Search Results (6,237)

Humpback whales (Megaptera novaeangliae) are well known for their complex acoustic communication, which plays a critical role in social interactions and reproduction. Understanding the variability in humpback whale songs is crucial to deciphering their communication strategies and the factors that influence these changes, which may affect reproductive success and population dynamics. While most studies of humpback whale song behavior have focused on annual variation, intra-seasonal changes remain underexplored. This study investigates intra-seasonal song variation in the Colombian Pacific humpback whale population, a unique and diverse breeding stock. We analyzed 37 h of recordings collected during two distinct periods of the 2019 breeding season (July and August–September) in the northern Colombian Pacific. Song repertoires were compared between periods, and the acoustic structure of a common song unit (Unit1) was analyzed using spectrographic cross-correlation. Results revealed a decrease in repertoire diversity over the course of the season, along with an increase in the song rate and the acoustic consistency of Unit1 during the second period. These findings highlight the dynamic nature of humpback whale song production and suggest potential influences of social learning and hormonal modulation. Such insights may be useful for the conservation and monitoring of humpback whale populations in breeding areas. Full article

Background/Objectives: Deep brain stimulation (DBS) is an effective surgical treatment for Parkinson’s Disease (PD) and other movement disorders. Despite its benefits, DBS explantation occurs in 5.6% of cases, with costs exceeding USD 22,000 per implant. Traditional statistical methods have struggled to identify reliable risk factors for explantation. We hypothesized that supervised machine learning would more effectively capture complex interactions among perioperative factors, enabling the identification of novel risk factors. Methods: The Medical Informatics Operating Room Vitals and Events Repository was queried for patients with DBS, adequate clinical data, and at least two years of follow-up (n = 38). Fisher’s exact test assessed demographic and medical history variables. Data were analyzed using Anaconda Version 2.3.1. with pandas, numpy, sklearn, sklearn-extra, matplotlin. pyplot, and seaborn. Recursive feature elimination with cross-validation (RFECV) optimized factor selection was used. A multivariate logistic regression model was trained and evaluated using precision, recall, F1-score, and area under the curve (AUC). Results: Fisher’s exact test identified chronic pain (p = 0.0108) and tobacco use (p = 0.0026) as risk factors. RFECV selected 24 optimal features. The logistic regression model demonstrated strong performance (precision: 0.89, recall: 0.86, F1-score: 0.86, AUC: 1.0). Significant risk factors included tobacco use (OR: 3.64; CI: 3.60–3.68), primary PD (OR: 2.01; CI: 1.99–2.02), ASA score (OR: 1.91; CI: 1.90–1.92), chronic pain (OR: 1.82; CI: 1.80–1.85), and diabetes (OR: 1.63; CI: 1.62–1.65). Conclusions: Our study suggests that supervised machine learning can identify risk factors for early DBS explantation. Larger studies are needed to validate our findings. Full article

Urban electricity power consumption (EPC) prediction plays a crucial role in urban management and sustainable development. Nighttime light (NTL) remote sensing imagery has demonstrated significant potential in estimating urban EPC due to its strong correlation with human activities and energy use. However, most existing models focus on annual-scale estimations, limiting their ability to capture month-scale EPC. To address this limitation, a novel monthly EPC prediction model that incorporates monthly average temperature, and the interaction between NTL data and temperature was proposed in this study. The proposed method was applied to cities within the Yangtze River Delta (YRD) urban agglomeration, and was validated using datasets constructed from NPP/VIIRS and SDGSAT-1 satellite imageries, respectively. For the NPP/VIIRS dataset, the proposed method achieved a Mean Absolute Relative Error (MARE) of 7.96% during the training phase (2017–2022) and of 10.38% during the prediction phase (2023), outperforming the comparative methods. Monthly EPC spatial distribution maps from VPP/VIIRS data were generated, which not only reflect the spatial patterns of EPC but also clearly illustrate the temporal evolution of EPC at the spatial level. Annual EPC estimates also showed superior accuracy compared to three comparative methods, achieving a MARE of 7.13%. For the SDGSAT-1 dataset, leave-one-out cross-validation confirmed the robustness of the model, and high-resolution (40 m) monthly EPC maps were generated, enabling the identification of power consumption zones and their spatial characteristics. The proposed method provides a timely and accurate means for capturing monthly EPC dynamics, effectively supporting the dynamic monitoring of urban EPC at the monthly scale in the YRD urban agglomeration. Full article

Polyvinyl alcohol (PVA) hydrogels have emerged as versatile materials due to their exceptional biocompatibility and tunable mechanical properties, showing great promise for flexible sensors, smart wound dressings, and tissue engineering applications. However, rational design remains challenging due to complex structure–property relationships involving multiple formulation parameters. This study presents an interpretable machine learning framework for predicting PVA hydrogel tensile strain properties with emphasis on mechanistic understanding, based on a comprehensive dataset of 350 data points collected from a systematic literature review. XGBoost demonstrated superior performance after Optuna-based optimization, achieving R² values of 0.964 for training and 0.801 for testing. SHAP analysis provided unprecedented mechanistic insights, revealing that PVA molecular weight dominates mechanical performance (SHAP importance: 84.94) through chain entanglement and crystallization mechanisms, followed by degree of hydrolysis (72.46) and cross-linking parameters. The interpretability analysis identified optimal parameter ranges and critical feature interactions, elucidating complex non-linear relationships and reinforcement mechanisms. By addressing the “black box” limitation of machine learning, this approach enables rational design strategies and mechanistic understanding for next-generation multifunctional hydrogels. Full article

Land–sea segmentation is a fundamental task in remote sensing image analysis, and plays a vital role in dynamic coastline monitoring. The complex morphology and blurred boundaries of coastlines in remote sensing imagery make fast and accurate segmentation challenging. Recent deep learning approaches lack the ability to model spatial continuity effectively, thereby limiting a comprehensive understanding of coastline features in remote sensing imagery. To address this issue, we have developed VMMT-Net, a novel dual-branch semantic segmentation framework. By constructing a parallel heterogeneous dual-branch encoder, VMMT-Net integrates the complementary strengths of the Mix Transformer and the Visual State Space Model, enabling comprehensive modeling of local details, global semantics, and spatial continuity. We design a Cross-Branch Fusion Module to facilitate deep feature interaction and collaborative representation across branches, and implement a customized decoder module that enhances the integration of multiscale features and improves boundary refinement of coastlines. Extensive experiments conducted on two benchmark remote sensing datasets, GF-HNCD and BSD, demonstrate that the proposed VMMT-Net outperforms existing state-of-the-art methods in both quantitative metrics and visual quality. Specifically, the model achieves mean F1-scores of 98.48% (GF-HNCD) and 98.53% (BSD) and mean intersection-over-union values of 97.02% (GF-HNCD) and 97.11% (BSD). The model maintains reasonable computational complexity, with only 28.24 M parameters and 25.21 GFLOPs, striking a favorable balance between accuracy and efficiency. These results indicate the strong generalization ability and practical applicability of VMMT-Net in real-world remote sensing segmentation tasks. Full article

Groundwater systems are intrinsically linked to climate, with changing conditions significantly altering recharge, storage, and discharge processes, thereby impacting water availability and ecosystem integrity. Critical knowledge gaps persist regarding groundwater equilibrium timescales, water table dynamics, and their governing factors. This study develops a novel remote sensing framework to quantify factor controls on groundwater–climate interaction characteristics in the Heihe River Basin (HRB). High-resolution (0.005° × 0.005°) maps of groundwater response time (GRT) and water table ratio (WTR) were generated using multi-source geospatial data. Employing Geographical Convergent Cross Mapping (GCCM), we established causal relationships between GRT/WTR and their drivers, identifying key influences on groundwater dynamics. Generalized Additive Models (GAM) further quantified the relative contributions of climatic (precipitation, temperature), topographic (DEM, TWI), geologic (hydraulic conductivity, porosity, vadose zone thickness), and vegetative (NDVI, root depth, soil water) factors to GRT/WTR variability. Results indicate an average GRT of ~6.5 × 10⁸ years, with 7.36% of HRB exhibiting sub-century response times and 85.23% exceeding 1000 years. Recharge control dominates shrublands, wetlands, and croplands (WTR < 1), while topography control prevails in forests and barelands (WTR > 1). Key factors collectively explain 86.7% (GRT) and 75.9% (WTR) of observed variance, with spatial GRT variability driven primarily by hydraulic conductivity (34.3%), vadose zone thickness (13.5%), and precipitation (10.8%), while WTR variation is controlled by vadose zone thickness (19.2%), topographic wetness index (16.0%), and temperature (9.6%). These findings provide a scientifically rigorous basis for prioritizing groundwater conservation zones and designing climate-resilient water management policies in arid endorheic basins, with our high-resolution causal attribution framework offering transferable methodologies for global groundwater vulnerability assessments. Full article

This study examines the complex interactions among soil moisture, evaporation, extreme weather events, and lightning, and their influence on fire activity across the extratropical and Pan-Arctic regions. Leveraging reanalysis and remote-sensing datasets from 2000 to 2020, we applied cross-correlation analysis, a modified Mann–Kendall trend test, and assessments of interannual variability to key variables including soil moisture, fire frequency and risk, evaporation, and lightning. Results indicate a significant increase in dry days (up to 40%) and heatwave events across Central Eurasia and Siberia (up to 50%) and Alaska (25%), when compared to the 1980–2000 baseline. Upward trends have been detected in evaporation across most of North America, consistent with soil moisture trends, while much of Eurasia exhibits declining soil moisture. Fire danger shows a strong positive correlation with evaporation north of 60° N (r ≈ 0.7, p ≤ 0.005), but a negative correlation in regions south of this latitude. These findings suggest that in mid-latitude ecosystems, fire activity is not solely driven by water stress or atmospheric dryness, highlighting the importance of region-specific surface–atmosphere interactions in shaping fire regimes. In North America, most fires occur in temperate grasslands, savannas, and shrublands (47%), whereas in Eurasia, approximately 55% of fires are concentrated in forests/taiga and temperate open biomes. The analysis also highlights that lightning-related fires are more prevalent in Eastern Europe and Southeastern Asia. In contrast, Western North America exhibits high fire incidence in temperate conifer forests despite relatively low lightning activity, indicating a dominant role of anthropogenic ignition. These findings underscore the importance of understanding land–atmosphere interactions in assessing fire risk. Integrating surface conditions, climate extremes, and ignition sources into fire prediction models is crucial for developing more effective wildfire prevention and management strategies. Full article

To solve the issue that existing yield prediction methods do not fully capture the interaction between multiple factors, we propose a winter wheat yield prediction framework with triple cross-attention for multi-source data fusion. This framework consists of three modules: a multi-source data processing module, a multi-source feature fusion module, and a yield prediction module. The multi-source data processing module collects satellite, climate, and soil data based on the winter wheat planting range, and constructs a multi-source feature sequence set by combining statistical data. The multi-source feature fusion module first extracts deeper-level feature information based on the characteristics of different data, and then performs multi-source feature fusion through a triple cross-attention fusion mechanism. The encoder part in the production prediction module adds a graph attention mechanism, forming a dual branch with the original multi-head self-attention mechanism to ensure the capture of global dependencies while enhancing the preservation of local feature information. The decoder section generates the final predicted output. The results show that: (1) Using 2021 and 2022 as test sets, the mean absolute error of our method is 385.99 kg/hm², and the root mean squared error is 501.94 kg/hm², which is lower than other methods. (2) It can be concluded that the jointing-heading stage (March to April) is the most crucial period affecting winter wheat production. (3) It is evident that our model has the ability to predict the final winter wheat yield nearly a month in advance. Full article

Compared to other multi-source image fusion tasks, visible and SAR image fusion faces a lack of training data in deep learning-based methods. Introducing structural priors to design fusion networks is a viable solution. We incorporated the feature hierarchy concept from computer vision, dividing deep features into low-, mid-, and high-level tiers. Based on the complementary modal characteristics of SAR and visible, we designed a fusion architecture that fully analyze and utilize the difference of hierarchical features. Specifically, our framework has two stages. In the cross-modal enhancement stage, a CycleGAN generator-based method for cross-modal interaction and input data enhancement is employed to generate pseudo-modal images. In the fusion stage, we have three innovations: (1) We designed feature extraction branches and fusion strategies differently for each level based on the features of different levels and the complementary modal features of SAR and visible to fully utilize cross-modal complementary features. (2) We proposed the Layered Strictly Nested Framework (LSNF), which emphasizes hierarchical differences and uses hierarchical characteristics, to reduce feature redundancy. (3) Based on visual saliency theory, we proposed a Gradient-weighted Pixel Loss (GWPL), which dynamically assigns higher weights to regions with significant gradient magnitudes, emphasizing high-frequency detail preservation during fusion. Experiments on the YYX-OPT-SAR and WHU-OPT-SAR datasets show that our method outperforms 11 state-of-the-art methods. Ablation studies confirm each component’s contribution. This framework effectively meets remote sensing applications’ high-precision image fusion needs. Full article

Background: The mechanisms linking loneliness to frailty in older adults with diabetes remain unclear. Guided by the Loneliness–Health Outcomes Model, this study is the first to simultaneously validate the dual mechanisms (mediation and moderation) of social support in the loneliness–frailty relationship among older Chinese adults with diabetes. Methods: A cross-sectional study enrolled 442 community-dwelling adults aged ≥60 years with type 2 diabetes in Beijing. Standardized scales assessed loneliness (UCLA Loneliness Scale), frailty (Tilburg Frailty Indicator), and social support (SSRS). Analyses included Pearson’s correlations, hierarchical regression, and PROCESS macro to evaluate mediating/moderating effects, after adjusting for demographics and comorbidities. Results: The frailty prevalence was 55.2%. Loneliness was positively correlated with frailty (r = 0.327, p < 0.01), while social support showed inverse associations with both loneliness (r = −0.496) and frailty (r = −0.315) (p < 0.01). Social support partially mediated loneliness’s effect on frailty (indirect effect: 30.86%; 95% CI: 0.028–0.087) and moderated this relationship (interaction β = −0.003, p = 0.011). High-risk clusters (e.g., aged ≥80 years, widowed, and isolated individuals) exhibited combined “high loneliness–low support–high frailty” profiles. Conclusions: Social support reduces the frailty risk through dual mechanisms. These findings advocate for tiered clinical interventions: (1) targeted home-visit systems and resource allocation for high-risk subgroups (e.g., solo-living elders aged ≥80 years); and (2) the integration of social support screening into routine diabetes care to identify individuals below the protective threshold (SSRS < 45.47). These findings advance psychosocially informed strategies for diabetes management in aging populations. Full article

Background/Objectives: Gastric cancer (GC) remains a significant global health burden due to its high mortality rate and frequent diagnosis at advanced stages. This study aimed to identify reliable diagnostic biomarkers and elucidate molecular mechanisms underlying GC by integrating transcriptomic data from independent platforms and applying machine learning techniques. Methods: Two transcriptomic datasets from the Gene Expression Omnibus were analyzed: GSE26899 (microarray, n = 108) as the discovery dataset and GSE248612 (RNA-seq, n = 12) for validation. Differential expression analysis was conducted using limma and DESeq2, selecting genes with |log2FC| > 1 and adjusted p < 0.05. The top 200 differentially expressed genes (DEGs) were used to develop machine learning models (random forest, logistic regression, neural networks). Functional enrichment analyses (GO, KEGG, Hallmark) were applied to explore relevant biological pathways. Results: In GSE26899, 627 DEGs were identified (201 upregulated, 426 downregulated), with key genes including CST1, KIAA1199, TIMP1, MSLN, and ATP4A. The random forest model demonstrated excellent classification performance (AUC = 0.952). GSE248612 validation yielded 738 DEGs. Cross-platform comparison confirmed 55.6% concordance among core genes, highlighting CST1, TIMP1, KRT17, ATP4A, CHIA, KRT16, and CRABP2. Enrichment analyses revealed involvement in ECM–receptor interaction, PI3K-Akt signaling, EMT, and cell cycle. Conclusions: This integrative transcriptomic and machine learning framework effectively identified high-confidence biomarkers for GC. Notably, CST1, TIMP1, KRT16, and ATP4A emerged as consistent, biologically relevant candidates with strong diagnostic performance and potential clinical utility. These findings may aid early detection strategies and guide future therapeutic developments in gastric cancer. Full article

This study investigated the effects of curdlan (CUR) with different helical conformations on the gelling behavior and mechanisms of heat-induced soy protein isolate (SPI) gels. The results demonstrated that CUR significantly improved the functional properties of SPI gels, including water-holding capacity (0.31–5.06% increase), gel strength (7.01–240.51% enhancement), textural properties, viscoelasticity, and thermal stability. The incorporation of CUR facilitated the unfolding and cross-linking of SPI molecules, leading to enhanced network formation. Notably, SPI composite gels containing CUR with an ordered triple-helix bundled structure exhibited superior gelling performance compared to other helical conformations, characterized by a more compact and uniform microstructure. This improvement was attributed to stronger hydrogen bonding interactions between the triple-helix CUR and SPI molecules. Furthermore, the entanglement of triple-helix CUR with SPI promoted the formation of a denser and more homogeneous interpenetrating polymer network. These findings indicate that triple-helix CUR is highly effective in optimizing the gelling characteristics of heat-induced SPI gels. This study provides new insights into the structure–function relationship of CUR in SPI-based gel systems, offering potential strategies for designing high-performance protein–polysaccharide composite gels. The findings establish a theoretical foundation for applications in the food industry. Full article

Accurate prediction of drug–drug interactions (DDIs) is essential for ensuring medication safety and optimizing combination-therapy strategies. However, existing DDI models face limitations in handling interactions related to stereochemistry and precisely locating drug interaction sites. These limitations reduce the prediction accuracy for conformation-dependent interactions and the interpretability of molecular mechanisms, potentially posing risks to clinical safety. To address these challenges, we introduce LSA-DDI, a Spatial-Contrastive-Attention-Based Drug–Drug Interaction framework. Our 3D feature extraction method captures the spatial structure of molecules through three features—coordinates, distances, and angles—and fuses them to enhance the model of molecular spatial structures. Concurrently, we design and implement a Dynamic Feature Exchange (DFE) mechanism that dynamically regulates the flow of information across modalities via an attention mechanism, achieving bidirectional enhancement and semantic alignment of 2D topological and 3D spatial structure features. Additionally, we incorporate a dynamic temperature-regulated multiscale contrastive learning framework that effectively aligns multiscale features and enhances the model’s generalizability. Experiments conducted on public drug databases under both warm-start and cold-start scenarios demonstrated that LSA-DDI achieved competitive performance, with consistent improvements over existing methods. Full article

The remote sensing monitoring of land use changes and future scenario simulation hold crucial significance for accurately characterizing urban expansion patterns, optimizing urban land use configurations, and thereby promoting coordinated regional development. Through the integration of multi-source data, this study systematically analyzes the spatiotemporal trajectories and driving forces of land use changes in the Pearl River Delta urban agglomeration (PRD) from 1990 to 2020 and further simulates the spatial patterns of urban land use under diverse development scenarios from 2025 to 2035. The results indicate the following: (1) During 1990–2020, urban expansion in the Pearl River Delta urban agglomeration exhibited a “stepwise growth” pattern, with an annual expansion rate of 3.7%. Regional land use remained dominated by forest (accounting for over 50%), while construction land surged from 6.5% to 21.8% of total land cover. The gravity center trajectory shifted southeastward. Concurrently, cropland fragmentation has intensified, accompanied by deteriorating connectivity of ecological lands. (2) Urban expansion in the PRD arises from synergistic interactions between natural and socioeconomic drivers. The Geographically and Temporally Weighted Regression (GTWR) model revealed that natural constraints—elevation (regression coefficients ranging −0.35 to −0.05) and river network density (−0.47 to −0.15)—exhibited significant spatial heterogeneity. Socioeconomic drivers dominated by year-end paved road area (0.26–0.28) and foreign direct investment (0.03–0.11) emerged as core expansion catalysts. Geographic detector analysis demonstrated pronounced interaction effects: all factor pairs exhibited either two-factor enhancement or nonlinear enhancement effects, with interaction explanatory power surpassing individual factors. (3) Validation of the Patch-generating Land Use Simulation (PLUS) model showed high reliability (Kappa coefficient = 0.9205, overall accuracy = 95.9%). Under the Natural Development Scenario, construction land would exceed the ecological security baseline, causing 408.60 km² of ecological space loss; Under the Ecological Protection Scenario, mandatory control boundaries could reduce cropland and forest loss by 3.04%, albeit with unused land development intensity rising to 24.09%; Under the Economic Development Scenario, cross-city contiguous development zones along the Pearl River Estuary would emerge, with land development intensity peaking in Guangzhou–Foshan and Shenzhen–Dongguan border areas. This study deciphers the spatiotemporal dynamics, driving mechanisms, and scenario outcomes of urban agglomeration expansion, providing critical insights for formulating regionally differentiated policies. Full article

Emotion recognition by wearable devices is essential for advancing emotion-aware human–computer interaction in real life. Earphones have the potential to naturally capture brain activity and its lateralization, which is associated with emotion. In this study, we newly introduced tympanic membrane temperature (TMT), previously used as an index of lateralized brain activation, for earphone-based emotion recognition. We developed custom earphones to measure bilateral TMT and conducted two experiments consisting of emotion induction by autobiographical recall and scenario imagination. Using features derived from the right–left TMT difference, we trained classifiers for both four-class discrete emotion and valence (positive vs. negative) classification tasks. The classifiers achieved 36.2% and 42.5% accuracy for four-class classification and 72.5% and 68.8% accuracy for binary classification, respectively, in the two experiments, confirmed by leave-one-participant-out cross-validation. Notably, consistent improvement in accuracy was specific to models utilizing right–left TMT and not observed in models utilizing the right–left wrist skin temperature. These findings suggest that lateralization in TMT provides unique information about emotional state, making it valuable for emotion recognition. With the ease of measurement by earphones, TMT has significant potential for real-world application of emotion recognition. Full article