxels. The results from both the simulated and experimental fMRI data demonstrate improvements of the proposed method over three complex-valued algorithms, namely, tensor-based spatial ICA, shift-invariant CPD and CPD without spatiotemporal constraints. When comparing with a real-valued algorithm combining shiftinvariant CPD and ICA, the proposed method detects 178.7% more contiguous task-related activations.

Tensor decomposition is a powerful tool for analyzing multiway data. Nowadays, with the fast development of multisensor technology, more and more data appear in higher-order (order > 4) and nonnegative form. However, the decomposition of higher-order nonnegative tensor suffers from poor convergence and low speed. In this study, we propose a new nonnegative CANDECOM/PARAFAC (NCP) model using proximal algorithm. The block principal pivoting method in alternating nonnegative least squares (ANLS) framework is employed to minimize the objective function. Our method can guarantee the convergence and accelerate the computation. The results of experiments on both synthetic and real data demonstrate the efficiency and superiority of our method.

Real-world data exhibiting high order/dimensionality and various couplings are linked to each other since they share some common characteristics. Coupled tensor decomposition has become a popular technique for group analysis in recent years, especially for simultaneous analysis of multi-block tensor data with common information. To address the multiblock tensor data, we propose a fast double-coupled nonnegative Canonical Polyadic Decomposition (FDC-NCPD) algorithm in this study, based on the linked CP tensor decomposition (LCPTD) model and fast Hierarchical Alternating Least Squares (Fast-HALS) algorithm. The proposed FDCNCPD algorithm enables simultaneous extraction of common components, individual components and core tensors from tensor blocks. Moreover, time consumption is greatly reduced without compromising the decomposition quality when handling large-scale tensor blocks. Simulation experiments of synthetic and real-world data are conducted to demonstrate the superior performance of the proposed algorithm.

The characterization of dynamic electrophysiological brain activity, which form and dissolve in order to support ongoing cognitive function, is one of the most important goals in neuroscience. Here, we introduce a method with tensor decomposition for measuring the task-induced oscillations in the human brain using electroencephalography (EEG). The time frequency representation of source-reconstructed single-trail EEG data constructed a third-order tensor with three factors of time · trails, frequency and source points. We then used a non-negative Canonical Polyadic decomposition (NCPD) to identify the temporal, spectral and spatial changes in electrophysiological brain activity. We validate this method using both simulation EEG data and real EEG data recorded during a task of irony comprehension. The results demonstrated that proposed method can track dynamics of the temporal-spectral modes of the rhythm in the brain on a timescale commensurate to the task they are undertaking.

Magnitude-only resting-state fMRI data have been largely investigated via independent component analysis (ICA) for exacting spatial maps (SMs) and time courses. However, the native complex-valued fMRI data have rarely been studied. Motivated by the significant improvements achieved by ICA of complex-valued task fMRI data than magnitude-only task fMRI data, we present an efficient method for de-noising SM estimates which makes full use of complex-valued resting-state fMRI data. Our two main contributions include: (1) The first application of a post-ICA phase de-noising method, originally proposed for task fMRI data, to resting-state data, which recognizes voxels within a specific phase range as desired voxels. (2) A new phase range detection strategy for a specific SM component based on correlation with its reference. We continuously change the phase range within a larger range, and compute a set of correlation coefficients between each de-noised SM and its reference. The phase range with the maximal correlation determines the final selection. The detected results by the proposed approach confirm the correctness of the post-ICA phase de-noising method in the analysis of resting-state complex-valued fMRI data.

Independent vector analysis (IVA) has exhibited great potential for the group analysis of magnitude-only fMRI data, but has rarely been applied to native complex-valued fMRI data. We propose an adaptive fixed-point IVA algorithm by taking into account the extremely noisy nature, large variability of the source component vector (SCV) distribution, and non-circularity of the complex-valued fMRI data. The multivariate generalized Gaussian distribution (MGGD) is exploited to match the SCV distribution based on nonlinearity, the shape parameter of MGGD is estimated using maximum likelihood estimation, and the nonlinearity is updated in the dominant SCV subspace to achieve denoising goal. In addition, the pseudo-covariance matrix is incorporated into the algorithm to represent the non-circularity. Experimental results from simulated and actual fMRI data demonstrate significant improvements of our algorithm over a complex-valued IVA-G algorithm and several circular and noncircular fixed-point IVA variants.

An extension of group independent component analysis (GICA) is introduced, where multi-set canonical correlation analysis (MCCA) is combined with principal component analysis (PCA) for three-stage dimension reduction. The method is applied on naturalistic functional MRI (fMRI) images acquired during task-free continuous music listening experiment, and the results are compared with the outcome of the conventional GICA. The extended GICA resulted slightly faster ICA convergence and, more interestingly, extracted more stimulus-related components than its conventional counterpart. Therefore, we think the extension is beneficial enhancement for GICA, especially when applied to challenging fMRI data.

Computer vision methods can benefit wood processing industry. We propose a method to detect wood surface quality and classify wood samples into sound and defective classes. Gray level histogram statistical features and gray level co-occurrence matrix (GLCM) texture features are extracted from wood surface images and combined for classification. A half circle template is proposed to generate GLCM, avoiding calculating distances at each pixel every time and speeding up the algorithm greatly. The proposed approach uses more pixel information than traditional four-angle method, resulting in a significantly higher classification accuracy. Moreover the running time demonstrates our algorithm is efficient and suitable for real-time applications.

Non-negative tensor factorization (NTF) has been successfully applied to analyze event-related potentials (ERPs), and shown superiority in terms of capturing multi-domain features. However, the time-frequency representation of ERPs by higher-order tensors are usually large-scale, which prevents the popularity of most tensor factorization algorithms. To overcome this issue, we introduce a non-negative canonical polyadic decomposition (NCPD) based on low-rank approximation (LRA) and hierarchical alternating least square (HALS) techniques. We applied NCPD (LRAHALS and benchmark HALS) and CPD to extract multi-domain features of a visual ERP. The features and components extracted by LRAHALS NCPD and HALS NCPD were very similar, but LRAHALS NCPD was 70 times faster than HALS NCPD. Moreover, the desired multi-domain feature of the ERP by NCPD showed a significant group difference (control versus depressed participants) and a difference in emotion processing (fearful versus happy faces). This was more satisfactory than that by CPD, which revealed only a group difference.

This study presents a method to analyze blood-oxygen-level-dependent (BOLD) functional magnetic resonance imaging (fMRI) signals associated with listening to continuous music. Semi-blind independent component analysis (ICA) was applied to decompose the fMRI data to source level activation maps and their respective temporal courses. The unmixing matrix in the source separation process of ICA was constrained by a variety of acoustic features derived from the piece of music used as the stimulus in the experiment. This allowed more stable estimation and extraction of more activation maps of interest compared to conventional ICA methods.

Non-negative Canonical Polyadic decomposition (NCPD) and non-negative Tucker decomposition (NTD) were compared for extracting the multi-domain feature of visual mismatch negativity (vMMN), a small event-related potential (ERP), for the cognitive research. Since signal-to-noise ratio in vMMN is low, NTD outperformed NCPD. Moreover, we proposed an approach to select the multi-domain feature of an ERP among all extracted features and discussed determination of numbers of extracted components in NCPD and NTD regarding the ERP context.

This study proposes a novel approach for the analysis of brain responses in the modality of ongoing EEG elicited by the naturalistic and continuous music stimulus. The 512-second long EEG data (recorded with 64 electrodes) are first decomposed into 64 components by independent component analysis (ICA) for each participant. Then, the spatial maps showing dipolar brain activity are selected in terms of the residual dipole variance through a single dipole model in brain imaging, and clustered into a pre-defined number (estimated by the minimum description length) of clusters. Subsequently, the temporal courses of the EEG theta and alpha oscillations of each component for each cluster are produced and correlated with the temporal courses of tonal and rhythmic features of the music. Using this approach, we found that the extracted temporal courses of the theta and alpha oscillations along central and occipital area of scalp in two of the selected clusters significantly correlated with the musical features representing progressions in the rhythmic content of the stimulus. We suggest that this demonstrates that with the proposed approach, we have managed to discover what kinds of brain responses were elicited when a participant was listening continuously to the long piece of naturalistic music.

Through exploiting temporal, spectral, time-frequency representations, and spatial properties of mismatch negativity (MMN) simultaneously, this study extracts a multi-domain feature of MMN mainly using non-negative tensor factorization. In our experiment, the peak amplitude of MMN between children with reading disability and children with attention deficit was not significantly different, whereas the new feature of MMN significantly discriminated the two groups of children. This is because the feature was derived from multi-domain information with significant reduction of the heterogeneous effect of datasets.

Nonnegative Matrix/Tensor factorization (NMF/NTF) have been used in the study of EEG, and the fit (explained variation) is often used to evaluate the performance of a nonnegative decomposition algorithm. However, this parameter only reveals the information derived from the mathematical model and just exhibits the reliability of the algorithms, and the property of EEG can not be reflected. If fits of two algorithms are identical, it is necessary to examine whether the desired components extracted by them are identical too. In order to verify this doubt, we performed NMF and NTF on the same dataset of an auditory event-related potentials (ERPs), and found that the identical fits of NMF and NTF under the hierarchical alternating least squares algorithms corresponded to different desired ERPs extracted by NMF and NTF, moreover, NTF contributed the ERP with much better timing and spectral properties. Such analysis implies that to combine the fit and property of the desired ERP component together helps evaluate the performance of NMF and NTF algorithms in the study of ERPs.

Time-frequency analysis is critical to study event-related potentials (ERPs) now. ERPs are usually generated through averaging over a number of trials, and such averaging limits the application of a nonlinear time-frequency analysis method-Hilbert-Huang transformation (HHT). This is because HHT usually requires very long recordings to sufficiently decompose the complicated signal into oscillations and the averaged ERP trace tends to possess only hundreds of samples. Thus, this study designs the concatenated trial based HHT to release the limitation on the decomposition. Such a paradigm may reveal better temporal and spectral properties of an ERP than the conventional wavelet transformation does. Moreover, under the proposed method, it is found that the children with attention deficit hyperactivity disorders may have more temporally, spectrally and spatially distributed brain activities than the control children do.

Independent component analysis (ICA) does not follow the superposition rule. This motivates us to study a negative event-related potential - mismatch negativity (MMN) estimated by the single-trial based ICA (sICA) and averaged trace based ICA (aICA), respectively. To sICA, an optimal digital filter (ODF) was used to remove low-frequency noise. As a result, this study demonstrates that the performance of the sICA+ODF and aICA could be different. Moreover, MMN under sICA+ODF fits better with the theoretical expectation, i.e., larger deviant elicits larger MMN peak amplitude.

In this presentation two event-related potentials, mismatch negativity (MMN) and P3a, are extracted from EEG by non-negative matrix factorization (NMF) simultaneously. Typically MMN recordings show a mixture of MMN, P3a, and responses to repeated standard stimuli. NMF may release the source independence assumption and data length limitations required by fast independent component analysis (FastICA). Thus, in theory NMF could reach better separation of the responses. In the current experiment MMN was elicited by auditory duration deviations in 102 children. NMF was performed on the time-frequency representation of the raw data to estimate sources. Support to absence ratio (SAR) of the MMN component was utilized to evaluate the performance of NMF and FastICA. To the raw data, FastICA-MMN component, and NMF-MMN component, SARs were 31, 34 and 49 dB respectively. NMF outperformed FastICA by 15 dB. This study also demonstrates that children with reading disability have larger P3a than control children under NMF.