Search Results (21,620)

Conventional multispectral photodetectors typically rely on multiple electrical terminals to discriminate different wavelengths, which inevitably increases structural complexity. Here, we break this paradigm by demonstrating a dual-color visible–infrared photodetector based on a simple two-terminal Au/MoS₂/Te heterostructure. The device operates through a bias-switching mechanism: reversing the voltage polarity selectively activates either the MoS₂/Au Schottky junction for visible-light detection (520 nm) or the Te/MoS₂ heterojunction for infrared detection (1550 nm). This bias-controlled wavelength selectivity is unambiguously verified by scanning photocurrent mapping. Beyond dual-color discrimination, an adaptive convolutional neural network is employed to decode the nonlinear current–voltage characteristics and enable precise spectral identification, achieving a reconstruction error of approximately 4.5%. Furthermore, high-fidelity dual-color imaging is demonstrated at room temperature. These results establish a hardware–algorithm co-design strategy based on a minimalist two-terminal architecture, providing a viable route toward compact and intelligent spectral-sensing systems. Full article

Image-based AI offers a low-cost alternative to traditional turbidity sensors in farm ponds, yet the prevailing shift toward Vision Transformers (ViTs) critically overlooks two field realities: the chronic scarcity of annotated data (Small Data) and the strict computational limits of edge hardware. This study presents a frugal computer vision framework that challenges the need for complex architectures in environmental screening. By systematically benchmarking six deep learning models across a calibrated high-turbidity dataset (200–800 NTU, 700 images) under standardized capture conditions, we demonstrate that traditional Convolutional Neural Networks (CNNs) possess a crucial inductive bias for this task. Specifically, ResNet-50 significantly outperformed modern ViTs in both accuracy (96.3% vs. 80.0%) and data efficiency, effectively capturing spatial scattering patterns without the massive data requirements that hindered transformer convergence. Deployed on a resource-constrained Raspberry Pi 4, the CNN-based system achieved an inference latency of 46 ms, demonstrated in an initial hardware-in-the-loop field proof-of-concept (82.4% agreement under baseline, calm-weather conditions, $n=17$). This edge-native approach not only provides actionable spatial turbidity maps to guide on-farm filtration and livestock management decisions but also establishes a critical architectural baseline: under controlled capture protocols, mature CNNs consistently outperform ViTs, establishing them as the optimal architecture for frugal, small-data agricultural Internet of Things (IoT) deployments. Full article

Understanding habitat use and connectivity in estuarine fishes is essential for effective conservation and management. In this study, otolith microchemistry was applied to investigate habitat use and connectivity of the estuarine catfish Genidens genidens across three estuaries in southeastern Brazil. A total of 58 individuals were analyzed using laser ablation inductively coupled plasma mass spectrometry, focusing on strontium-to-calcium (Sr:Ca) and barium-to-calcium (Ba:Ca) ratios. Variations in elemental ratios along otolith transects were used to infer individual ontogenetic patterns along the estuarine–marine gradient. Most individuals exhibited combined use of estuarine and marine environments, while trajectories restricted to freshwater were rare. The apparent complexity of chemical profiles tended to increase with age; however, this pattern disappeared after correction for size-related bias, suggesting that age itself did not significantly influence habitat-use transitions. These patterns are consistent with ecological plasticity and partial migration within populations of G. genidens, although they may also reflect exposure to variable environmental conditions. Sr:Ca ratios were useful indicators of salinity-related transitions, whereas Ba:Ca ratios provided complementary information associated with continental influence. Overall, this study highlights the applicability of otolith microchemistry for investigating habitat-use patterns in estuarine fishes and reinforces the ecological importance of estuaries for feeding, growth, and recruitment in G. genidens, while acknowledging inherent limitations related to environmental variability and proxy interpretation. Full article

Quantitative assessment of mitral regurgitation (MR) in dogs with myxomatous mitral valve disease (MMVD) is influenced by the method used to estimate left ventricular volume. This study aimed to evaluate the impact of different left ventricular volume estimation methods on quantitative MR assessment, using the modified Simpson’s method of discs (Disc method) as a reference. Echocardiographic data from 167 dogs with MMVD and 19 healthy control dogs were analyzed. Regurgitant volume (RVol), body size-normalized RVol, and regurgitant fraction (RF) were calculated using diameter-based methods (Cube, Gibson, Meyer, and Teichholz) and compared with values obtained using the Disc method. All diameter-based methods showed significant positive correlations with the Disc method. However, Bland–Altman analyses demonstrated wide limits of agreement and systematic bias. Between-method discrepancies increased with advancing disease stage, with diameter-based methods tending to overestimate RVol and RF, particularly in dogs classified as American College of Veterinary Internal Medicine (ACVIM) stages B2 and C/D. Although relative trends in regurgitant indices were consistent across methods, substantial differences were observed in absolute values. These findings indicate that diameter-based methods are not interchangeable with the Disc method for absolute quantification of MR severity in dogs with MMVD, especially in advanced disease stages. Full article

Introduction: Three-dimensional (3D) bioprinting is a promising approach for endodontic tissue engineering, enabling scaffolds with controlled architecture and bioactivity to support pulp regeneration. Objectives: This systematic review assessed the following: “What 3D bioprinting applications are reported in endodontics-related studies?” Materials and Methods: Following PRISMA 2020 guidelines, PubMed/MEDLINE, Scopus, Embase, Cochrane Library, Web of Science, SciELO, LILACS, and Google Scholar were searched up to January 2026 with no date or language limits. Two reviewers independently screened studies; risk of bias in in vitro studies was assessed with the QUIN tool. As only one study reported complete antimicrobial outcomes, an intra-study quantitative comparison (MD, 95% CI) of inhibition halos was performed (not a meta-analysis). Results: From 518 records, nine studies were included. Outcomes mainly addressed physicochemical properties (n = 9), cell viability (n = 7), biocompatibility (n = 5), and cell differentiation (n = 5); antimicrobial activity was evaluated in two studies. Most used hDPSCs and extrusion-based printing, testing calcium silicate composites, alginate hydrogels, functionalized PCL, and modified PLA. Modified PLA scaffolds showed greater antimicrobial activity, strongest with naringin and nHA formulations. Overall risk of bias was moderate (58.33%), largely due to limited reporting of randomization, blinding, and sampling. Conclusion: 3D-bioprinted scaffolds/bioinks generally improved cellular responses and bioactivity, especially with MTA, Biodentine, nHA, or naringin; antimicrobial effects were most evident in functionalized PLA (PLA/NAR and PLA/nHA/NAR). Full article

Large-scale integration of photovoltaics (PV) introduces complex source-load temporal volatility and grid-connection/off-grid transitions. Traditional static reliability assessments fail to capture these dynamics, resulting in “considerable deviations” in system indices. This paper proposes a reliability evaluation framework that couples temporal source-load trajectories with a multi-stage fault recovery process. Unlike traditional methods that rely on a single static snapshot, the proposed model evaluates the system state across a continuous 5-h restoration window. The novelty lies in the unique integration of a Dynamic Time Warping (DTW)–Kmedoids method to preserve temporal phase-shifts and a multi-stage Mixed-Integer Linear Programming (MILP) model to simulate PV grid-connection transitions throughout this window. By capturing the intra-outage evolution of sources and loads, the framework fundamentally corrects the “considerable deviations” of static assessments. Case studies demonstrate high precision with an error of less than 0.71% and a 20-fold speedup. Crucially, the framework corrects the 22.31% risk underestimation bias inherent in static models by tracking real-time source-load evolution. This confirms that temporal coordination performance is the primary determinant of the reliability ceiling in active distribution networks. The findings reveal that the precise alignment of intermittent generation and fluctuating demand defines the actual operational safety margin, providing a superior quantitative foundation for grid resilience enhancement. Full article

Macroeconomic conditions are widely believed to influence the direction of equity markets, yet most forecasting models either ignore macroeconomic information or incorporate it through a small set of ad hoc predictors. We propose XAttnFusion, a macro–market fusion architecture that jointly learns from high-frequency market data and lower-frequency macroeconomic time series for equity return prediction. The model comprises three branches: a 1D convolutional network that encodes 40-day market windows (price, volume, and technical indicators), a temporal convolutional network that encodes 24-month macro sequences, and a feedforward branch for volume-at-price structure features. These representations are integrated through multi-head cross-attention, in which the current market state queries the macro sequence to produce a fused representation for directional forecasting. We evaluate XAttnFusion on daily SPY returns from 2012 to 2024 using purged cross-validation with a 5-day embargo to prevent information leakage. To address potential look-ahead bias from macroeconomic publication lags, all macro inputs are lagged by two months. The model achieves a mean out-of-sample AUROC of $0.63\pm 0.05$, representing a 27% improvement over random and an 8.1% improvement over the best concatenation baseline. In a fair comparison where each model is independently hyperparameter-tuned, cross-attention fusion improves AUROC by 0.047 over concatenation ($p=0.031$, Wilcoxon signed-rank test). The model also generalizes to QQQ and IWM, where cross-attention consistently outperforms concatenation fusion. Crucially, the model’s discriminative ability is state-dependent, indicating that the value of macro–market fusion is itself conditioned on market structure. Permutation-based feature importance shows that macro and market branches contribute on a comparable scale (approximately 48% and 36%, respectively), so the gains come from jointly fusing two comparably weighted sources rather than from a single dominant input. Our results show that explicitly modeling macro–market interactions with interpretable attention improves predictive accuracy over naive fusion strategies and provides insight into the time-varying relevance of macroeconomic information in financial forecasting and equity market prediction. Full article

Rapid urbanisation and climate-induced extreme weather events have intensified urban stormwater runoff challenges. Biofiltration systems have emerged as effective, sustainable urban drainage solutions for mitigating these impacts. A total of 78 peer-reviewed studies were assessed to synthesise findings on how design parameters influence pollutant removal performance in biofiltration systems treating urban stormwater runoff. Peer-reviewed articles published from 1 January 1995 to 3 June 2025 were retrieved from Scopus and Web of Science (WoS). Non-peer-reviewed, non-empirical, incomplete, or non-relevant studies were excluded. Rigorous application of a standardised review protocol and predefined criteria was employed to mitigate bias. The findings reveal high removal efficiencies for certain trace metals, ammonium, Escherichia coli (E. coli), hydrocarbons, and microplastics, with inconsistent removal for total nitrogen, nitrates, and phosphorus. The primary pollutant removal mechanisms were adsorption, ion exchange with select media, and denitrification in saturated zones. Only 22% of the reviewed studies incorporated a saturated zone, while 18% included a protective surface layer, despite both design elements being associated with improved pollutant removal performance. Variations in media composition and stormwater quality limit comparability across studies. This review highlights the need for context-specific design guidance and further exploration of multi-functional media to enhance multi-pollutant removal. Full article

Accurate wind speed forecasting is critical for reliable wind-power integration, yet it remains challenging due to the strongly non-stationary and inherently multi-scale nature of atmospheric processes. While deep learning models—such as LSTM, GRU, and Transformer architectures—achieve competitive short- and medium-term performance, they frequently suffer from spectral bias, hyperparameter sensitivity, and reduced generalization under heterogeneous operating regimes. To address these limitations, we propose a multi-scale spectral–recurrent framework, termed RFF-RNN, which integrates multi-band Random Fourier Feature (RFF) encodings with parameterizable recurrent backbones. A key innovation of our approach is the deliberate relaxation of strict shift-invariance constraints; by jointly optimizing spectral frequencies, phase biases, and bandwidth scales alongside the neural weights, the framework dynamically shapes a fully data-driven spectral embedding. To ensure robust adaptation, we employ a two-stage optimization strategy combining gradient-based inner-loop learning with outer-loop Bayesian hyperparameter tuning. Our extensive evaluations on a controlled synthetic benchmark and six geographically diverse real-world wind datasets (spanning the USA, China, and the Netherlands) demonstrate the superiority of the proposed framework. Statistical validation via the Friedman test confirms that RFF-enhanced models—particularly RFF-GRU and RFF-LSTM—systematically outperform standard recurrent networks and state-of-the-art Transformer architectures (Autoformer and FEDformer). The proposed approach yields significantly lower error metrics (MAE and RMSE) and higher explained variance ($R^2$), while exhibiting remarkable resilience against error accumulation at extended forecasting horizons. Full article

Accurate rainfall information is essential for rainfall–runoff modeling in monsoon-influenced basins, where pronounced spatial variability and limited gauge coverage introduce significant uncertainty. Satellite precipitation products provide spatially continuous estimates but are affected by systematic biases, and improvements in statistical rainfall accuracy do not necessarily translate into hydrologically consistent model forcing. This study interpreted satellite rainfall bias correction through a rainfall–runoff framework in the Phetchaburi River Basin, Thailand, using the DWCM-AgWU hydrological model. Simulations were driven by gauge observations and multiple satellite-based rainfall products (GSMaP, CMORPH, CHIRPS, and PERSIANN-CCS), with bias correction applied using Linear Scaling and Quantile Mapping under rainfall-specific calibration. Results showed that bias correction significantly modified rainfall characteristics in distinct ways. Linear Scaling primarily preserved temporal and spatial structure while adjusting rainfall magnitude, whereas Quantile Mapping improved the distributional representation of rainfall intensities. These differences propagated through hydrological processes, leading to systematic variations in runoff responses across multiple metrics, including water balance consistency, peak magnitude, and timing errors. This suggests that each method performs differently depending on the aspect of system response. Rather than identifying a universally optimal method, the findings highlight trade-offs in how rainfall correction strategies influence hydrological system response. Runoff behavior is interpreted as a process-level indicator of rainfall representation, emphasizing that hydrological consistency depends not only on rainfall accuracy but also on its interaction with model structure. These results suggest a process-oriented perspective for interpreting the role of satellite rainfall products in regulated and monsoon-affected basins. Full article

In this study, the indium (In) composition in amorphous indium gallium zinc oxide (a-IGZO) thin films was systematically varied from 33% to 84% using a sol–gel process. Subsequently, aluminum/IGZO/aluminum (Al/IGZO/Al) metal–semiconductor–metal (MSM) UV photodetectors were fabricated to investigate the influence of composition on the structural, optical, and photoelectric properties. The results indicate that all films maintain an amorphous structure despite the increasing In content, while the ratio of oxygen vacancies, O_vac/(M-O + O_vac), rises from 36% to 52%. Concurrently, the optical bandgap decreases from 2.92 eV to 2.32 eV. Under a bias of 20 V, the dark current increases from 2.11 × 10⁻⁹ A to 1.90 × 10⁻⁵ A as the In content rises. When illuminated by a 360 nm LED with a power density of 8.6 mW/cm², the device with 60% In exhibits a photocurrent-to-dark-current ratio of approximately 10⁴, a responsivity of 19.45 A/W, and a specific detectivity of 8.19 × 10¹² Jones. The response time and recovery time of this device are 39.8 s and 577.4 s, respectively. These findings reveal a competitive relationship between enhanced optical absorption and defect generation induced by In composition, providing valuable guidance for the performance optimization of a-IGZO UV photodetectors through compositional engineering. Full article

Hydrogen production is gaining increasing importance as a key component of the transition toward carbon-neutral energy systems. In this study, the prediction of hydrogen generation in membraneless alkaline water electrolyzers (MAWEs) is investigated using deep-learning-based time-series modeling. A single-input modeling framework is adopted, where only the system current is used as the input variable. Experimental current signals obtained from long-duration tests conducted at electrolyte concentrations between 5 and 35 g KOH (7200 s per experiment) are employed as the model inputs, while mass-based hydrogen production (in grams) is used as the output variable. Two recurrent neural network architectures, namely Long Short-Term Memory (LSTM) and Gated Recurrent Unit (GRU), are implemented, and their predictive performance is comparatively evaluated using RMSE, MAE, and R² metrics. In addition to deep learning models, classical approaches including Linear Regression, ARIMA, and Naïve Forecast are also considered for comparison. The results show that both models are capable of accurately reproducing the hydrogen-production dynamics across the entire concentration range. In particular, the prediction accuracy improves notably at medium and high electrolyte concentrations, where the coefficient of determination (R²) approaches 0.98. The residual distributions remain narrow and symmetric around zero, indicating the absence of systematic estimation bias. The results also show that classical models can achieve comparable performance under stable operating conditions, while deep learning models provide advantages in capturing nonlinear and dynamic behavior. While LSTM and GRU exhibit comparable accuracy, each architecture provides complementary advantages under different operating conditions. These findings indicate that deep-learning-based time-series modeling constitutes a lightweight and reliable framework for prediction and control applications in MAWE systems. Overall, this study demonstrates the applicability of data-driven models for the dynamic characterization of membraneless water electrolysis. Full article

Background: The IQVIA Disease Analyzer (DA) database is a major outpatient electronic health record dataset in Germany. Over recent years, it has been increasingly used to study neurological diseases, comorbidities, treatment patterns, and long-term sequelae. We narratively summarized neurology-related studies using the German IQVIA Disease Analyzer (DA) database published since 2020 and to highlight methodological considerations relevant for interpreting DA-based neurological research. Methods: We conducted a narrative review of DA-based studies published between January 2020 and December 2025. PubMed was searched using DA-related keywords and major neurological disease terms. Eligible articles included peer-reviewed cohort, case–control, or descriptive studies using DA outpatient data. Results: The review identified studies covering epilepsy, cerebrovascular outcomes, Parkinson’s disease, dementia, multiple sclerosis, migraine, and sensory disorders. Most used retrospective cohort or nested case–control designs with regression or propensity score methods. Follow-up durations ranged from 3 to 10 years. Results consistently reflected routine care outpatient diagnostic and prescribing patterns. Discussion: Strengths of DA studies include large patient populations, long follow-up, and detailed prescription information. Limitations include reliance on outpatient ICD-10 coding, lack of detailed neurological phenotyping, and potential residual confounding and bias. Conclusions: DA-based analyses generate clinically relevant routine care evidence on neurological conditions in the German outpatient setting. Proper methodological safeguards and complementary data sources are required to contextualize findings for clinical and epidemiological use. Full article

Objective: This study investigates the prognostic value and clinical utility of the neutrophil percentage-to-albumin ratio (NPAR) in patients with resected non-small-cell lung cancer (NSCLC). Methods: We retrospectively included 335 patients with NSCLC who underwent lung resection at our institution between January 2017 and October 2018. Optimal cutoffs for preoperative and postoperative day 1 (D1) NPAR were determined using X-tile (version 3.6.1; Yale University, New Haven, CT, USA) to define high and low groups. Overall survival (OS) was evaluated using Kaplan–Meier analysis and Cox proportional hazards models. A perioperative NPAR trajectory (low–low, low–high, high–low, high–high) was constructed to characterize dynamic risk patterns. To mitigate potential bias associated with postoperative measurements, a D1 landmark analysis was performed. A nomogram was developed based on the multivariable model and assessed by calibration at 1, 3, and 5 years. Incremental clinical value beyond TNM stage and surgical approach was evaluated using decision curve analysis (DCA), as well as by 5-year continuous net reclassification improvement (NRI) and integrated discrimination improvement (IDI). Results: The optimal cutoffs for preoperative and postoperative D1 NPAR were 14.5 and 23.1, respectively. In univariate analyses, sex, smoking history, preoperative NPAR, postoperative D1 NPAR, pathologic type, TNM stage, surgical approach, and adjuvant therapy were associated with OS (all p < 0.01). In multivariable Cox regression, high preoperative NPAR (HR 1.896, 95% CI 1.135–3.168; p = 0.014) and high postoperative D1 NPAR (HR 1.905, 95% CI 1.097–3.305; p = 0.014) were independent risk factors, along with TNM stage (Stage II: HR 2.824, 95% CI 1.209–6.595; p = 0.016; Stage III: HR 9.470, 95% CI 4.935–18.171; p < 0.001) and open surgery (HR 2.350, 95% CI 1.341–4.117; p = 0.003). Trajectory analysis further stratified risk, with the high–high group showing the poorest survival (adjusted HR 3.48, 95% CI 1.43–8.47; p = 0.006). The association of postoperative NPAR persisted in the D1 landmark analysis (HR 1.836, 95% CI 1.071–3.148; p = 0.027). Adding NPAR to TNM stage and surgical approach improved 5-year risk reclassification (continuous NRI 0.377, 95% CI 0.094–0.659; IDI 0.028, 95% CI −0.002–0.054) and increased net benefit on DCA. The nomogram demonstrated acceptable calibration at 1, 3, and 5 years. Conclusions: This study demonstrates that NPAR serves as an independent prognostic marker for long-term outcomes in patients with NSCLC. The use of NPAR offers clinicians a comprehensive and precise tool for assessing patient prognosis. Full article

Background/Objectives: Measures of excess body fat are often more informative for predicting health risk than body mass index (BMI) alone. With obesity prevalence increasing among young adults, this study evaluated whether adding dominant handgrip strength improves prediction of body fat percentage (BF%) and BF%-defined obesity in a university population. Methods: Cross-sectional data from 895 students (401 women, 494 men; mean age 19.9 years; fall 2015–spring 2016) in Kentucky, USA were analyzed. BMI was calculated from self-reported height and weight. BF% was estimated using bioelectrical impedance analysis (BIA), and dominant handgrip strength was measured with a hydraulic hand grip dynamometer. Sex-specific linear and logistic regression models assessed associations among BMI, grip strength, relative grip strength, and BF%. Results: BMI was a strong predictor of BF% in linear models (R² = 0.74 in women; 0.68 in men). Grip strength alone was not associated with BF% but showed an inverse association when combined with BMI. For BF%-defined obesity, BMI remained the most influential predictor, with grip strength contributing additional predictive value. Among men, age significantly modified these relationships, with marked differences between those aged 18–19 years versus older participants. Conclusions: BMI strongly predicted BF% and BF%-based obesity in this cross-sectional study of a predominantly white young adult population. Incorporating handgrip strength modestly improved classification, particularly among women, suggesting that a functional measure like hand grip strength may enhance obesity screening and health communication in young adults. Full article