Search Results (128)

The expansion of fifth-generation (5G) wireless networks has increased environmental exposure to mid-band and millimeter-wave radiofrequency electromagnetic fields (RF-EMF), but their molecular effects on male reproductive tissues remain insufficiently understood. This study evaluated whether repeated exposure to 3.5 GHz and 24 GHz RF-EMF alters testicular stress-associated molecular responses by integrating electromagnetic dosimetry with an in vivo rat model. Whole-body specific absorption rate (SAR) and 10 g peak SAR were estimated using a rat voxel model and scaled to the 20 cm antenna-to-cage geometry used during exposure. Thirty-six adult male Sprague Dawley rats were allocated to sham, 3.5 GHz, or 24 GHz groups and exposed for 1 h/day or 7 h/day over 60 days. Testes were examined histologically and assessed for HSP27, HSP70, and HSP90 protein expression. SAR values were low overall, although absorption was higher at 3.5 GHz than at 24 GHz. Histological evaluation showed preserved seminiferous tubule architecture without consistent structural injury. In contrast, molecular analysis demonstrated frequency- and duration-dependent modulation of heat shock proteins, including early HSP70 downregulation at both frequencies, followed by HSP90 upregulation at 3.5 GHz and HSP27 upregulation at 24 GHz. These findings indicate that low-level 5G-relevant RF-EMF exposure can modify molecular stress responses in testicular tissue even in the absence of overt histological damage. Full article

To support the development and computer simulation of radio frequency (RF) and microwave (MW) thawing processes, this study characterized the dielectric properties and penetration depth of chicken breast across a frequency range of 10–3000 MHz and temperatures from −20 °C to 10 °C. The influence of three RF anode voltages and four MW power levels on heating rates was also evaluated. Results showed that both the dielectric constant and loss factor decreased with increasing frequency, with the most significant reduction occurring between 10 and 60 MHz. In contrast, these properties increased with temperature, exhibiting a sharp rise during the phase transition zone (−5 to 0 °C). Penetration depth decreased with frequency and was consistently higher under RF than MW exposure. High-precision regression models (R² > 0.97) were established to describe these relationships. RF heating achieved more uniform temperature distribution compared to MW, which showed pronounced center-corner temperature differences. By integrating experimental measurements with mathematical modeling, this work provides key insights and reliable data for optimizing RF and MW thawing strategies in industrial applications. Full article

Despite concern over air pollution, fine-scale spatial and demographic disparities in exposure remain largely unquantified in Australian cities due to sparse monitoring and coarse models. In Greater Melbourne, this gap limits neighbourhood-level assessment of PM_2.5 exposure and associated environmental inequalities. To address this gap, we integrated 6-month averaged PM_2.5 observations (October 2023 to March 2024) from 5 regulatory monitoring stations and 13 low-cost sensors (LCSs) to develop a land use regression (LUR) model estimating concentrations at a 100 m resolution. These estimates were used to calculate population-weighted PM_2.5 exposure (PWE) at the mesh block level across Melbourne. To examine factors associated with spatial heterogeneity in PWE, we applied a hybrid modeling framework combining Spatially Explicit Random Forest (Spatial-RF) and Geographically Weighted Regression (GWR), incorporating physical, built-environment, and socio-demographic variables from the Synthesized Multi-Dimensional Environmental Exposure Database (SEED). The Spatial-RF model initially exhibited an R² of 0.56. After multicollinearity diagnostics using the Variance Inflation Factor (VIF), three key explanatory variables were selected for GWR modeling: the Normalized Difference Vegetation Index (NDVI), the Index of Education and Occupation (IEO), and the proportion of culturally and linguistically diverse populations (CALDP). The developed GWR model achieved higher model performance (R² = 0.65) than Spatial-RF and global Ordinary Least Squares (OLS) regression (R² = 0.38), revealing strong spatial non-stationarity. Results show that PWE generally ranged from 5 to 7 µg/m³, exceeding the 2021 WHO air quality guideline, with hotspots in the urban core and along major transport corridors. Elevated exposure occurred in both socioeconomically disadvantaged areas and residents in urban centers with higher socio-economic status, reflecting complex, spatially contingent exposure inequalities. These findings support fine-scale, equity-oriented air quality management. Full article

Background: Falls are a leading cause of injury and mortality worldwide. Higher physical activity levels in young adults may increase exposure to fall-related situations. Understanding their neuromuscular adaptations is critical for balance control research and perturbation-based training. This study examined proactive and reactive adaptations in lower-limb muscle activity during repeated simulated trips among young adults. Methods: Twenty participants experienced five treadmill-induced standing-trips. Bilateral electromyography (EMG) activities of the rectus femoris (RF), vastus lateralis (VL), tibialis anterior (TA), medial gastrocnemius (MG), and biceps femoris (BF) were recorded. Muscle activity magnitude at perturbation onset (ON), EMG peak amplitude, and time-to-peak from ON were extracted and compared across trials. Results: Proactive activation at ON increased across trials in TA and RF on the recovery side (p = 0.012–0.023) and in TA, VL, and BF on the stance side (p = 0.002–0.034). Reactive peak amplitudes decreased in RF, VL, and BF on the recovery side (p < 0.001–0.014) and in RF, VL, and BF on the stance side (p < 0.001–0.016). Time-to-peak shortened in MG, RF, VL, and BF on the recovery side (p < 0.001–0.030) and in RF, VL, TA, and BF on the stance side (p < 0.001–0.050). Conclusions: Repeated simulated trips elicited proactive adaptations in muscle activity and reactive changes in time-to-peak, which may suppress the need for increased reactive muscle activations to recover balance post-perturbation over trials in young adults. The findings augment our understanding of the intercorrelation between proactive and reactive adaptations to repeated perturbations. Full article

Assessment of landslide risk is crucial given the substantial related economic losses and infrastructure damage in mountain areas every year. Particularly, the Sagarmatha National Park (SNP), a key destination for Himalayan glacier tourism, remains relatively understudied in this context. Existing studies primarily focus on regional inventories or simply inventory landslides and lack tourism-specific hazard assessment. This study evaluates landslide distribution, its controlling factors, and the exposure of infrastructure to varying degrees of landslide susceptibility in SNP. A blind inventory of 680 landslides and twelve conditioning factors, including six topographic and six non-topographic variables, were analyzed using Frequency Ratio (FR), Logistic Regression (LR), and Random Forest (RF) models. In addition, spatial overlay analysis was employed to assess the degree of infrastructure exposure. Results indicate that Land Surface Temperature (LST) is the most dominant factor influencing landslides occurrence, followed by rainfall, elevation, and slope, along with specific aspects like south and west and, land cover class like Barren land and Alpine meadows. Random Forest achieved the highest predictive accuracy (91%), outperforming both Logistic Regression (87%) and Frequency Ratio (84%). Exposure assessment of key tourism infrastructure indicates that trekking routes, helipads, buildings, campsites, and bridges are subject to varying levels of landslide risk. Although only 2.73 km (0.52%) of trekking routes intersect active landslide scars, 147 km (28%) lie within high-exposure zones. Consequently, both typical and paraglacial landslides threaten access to glacier tourism destinations, highlighting significant implications for Nepal’s tourism. Full article

Nowadays, air quality is a major issue, especially in large cities. Apart from air pollution, particulate matter (PM), especially PM2.5, poses serious health risks to individuals with respiratory conditions. Accurate forecasting of PM levels is crucial to warn vulnerable populations and reduce exposure. Machine learning models can effectively predict PM concentrations based on historical data and barometric conditions such as temperature and humidity. Such predictions can support timely public health interventions and environmental policy decisions. The selection of the optimal machine learning model for time series forecasting requires a careful balance between predictive accuracy and computational efficiency. This study evaluates a number of widely used models, such as Random Forest (RF), Long Short-Term Memory (LSTM), Convolutional Neural Network-LSTM (CNN–LSTM), Extreme Gradient Boosting (XGB/HistGradientBoosting), and hybrid approaches (LSTM embeddings + RF), in the context of time series forecasting for particulate matter (PM) concentrations. Performance is assessed using three key error metrics: Mean Squared Error (MSE), Root Mean Squared Error (RMSE), and Mean Absolute Scaled Error (MASE). Additionally, the computational demands and development complexity of each model are analyzed. The overall results are of great interest for each application model, and in more detail, it is shown that the best compromise between accuracy and efficiency can be achieved, while a corresponding prediction model with satisfactory predictive performance can be implemented. The results show that CNN–LSTM and hybrid approaches provide high accuracy, while tree-based models are computationally efficient, offering practical options for real-time forecasting systems. Full article

Air pollution remains a major public-health concern, and exposure to particulate matter (PM), particularly PM₁₀ (with a diameter ≤ 10 µm), is associated with adverse respiratory and cardiovascular outcomes. Most research relies on a singular model for PM₁₀ surface estimation. This study is an assessment of a national-scale, daily PM₁₀ estimation framework for Lithuania (2019–2024), using a hybrid machine-learning method that combines Random Forest (RF) and extreme gradient boosting (XGBoost) algorithms. Hourly PM₁₀ observations were aggregated from 18 monitoring stations to obtain daily means and temporal means. The predictors integrated meteorological factors, such as temperature, wind, humidity, and precipitation, to determine satellite-based atmospheric composition from Sentinel-5P Tropospheric Monitoring Instruments (TROPOMI). Atmospheric components include nitrogen dioxide (NO₂), carbon monoxide (CO), sulfur dioxide (SO₂), ozone (O₃), formaldehyde (HCHO), and the absorbing aerosol index (AI). Moderate-Resolution Imaging Spectroradiometers (MODIS) were used to record land-surface temperature and static spatial descriptors, such as elevation, land cover, Normalized Difference Vegetation Index (NDVI), population, and road proximity. The dataset was partitioned temporally into training (70%), validation (20%), and testing (10%). The hybrid model achieved an improved accuracy, compared with single-model baselines, reaching a coefficient of determination (R²) of 0.739 in validation and R² = 0.75 in the tested dataset. Mean absolute error (MAE) was 3.15 µg/m³, and root mean square error (RMSE) was 3.98 µg/m³. The results indicate a slight tendency to overestimate PM₁₀ concentrations at lower concentration levels. Feature-importance analysis revealed that short-term temporal persistence is the key to daily PM₁₀ prediction, while meteorological variables provide secondary contributions. Temporal evaluation, using consecutive two-year windows, revealed a consistent improvement in predictive performance from 2019–2020 to 2023–2024, while station-level analysis showed moderate-to-strong agreement between the predicted and observed PM₁₀ concentrations across monitoring stations, with R² ranging from 0.455 to 0.760. This provides decision-support capabilities for air-quality management, the evaluation of mitigation measures, and integration of air-pollution considerations into sustainable urban planning strategies assessing public-health protection. Full article

Background: Rheumatoid arthritis (RA) is a chronic, systemic autoimmune disease characterized by persistent synovial inflammation and progressive joint destruction. Despite the implementation of the treat-to-target (T2T) strategy and the introduction of several classes of biologic and targeted synthetic disease-modifying antirheumatic drugs (b/tsDMARDs), a considerable proportion of patients continues to exhibit active, refractory disease. In 2021, the European Alliance of Associations for Rheumatology (EULAR) defined this condition as Difficult-to-Treat Rheumatoid Arthritis (D2T-RA). This study aimed to identify clinical, laboratory, and therapeutic factors associated with D2T-RA. Methods: A total of 344 patients with established RA were retrospectively evaluated. Among them, 164 fulfilled the 2021 EULAR criteria for D2T-RA (D2T group), while 180 did not (NO-D2T group). Clinical (age, sex, disease duration, BMI, smoking, comorbidities), laboratory (RF, ACPA, ESR, CRP), clinimetric (DAS28, CDAI, PhGA, PGA, HAQ), and therapeutic data (glucocorticoid use, methotrexate treatment and dose, monotherapy, advanced therapy exposure, number of failed advanced therapies, current DMARD regimen) were analyzed. Results: Factors significantly associated with D2T-RA included female sex, longer disease duration, higher RF and ACPA titers, elevated ESR levels, glucocorticoid therapy, and a greater number of failed advanced therapies. Although both groups achieved low disease activity or remission by DAS28 and CDAI, JAK inhibitors—particularly Filgotinib and Upadacitinib—were significantly more common in the D2T cohort and appeared associated with clinical stabilization. Conclusions: This study strengthens the understanding of the predictive profile of D2T-RA, confirming the role of disease chronicity and persistent inflammation in the development of treatment resistance. Importantly, the observed trend toward clinical stabilization achieved under JAK inhibitor therapy reinforces their potential to address unmet therapeutic needs in D2T-RA, providing a mechanistically grounded strategy for patients refractory to conventional and biologic DMARDs. Full article

Traditional chemical analysis of plant tissue is time-consuming, costly, and poses risks due to exposure to toxic gases, highlighting the need for faster, low-cost, and safer alternatives. Vis-NIR spectroscopy, combined with machine learning, offers a promising method for estimating leaf nutrient levels without chemical reagents. This study evaluated the potential of Vis-NIR spectroscopy for nutrient estimation in leaf samples of banana (n = 363), mango (n = 239), and grapevine (n = 336) by applying spectral pre-processing techniques—smoothing (SMO) and first derivative Savitzky–Golay (SGD1d) alongside two machine learning methods: Partial Least Squares Regression (PLSR) and Random Forest (RF). Plant tissue samples were analyzed using sulfuric and nitroperchloric wet digestion and hyperspectral sensors. The prediction models were assessed using concordance correlation coefficient (CCC) and mean squared error (MSE). The highest accuracy (CCC > 0.80 and MSE < 2 g kg⁻¹) was achieved for Ca in banana, P in mango, and N and Ca in grapevine across both machine learning methods and pre-processing techniques. The predictive models calibrated for ‘Grapevine’ exhibited the highest accuracy—characterized by higher CCC values and lower MSE values—when compared with the models developed for ‘Mango’ and ‘Banana’. Models using SMO and SGD1d showed better performance than those using raw spectra (RAW). The high amplitudes and variations in nutrient levels, combined with large standard deviations, negatively affected the predictive performance of the models. Full article

The deployment of 5G networks has transformed the landscape of radiofrequency electromagnetic field (RF-EMF) exposure patterns, shifting from high-power macro base stations to dense networks of small, beamforming cells. This review critically assesses the validity, challenges, and research gaps of low-cost RF-EMF sensors used for 5G exposure monitoring. An analysis of over 60 studies covering Sub-6 GHz and emerging mmWave systems shows that well-calibrated sensors can achieve measurement deviations of ±3–6 dB compared to professional instruments like the Narda SRM-3006, with long-term calibration drift less than 0.5 dB per month and RMS reproducibility around 5%. Typical outdoor 5G FR1 exposure levels range from 0.01 to 0.5 W/m² near small cells, while personal device use can cause transient exposures 10–30 dB higher. Although mmWave (24–100 GHz) and Wi-Fi 7/8 (~60 GHz) are underrepresented due to antenna and component limitations, Sub-6 GHz sensing platforms, including software-defined radio (SDR)-based and triaxial isotropic designs, provide sufficient sensitivity for both citizen and institutional monitoring. Major challenges involve calibration drift, frequency band gaps, data interoperability, and ethical management of participatory networks. Addressing these issues through standardized calibration protocols, machine learning-assisted drift correction, and open data frameworks will allow affordable sensors to complement professional monitoring, improve spatial coverage, and enhance public transparency in 5G RF-EMF exposure governance. Full article

Background: Total hip arthroplasty (THA) is a well-tolerated and effective procedure that can improve a patient’s mobility and quality of life. A main concern, however, is the release of metal ions into the body due to wear and corrosion. Commonly reported ions are Co and Cr, while others, such as Ti, Mo, and Ni, are less frequently studied. The objective of this study was to characterize compartmentalization and time-dependent ion behaviors across serum, whole blood, and urine after hip prosthetic implantation. The goal of using Random Forest (RF) was to determine whether machine learning modeling could support temporal trends across data. Methods: Data was gathered from the literature of clinical studies, and we conducted a pooled analysis of the temporal kinetics from cohorts of patients who received hip prosthetics. Mean ion concentrations were normalized to µg/L across each fluid and weighted by cohort sample size. RF was used as a study-level test of predictive accuracy across ions. Results: For serum and whole blood, Co and Cr displayed one-phase association models, while Ti showed an exponential rise and decay. Ions typically rose quickly within the first 24 months postoperatively. Serum Co and whole blood had similar patterns, tapering off just under 2 µg/L, but serum Cr (~2.02 µg/L) was generally higher than that of whole blood (~0.99 µg/L). Mean urinary Co levels were greater than those of Cr, suggesting a larger, freely filterable fraction for Co. RF was implemented to determine predictive accuracy for each ion, showing a stronger fit for Co (R² = 0.86, RMSE = 0.57) compared to Cr (R² = 0.52, RMSE = 0.50). Conclusions: Sub-threshold exposure was prevalent across cohorts. Serum and whole blood Co and Cr displayed distinct kinetic profiles and, if validated, could support fluid-specific monitoring strategies. We present a methodology for interpreting ion kinetics and show potential for machine learning applications in postoperative monitoring. Full article

Due to widespread industrial applications and increased discharges, concentrations of polyfluoroalkyl substances (PFAS) in potential drinking water sources have risen significantly, putting more people at risk of PFAS exposure. This study aimed to systematically clarify the occurrence characteristics (concentrations, detection frequencies, and temporal trends) of PFAS in global potential drinking water sources over the past decade, assess their oral exposure risks, and identify key PFAS species with high detection frequencies, high contamination levels, or high toxicity risks, thereby providing scientific support for the development of targeted control technologies and management strategies. This study systematically searched and reviewed the relevant literature published between 2014 and 2024 on PFAS levels in global potential drinking water sources, extracting data on PFAS concentrations, detection information, and sampling characteristics. Using the U.S. Environmental Protection Agency (EPA) Reference Dose (RfD) for oral exposure as the Acceptable Daily Intake (ADI), we evaluated the exposure risks of eight specific PFAS via the Risk Quotient for Specific Contaminants (RQ_RSC) model and analyzed the annual detection trends of the top thirty PFAS with the highest detection frequencies. Regarding total PFAS contamination, China, Brazil, Germany, South Africa, and the Danube River Basin exhibited particularly high levels, with China being the most severely contaminated. Risk assessment indicated that 45.6% of global potential drinking water sources were at high risk (RQ_RSC > 1), while 48.4% were at low risk (RQ_RSC < 0.2). Among the evaluated PFAS, PFOA, PFOS, PFDA, and GenX were associated with higher toxicity exposure risks. For the identified key concern PFAS, it is necessary to simplify detection techniques, promote targeted large-scale safe treatment technologies, and explore intelligent monitoring tools to reduce regulatory lag, thereby effectively monitoring, preventing, and controlling PFAS contamination. Full article

This paper presents the results of personal exposure measurements to Radiofrequency Electromagnetic Fields from 2.4 GHz and 5.85 GHz Wi-Fi frequency bands. Measurements were taken in several specific scenarios: within international airports terminals, during takeoff, inside airplanes while flying with and without onboard Wi-Fi service (including while actively using a Wi-Fi connection), and during landing. Data were recorded onboard four international flights (two-round trip flights), from Spain to Mexico, and from Spain to Belgium. Two personal exposimeters, EME SPY 140 and EME Spy Evolution, were used to collect intensity level measurements in each scenario. During the outbound, the mean exposure value inside the airplane flight was 93.9 µW/m² in the 2.4 GHz Wi-Fi frequency band and 46.4 µW/m² in the 5.85 GHz Wi-Fi band (Spain to Mexico), and 7.29 µW/m² in the 2.4 GHz Wi-Fi band and 2.40 µW/m² in the 5.85 GHz Wi-Fi band (Spain to Belgium). For the return flight, the average value was 26.7 µW/m² in the 2.4 GHz Wi-Fi band and an average of 9.87 µW/m² in the 5.85 GHz Wi-Fi band (Mexico to Spain), and 3.24 µW/m² in the 2.4 GHz Wi-Fi band and 1.23 µW/m² in the 5.85 GHz Wi-Fi band (Belgium to Spain). Personal exposure levels to RF-EMFs from the Wi-Fi frequency band inside an airplane, even at the airport, are very low and well below the reference levels established by the international guidelines (10 W/m²). Full article

Background/Objectives: Swing leg resistance may stimulate propulsive force, required for forward progression and leg swing, in post-stroke patients. To assess the potential of swing leg resistance in rehabilitation, more knowledge is needed on how this unilateral manipulation affects gait. Therefore, we explored the bilateral effects of a unilateral swing leg resistance on muscle activity, kinematics, and kinetics of gait in able-bodied individuals. Methods: Fourteen able-bodied participants (8 female, aged 20.7 ± 0.8 years, BMI 23.5 ± 1.9) walked on an instrumented treadmill at 0.28 m/s, 0.56 m/s, and 0.83 m/s with and without unilateral swing leg resistance provided by a weight (0 kg, 0.5 kg, 1.25 kg, and 2 kg) attached to the leg through a pulley system. Propulsion and braking forces, swing time, step length, transverse ground reaction torques, and muscle activity in the gluteus medius (GM), biceps femoris (BF), rectus femoris (RF), vastus medialis (VM), medial gastrocnemius (MG), and soleus (SOL) were compared between conditions. Statistical analyses were performed using repeated measures ANOVAs, with a significance level of 5%. Results: Peak propulsive force and propulsive duration increased bilaterally, while peak braking force decreased bilaterally with unilateral swing leg resistance. In addition, the swing time of the perturbed leg increased with swing leg resistance. Muscle activity in the perturbed leg (GM, BF, RF, VM, MG) and the unperturbed leg (GM, BF, VM, MG, SOL) increased. Only in the BF (perturbed leg, late swing) and MG (unperturbed leg, early stance) did the muscle activity decrease with swing leg resistance. No adaptations in step length and transverse ground reaction torques were observed. Specific effects were enhanced by gait speed. Conclusions: Unilateral swing leg resistance can evoke effects that might stimulate the training of propulsion. A study in post-stroke patients should be conducted to test whether prolonged exposure to unilateral swing leg resistance leads to functional training effects. Full article

Virtual reality (VR) is often regarded as the “ultimate empathy machine” because of its ability to immerse users in alternative perspectives and environments beyond physical reality. In this study, 105 participants (average age 22.43 ± 5.31 years, range 19–45, 75% female) with diverse educational and professional backgrounds experienced three-dimensional 360° VR videos featuring actors expressing different emotions. Despite the availability of established methodologies in both research and clinical domains, there remains a lack of a universally accepted “gold standard” for empathy assessment. The primary objective was to explore the relationship between the empathy levels of the participants and the changes in their physiological responses. Empathy levels were self-reported using questionnaires, while physiological attributes were recorded through various sensors. The main outcomes of the study are machine learning (ML) models capable of predicting state empathy levels and trait empathy scores during VR video exposure. The Random Forest (RF) regressor achieved the best performance for trait empathy prediction, with a mean absolute percentage error (MAPE) of 9.1%, and a standard error of the mean (SEM) of 0.32% across folds. For classifying state empathy, the RF classifier achieved the highest balanced accuracy of 67%, and a standard error of the proportion (SE) of 1.90% across folds. This study contributes to empathy research by introducing an objective and efficient method for predicting empathy levels using physiological signals, demonstrating the potential of ML models to complement self-reports. Moreover, by providing a novel dataset of VR empathy-eliciting videos, the work offers valuable resources for future research and clinical applications. Additionally, predictive models were developed to detect non-empathic arousal (78% balanced accuracy ± 0.63% SE) and to distinguish empathic vs. non-empathic arousal (79% balanced accuracy ± 0.41% SE). Furthermore, statistical tests explored the influence of narrative context, as well as empathy differences toward different genders and emotions. We also make available a set of carefully designed and recorded VR videos specifically created to evoke empathy while minimizing biases and subjective perspectives. Full article