Search Results (180)

Voltage-sensing phosphatases (VSPs) provide a conserved framework for dissecting the mechanics of voltage sensing and for engineering genetically encoded voltage indicators (GEVIs). To evaluate how natural sequence diversity shapes function, we compared VSP voltage-sensing domains (VSDs) from multiple species by replacing the phosphatase domain with a fluorescent protein to enable optical detection of VSD responses. Every construct that reached the plasma membrane produced a voltage-dependent optical signal, underscoring the deep conservation of voltage sensing across VSP orthologs. Yet lineage-specific substitutions generated strikingly different phenotypes. A plankton VSP ortholog from Eurytemora carolleeae and the Sea Hare (Aplysia californica) VSP exhibited left-shifted activation ranges, producing robust fluorescence transitions during modest depolarizations of the plasma membrane. The human VSD of hVSP2 yielded weak, sluggish responses with poor recovery, but reintroduction of a conserved arginine in S1 (G95R) partially restored reversibility, implicating lipid-facing residues in conformational stability. The Chinese hamster (Cricetulus griseus) VSD, with atypical S4 sensing charges (RWIR), generated a slow fluorescence increase during depolarization, while reverting to the consensus arginine (RRIR) inverted the polarity to a decrease. These contrasting behaviors show that single residue changes can reshape how VSD movements influence the fluorescent reporter, highlighting the molecular precision revealed by GEVI measurements. Together, these results show that voltage-dependent signaling is deeply conserved across VSPs but shaped by lineage-specific sequence variation, establishing VSPs as powerful models for probing voltage sensing and guiding GEVI design. Full article

Event cameras provide microsecond temporal resolution, high dynamic range, and low latency by asynchronously capturing per-pixel luminance changes, thereby introducing a novel sensing paradigm. These advantages render them well-suited for high-speed applications such as autonomous vehicles and dynamic environments. Nevertheless, the sparsity of event data and the absence of dense annotations are significant obstacles to supervised learning for motion segmentation from event streams. Domain adaptation is also challenging due to the considerable domain shift in intensity images. To address these challenges, we propose a two-phase cross-modality adaptation framework that translates motion segmentation knowledge from labeled RGB-flow data to unlabeled event streams. A dual-branch encoder extracts modality-specific motion and appearance features from RGB and optical flow in the source domain. Using reconstruction networks, event voxel grids are converted into pseudo-image and pseudo-flow modalities in the target domain. These modalities are subsequently re-encoded using frozen RGB-trained encoders. Multi-level consistency losses are implemented on features, predictions, and outputs to enforce domain alignment. Our design enables the model to acquire domain-invariant, semantically rich features through the use of shallow architectures, thereby reducing training costs and facilitating real-time inference with a lightweight prediction path. The proposed architecture, alongside the utilized hybrid loss function, effectively bridges the domain and modality gap. We evaluate our method on two challenging benchmarks: EVIMO2, which incorporates real-world dynamics, high-speed motion, illumination variation, and multiple independently moving objects; and MOD++, which features complex object dynamics, collisions, and dense 1kHz supervision in synthetic scenes. The proposed UDA framework achieves 83.1% and 79.4% accuracy on EVIMO2 and MOD++, respectively, outperforming existing state-of-the-art approaches, such as EV-Transfer and SHOT, by up to 3.6%. Additionally, it is lighter and faster and also delivers enhanced mIoU and F1 Score. Full article

Background/Purpose: Reproductive health and fertility outcomes are essential but often overlooked aspects of occupational well-being. Organizational dysfunction, demanding workloads, and limited workplace accommodations may negatively affect fertility, while supportive policies and inclusive cultures can mitigate risks. This review aimed to map current evidence on these relationships and their economic consequences. Methodology/Approach: A scoping review was conducted using the PCC (Population–Concept–Context) framework. Systematic searches across multiple databases identified 30 eligible studies, including quantitative, qualitative, and mixed-method designs, spanning different sectors and international contexts. Findings: Four main domains emerged: shift work and circadian disruption, organizational stress and burnout, workplace flexibility and accommodations, and fertility-related policies and organizational support. Hazardous working conditions, long hours, and psychosocial stressors were consistently associated with impaired fertility, reduced fecundability, and pregnancy complications. Conversely, flexible scheduling, fertility benefits, and supportive organizational cultures were linked to improved well-being, retention, and productivity. Originality/Value: This review integrates evidence across occupational health, organizational psychology, and labor economics, offering a comprehensive overview of workplace influences on reproductive health. It highlights gaps in equity and representation—particularly for men, LGBTQ+ employees, and workers in precarious jobs—and calls for longitudinal, interdisciplinary, and intervention-based studies to inform effective workplace policies. Full article

Objective: As the survival of individuals born with congenital heart disease (CHD) improves into adulthood, the focus has shifted from traditional clinical outcomes to patient-reported outcome measures that better reflect the impact of the disease on daily life. Our aim was to assess the quality of life (QoL) of adult patients with congenital heart disease (ACHD) followed in a tertiary centre and to evaluate the parameters that influence QoL in this population. Methods: This cross-sectional observational study included patients followed up at the national referral ACHD centre between April and September 2022. Sociodemographic and clinical data were collected from medical records and self-report questionnaires. Quality of life (QoL) was assessed using the validated Short Form–36 (SF-36) and Euro Quality of Life–5 Dimension (EQ-5D) questionnaires, including the EQ Visual Analogue Scale (VAS). Results: A total of 123 ACHD patients were included (median age 34 (29–41) years; 43.9% male). Most participants had moderate CHD (61%), and 14.6% were cyanotic. Overall, SF-36 Physical Component Summary scores were higher than Mental Component Summary scores. Almost half of the patients (48.8%) reported no problems in all five domains of the EQ-5D, with most problems reported in anxiety/depression domain. Patients with severe CHD, cyanosis, or HF reported lower QoL scores across multiple SF-36 domains, particularly general health, role–physical, and physical functioning domains. Conclusions: QoL among ACHD patients in our cohort was generally high in most domains as assessed by the SF-36 and EQ-5D. Patients with HF reported lower QoL scores, emphasizing the importance of close clinical follow-up and the need for tailored QoL assessment tools for this complex population. Full article

As wind power penetration continues to increase, the sub-synchronous control interaction (SSCI) problem caused by the interaction between doubly fed induction generators (DFIGs) and series-compensated transmission lines has become increasingly prominent, posing a serious threat to power system stability. To address this problem, this research proposes a centralized sub-synchronous oscillation damping controller (CSODC) for wind farms. First, a DFIG impedance model was constructed based on multi-operating-point impedance scanning and a Taylor series expansion, achieving impedance prediction with an error of less than 2% under various power conditions. Subsequently, a CSODC comprising a sub-synchronous damping calculator (SSDC) and a power electronic converter is designed. By optimizing feedback signals, phase shift angles, gain parameters, and filter parameters, dynamic adjustment of controllable impedance in the sub-synchronous frequency band is achieved. Frequency-domain impedance analysis demonstrates that the CSODC significantly enhances the system’s equivalent resistance, reversing it from negative to positive at the resonance frequency point. Time-domain simulations validated the CSODC’s effectiveness in scenarios involving series capacitor switching and wind speed disturbances, demonstrating rapid sub-synchronous current decay. The results confirm that the proposed strategy effectively suppresses sub-synchronous oscillations across multiple scenarios, offering an economical and efficient solution to stability challenges in high-penetration renewable energy grids. Full article

Multiple time series forecasting is critical in domains such as energy management, economic analysis, web traffic prediction and air pollution monitoring to support effective resource planning. Traditional statistical learning methods, including Vector Autoregression (VAR) and Vector Autoregressive Integrated Moving Average (VARIMA), struggle with nonstationarity, temporal dependencies, inter-series correlations, and data anomalies such as trend shifts, seasonal variations, and missing data. Furthermore, their effectiveness in multi-step ahead forecasting is often limited. This article presents an Automated Machine Learning (AutoML) framework that provides an end-to-end solution for researchers who lack in-depth knowledge of time series forecasting or advanced programming skills. This framework utilizes Gated Recurrent Unit (GRU) networks, a variant of Recurrent Neural Networks (RNNs), to tackle multiple correlated time series forecasting problems, even in the presence of anomalies. To reduce complexity and facilitate the AutoML process, many model parameters are pre-specified, thereby requiring minimal tuning. This design enables efficient and accurate multi-step forecasting while addressing issues including missing values and structural shifts. We also examine the advantages and limitations of GRU-based RNNs within the AutoML system for multivariate time series forecasting. Model performance is evaluated using multiple accuracy metrics across various forecast horizons. The empirical results confirm our proposed approach’s ability to capture inter-series dependencies and handle anomalies in long-range forecasts. Full article

Background: End-stage renal disease (ESRD) presents a substantial and growing global health challenge, where hemodialysis serves as an essential life-sustaining therapy for countless individuals. Despite its physiological necessity, the demanding treatment regimen can profoundly impact mental health and overall well-being, though gender-specific data and correlates within the Saudi population remain insufficiently explored. Methods: This cross-sectional study aimed to investigate this gap by assessing the prevalence of anxiety and depression, evaluating health-related quality of life (HRQoL), and analyzing associations with gender and treatment duration in a cohort of 250 hemodialysis patients from multiple centers in Madinah, Saudi Arabia. Validated instruments, namely, the Hospital Anxiety and Depression Scale (HADS) and the 36-Item Short Form Health Survey (SF-36), were employed. Results: The findings revealed a significant psychological burden, with 38% of patients exhibiting anxiety and 32% depression, with females disproportionately affected. HRQoL scores were severely diminished across all domains compared to healthy population norms. Furthermore, a longer dialysis vintage demonstrated a significant positive correlation with worsening psychological scores and a decline in physical HRQoL. Conclusions: These results underscore the critical need for a paradigm shift in standard care, advocating for the systematic integration of routine mental health screenings and the development of tailored, gender-sensitive psychosocial interventions to mitigate this considerable burden. Full article

Background: Neuroimaging-based diagnostic approaches are of critical importance for the accurate diagnosis and treatment of major depressive disorder (MDD). However, multisite neuroimaging data often exhibit substantial heterogeneity in terms of scanner protocols and population characteristics. Moreover, concerns over data ownership, security, and privacy make raw MRI datasets from multiple sites inaccessible, posing significant challenges to the development of robust diagnostic models. Federated learning (FL) offers a privacy-preserving solution to facilitate collaborative model training across sites without sharing raw data. Methods: In this study, we propose the personalized Federated Gradient Matching and Contrastive Optimization (pF-GMCO) algorithm to address domain shift and support scalable MDD classification using multimodal MRI. Our method incorporates gradient matching based on cosine similarity to weight contributions from different sites adaptively, contrastive learning to promote client-specific model optimization, and multimodal compact bilinear (MCB) pooling to effectively integrate structural MRI (sMRI) and functional MRI (fMRI) features. Results and Conclusions: Evaluated on the Rest-Meta-MDD dataset with 2293 subjects from 23 sites, pF-GMCO achieved accuracy of 79.07%, demonstrating superior performance and interpretability. This work provides an effective and privacy-aware framework for multisite MDD diagnosis using federated learning. Full article

In this paper, we present a unique method to determine the level of bot contamination of web-based user agents. It is common practice for bots and robotic agents to masquerade as human-like to avoid content and performance limitations. This paper continues our previous work, using over 600 million web log entries collected from over 4000 domains to derive and generalize how the prominence of specific web browser versions progresses over time, assuming genuine human agency. Here, we introduce a parametric model capable of reproducing this progression in a tunable way. This simulation allows us to tag human-generated traffic in our data accurately. Along with the highest confidence self-tagged bot traffic, we train a Transformer-based classifier that can determine the bot contamination—a botness metric of user-agents without prior labels. Unlike traditional syntactic or rule-based filters, our model learns temporal patterns of raw and heuristic-derived features, capturing nuanced shifts in request volume, response ratios, content targeting, and entropy-based indicators over time. This rolling window-based pre-classification of traffic allows content providers to bin streams according to their bot infusion levels and direct them to several specifically tuned filtering pipelines, given the current load levels and available free resources. We also show that aggregated traffic data from multiple sources can enhance our model’s accuracy and can be further tailored to regional characteristics using localized metadata from standard web server logs. Our ability to adjust the heuristics to geographical or use case specifics makes our method robust and flexible. Our evaluation highlights that 65% of unclassified traffic is bot-based, underscoring the urgency of robust detection systems. We also propose practical methods for independent or third-party verification and further classification by abusiveness. Full article

Older adults often face significant challenges in understanding medical information due to cognitive aging and limited health literacy. Existing simplification models, while effective in general domains, cannot adapt content for elderly users, frequently overlooking narrative tone, readability constraints, and semantic fidelity. In this work, we propose GENSIM—a Generative Expert-Narrated Simplification Model tailored for age-adapted medical text simplification. GENSIM introduces a modular architecture that integrates a Dual-Stream Encoder, which fuses biomedical semantics with elder-friendly linguistic patterns; a Persona-Tuned Narrative Decoder, which controls tone, clarity, and empathy; and a Reinforcement Learning with Human Feedback (RLHF) framework guided by dual discriminators for factual alignment and age-specific readability. Trained on a triad of corpora—SimpleDC, PLABA, and a custom NIH-SeniorHealth corpus—GENSIM achieves state-of-the-art performance on SARI, FKGL, BERTScore, and BLEU across multiple test sets. Ablation studies confirm the individual and synergistic value of each component, while structured human evaluations demonstrate that GENSIM produces outputs rated significantly higher in faithfulness, simplicity, and demographic suitability. This work represents the first unified framework for elderly-centered medical text simplification and marks a paradigm shift toward inclusive, user-aligned generation for health communication. Full article

Background: Post-traumatic epilepsy (PTE) is a recognized complication of traumatic brain injury (TBI), yet its risk following mild and moderate TBI remains underappreciated. Although mild TBI represents the majority of cases in clinical practice, a subset of patients develop unprovoked seizures months or even years post-injury. This review aims to synthesize current evidence on the incidence and predictors of PTE in mild and moderate TBI and to propose a clinically actionable decision-support tool for early risk stratification. Methods: We performed a narrative review of peer-reviewed studies published between 1985 and 2024 that reported on the incidence, risk factors and predictive models of PTE in patients with mild (Glasgow Coma Scale [GCS] 13–15) and moderate (GCS 9–12 or imaging-positive) TBI. Data from 24 studies were extracted, focusing on neuroimaging findings, early post-traumatic seizures, EEG abnormalities and clinical risk factors. These variables were integrated into a rule-based algorithm, which was implemented using Streamlit to enable real-time clinical decision-making. The decision-support tool incorporated five domains: injury severity, early post-traumatic seizures, neuroimaging findings (including contusion location and hematoma type), clinical and demographic variables (age, sex, psychiatric comorbidities, prior TBI, neurosurgical intervention) and EEG abnormalities. Results: PTE incidence following mild TBI ranged from <1% to 10%, with increased risk observed in patients presenting with intracranial hemorrhage or early seizures. From moderate TBI, incidence rates were consistently higher (6–12%). Key predictors included early seizures, frontal or temporal contusions, subdural hematoma, multiple contusions and midline shift. Additional risk-enhancing factors included prolonged loss of consciousness, male sex, psychiatric comorbidities and abnormal EEG patterns. Based on these features, we developed a decision-support tool that stratifies patients into low-, moderate- and high-risk categories for developing PTE. Conclusions: Even in non-severe cases, patients with mild and moderate TBI who exhibit high-risk features remain vulnerable to long-term epileptogenesis. Our proposed tool provides a pragmatic, evidence-based framework for early identification and follow-up planning. Prospective validation studies are needed to confirm its predictive accuracy and optimize its clinical utility. Full article

Convolutional neural network (CNN) models are widely used for environmental sound classification (ESC). However, 2-D convolutions assume translation invariance along both time and frequency axes, while in practice the frequency dimension is not shift-invariant. Additionally, single-scale convolutions limit the receptive field, leading to incomplete feature representation. To address these issues, we introduce a parallel time-frequency multi-scale attention (PTFMSA) module that integrates local and global attention across multiple scales to improve dynamic convolution in order to overcome these problems. We also introduce the parallel branch structure to avoid mutual interference of information in case of extracting time and frequency domain features. Additionally, we utilize learnable parameters that can dynamically adjust the weights of different branches during network training. Building on this module, we develop PTFMSAN, a compact network that processes raw waveforms directly for ESC. To further strengthen learning, between-class (BC) training is applied. Experiments on the ESC-50 dataset show that PTFMSAN outperforms the baseline model, achieving a classification accuracy of 90%, competitive among CNN-based networks. We also performed ablation experiments to verify the effectiveness of each module. Full article

Endothelial hyperpermeability is a hallmark of diverse inflammatory and vascular pathologies, including sepsis, acute respiratory distress syndrome (ARDS), ischemia–reperfusion injury, and atherosclerosis. Traditionally considered a passive return to baseline following stimulus withdrawal, barrier recovery is now recognized as an active, endothelial-driven process. Earlier work identified individual components of this restorative phase, such as cyclic adenosine monophosphate (cAMP)/exchange protein directly activated by cAMP 1 (Epac1) signaling, Rap1/Rac1 activation, vasodilator-stimulated phosphoprotein (VASP) phosphorylation, and targeted cytoskeletal remodeling, as well as kinase pathways involving PKA, PKG, and Src. However, these were often regarded as discrete events lacking a unifying framework. Recent integrative analyses, combining mechanistic insights from multiple groups, reveal that nitric oxide (NO) generated early during hyperpermeability can initiate a delayed cAMP/Epac1 cascade. This axis coordinates Rap1/Rac1-mediated cortical actin polymerization, VASP-driven junctional anchoring, retro-translocation of endothelial nitric oxide synthase (eNOS) to caveolar domains, PP2A-dependent suppression of actomyosin tension, and Krüppel-like factor 2 (KLF2)-driven transcriptional programs that sustain endothelial quiescence. Together, these pathways form a temporally orchestrated, multi-tiered “inactivation” program capable of restoring barrier integrity even in the continued presence of inflammatory stimuli. This conceptual shift reframes NO from solely a barrier-disruptive mediator to the initiating trigger of a coordinated, pro-resolution mechanism. The unified framework integrates cytoskeletal dynamics, junctional reassembly, focal adhesion turnover, and redox/transcriptional control, providing multiple potential intervention points. Therapeutically, Epac1 activation, Rap1/Rac1 enhancement, RhoA/ROCK inhibition, PP2A activation, and KLF2 induction represent strategies to accelerate endothelial sealing in acute microvascular syndromes. Moreover, applying these mechanisms to arterial endothelium could limit low-density lipoprotein (LDL) entry and foam cell formation, offering a novel adjunctive approach for atherosclerosis prevention. In this review, we will discuss both the current understanding of endothelial hyperpermeability mechanisms and the emerging pathways of its active inactivation, integrating molecular, structural, and translational perspectives. Full article

Deep neural network-based fault diagnosis is gaining significant attention within the Industry 4.0 framework, yet practical deployment is still hindered by domain shift, partial label mismatch, and class imbalance. In this regard, this paper proposes a Multi-Target Adversarial Learning for Partial Fault Diagnosis (MTAL-PFD), an extension of adversarial and discrepancy-based domain adaptation tailored to single-source, multi-target (1SmT) partial fault diagnosis in electric motor-driven systems. The framework transfers knowledge from a labeled source to multiple unlabeled target domains by combining dual 1D-CNN feature extractors with adversarial domain discriminators, an inconsistency-based regularizer to stabilize learning, and class-aware weighting to mitigate partial label shift by down-weighting outlier source classes. Thus, the proposed scheme combines a multi-objective approach with partial domain adaptation applied to the diagnosis of electric motor-driven systems. The proposed model is evaluated across 24 cross-domain tasks and varying operating conditions on two motor test benches, showing consistent improvements over representative baselines. Full article

Acetylcholinesterase (AChE) has emerged not only as a cholinergic enzyme but also as a modulator of β-amyloid (Aβ) aggregation via its peripheral anionic site (PAS), making it a dual-purpose target in Alzheimer’s disease. While classical AChE inhibitors provide symptomatic relief, they lack efficacy against the amyloidogenic cascade. This review highlights recent advances in multifunctional AChE pharmacophores that inhibit enzymatic activity while simultaneously interfering with Aβ aggregation, oxidative stress, metal dyshomeostasis, and neuroinflammation. Particular emphasis is placed on dual-site inhibitors targeting both catalytic and peripheral domains, multi-target-directed ligands (MTDLs) acting on multiple neurodegenerative pathways, and metal-chelating hybrids that address redox-active metal ions promoting Aβ fibrillization. We also discuss enabling technologies such as AI-assisted drug design, high-resolution structural tools, and human induced pluripotent stem cell (iPSC)-derived neuronal models that support physiologically relevant validation. These insights reflect a paradigm shift towards disease-modifying therapies that bridge molecular pharmacology and pathophysiological relevance. Full article