Search

Complex-valued neural networks process both amplitude and phase information, in contrast to conventional artificial neural networks, achieving additive capabilities in recognizing phase-sensitive data inherent in wave-related phenomena. The ever-increasing data capacity and network scale place substantial demands on underlying computing hardware. In parallel with the successes and extensive efforts made in electronics, optical neuromorphic hardware is promising to achieve ultra-high computing performances due to its inherent analog architecture and wide bandwidth. Here, we report a complex-valued optical convolution accelerator operating at over 2 Tera operations per second (TOPS). With appropriately designed phasors we demonstrate its performance in the recognition of synthetic aperture radar (SAR) images captured by the Sentinel-1 satellite, which are inherently complex-valued and more intricate than what optical neural networks have previously processed. Experimental tests with 500 images yield an 83.8% accuracy, close to in-silico results. This approach facilitates feature extraction of phase-sensitive information, and represents a pivotal advance in artificial intelligence towards real-time, high-dimensional data analysis of complex and dynamic environments., Competing Interests: Competing interests: The authors declare no competing interests., (© 2024. The Author(s).)

Background: Inpatients with high risk of venous thromboembolism (VTE) usually face serious threats to their health and economic conditions. Many studies using machine learning (ML) models to predict VTE risk overlook the impact of class-imbalance problem due to the low incidence rate of VTE, resulting in inferior and unstable model performance, which hinders their ability to replace the Padua model, a widely used linear weighted model in clinic. Our study aims to develop a new VTE risk assessment model suitable for Chinese medical inpatients., Methods: 3284 inpatients in the medical department of Peking Union Medical College Hospital (PUMCH) from January 2014 to June 2016 were collected. The training and test set were divided based on the admission time and inpatients from May 2016 to June 2016 were included as the test dataset. We explained the class imbalance problem from a clinical perspective and defined a new term, "fuzzy population", to elaborate and model this phenomenon. By considering the "fuzzy population", a new ML VTE risk assessment model was built through population splitting. Sensitivity and specificity of our method was compared with five ML models (support vector machine (SVM), random forest (RF), gradient boosting decision tree (GBDT), logistic regression (LR), and XGBoost) and the Padua model., Results: The 'fuzzy population' phenomenon was explained and verified on the VTE dataset. The proposed model achieved higher specificity (64.94% vs. 63.30%) and the same sensitivity (90.24% vs. 90.24%) on test data than the Padua model. Other five ML models couldn't simultaneously surpass the Padua's sensitivity and specificity. Besides, our model was more robust than five ML models and its standard deviations of sensitivities and specificities were smaller. Adjusting the distribution of negative samples in the training set based on the 'fuzzy population' would exacerbate the instability of performance of five ML models, which limited the application of ML methods in clinic., Conclusions: The proposed model achieved higher sensitivity and specificity than the Padua model, and better robustness than traditional ML models. This study built a population-split-based ML model of VTE by modeling the class-imbalance problem and it can be applied more broadly in risk assessment of other diseases., Competing Interests: Declarations. Ethics approval and consent to participate: This study was approved by the Ethics Committee of PUMCH in Chinese Academy of Medical Sciences (reference number for ethics approval: B164). All methods were carried out according to the Declaration of Helsinki. All participants gave informed consent. Consent for publication: Not applicable. Competing interests: The authors declare no competing interests., (© 2024. The Author(s).)

Rapid and accurate determination of target proteins in cells provide essential diagnostic information for early detection of diseases, evaluation of drug responses, and the study of pathophysiological mechanisms. Traditional Western blotting method has been used for the determination, but it is complex, time-consuming, and semi-quantitative. Here, a tapered seven-core fiber (TSCF) biosensor was designed and fabricated. By immobilizing protein kinase B (PKB), also known as AKT, antibody onto TSCF surface, the microfiber biosensor can be used for quantitatively detecting the AKT level in solution concentrations as low as 0.26 ng/mL. To test the reliability of the TSCF sensing method in a medical application, the TSCF biosensor was used to study the relationship between chrysin's anticancer effect and the concentration of AKT in a human colorectal cancer cell line (LoVo cells). The results reveal that the inhibitory effect of chrysin on LoVo cells is positively correlated with the dose, agreeing well with the equivalent results using the traditional Western blotting method., Competing Interests: Competing interests: The authors declare no competing interests., (© 2024. Crown.)

Understanding the heterogeneity of Alzheimer's disease (AD) is crucial for advancing precision medicine specifically tailored to this disorder. Recent research has deepened our understanding of AD heterogeneity, yet translating these insights from bench to bedside via neuroimaging heterogeneity frameworks presents significant challenges. In this review, we systematically revisit prior studies and summarize the existing methodology of data-driven neuroimaging studies for AD heterogeneity. We organized the present methodology into (i) a subtyping cluster strategy for AD patients, and we also subdivided it into subtyping analysis based on cross-sectional multimodal neuroimaging profiles, and the identification of long-term disease progression from short-term datasets; (ii) a stratified strategy that integrates neuroimaging measures with biomarkers; (iii) individual-specific abnormal patterns based on the Normative model. We then evaluated the characteristics of these studies along two dimensions: (i) the understanding of pathology and (ii) clinical application. We systematically address the limitations, challenges, and future directions of research into AD heterogeneity. Our goal is to enhance the neuroimaging heterogeneity framework for AD, facilitating its transition from bench to bedside., (Copyright © 2024. Published by Elsevier Inc.)

Neuromodulation plays a critical role in the normal physiological functions of organisms. With advancements in science and technology, neuromodulation has expanded into various fields. For instance, in the field of engineering, in vitro-cultured neural networks are utilized to perform closed-loop control for achieving complex functionalities. Conducting pioneering theoretical research using mathematical models is particularly essential for enhancing efficiency and reducing costs. This study focuses on examining the relationship between input and output in order to establish a groundwork for more advanced closed-loop regulation applications in engineering. Using a constructed neural network model, Poisson, square wave and direct current (DC) stimulation are applied. The results show that the network's firing rate increases with the frequency or amplitude of these stimulations. And the network's firing rate could reach to a stable state after the stimulation is applied for 0.8s and return to initial states when the stimulus is removed for 1s. To ascertain if the system exhibits a memory effect from the previous stimulus, we conduct independent and continuous stimulation schemes. Comparing the firing rate of neuronal networks under these two stimulation schemes reveals a memory effect of the system on the previous stimulus, which is independent of network properties and stimulus types. Finally, by applying square wave stimulation to the in vitro cultured neural network, we have confirmed that cultured neural network actually can reach to a steady state and have memory effects on the previous stimulus. Our research results have important theoretical significance and reference value for designing the closed-loop regulation strategy of in vitro cultured neuronal networks., Competing Interests: The authors have declared that no competing interests exist., (Copyright: © 2024 Xue et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.)

Automated clinical coding (ACC) has emerged as a promising alternative to manual coding. This study proposes a novel human-in-the-loop (HITL) framework, CliniCoCo. Using deep learning capacities, CliniCoCo focuses on how such ACC systems and human coders can work effectively and efficiently together in real-world settings. Specifically, it implements a series of collaborative strategies at annotation, training and user interaction stages. Extensive experiments are conducted using real-world EMR datasets from Chinese hospitals. With automatically optimised annotation workloads, the model can achieve F1 scores around 0.80-0.84. For an EMR with 30% mistaken codes, CliniCoCo can suggest halving the annotations from 3000 admissions with an ignorable 0.01 F1 decrease. In human evaluations, compared to manual coding, CliniCoCo reduces coding time by 40% on average and significantly improves the correction rates on EMR mistakes (e.g., three times better on missing codes). Senior professional coders' performances can be boosted to more than 0.93 F1 score from 0.72., Competing Interests: Competing interests: The authors declare no competing interests., (© 2024. The Author(s).)

This paper investigates the safety control problem of a bicycle robot with front-wheel drive and without a trail or mechanical regulator during circular motion. Constraints on the drive angular speed necessary for the bicycle to achieve circular motion are proposed. In practical robot systems, bounded input disturbances are inevitable. To address this, we propose a safe controller that integrates the control Lyapunov function (CLF) constraints for input-to-state stability (ISS) and the control barrier function (CBF) constraints for input-to-state safety (ISSf), implemented using quadratic programming (QP). Our controller achieves enhanced safety in control while reducing control effort. The effectiveness of this controller is verified through simulation comparative experiments. Furthermore, circular motion is achieved through physical experiments with a real robot, and the effectiveness of the ISS-CLF-ISSf-CBF-QP controller is validated through comparative experiments., Competing Interests: Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper., (Copyright © 2024. Published by Elsevier Ltd.)

This study proposes a rapid method for determining pregnancy status based on fingertip pulse signals. A finger pulse sensor collects data, which is processed into unified multimodal signals. The Bamboo-Net model, combining ResNet, LSTM, and 1D-CNN, extracts key features from time, frequency, and time-frequency domains. Tested on 346 training and 138 testing samples, the model achieves 91% accuracy with 6 s input, outperforming mainstream methods. Recognition rates for mid and late pregnancy are higher than for early pregnancy, highlighting its potential for practical applications.

Hemodynamics significantly influences the clinical outcomes of brain arteriovenous malformations (AVM). This study aimed to determine if the lesion filling index (LFI), obtained via quantitative digital subtraction angiography (QDSA), can predict complete complete obliteration after Gamma knife radiosurgery (GKRS). We retrospectively reviewed AVM patients who underwent GKRS and DSA exams from 2011 to 2021. Clinical, angioarchitectural, and QDSA hemodynamic features were analyzed. The LFI, derived from QDSA, was evaluated as a predictor of complete complete obliteration post-SRS using Cox proportional hazards and Kaplan-Meier analyses. Among 118 AVMs with a mean follow-up of 5.76 ± 2.76 years, post-SRS complete obliteration was linked to reduced nidus volume (7.27 ± 12.3 vs. 19.2 ± 35.7 mm³, p = 0.049), smaller nidus diameter (26.0 ± 14.9 vs. 34.1 ± 19.8 mm, p = 0.015), and absence of feeding artery dilation (21.1% vs. 46.3%, p = 0.008). Higher Arterial Diagnostic Window (ADW) (972.27 ± 1615.53 vs. 515.29 ± 730.26, p = 0.036), higher LFI (905.31 ± 2288.37 vs. 249.65 ± 1092.46, p = 0.037), and lower Transnidal Relative Velocity (TRV), a parameter defined as the maximum diameter of AVM divided by the full width at half maximum (74.31 ± 95.67 vs. 137.80 ± 152.01, p = 0.021), were also associated with complete obliteration. After adjusting for confounders, only two variables-absence of feeding artery dilation (HR 0.35, 95%CI 0.16-0.78, p = 0.010) and higher LFI (HR 1.00, 95%CI 1.00-1.00, p = 0.006) remained significant predictors. The ROC curve identified 188.4 as the LFI cutoff, and Kaplan-Meier analysis confirmed LFI's predictive value (log-rank test, χ² = 12.776, p < 0.001). Elevated LFI and absence of feeding artery dilation predict AVM complete obliteration after GKRS, indicating that overfilling and low blood flow in the nidus may promote complete obliteration., Competing Interests: Declarations. Ethical approval: This retrospective study was approved by Beijing Tiantan Hospital's Institutional Review Board (KY 2020-003-01), adhering to Helsinki Declaration guidelines and STROBE reporting for obser-vational case-control studies. Consent to participate: Written informed consent was obtained from the parents. Competing interests: The authors declare no competing interests., (© 2024. The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature.)

Background and Aim: Pyogenic liver abscess (PLA) is a devastating and potentially life-threatening disease globally, with Klebsiella pneumoniae liver abscess (KPLA) being the most prevalent in Asia. This study aims to develop an effective and comprehensive nomogram combining clinical and radiomics features for early prediction of KPLA., Methods: 255 patients with PLA from 2013 to 2023 were enrolled and randomly divided into the training and validation cohorts at a 7:3 ratio. The differences between the two cohorts of patients were assessed via univariate analysis. The radiomics features were extracted from imaging data from enhanced CT of liver abscesses. The optimal radiomics features were filtered using the independent sample t-test and least absolute shrinkage and selection operator, and a radiomics score (Rad-score) was calculated by weighting their respective coefficients. Clinically independent predictors were identified from the clinical data and combined with the Rad-score to develop a nomogram by multivariate logistic regression. The predictive performance was evaluated using the area under the receiver operating characteristic curve (AUC), calibration curve, and clinical decision curve., Results: The nomogram incorporated four clinical features of diabetes mellitus, cryptogenic liver abscess, C-reactive protein level, and splenomegaly, and the Rad-score that was constructed based on seven optimal radiomics features. It had an AUC of 0.929 (95% CI, 0.894-0.964) and 0.923 (95% CI, 0.864-0.981) in the training and validation cohorts, respectively. The calibration and decision curves showed that the nomogram had good agreement and clinical applicability., Conclusions: The clinical-radiomics nomogram performed well in predicting KPLA, hopefully serving as a reference for early diagnosis of KPLA.

Background: Growing evidence of neuroimaging has indicated brain microstructural abnormalities in comitant strabismus. Nonetheless, few studies have investigated neuropathological alterations in patients with intermittent exotropia (IXT). This study aimed at examining the characteristics of brain microstructure along major fiber tracts in IXT patients using an automated fiber quantification analysis., Methods: A total of 25 patients with IXT as well as 25 healthy participants matched for age and gender finished the diffusion tensor imaging scanning and the ophthalmic examination. Automated fiber quantification analysis of 20 major fiber tracts was carried out for IXT patients and healthy subjects, respectively. Diffusion metrics of 100 equidistant nodes resampled along each tract were measured for every subject and compared between two groups. Effect size analysis was performed to identify the most affected fiber tracts in IXT., Results: Widely declined mean diffusivity was noted in IXT along major tracts containing bilateral thalamic radiations, bilateral corticospinal fasciculi, bilateral cingulum cingulate, left inferior fronto-occipital fasciculus, right arcuate fasciculus and superior longitudinal fasciculus. Local reduction in fractional anisotropy was observed in IXT along left cingulum hippocampus, right inferior longitudinal fasciculus and right uncinate fasciculus, in contrast to the regionally increased fractional anisotropy along bilateral thalamic radiation, left corticospinal tract and left arcuate fasciculus. Among the tracts with significantly changed diffusion metrics in IXT, right inferior longitudinal fasciculus was the most affected one in fractional anisotropy while left thalamic radiation was the most influenced one in mean diffusivity., Conclusions: Abnormalities in microstructural properties along visual-related fiber tracts are likely to contribute to difficulties in visual information processing in IXT patients, which could serve as the neural basis of underlying pathological mechanism of IXT., Competing Interests: Declaration of Competing Interest None declared., (Copyright © 2024 The Authors. Published by Elsevier Inc. All rights reserved.)

Bottlenecks of the freeway generated especially by traffic accidents or temporary work zones contribute to significant reductions in system throughput and hinder the efficient traffic operations. It is imperative to take proactive measures to improve traffic state. With the rapid advancements in intelligent transportation, connected and autonomous vehicles (CAVs) have attracted much attention by its speculated capabilities in improving traffic safety and well-organized operational coordination. Therefore, reasonably utilizing the advantages of CAVs is possible to reduce the impact induced by bottlenecks. In this research, we propose a novel algorithm called CAV-Lead to obtain the CAV's regulated speed under mixed CAVs and human-driven vehicles (HVs) environment to improve the overall utilization of the freeway capacity near bottlenecks. Firstly, we illustrate the basic principle of the CAV-Lead algorithm that takes both microscopic and macroscopic traffic characteristics into account. Then, based on the local spatiotemporal traffic state, the CAV-Lead algorithm is proposed to determine each CAV's speed under mixed flow. Furthermore, a real-time simulation control framework considering the random behavior of HVs is presented. Moreover, several simulation evaluations including comparisons with basic scenarios and similar research are conducted under various CAV market penetration rates (MPRs). The results demonstrate that the CAV-Lead could improve the traffic performance, especially for the high traffic demand with certain MPRs., Competing Interests: Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper., (Copyright © 2024 Elsevier Ltd. All rights reserved.)

Objectives: This study aims to develop a multimodal deep learning-based algorithm for detecting specific fetal heart rate (FHR) events, to enhance automatic monitoring and intelligent assessment of fetal well-being., Methods: We analyzed FHR and uterine contraction signals by combining various feature extraction techniques, including morphological features, heart rate variability features, and nonlinear domain features, with deep learning algorithms. This approach enabled us to classify four specific FHR events (bradycardia, tachycardia, acceleration, and deceleration) as well as four distinct deceleration patterns (early, late, variable, and prolonged deceleration). We proposed a multi-model deep neural network and a pre-fusion deep learning model to accurately classify the multimodal parameters derived from Cardiotocography signals., Results: These accuracy metrics were calculated based on expert-labeled data. The algorithm achieved a classification accuracy of 96.2 % for acceleration, 94.4 % for deceleration, 90.9 % for tachycardia, and 85.8 % for bradycardia. Additionally, it achieved 67.0 % accuracy in classifying the four distinct deceleration patterns, with 80.9 % accuracy for late deceleration and 98.9 % for prolonged deceleration., Conclusions: The proposed multimodal deep learning algorithm serves as a reliable decision support tool for clinicians, significantly improving the detection and assessment of specific FHR events, which are crucial for fetal health monitoring., (© 2024 Walter de Gruyter GmbH, Berlin/Boston.)

As new quality productivity (NQP) emerges as a rising star of productivity that can effectively leverage technological innovation and sustainability, this study aims to explore the relationship between NQP and environmental innovation, with a particular focus on the roles of managerial empowerment and board centralization within the context of China A-share listed companies. Utilizing the entire sample of China A-share market from 2013 to 2022, the study analyses the effectiveness of various dimensions reflecting innovation engagement among Chinese listed companies. For measuring NQP, the entropy method is employed to calculate the weights. By controlling for industry and year effects, the study examines both the main and moderating effects of managerial empowerment and board centralization. Additionally, heterogeneity tests, robustness checks, and two-stage least squares (2SLS) estimation were conducted to address endogeneity concerns. The results demonstrate that NQP significantly enhances environmental innovation, with managerial empowerment supporting this positive effect and board centralization obstructing it. The positive effect of NQP is particularly evident in state-owned enterprises, while in heavily polluting industries, the anticipated positive moderating effect of top managers disappears due to strict regulatory environments. Furthermore, board centralization negatively moderates environmental innovation, especially in lightly polluting sectors where internal governance is more sensitive. The study underscores the importance for policymakers to tailor regulations that balance managerial empowerment and board centralization to enhance the transformation of NQP into environmental innovation. Future research is needed to further investigate different background settings and the mechanisms through which NQP influences sustainability., Competing Interests: Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper., (Copyright © 2024 Elsevier Ltd. All rights reserved.)

In the last decade, extensive research has emerged into understanding the impact of risk factors for Alzheimer's Disease (AD) on brain in pre-symptomatic stages. We investigated the neuroimaging correlates of the APOEe4 genetic risk factor for AD in young adulthood, its relationship with cognition, and potential effects of other variables on the findings. While conventional volumetric analyses revealed no consistent differences, more sophisticated analyses identified subtle structural differences between APOEe4 carriers and non-carriers. Findings from diffusion studies were limited, but functional studies demonstrated consistent alterations in connectivity and activity. The complex relationship between APOE genotype, neuroimaging variables, and cognition revealed no consensus on the directionality of findings. Methodological choices, including analytical approaches, sample size, and the influence of other genes, gender, and ethnicity, varied across studies, impacting comparability and generalizability. Recommendations for future research include multimodal and longitudinal imaging, standardisation of pipelines, advanced analytical techniques, and collaborative data pooling., Competing Interests: Declaration of Competing Interest LK, XX, PR, OT, EJ, JM, WH, MD have no competing interests or financial disclosures related to present article. LS was funded by Alzheimer’s Research UK Senior Research Fellowship (ARUK-SRF2017B-1) and the Lewy Body Society (LS002/2019). This research was supported by the NIHR Sheffield Biomedical Research Centre (BRC) / NIHR Sheffield Clinical Research Facility (CRF). The views expressed are those of the authors and not necessarily those of the NHS, the NIHR or the Department of Health and Social Care (DHSC)., (Copyright © 2024 The Authors. Published by Elsevier B.V. All rights reserved.)

Objective: This study endeavors to clarify the impact of venous aneurysms (VA) on hemorrhagic risk in brain arteriovenous malformations (AVMs) and uncover potential hemodynamic mechanisms, utilizing quantitative digital subtraction angiography (QDSA) technology and survival dataset., Methods: Patients were enrolled in a multicenter prospective collaboration registry between August 2011 and August 2021, and subsequently categorized into the VA and non-VA cohorts. Using propensity score-matched survival analysis, we quantitatively assessed the natural risk of hemorrhagic stroke in these two cohorts. Additionally, a quantitative hemodynamic analysis was conducted to explore the distinctions in hemodynamic characteristics between these two cohorts., Results: Among 3758 consecutive AVMs documented at a single center from the registry, 820 unruptured AVMs who maintained conservation management over 1 month were identified. Following a two-step matching process, 504 cases were retained for survival analysis and 408 cases for hemodynamic analysis. Overall, the presence of VA emerged as a protective factor, associated with a decreased risk of hemorrhagic stroke (HR, 0.21 [95% CI: 0.07-0.62], p = 0.004). Distinct hemodynamic characteristics were observed in AVMs with VA, showing a lower stasis index in two components of AVMs-the nidus (p = 0.014) and the main draining vein (p = 0.018)., Conclusion: In this observational prospective cohort study, the presence of VA is associated with a decreased risk of hemorrhagic stroke in AVMs, suggesting an underlying hemodynamic mechanism involving the redistribution of excessive pressure loads within the AVM nidus by the VA., Key Points: Questions What impact, if any, does VA have on the hemorrhagic risk in brain AVMs? Findings Presence of VA is associated with a decreased hemorrhagic stroke risk through the redistribution of pressure loads. Critical relevance VA in brain AVMs emerges as a protective factor against hemorrhagic stroke. Understanding this association and the underlying hemodynamic mechanisms offers valuable guidance for preventive strategies and informs clinical decision-making, improving overall patient care., (© 2024. The Author(s), under exclusive licence to European Society of Radiology.)

Hypertension is the most prevalent cardiovascular disease, affecting one-third of adults. All antihypertensive drugs have potential side effects. Gut metabolites influence hypertension. The objective of this study was to identify antihypertensive gut metabolites through network pharmacology and molecular docking techniques and to validate their antihypertensive mechanisms via in vitro experiments. A total of 10 core antihypertensive targets and 18 gut metabolites that act on hypertension were identified. Four groups of protein metabolites, namely, CXCL8-baicalein, CXCL8-baicalin, CYP1A1-urolithin A, and PTGS2-equol, which have binding energies of - 7.7, - 8.5, - 7.2, and - 8.8 kcal-mol -1 , respectively, were found to have relatively high affinities. Based on its drug-likeness properties in silico and toxicological properties, equol was identified as a potential antihypertensive metabolite. On the basis of the results of network pharmacology and molecular docking, equol may exert antihypertensive effects by regulating the IL-17 signaling pathway and PTGS2. A phenylephrine-induced H9c2 cell model was subsequently utilized to verify that equol inhibits cell hypertrophy (P < 0.05) by inhibiting the IL-17 signaling pathway and PTGS2 (P < 0.05). This study demonstrated that equol has the potential to be developed as a novel therapeutic agent for the treatment of hypertension., (© 2024. The Author(s).)

Despite being a relatively new cultural phenomenon, the ability to perform letter-sound integration is readily acquired even though it has not had time to evolve in the brain. Leading theories of how the brain accommodates literacy acquisition include the neural recycling hypothesis and the assimilation-accommodation hypothesis. The neural recycling hypothesis proposes that a new cultural skill is developed by "invading" preexisting neural structures to support a similar cognitive function, while the assimilation-accommodation hypothesis holds that a new cognitive skill relies on direct invocation of preexisting systems (assimilation) and adds brain areas based on task requirements (accommodation). Both theories agree that letter-sound integration may be achieved by reusing pre-existing functionally similar neural bases, but differ in their proposals of how this occurs. We examined the evidence for each hypothesis by systematically comparing the similarities and differences between letter-sound integration and two other types of preexisting and functionally similar audiovisual (AV) processes, namely object-sound and speech-sound integration, by performing an activation likelihood estimation (ALE) meta-analysis. All three types of AV integration recruited the left posterior superior temporal gyrus (STG), while speech-sound integration additionally activated the bilateral middle STG and letter-sound integration directly invoked the AV areas involved in speech-sound integration. These findings suggest that letter-sound integration may reuse the STG for speech-sound and object-sound integration through an assimilation-accommodation mechanism., (© 2024 The Author(s). Human Brain Mapping published by Wiley Periodicals LLC.)

Macroscale functional gradient techniques provide a continuous coordinate system that extends from unimodal regions to transmodal higher-order networks. However, the alterations of these functional gradients in AD and their correlations with cognitive terms and gene expression profiles remain to be established. In the present study, we directly studied the functional gradients with functional MRI data from seven scanners. We adopted data-driven meta-analytic techniques to unveil AD-associated changes in the functional gradients. The principal primary-to-transmodal gradient was suppressed in AD. Compared to NCs, AD patients exhibited global connectome gradient alterations, including reduced gradient range and gradient variation, increased gradient scores in the somatomotor, ventral attention, and frontoparietal regions, and decreased in the default mode network. More importantly, the Gene Ontology terms of biological processes were significantly enriched in the potassium ion transport and protein-containing complex remodeling. Our compelling evidence provides a new perspective in understanding the connectome alterations in AD., (© 2024 The Author(s). Human Brain Mapping published by Wiley Periodicals LLC.)

The intricate spatial configurations of brain networks offer essential insights into understanding the specific patterns of brain abnormalities and the underlying biological mechanisms associated with Alzheimer's disease (AD), normal aging, and other neurodegenerative disorders. This study investigated alterations in the topographical structure of the brain related to aging and neurodegenerative diseases by analyzing brain gradients derived from structural MRI data across multiple cohorts ( n = 7323). The analysis identified distinct gradient patterns in AD, aging, and other neurodegenerative conditions. Gene enrichment analysis indicated that inorganic ion transmembrane transport was the most significant term in normal aging, while chemical synaptic transmission is a common enrichment term across various neurodegenerative diseases. Moreover, the findings show that each disorder exhibits unique dysfunctional neurophysiological characteristics. These insights are pivotal for elucidating the distinct biological mechanisms underlying AD, thereby enhancing our understanding of its unique clinical phenotypes in contrast to normal aging and other neurodegenerative disorders.

Parity-time (PT) symmetry is an active research area that provides a variety of new opportunities for different systems with novel functionalities. For instance, PT symmetry has been used in lasers and optoelectronic oscillators to achieve single-frequency lasing or oscillation. A single-frequency system is essentially a static PT-symmetric system, whose frequency is time-invariant. Here we investigate time-variant PT symmetry in frequency-scanning systems. Time-variant PT symmetry equations and eigenfrequencies for frequency-scanning systems are developed. We show that time-variant PT symmetry can dynamically narrow the instantaneous linewidth of frequency-scanning systems. The instantaneous linewidth of a produced frequency-modulated continuous-wave (FMCW) waveform is narrowed by a factor of 14 in the experiment. De-chirping and radar imaging results also show that the time-variant PT-symmetric system outperforms a conventional frequency-scanning one. Our study paves the way for a new class of time-variant PT-symmetric systems and shows great promise for applications including FMCW radar and lidar systems., (© 2024. The Author(s).)

A multiscale regulation strategy has been demonstrated for synthetic energy storage enhancement in a tetragonal tungsten bronze structure ferroelectric. Grain refining and second-phase precipitation (perovskite phase) are introduced in the BaSrTiNb 2-x Ta x O 9 ceramics by regulating the composition and sintering process. Disordered polarization and distribution, chemical inhomogeneity, and insulating boundary layers are achieved to provide the fundamental structural origin of the relaxation characteristic, high breakdown strength, and superior energy storage performance. Thus, an ultrahigh energy storage density of 12.2 J cm -3 with an low energy consumption was achieved at an electric field of 950 kV cm -1 . This is the highest known energy storage performance in tetragonal tungsten bronze-based ferroelectric. Notably, this ceramic shows remarkable stability over frequency, temperature, and cycling electric fields. This work brings new material candidates and structure design for developing of energy storage capacitors apart from the predominant perovskite ferroelectric ceramics., (© 2024. The Author(s).)

Task functional magnetic resonance imaging research has generally shielded away from studying individuals due to the low reproducibility. Here, we propose that heterogeneous brain activations across individuals localize to a common network. To test this hypothesis, we use working memory (WM) as our example. First, we showed that discrete-brain-based reproducibility of brain activation during WM across individuals was low. Then, we used activation network mapping (ANM) technique to identify each individual's brain network of WM and found that network-based reproducibility was rather high. Prediction analyses using machine learning algorithms indicated that individual WM networks identified via ANM can predict WM behavioral performance. This predictive ability even outperformed that of brain activations. Our study provides a new explanation on the low reproducibility of brain activations across individuals. The results suggest that ANM can be used to identify individual brain networks of cognitive processes, thus promising broad potential applications., (© 2024. The Author(s).)

The automated segmentation of Intracranial Arteries (IA) in Digital Subtraction Angiography (DSA) plays a crucial role in the quantification of vascular morphology, significantly contributing to computer-assisted stroke research and clinical practice. Current research primarily focuses on the segmentation of single-frame DSA using proprietary datasets. However, these methods face challenges due to the inherent limitation of single-frame DSA, which only partially displays vascular contrast, thereby hindering accurate vascular structure representation. In this work, we introduce DIAS, a dataset specifically developed for IA segmentation in DSA sequences. We establish a comprehensive benchmark for evaluating DIAS, covering full, weak, and semi-supervised segmentation methods. Specifically, we propose the vessel sequence segmentation network, in which the sequence feature extraction module effectively captures spatiotemporal representations of intravascular contrast, achieving intracranial artery segmentation in 2D+Time DSA sequences. For weakly-supervised IA segmentation, we propose a novel scribble learning-based image segmentation framework, which, under the guidance of scribble labels, employs cross pseudo-supervision and consistency regularization to improve the performance of the segmentation network. Furthermore, we introduce the random patch-based self-training framework, aimed at alleviating the performance constraints encountered in IA segmentation due to the limited availability of annotated DSA data. Our extensive experiments on the DIAS dataset demonstrate the effectiveness of these methods as potential baselines for future research and clinical applications. The dataset and code are publicly available at https://doi.org/10.5281/zenodo.11401368 and https://github.com/lseventeen/DIAS., Competing Interests: Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper., (Copyright © 2024 Elsevier B.V. All rights reserved.)

Background: Extracellular free water (FW) resulting from white matter degeneration limits the sensitivity of diffusion tensor imaging (DTI) in predicting Alzheimer's disease (AD)., Purpose: To evaluate the sensitivity of FW-DTI in detecting white matter microstructural changes in AD. To validate the effectiveness of FW-DTI indices to predict amyloid-beta (Aβ) positivity in mild cognitive impairment (MCI) subtypes., Study Type: Retrospective., Population: Thirty-eight Aβ-negative cognitively healthy (CH) controls (68.74 ± 8.28 years old, 55% female), 15 Aβ-negative MCI patients (MCI-n) (68.87 ± 8.83 years old, 60% female), 29 Aβ-positive MCI patients (MCI-p) (73.03 ± 7.05 years old, 52% female), and 29 Aβ-positive AD patients (72.93 ± 9.11 years old, 55% female)., Field Strength/sequence: 3.0T; DTI, T 1 -weighted, T 2 -weighted, T 2 star-weighted angiography, and Aβ PET ( 18 F-florbetaben or 11 C-PIB)., Assessment: FW-corrected and standard diffusion indices were analyzed using trace-based spatial statistics. Area under the curve (AUC) in distinguishing MCI subtypes were compared using support vector machine (SVM)., Statistical Tests: Chi-squared test, one-way analysis of covariance, general linear regression analyses, nonparametric permutation tests, partial Pearson's correlation, receiver operating characteristic curve analysis, and linear SVM. A P value <0.05 was considered statistically significant., Results: Compared with CH/MCI-n/MCI-p, AD showed significant change in tissue compartment indices of FW-DTI. No difference was found in the FW index among pair-wise group comparisons (the minimum FWE-corrected P = 0.114). There was a significant association between FW-DTI indices and memory and visuospatial function. The SVM classifier with tissue radial diffusivity as an input feature had the best classification performance of MCI subtypes (AUC = 0.91), and the classifying accuracy of FW-DTI was all over 89.89%., Data Conclusion: FW-DTI indices prove to be potential biomarkers of AD. The classification of MCI subtypes based on SVM and FW-DTI indices has good accuracy and could help early diagnosis., Evidence Level: 4 TECHNICAL EFFICACY: Stage 2., (© 2023 International Society for Magnetic Resonance in Medicine.)

Open-source communities(OSCs) are gaining significant attention in the current business environment of information technology(IT). More and more IT companies and individuals are exploring how to achieve innovation through open-source collaboration, and value co-creation(VCC) in the OSCs has become a trend. Therefore, it is particularly important to examine the mechanism of OSCs under the background of VCC theory. This study proposes a conceptual framework of open-source value co-creation (OSVCC), which is characterized by openness, sharing, collaboration, and freedom, for understanding the internal mechanisms and contextual conditions in the relationship between OSCs participants. This study constructed a pairwise combined four-category classification model combining the perspectives of the commercialization level (low and high) and the maturity stage (developmental and mature) of the OSCs. Based on the model, this study selects and analyzes four presentive cases of OSCs using a multiple case study approach. Then, this study proposes a framework for OSVCC to identify the crucial factors that promote the successful implementation of innovation and value creation. The OSVCC framework encompasses three primary participants, effective VCC processes, and key open-source principles. This study offers valuable managerial implications for enterprises that plan to participate in OSCs., Competing Interests: The authors have declared that no competing interests exist., (Copyright: © 2024 Luo et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.)

The impact of CEO ability on corporate behavior is dual-faceted. Based on the "rent-seeking hypothesis", this paper focuses on the "dark side" of CEO ability and suggests that CEO ability may hurt corporate behavior. Using the DEA method to measure CEO ability and selecting Chinese A-share listed companies from 2007 to 2017 as a sample, we examine the relationship, mechanisms, and boundary conditions between CEO ability and stock price crash risk. Our study finds that CEO ability exacerbates stock price crash risk positively. Mechanism tests based on CEO tenure, agency conflict, and litigation risk reveal that CEO ability intensifies stock price crash risk for alleviating career concerns, seeking private gains, and risk avoidance, supporting the "rent-seeking hypothesis". Boundary condition tests based on internal and external governance environments show that internal control quality and external audit supervision can play a governance role in mitigating the exacerbating effect of CEO ability on stock price crash risk. The conclusions are of significant importance for companies, relevant departments, and global market participants to strengthen governance and supervision to prevent the negative impact of CEO ability on stock prices, thereby promoting the stable development of global financial markets., Competing Interests: The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper., (© 2024 The Authors.)

Ferroelectrics are essential in memory devices for multi-bit storage and high-density integration. Ferroelectricity mainly exists in compounds but rare in single-element materials due to their lack of spontaneous polarization in the latter. However, we report a room-temperature ferroelectricity in quasi-one-dimensional Te nanowires. Piezoelectric characteristics, ferroelectric loops and domain reversals are clearly observed. We attribute the ferroelectricity to the ion displacement created by the interlayer interaction between lone-pair electrons. Ferroelectric polarization can induce a strong field effect on the transport along the Te chain, giving rise to a self-gated ferroelectric field-effect transistor. By utilizing ferroelectric Te nanowire as channel, the device exhibits high mobility (~220 cm 2 ·V -1 ·s -1 ), continuous-variable resistive states can be observed with long-term retention (>10 5 s), fast speed (<20 ns) and high-density storage (>1.92 TB/cm 2 ). Our work provides opportunities for single-element ferroelectrics and advances practical applications such as ultrahigh-density data storage and computing-in-memory devices., (© 2024. The Author(s).)

Background: Alzheimer's disease (AD) is a neurodegenerative condition characterized by cognitive decline. To date, the specific dysfunction in the brain's hierarchical structure in AD remains unclear., Methods: We introduced the structural decoupling index (SDI), based on a multi-site data set comprising functional and diffusion-weighted magnetic resonance imaging data from 793 subjects, to assess their brain hierarchy., Results: Compared to normal controls (NCs), individuals with AD exhibited increased SDI within the posterior superior temporal sulcus, insular gyrus, precuneus, hippocampus, amygdala, postcentral gyrus, and cingulate gyrus; meanwhile, the patients with AD demonstrated decreased SDI in the frontal lobe. The SDI in those regions also showed a significant correlation with cognitive ability. Moreover, the SDI was a robust AD neuroimaging biomarker capable of accurately distinguishing diagnostic status (area under the curve [AUC] = 0.86)., Discussion: Our findings revealed the dysfunction of the brain's hierarchical structure in AD. Furthermore, the SDI could serve as a promising neuroimaging biomarker for AD., Highlights: This study utilized multi-center, multi-modal data from East Asian populations. We found an increased spatial gradient of the structure decoupling index (SDI) from sensory-motor to higher-order cognitive regions. Changes in SDI are associated with energy metabolism and mitochondria. SDI can identify Alzheimer's disease (AD) and further uncover the disease mechanisms of AD., (© 2024 The Author(s). Alzheimer's & Dementia published by Wiley Periodicals LLC on behalf of Alzheimer's Association.)

Background and Objective: The apnoea-hypopnoea index (AHI) and oxygen desaturation index (ODI) encounter challenges in capturing the intricate relationship between obstructive sleep apnoea (OSA) and cardiovascular disease (CVD) risks. Although novel hypoxic indices have been proposed to tackle these limitations, there remains a gap in comprehensive validation and comparisons across a unified dataset., Methods: Samples were derived from the Sleep Heart Health Study (SHHS), involving 4485 participants aged over 40 years after data quality screening. The study compared several key indices, including AHI, ODI, the reconstructed hypoxic burden (rHB), the percentage of sleep time with the duration of respiratory events causing desaturation (pRED&#95;3p) and the sleep breathing impairment index (SBII), in relation to CVD mortality and morbidity risks. Adjusted Cox proportional models were employed to calculate hazard ratios (HRs) for each index, and comparisons were performed., Results: SBII and pRED&#95;3p exhibited significant correlations with both CVD mortality and morbidity, with SBII showing the highest adjusted HR (95% confidence interval) for mortality (2.04 [1.25, 3.34]) and pRED&#95;3p for morbidity (1.43 [1.09-1.88]). In contrast, rHB was only significant in predicting CVD mortality (1.63 [1.05-2.53]), while AHI and ODI did not show significant correlations with CVD outcomes. The adjusted models based on SBII and pRED&#95;3p exhibited optimal performance in the CVD mortality and morbidity datasets, respectively., Conclusion: This study identified the optimal indices for OSA-related CVD risks prediction, SBII for mortality and pRED&#95;3p for morbidity. The open-source online platform provides the computation of the indices., (© 2024 Asian Pacific Society of Respirology.)

Competing Interests: The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Network platforms have ushered in a novel propagation model for feminist discourses. The emergence of oriental feminism in society has led to gender-based public opinions surrounding public events becoming a trending topic on Chinese social media. This study uses the 2022 Tangshan restaurant attack as a case study, an incident that sparked widespread discussions across China in 2022. The research gathered 366,602 network communication nodes within a week and examined the communication networks of three types of content nodes (information, opinion, and appeasement) using the complex network modeling method. The findings revealed that all three types of information communication networks exhibit an apparent scale-free characteristic, and the "key minority" of nodes significantly affects information communication. Information-type and appeasement-type Weibo display notable similarities in the quantity and degree distribution of nodes within the communication networks and in the information decay rate. Moreover, authoritative information issuers have become the primary catalyst for information propagation. Conversely, opinion-type Weibo has the widest communication network diameter and features a high degree of participation, multilevel propagation, and a slow decay rate. This indicates that the interaction between opinion leaders and netizens has enhanced the depth and breadth of information diffusion for opinion-type Weibo., Competing Interests: The authors have declared that no competing interests exist., (Copyright: © 2024 Fang et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.)

Major depressive disorder (MDD) is currently the most common psychiatric disorder in the world. It characterized by a high incidence of disease with the symptoms like depressed mood, slowed thinking, and reduced cognitive function. Without timely intervention, there is a 20-30% risk of conversion to treatment-resistant depression (TRD) and a high burden for the patient, family and society. Numerous studies have shown that physical activity (PA) is a non-pharmacological treatment that can significantly improve the mental status of patients with MDD and has positive effects on cognitive function, sleep status, and brain plasticity. However, the physiological and psychological effects of different types of PA on individuals vary, and the dosage profile of PA in improving symptoms in patients with MDD has not been elucidated. In most current studies of MDD, PA can be categorized as continuous endurance training (ECT), explosive interval training (EIT), resistance strength training (RST), and mind-body training (MBT), and the effects on patients' depressive symptoms, cognitive function, and sleep varied. Therefore, the present study was based on a narrative review and included a large number of existing studies to investigate the characteristics and differences in the effects of different PA interventions on MDD. The study also investigated the characteristics and differences of different PA interventions in MDD, and explained the neural mechanisms through the results of multimodal brain function monitoring, including the intracranial environment and brain structure. It aims to provide exercise prescription and theoretical reference for future research in neuroscience and clinical intervention in MDD., Competing Interests: The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest., (Copyright © 2024 Guan, Sun, Fan, Liang, Liu, Yu and Liu.)

In recent years, self-supervised learning has emerged as a powerful approach to learning visual representations without requiring extensive manual annotation. One popular technique involves using rotation transformations of images, which provide a clear visual signal for learning semantic representation. However, in this work, we revisit the pretext task of predicting image rotation in self-supervised learning and discover that it tends to marginalise the perception of features located near the centre of an image. To address this limitation, we propose a new self-supervised learning method, namely FullRot, which spotlights underrated regions by resizing the randomly selected and cropped regions of images. Moreover, FullRot increases the complexity of the rotation pretext task by applying the degree-free rotation to the region cropped into a circle. To encourage models to learn from different general parts of an image, we introduce a new data mixture technique called WRMix, which merges two random intra-image patches. By combining these innovative crop and rotation methods with the data mixture scheme, our approach, FullRot + WRMix, surpasses the state-of-the-art self-supervision methods in classification, segmentation, and object detection tasks on ten benchmark datasets with an improvement of up to +13.98% accuracy on STL-10, +8.56% accuracy on CIFAR-10, +10.20% accuracy on Sports-100, +15.86% accuracy on Mammals-45, +15.15% accuracy on PAD-UFES-20, +32.44% mIoU on VOC 2012, +7.62% mIoU on ISIC 2018, +9.70% mIoU on FloodArea, +25.16% AP 50 on VOC 2007, and +58.69% AP 50 on UTDAC 2020. The code is available at https://github.com/anthonyweidai/FullRot&#95;WRMix., Competing Interests: Declaration of competing interest The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: Jun Liu reports financial support was provided by Research Grant Council (RGC) of Hong Kong. Jun Liu reports financial support was provided by Guangdong Province Basic and Applied Basic. Jun Liu reports financial support was provided by Science and Technology Innovation Committee of Shenzhen., (Copyright © 2024 Elsevier Ltd. All rights reserved.)

In this paper, the degradation of PEMFC under different operating conditions in dynamic cycle condition is studied. Firstly, according to the failure mechanism of PEMFC, various operating conditions in dynamic cycle condition are classified, and the health indexes are established. Simultaneously, the rates and degrees of the output voltage decline of the PEMFC under different operating conditions during the dynamic cycling process were compared. Then, a model based on variational mode decomposition and long short-term memory with attention mechanism (VMD-LSTM-ATT) is proposed. Aiming at the performance of PEMFC is affected by the external operation, VMD is used to capture the global information and details, and filter out interference information. To improve the prediction accuracy, ATT is used to assign weight to the features. Finally, in order to verify the effectiveness and superiority of VMD-LSTM-ATT, we respectively apply it to three current conditions under dynamic cycle conditions. The experimental results show that under the same test conditions, RMSE of VMD-LSTM-ATT is increased by 56.11 % and MAE is increased by 28.26 % compared with GRU attention. Compared with SVM, RNN, LSTM and LSTM-ATT, RMSE of VMD-LSTM-ATT is at least 17.26 % higher and MAE is at least 5.65 % higher., Competing Interests: The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper., (© 2024 The Authors.)

The field of regenerative medicine for sports injuries has grown significantly in the 21st century. This study attempted to provide an overview of the current state of research and key findings regarding the relationship between sport and regenerative medicine in general, identifying trends and hotspots in research topics. We gathered the literature from the Web of Science (WOS) database covering the last 10 years (2013-2023) pertaining to regenerative medicine for sporter and applied Citespace to assess the knowledge mapping. The findings demonstrated that there were 572, with a faster increase after 2018. The country, institution, and author with the most publications are the USA, Harvard University, and Maffulli Nicola. In addition, the most co-cited reference is J Acad Nutr Diet (2016) (199). Adipose tissue, high tibial osteotomy, and bone marrow are the hot spots in this field in the next few years., Competing Interests: The authors have no conflicts of interest to disclose., (Copyright © 2024 the Author(s). Published by Wolters Kluwer Health, Inc.)

Diffusion magnetic resonance imaging (dMRI) tractography is a critical technique to map the brain's structural connectivity. Accurate segmentation of white matter, particularly the superficial white matter (SWM), is essential for neuroscience and clinical research. However, it is challenging to segment SWM due to the short adjacent gyri connection in a U-shaped pattern. In this work, we propose an Anatomically-guided Superficial Fiber Segmentation (Anat-SFSeg) framework to improve the performance on SWM segmentation. The framework consists of a unique fiber anatomical descriptor (named FiberAnatMap) and a deep learning network based on point-cloud data. The spatial coordinates of fibers represented as point clouds, as well as the anatomical features at both the individual and group levels, are fed into a neural network. The network is trained on Human Connectome Project (HCP) datasets and tested on the subjects with a range of cognitive impairment levels. One new metric named fiber anatomical region proportion (FARP), quantifies the ratio of fibers in the defined brain regions and enables the comparison with other methods. Another metric named anatomical region fiber count (ARFC), represents the average fiber number in each cluster for the assessment of inter-subject differences. The experimental results demonstrate that Anat-SFSeg achieves the highest accuracy on HCP datasets and exhibits great generalization on clinical datasets. Diffusion tensor metrics and ARFC show disorder severity associated alterations in patients with Alzheimer's disease (AD) and mild cognitive impairments (MCI). Correlations with cognitive grades show that these metrics are potential neuroimaging biomarkers for AD. Furthermore, Anat-SFSeg could be utilized to explore other neurodegenerative, neurodevelopmental or psychiatric disorders., Competing Interests: Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper., (Copyright © 2024 Elsevier B.V. All rights reserved.)

To conduct a comprehensive review of international publications, this study focuses on discussing the development status and future trends of Satellite Internet. In accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines, all peer-reviewed publications in English from journals indexed with the SSCI and SCIE categories of the Web of Science are reviewed, with an emphasis on studies that focused only on the architecture, functions, technologies, and applications of Satellite Internet across the entire period. A total of 2085 records were identified, of which 69 met the inclusion criteria. Through a systematic literature review, the results show that three main research perspectives were summarised and a unified Satellite Internet definition and technical architecture were innovatively provided. Additionally, Satellite Internet application industries and scenarios were comprehensively sorted, and four types of potential impacts were categorised and discussed. This study provides an insightful analysis of recent trends in Satellite Internet research, which may provide guidance for planning and supporting the forthcoming wave of information infrastructure growth. Future research should broaden the scope of theoretical studies when assessing the strategic importance of Satellite Internet. Further quantitative research is needed to provide a more scientific and comprehensive perspective., Competing Interests: The authors declare the following financial interests/personal relationships which may be considered as potential competing interests:Yan Chen reports financial support was provided by 10.13039/501100001809National Natural Science Foundation of China.Yan Chen reports financial support was provided by 10.13039/501100019069Chinese Academy of Engineering. If there are other authors, they declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper., (© 2024 The Authors.)

Time series is a typical data type in numerous domains; however, labeling large amounts of time series data can be costly and time-consuming. Learning effective representation from unlabeled time series data is a challenging task. Contrastive learning stands out as a promising method to acquire representations of unlabeled time series data. Therefore, we propose a self-supervised time-series representation learning framework via Time-Frequency Fusion Contrasting (TF-FC) to learn time-series representation from unlabeled data. Specifically, TF-FC combines time-domain augmentation with frequency-domain augmentation to generate the diverse samples. For time-domain augmentation, the raw time series data pass through the time-domain augmentation bank (such as jitter, scaling, permutation, and masking) and get time-domain augmentation data. For frequency-domain augmentation, first, the raw time series undergoes conversion into frequency domain data following Fast Fourier Transform (FFT) analysis. Then, the frequency data passes through the frequency-domain augmentation bank (such as low pass filter, remove frequency, add frequency, and phase shift) and gets frequency-domain augmentation data. The fusion method of time-domain augmentation data and frequency-domain augmentation data is kernel PCA, which is useful for extracting nonlinear features in high-dimensional spaces. By capturing both the time and frequency domains of the time series, the proposed approach is able to extract more informative features from the data, enhancing the model's capacity to distinguish between different time series. To verify the effectiveness of the TF-FC method, we conducted experiments on four time series domain datasets (i.e., SleepEEG, HAR, Gesture, and Epilepsy). Experimental results show that TF-FC significantly improves in recognition accuracy compared with other SOTA methods., Competing Interests: The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest., (Copyright © 2024 Zhao and Fan.)