Hydro-environmental dynamics of Kaptai Lake using satellite derived biophysical metrics and an ensemble Machine Learning Framework.

FUSION-CDW: Automated curation of scalable datasets for enhancing construction and demolition waste valorisation.

Increasing Technical Complexity of EVAR Conversion Over Time Is Not Associated With Worse Outcomes at a High-Volume Aortic Treatment Center.

Arterial thrombosis remains a leading cause of morbidity and mortality worldwide, while its mechanistic understanding and clinical management remain limited. In this review, we discuss how recent advances in microfluidic thrombosis models, mechanobiology, and microrobotic technologies may enable the development of clinically actionable and personalized thrombosis platforms. Recent findings demonstrate that arterial thrombus formation is strongly regulated by dynamic shear stress, platelet-rich aggregation, and von Willebrand factor (vWF)-mediated interactions. Emerging evidence further shows that shear-induced platelet aggregates, also known as SIPA clots, can form mechanically robust thrombi independently of classical coagulation pathways, highlighting thrombosis as a highly mechanosensitive process. Although microfluidic and flow-based systems have improved the physiological modeling of thrombosis, current platforms still face major limitations in capturing multidimensional shear dynamics, mechanobiological complexity, and patient-specific variability. Recent progress in vessel-on-a-chip technologies, computational modeling, artificial intelligence, and microrobotic systems suggests a pathway toward integrated and feedback-driven thrombosis management. These approaches may enable not only the measurement and prediction of thrombotic behavior but also its active modulation through targeted interventions. Collectively, this perspective supports a transition from static thrombosis assays toward dynamic and controllable mechanobiological platforms for precision cardiovascular medicine.

Accurate identification of cerebrovascular stenosis is essential for early stroke prevention and effective clinical management. Magnetic resonance angiography provides non-invasive 3D visualization of cerebral vessels, but reliable automated stenosis detection remains challenging due to anatomical complexity and imaging variability. This study aims to develop an automated, robust, and clinically useful transformer-based deep learning framework for detecting stenosis in 3D brain MRA scans. We propose a framework designed for cerebrovascular stenosis detection. It first automatically localizes the centerlines of all arteries and veins within the 3D MRA volume. Each resulting vessel-centered 3D region is then analyzed and classified as normal or narrowed using our proposed transformer-based model. The model was trained and validated on a manually curated, expert-annotated dataset from Far Eastern Memorial Hospital, Taiwan, to ensure high-quality ground-truth labels. Our proposed framework demonstrated strong and stable performance across five-fold cross-validation. Specially, under the imbalanced data setting, the model achieved an average accuracy of 0.9339, F1-score of 0.7998, AUC of 0.9488, and Precision-Recall AUC of 0.8313-indicating robust discrimination capability and effective detection. The experimental results underscore the capability of the proposed framework as a dependable tool for automated cerebrovascular evaluation. Its superior performance indicates significant utility in clinical settings, supporting early detection and risk reduction for stroke.

Estimating dependency relationships between variables is an important issue in statistics. Mutual information (MI) is a measure of dependency which quantifies the amount of shared information between two variables. It is free of distribution assumption and captures both linear and non-linear dependencies. MI estimation methods were primarily developed for datasets with exclusively discrete variables, exclusively continuous variables, or a mixture of both. In practice, complex variables containing both discrete and continuous values (discrete-continuous variables), specifically continuous censored variables, are often present in real datasets (e.g., biological measures from analytical tools with lower detection limit). Methods have been developed to handle discrete-continuous data, but their effectiveness on the specific case of continuous censored data has not yet been evaluated. We propose a new estimation method based on the decomposition of the MI formula, with a first part handling the censoring status of the data, and a second part handling its continuous section. This estimation method works as a correction, as it takes in parameter one MI estimator for continuous data, and makes it able to handling censoring. We constructed different simulation scenarios of pairs of correlated censored log-normal variables, by varying the censoring rate, correlation, and sample size. We evaluated our correction on a few existing estimators previously developed for continuous, mixed or discrete-continuous data. We compared the selected estimators, with and without the correction, on these different scenarios. We found that the correction globally enables to reduce bias, and allows convergence towards the true MI value as the number of observations increases.

Multiple roles of circRNAs in cervical cancer: From fundamental carcinogenic mechanisms to clinical application prospects.

The zygomaticus major muscle (ZM) is pivotal in midfacial expression and smile dynamics and serves as a key anatomic landmark in craniofacial surgery. Although traditionally described as a single, uniform muscle, accumulating evidence reveals significant morphologic variability, including bifid configurations and atypical relationships with neurovascular structures. A meticulous lateral facial dissection was performed on a 74-year-old male cadaver to assess ZM morphology and its spatial associations with surrounding anatomic structures. Special attention was given to perioral musculature, branches of the facial nerve, the parotid duct, and facial vasculature. The ZM demonstrated a bifid configuration, comprising distinct superior and inferior slips converging at the modiolus. The risorius muscle was absent. The parotid duct coursed deep to the ZM inferior slip. At the mandibular lower border, the facial artery showed an atypical trifurcation into submental, ascending facial, and premasseteric branches. Zygomatic and buccal branches of the facial nerve were identified traveling deep to the bifid muscle slips. These observations underscore the coexistence of muscular and vascular abnormalities in the midfacial region. This case illustrates the complex anatomic variability of the ZM and its intimate relationship with critical neurovascular structures. Recognition of such variants is essential for optimizing surgical outcomes in facial reanimation, rhytidectomy, and minimally invasive aesthetic procedures.

Background:Contrast-induced nephropathy (CIN) is a major complication following coronary interventions, contributing to increased morbidity and healthcare costs. Machine learning (ML) models provide innovative approaches for predicting CIN by integrating complex clinical variables, potentially improving risk stratification and patient outcomes. This meta-analysis evaluates the predictive performance of ML models for CIN, focusing on the best-performing models.Methods:Seventeen studies encompassing 21,69,263 patients were analyzed. The predictive accuracy of ML models was synthesized using pooled area under the curve (AUC) estimates and heterogeneity metrics.Results:The pooled incidence of CIN was 11% (95% CI: 9–13%). Overall, ML models achieved a pooled AUC of 0.74 (95% CI: 0.72–0.75). Random forest (RF) model demonstrated the highest performance with an AUC of 0.86 (95% CI: 0.85–0.87), followed by gradient boosting machines (GBM) and Extreme Gradient Boosting (XGBoost), both achieving an AUC of 0.79. In training datasets, RF and XGBoost achieved the highest AUCs of 0.98 (95% CI: 0.97–0.99), with GBM following at 0.88 (95% CI: 0.85–0.90). In test datasets, Ensemble models achieved the best performance with an AUC of 0.80 (95% CI: 0.66–0.94), followed by RF and XGBoost with AUCs of 0.75. External validation results showed an overall pooled AUC of 0.77 (95% CI: 0.71–0.84), indicating strong generalizability of the models. Among CIN definitions, the European Society of Urogenital Radiology (ESUR) criteria yielded the best predictive performance, with an AUC of 0.77 (95% CI: 0.72–0.82).Conclusion:RF, Ensemble models, and XGBoost emerged as the most effective ML models for predicting CIN, with RF showing consistent superiority in training datasets and Ensemble models excelling in test datasets. The pooled CIN incidence emphasizes the clinical burden, and the ESUR definition provided the highest predictive accuracy, supporting its utility in CIN risk stratification.

Cokriging is a multivariate spatial method used to predict the observed value for a primary variable in an unknown location with the help of a spatially correlated secondary variable. The existence of two or more nonlinear secondary variables in predicting spatial data usually arises, especially in cokriging. Therefore, a method that can improve the model's predictive power by adding the interaction of variables is proposed. The proposed method can be effectively used, especially when the primary and secondary variables have a nonlinear relationship. By transforming the nonlinear variables, a higher correlation can be attained. This study used principal component analysis with interaction (PCAI) method among secondary variables to reduce two or more secondary variables into one dimension as a secondary variable in the cokriging technique. The proposed method was tested and verified through simulation and real data using the 2015 South Korea Air Pollution, a dataset known for its complex spatial patterns and high variability, to prove its validity and usefulness. The predicted residual error sum of squares (PRESS) statistic was used for cross-validations. Computations were done using the R Project for Statistical Computing software. PCAI as a secondary variable gives the lowest PRESS value compared to only one secondary variable or principal component analysis (PCA). Considering the criterion, the lowest value of PRESS indicates the best model. Thus, PCAI cokriging outperformed PCA cokriging. Using PCAI as a secondary variable may be a better method than PCA for cokriging with nonlinear multicovariates.

Occipital nerves: A concise encyclopedic review with clinical relevance in physiotherapy

• Surveyed 1,000 rural households in Mahendragarh (Haryana) and Jhunjhunu (Rajasthan) to assess health impacts of biomass fuel use. • Identified high prevalence of respiratory symptoms, especially among women and children exposed to smoke from wood, dung cakes, and crop residues. • Developed a comprehensive analytical system integrating PM monitoring, weather data, and health surveys to model exposure-risk relationships. • Applied machine learning (Random Forest, XGBoost, LSTM) and causal inference methods (Granger causality, Causal Impact) to predict pollution and health outcomes. • Designed real-time API and dashboard tools for public health alerts and policy support in rural air quality management. This study investigates the spatiotemporal dynamics of ambient air pollution and its health impacts across two semi-urban districts in India- Jhunjhunu and Mahendragarh, using a multidisciplinary approach combining statistical analysis, machine learning, and causal inference. A one-year high-resolution monitoring dataset of PM₁, PM₂.₅, PM₄, and PM₁₀ was integrated with structured household health surveys covering over 1,000 households. High-resolution monitoring of PM₁, PM 2.5 , PM₄, and PM₁₀, along with survey-based health data, was analyzed to explore pollutant behavior, exposure-response relationships, and symptom prevalence. Linear regression models effectively predicted PM 2.5 trends in Jhunjhunu, while advanced models such as Random Forest, XGBoost, and Long Short-Term Memory (LSTM) captured complex variability in Mahendragarh. Models were trained using a 70:30 train–test split with k-fold cross-validation and evaluated using RMSE, MAE, and R² metrics. LSTM and XGBoost achieved the best performance (R² up to 0.87; RMSE reduced by approximately 30% compared to linear regression). SHAP analysis highlighted PM₁ and PM₄ as critical predictors, underscoring the need to expand national air quality standards beyond PM 2.5 and PM₁₀. Explainable machine learning using SHAP identified PM₁ and PM₄ as influential predictors of health-related outcomes, underscoring the need to expand national air quality standards beyond PM2.5 and PM₁₀. Granger-causal links, residual diagnostics, and health symptom anomalies revealed significant associations between particulate pollution and respiratory, cardiovascular, and visual symptoms, particularly in Mahendragarh. Policy insights emphasize cleaner fuel adoption, improved ventilation, and awareness campaigns to mitigate risk among vulnerable, low-income households. By integrating machine learning with epidemiological modeling, this study provides robust, location-specific evidence to support targeted environmental health interventions in under-monitored regions. A key innovation of this study lies in the joint monitoring and modeling of PM₁ and PM₄ alongside conventional PM₂.₅ and PM₁₀ using explainable ML and causal inference. This framework captures nonlinear exposure–response patterns and improves predictive accuracy while providing mechanistic insight into particle-size-specific health risks. The results offer actionable evidence for clean fuel transition, household ventilation improvements, and community-level air quality management in semi-urban and rural settings. Integration of high-resolution particulate monitoring, machine learning, and causal inference reveals strong links between PM₁-PM₁₀ exposure and cardiopulmonary and ocular symptoms in semi-urban India, highlighting PM₁ and PM₄ as key predictors for targeted interventions.

Time series forecasting is a critical task with widespread applications in industrial domains and daily life, including weather prediction, long-term energy consumption planning, and marketing analysis. Nevertheless, effectively extracting salient temporal patterns and exploring dependencies within multivariate time series remains a challenge. This paper focuses on multivariate time series forecasting, a common and pivotal issue in numerous analytical tasks. To address the complexity and high variability inherent in multivariate time series, we propose a scale-adaptive multi-head attention mechanism based on a hybrid mixture of experts network. Building on this mechanism, we develop MixNet, a novel architecture designed to achieve flexible feature extraction across diverse types of time series data. Furthermore, to tackle the difficulty in capturing inter-variable dependencies, we introduce a dedicated multivariate time series embedding (MTSE) scheme integrated with learnable positional encoding. This approach aims to comprehensively model the dependencies among variables, thereby enhancing overall forecasting performance. Experimental results demonstrate that MixNet outperforms several state-of-the-art methods on seven benchmark datasets from primary domains.

Study DesignRetrospective Cohort Study.ObjectivesTo evaluate the accuracy of the ACS-NSQIP Pediatric Surgical Risk Calculator in predicting postoperative complications and mortality following pediatric spinal deformity surgery. Predicted risks were compared with observed outcomes from the ACS-NSQIP Pediatric database, stratified by scoliosis etiology, fusion level, and surgical approach.MethodsWe performed a retrospective analysis of pediatric patients who underwent spinal deformity correction between 2012 and 2023 using the ACS-NSQIP Pediatric database. Patients were categorized as idiopathic or neuromuscular scoliosis. Predicted risks were compared with observed 30-day outcomes including mortality, surgical site infection, pneumonia, and urinary tract infection. Predictive performance was assessed using the Brier score across discrimination and calibration dimensions.ResultsA total of 58,010 patients were included (45,211 idiopathic; 12,799 neuromuscular). Overall, the calculator predicted a 2.74% complication rate vs an observed rate of 9.54% (Brier: 0.00462; 5.35% of maximum), reflecting poor discrimination and substantially underestimated absolute risk. The most frequent complications were surgical site infection (2.12%), pneumonia (0.99%), and urinary tract infection (0.64%), each demonstrating adequate individual-level discrimination and calibration. Stratified analyses showed adequate performance for idiopathic scoliosis patients undergoing 0-12 level fusions, while discrimination was poor for ≥13 level fusions. Performance was substantially worse among neuromuscular scoliosis patients across all fusion levels. Surgical approach did not meaningfully affect performance.ConclusionsThe calculator reliably predicts select individual complications but underestimates overall risk in high-complexity cases, particularly extensive fusions and neuromuscular scoliosis. Incorporating deformity-specific and surgical complexity variables may improve preoperative risk stratification and counseling.

Vegetation spatial heterogeneity is fundamental to biodiversity management and ecosystem service provision, yet detailed phytosociological mapping of forest vegetation remains largely unresolved in the Colombian Orinoquia. This study characterized the geographic distribution of forest vegetation through the integration of 178 field surveys, environmental complex variables defined by geomorphological and bioclimatic gradients, and multi-sensor satellite imagery combining Landsat-8 optical bands and Sentinel-1 dual-polarization data, processed within a Random Forest classification framework in Google Earth Engine. Classifications achieved overall accuracies between 0.910 and 0.975 and Kappa coefficients above 0.93, identifying 24 phytosociological alliances or geobotanical formations distributed across approximately 7,565,696 ha, representing 34.63% of the region. Forest cover ranges from 10.95% in the Floodplain to 55.22% in La Macarena, with the High Plain concentrating the greatest formation diversity. The spatial organization of forest vegetation is primarily governed by the geomorphological gradient—fluvial, denudational, and structural—and limiting bioclimatic factors, together with their associated edaphic−hydrological regimes, with anthropic transformation driven by cattle ranching and agricultural expansion constituting the principal threat to forest cover. These results advance beyond existing land cover surrogates, providing an empirically validated cartographic framework for biodiversity assessment, habitat modeling, and natural capital management in the Colombian Orinoquia.

Parkinson's disease (PD) is a growing neurological challenge. The Sit-to-Stand (STS) transition is a key proxy for functional independence and fall risk. While sensor technology offers objective STS assessment, the methodological landscape is highly heterogeneous, lacking standardized protocols. This scoping review systematically maps the literature on the technological assessment of the STS transition in PD, focusing on sensors, tasks, settings, and variables. Additionally, the review details device accessibility and ecological validity in home-based settings, as these are two important elements for large-scale application and implementation in standardised protocols for both clinical assessment and real-world monitoring. A scoping review following Joanna Briggs Institute and PRISMA-ScR guidelines. Five major databases (PubMed, Embase, CINAHL, Scopus, Web of Science) and citation tracking were used to identify studies from 2015 to 2025. Eligibility criteria included studies with PD patients of any stage undergoing instrumental STS transition assessment. We mapped data according to sensors, tasks, settings, and variables. We performed specific subanalyses on Inertial Measurement Units (IMUs)-distinguishing research-grade 'kinematic' sensors from consumer 'Mobile' devices-and assessed ecological validity (unsupervised vs. supervised) in home-based settings. From 7,368 records, 77 studies were included. Four dominant trends emerged: (1) Technology: IMUs were used in 87% of the studies. (2) Protocol: the Timed-Up-and-Go (TUG) test (60%) was more prevalent than pure STS tasks (22%), (3) Setting: Most assessments occurred in outpatient settings (79%), followed by home-based settings (25%). (4) Functional variables: despite technological sophistication, 'duration' remained the primary variable followed by linear acceleration and angular velocity. Moreover in the home setting, while 74% of studies achieved high ecological validity (unsupervised), a "scalability gap" emerged: 63% still utilized complex research-grade IMUs. Instrumental STS transition assessment in PD is a rapidly expanding field. It is dominated by IMU sensors, but fragmented by protocol heterogeneity and a critical "scalability gap" in home monitoring. Complex variables still require non-scalable sensors, while scalable mobile devices have not yet been validated for advanced metrics. Future work must develop robust algorithms for mobile devices and validate metrics to standardise protocols for clinical assessment versus ecological monitoring.

Researchers understand that conducting numerous pairwise comparisons between group means increases the Type I error rate, prompting the use of planned contrasts like orthogonal contrast sets. Implicit to orthogonal contrast sets is the principal assumption that groups are balanced in size. Further, when dealing with complex variables like latent constructs, specialized modeling is necessary. Understanding how violating the assumptions of orthogonal contrasts, specifically under conditions of sample imbalance, can help identify variability in parameter recovery. This study examines the effect of sample size imbalance and modeling approach on the accuracy of latent group mean difference estimates when using orthogonal contrasts. Monte Carlo simulations compared the Multiple Indicators Multiple Causes (MIMIC) and re-parameterized multigroup confirmatory factor analysis models while manipulating sample sizes, group proportions, and effect size. Results suggest declining parameter recovery as group imbalance increased, particularly in small samples, with some estimates falling below acceptable thresholds for power, Type I error, and bias. The MIMIC model consistently produced more accurate estimates, though is replete with implicit measurement assumptions that are seldom tested. These findings suggest that researchers using orthogonal contrasts when comparing groups on a latent variable continuum must (a) be aware of examined group's sample size proportions and the impact of group size inequalities on estimate accuracy, and (b) carefully consider the costs and benefits of the latent variable modeling approach, including how the model addresses measurement non-invariance.

Neonatal brain injuries, such as stroke, cause focal ischaemic lesions that often result in lifelong neurological disabilities, yet effective treatments remain limited. Early-phase therapeutic screening requires models that can reliably reproduce injury severity while minimising confounding variables, including prolonged or variable anaesthesia, surgical stress, and invasive procedures that themselves affect injury progression. Existing models of neonatal focal ischaemia often exhibit high mortality, technical complexity, and substantial variability in lesion location and volume. As a result, there is a critical need for a rapid, ethically refined, and scalable neonatal model that produces consistent cortical injury suitable for screening neuroprotective, biomaterial-based, and regenerative therapies. We established a minimally invasive photothrombotic ischaemia model in postnatal day 10 rats by administering intraperitoneal Rose Bengal (25, 40, or 60 mg/kg) and activating it with a fixed 10-minute exposure to 565-nm light through the intact scalp and skull. This incision-free protocol allowed a total procedure duration of 19 min. We characterized dose-dependent effects on infarct volume and anatomical distribution, cortical atrophy, ventricular enlargement, apoptosis (cleaved caspase-3), astrocytic and microglial reactivity (glial fibrillary acidic protein, GFAP; ionized calcium-binding adapter molecule 1, Iba1), and sensorimotor outcomes (wire hang, cylinder rearing, adhesive tape removal) at 1, 7, and 14 days after injury. Additional analyses assessed the reproducibility of lesion size across litters and explored sex-specific differences. A 25 mg/kg dose induced a reliable and well-localized motor cortex infarct with no mortality. Higher doses of Rose Bengal produced proportionally larger infarcts with greater subcortical involvement and more pronounced secondary atrophy. Across all groups, apoptotic signalling and glial reactivity remained elevated through 14 days, indicating persistent tissue injury. Sensorimotor impairments were robust at all stages, with deficits in forepaw use, endurance, and tactile response correlating with lesion volumes in the 25 mg/kg group. No significant sex differences were observed for any histological or behavioural outcomes. This refined neonatal photothrombotic model provides a reproducible, simple, scalable, and ethically optimized platform for inducing severe focal cortical injury. The model’s stable injury territory, short, standardized procedure, and consistent functional readouts fill a major gap in current research tools and provide a practical foundation for early-phase testing of neuroprotective and regenerative interventions.

Sensory perception in autism is strikingly heterogeneous, with individuals showing both hypo- and hypersensitivity across different sensory domains. While sensory differences are widely recognized as a core feature of autism, the structure and underlying patterns of this variability remain poorly understood. Previous studies have yielded mixed findings, often examining sensory processing in isolation within single domains, thereby limiting a comprehensive understanding of sensory sensitivity in autism. We compiled psychophysical data from 107 autistic and 408 age- and IQ-matched non-autistic individuals across 32 experimental conditions spanning multiple perceptual domains, including size, brightness, orientation, pitch, and face processing. Two complementary statistical approaches were used: segmented regression and a Bayesian hierarchical model. Despite substantial inter- and intra-individual variability, both models revealed a consistent domain-specific pattern: on average, autistic individuals showed reduced sensitivity to faces and speech, while performance on basic non-social tasks was comparable to or exceeded that of the comparison group. Bayesian modelling further indicated that social relevance, rather than domain alone, accounted for the primary source of divergence between groups. This study focused on sensory sensitivity thresholds and did not assess perceptual biases or changes in subjective appearance of the stimuli. A full account of perception in autism requires considering these broader alterations. The current findings suggest that sensory differences in autism reflect a structured perceptual profile shaped by social relevance, stimulus complexity, and individual variability. The results highlight the importance of individualized sensory profiling and may inform both theoretical models and personalized approaches to intervention in autism.

Diagnostic test for the circular-circular additive regression model