Rapid population growth in the town of Mila, in northeastern Algeria, has made urban expansion essential. However, unfavorable soil conditions pose major challenges to urban development. This study aims to characterize subsurface conditions, assess soil mechanical behavior, and establish a geotechnical zoning framework together with an Engineering Ground Model to support safe urban development. An integrated site investigation program was conducted, including 15 Core Drillings, 49 dynamic penetration tests, laboratory analyses, and hydrogeological monitoring. The subsurface stratigraphy consists of clays and marls containing limestone blocks at depths exceeding 20 m, together with two slip surfaces identified at depths of - 4 m and - 15 m. The results of the dynamic penetration tests divide the study area into two zones: one characterized by low peak resistance (Pr = 4.38 MPa) and shallow bedrock (BR = - 1.6 m), and the other by moderate conditions (BR = - 6 m and Pr = 18.48 MPa). The soils show low chemical aggressiveness (SO42⁻ < 5.5 mg/kg), very high clay content (> 75%), high plasticity (21 < Pi < 41.94%), and significant compressibility (Cr = 27.80%, 15.18% < w < 24.54%). X-ray diffraction analysis revealed clay and interstratified minerals dominated by illite (10-30%) and montmorillonite/smectite (≈12.5%). Most of the site is characterized by moderate to low admissible bearing capacity (0.164 < qad(max) < 0.842 MPa) for shallow foundations and is susceptible to significant volumetric changes, with settlements exceeding 5cm across large areas. Based on the combined analysis of bearing capacity, settlement potential, groundwater depth, and soil heterogeneity, three geotechnical zones were identified, ranging from highly unfavorable to relatively competent foundation conditions. The spatial distribution of areas with low bearing capacity and high settlement potential closely correlates with the observed patterns of structural damage. The results demonstrate that shallow foundations are largely unsuitable throughout the study area unless ground improvement measures or deep foundation systems are adopted. This study provides a robust Engineering Ground Model and geotechnical zoning framework to guide sustainable urban planning and foundation design in similar clay-dominated, hydro-mechanically sensitive environments.
Multidisciplinary tumor boards integrate longitudinal treatment histories, molecular profiling and rapidly evolving evidence to guide decisions in hematological malignancies, yet access to this level of subspecialty deliberation is increasingly uneven. Here we develop HemaGuide, a locally deployable, modular large language model agent that converts unstructured clinical documents into structured case representations, autonomously routes cases to specialized decision modes ('guideline', 'advanced' and 'molecular') and grounds recommendations in disease-specific guideline flowcharts and a clinical decision memory of >2,000 real-world tumor board cases. In expert-blinded benchmarking on 45 high-complexity cases across six foundation models, HemaGuide substantially improved concordance with tumor board decisions. A systematic ablation study across 11 layers confirmed that performance gains were routing-type-dependent, with no single component sufficient across case types. Automated classification of 70 clinically relevant missense variants showed high concordance with expert standards; no oncogenic variant was downgraded to benign and the whole workflow was completed under real-time conditions on commodity hardware with a median latency of 39 s rather than the hours typically required for manual molecular board workflows. In a simulated practice study, agent-assisted resident physicians achieved near-senior concordance and partially outperformed senior physicians in their subspecialty. External validation on 555 independent cases from a second academic center yielded 81.8% concordance across 47 entities, and a prospective 1-month silent trial on 64 consecutive, unselected cases achieved 82.8% concordance. Hallucinations occurred in 2 of 664 evaluated cases (0.3%). Together these data provide evidence that locally deployable, case-grounded large language model agents can deliver auditable clinical decision support across hematological malignancies, with concordance maintained across institutions and under real-time conditions on commodity hardware.
Breast milk reduces complications among very low birth weight infants (VLBWIs) admitted to neonatal intensive care units (NICUs). When mothers' own milk (MOM) is temporarily unavailable, donor human milk (DHM) is used as an alternative. In Japan, the first human milk bank was established in 2017, and the use of DHM remains relatively new. However, little is known about how mothers experience emotional changes and social interactions related to DHM use while continuing to express their own milk. Semi-structured interviews were conducted with 10 mothers of VLBWIs who had used DHM immediately after birth, and data were analysed using the Modified Grounded Theory Approach. Mothers experienced emotional conflict between their desire to provide their own milk and the need to prioritize their infant's health and survival; however, support from family members and healthcare professionals facilitated emotional reconciliation through continued milk expression and observing their infants' growth. Mothers experienced emotional conflict regarding DHM use; however, support from spouses, family members, and healthcare professionals helped mothers gradually come to terms with these feelings through continued milk expression and observing their infants' growth.
The necessity of adjuvant targeted therapy in stage IB epidermal growth factor receptor (EGFR)-mutant non-small cell lung cancer (NSCLC) remains controversial. We aimed to evaluate the efficacy of adjuvant EGFR tyrosine kinase inhibitors in patients with high-risk pathologic features and sought to explore potential subgroups that may derive benefit. This study reviewed patients with pathologic-stage IB EGFR-mutant NSCLC and high-risk factors between 2018 and 2020. Propensity score matching was used to match patients who received adjuvant EGFR tyrosine kinase inhibitors (adjuvant targeted therapy, or ATT) with those who did not (non-ATT). Disease-free survival (DFS) and overall survival (OS) were estimated by Kaplan-Meier method and compared by log-rank test. A total of 361 patients were included, of whom 41 received ATT and 320 did not. The median follow-up was 63.6 months. In the overall cohort, post-matched DFS (hazard ratio [HR], 0.60; 95% CI, 0.31-1.17) and OS (HR, 1.06; 95% CI, 0.41-2.71) did not differ between the ATT and non-ATT groups. Multivariable analysis identified the presence of a ground-glass opacity component as a favorable prognostic factor. Subsequent subgroup analyses demonstrated that ATT significantly prolonged DFS in patients with pure-solid tumors (HR, 0.33; 95% CI, 0.12-0.92), but not OS (HR, 0.55; 95% CI, 0.13-2.28). In contrast, ATT was not associated with improved DFS or OS in patients with part-solid tumors. Adjuvant targeted therapy did not improve long-term survival in stage IB EGFR-mutant NSCLC with high-risk features, although it prolonged DFS in pure-solid tumors. A de-escalation strategy may be considered for postoperative management, particularly in part-solid tumors.
Small microplastics are an increasing environmental concern due to their high mobility, large surface area-to-volume ratio, and ability to transport co-contaminants and microorganisms. Yet, particles in the 1-μm range remain difficult to quantify and remove, and standardized approaches to evaluate treatment performance are still lacking. This study addresses this methodological gap by combining controlled microplastic spiking in real secondary-treated wastewater with high-throughput flow cytometry for rapid quantification of 1-μm fluorescent microplastics and SYBR Green I-stained bacteria. Spent coffee grounds biochar produced at 300, 500, and 700 °C was evaluated in rapid small-scale column tests (1 mL min-1; sampling every 2 min) to investigate how pyrolysis temperature influences structure and removal efficiency. Biochar produced at 500 °C achieved nearly complete removal of 1-μm microplastics and retained approximately 90% of bacteria, outperforming materials produced at 300 and 700 °C, which showed reduced performance consistent with lower surface area and pore development. Breakthrough data from a long-term test were fitted using a lag-corrected Yoon-Nelson model, accurately predicting 50% breakthrough at 500 min and an estimated sorption capacity of 200 mg g-1. These results demonstrate that tuning biochar production and integrating rapid analytical tools enable systematic optimization of sustainable filtration strategies for emerging micropollutants.
A global comparison of dairy cow breeding objectives provides valuable insight into areas of convergence and divergence, helping identify populations with compatible breeding goals for sourcing germplasm to strengthen domestic genetic improvement programs. The objective of this study was primarily to compare the rankings of Holstein-Friesian artificial insemination sires across countries on the total merit index value of each country. The study also examined how different countries present estimates of genetic merit and how frequently they update their base populations; Denmark, Finland, and Sweden (DFS) were treated as a single group. The partial correlation among 22 indexes (i.e., 21 national indexes plus the Holstein Association USA Total Performance Index) was estimated for 49,450 Holstein(-Friesian) sires born post-2000 with a reliability >70%, after adjustment for genetic trends. The partial correlations among the ranking of sires on different indexes varied from 0.24 to 0.87 with 41% of the pairwise correlations being stronger than 0.70 but just 11% being stronger than 0.80. Notably, indexes designed for indoor confinement-based systems had, on average, weaker correlations with grazing-focused indexes from Ireland and New Zealand (0.48 to 0.50). Stronger average correlations (0.69) were observed among the ranking of sires on confinement-based indexes; the correlation between the Irish Economic Breeding Index and New Zealand Breeding Worth was 0.56. When limited to just the milk production components of the different indexes, the partial correlations among the ranking of sires on subindexes were, on average, 0.78, varying from 0.29 (Uruguay with Poland) to 0.96 (Japan with Spain). Countries differ in their approach to trait weighting. Across the 21 countries examined in this study, 38% derived the weights assigned to traits solely from economic models or functions, whereas an additional 29% of the countries adopted only a desired-gains approach. The remainder of countries adopted a hybrid strategy, applying economic values to certain traits, typically the production traits, while using desired gains for others (e.g., the Netherlands, DFS). In some cases, initial weights were derived from economic principles but subsequently adjusted to achieve specific desired gains (e.g., Canada, Australia, Uruguay). Of the 21 milk production subindexes compared, 8 had a negative weight on milk yield (i.e., Canada, DFS, United Kingdom, New Zealand, Australia, Belgium, Ireland, Uruguay), 4 had a positive value (i.e., United States, Poland, Spain, South Africa), whereas Germany, the Netherlands, Japan, France, Slovenia, Italy, Israel, Switzerland, and Hungary did not consider milk yield in their total merit indexes. With the exception of Ireland, United Kingdom, United States, Israel, and Uruguay who presented genetic evaluations of individual animals as PTA, genetic evaluations of milk production traits are presented as EBVs by all other countries. Of the 21 countries, 10% update their genetic evaluation base population more than once per year, 38% update it annually, 38% update it every 5 yr, and the remaining 14% update it periodically. Health traits are increasingly being considered for inclusion in future breeding objectives as are traits associated with both environmental impact and feed efficiency. The findings underscore how economic, biological, genetic, and policy factors shape national breeding objectives, helping interpret international differences in genetic trends and performance outcomes.
Lipid oxidation produces an unpleasant warmed-over flavor (WOF) in meat. In this study, the antioxidant peptide carnosine (Car) was added to minced chicken breast and thigh meat, and its effect on WOF, volatile compounds, and meat quality during refrigerated storage was assessed. Breast and thigh meat patties were supplemented with four levels of Car (0, 1.0, 2.5, and 15 mg/g meat) and stored under refrigerated conditions for 1, 3, and 7 days. After that, WOF, volatile components, and meat quality factors were analyzed. The sensory attribute 'oxidized oil odor', a WOF indicator, was affected by Car supplementation and was lowest in samples supplemented with Car at 15 mg/g meat (Car15). In addition, some volatile aldehydes and thiobarbituric acid reactive substances (TBARS) in cooked meat causing WOF were lowest in Car15. Instead, samples supplemented with Car at 1.0 or 2.5 mg/g meat did not show significant differences in WOF, volatiles, or TBARS compared to controls without Car supplementation. Other meat quality parameters, such as pH, increased in Car15 samples; whereas L* and cooking loss decreased. In summary, supplementation of chicken breast and thigh meat with 15 mg/g Car reduced WOF intensity through inhibition of lipid peroxidation and influenced multiple meat quality parameters.
In contemporary China, increasing numbers of Ph.D.-holding corporate professionals and new-generation doctoral-stage participants (Ph.D. candidates or Ph.D. graduates within 1-2 years) prioritize university faculty or research posts ("reverse mobility"). Prior accounts have focused mainly on economic and institutional push-pull forces; consequently, the psychological and meaning-making processes driving this shift remain under-specified. Accordingly, reverse mobility is conceptualized as a form of setting-seeking, whereby individuals pursue work settings perceived as more restorative, predictable, and boundary-controllable. Using constructivist grounded theory, 47 participants (25 transitioners and 22 new-generation doctoral-stage participants) were interviewed, and iterative open-axial-selective coding with constant comparison and triangulation was conducted until theoretical saturation. Participants depict marketized workplaces as a "performance competition arena" characterized by KPI-driven value alienation, chronic exhaustion, and anticipatory insecurity. When perceived risk crosses a subjective threshold, psychological safety needs become salient and shift decisions from escape toward shelter seeking. Universities are then symbolically constructed as a "psychological exemption zone," namely a perceived work setting that can reduce exposure to marketized performance risks and help restore recovery time, predictability, and control over work-life boundaries. This construction is supported by institutional shelter, time sovereignty, meaningful work, and community belonging. Participants then enact identity reconstruction, narrative management, and boundary work to turn this imagined safer setting into an actual career move and a subjective sense of relocation from one work world to another. Transitioners emphasize restoration after accumulated strain, whereas new-generation doctoral-stage participants foreground preventive risk avoidance. This study develops a dynamic process model integrating need-based mechanisms with risk perceptions, extending reverse-mobility research by foregrounding psychological safety and defensive career strategies.
This dataset presents survey data of 71,578 ancient and notable trees in Sichuan Province, China. The data cover all 183 county-level administrative units within Sichuan Province. Data collection was conducted from May 2016 to March 2023. Twenty-four indicators were recorded through field surveys: tree ID, detailed address, tree category, species, coordinates, ancient tree grade, tree age, tree height, girth at breast height / girth at ground level, mean crown width, growth site, distribution pattern, elevation, aspect, slope, slope position, soil name, growth vigour, growth environment, factors affecting growing environment, historical and cultural description, description of special tree condition, existing above-ground protection, and maintenance and restoration measures. All 24 indicators and their corresponding survey methods strictly follow the national standard LY/T 2738-2016. The dataset comprises three files: (1) the survey form template with completion instructions; (2) raw survey data containing 71,578 individual records; and (3) summary statistics presenting the distribution of tree counts across 13 indicators, including city (prefecture), growth site, and tree age. This dataset can support regional comparative analyses of ancient and notable trees, multi-dimensional interaction analyses among trees, environment, and humans, as well as longitudinal studies of tree population changes.
Arterial spin labeled (ASL) perfusion MRI is the only non-invasive and non-radioactive technique for measuring regional tissue perfusion. Perfusion signal in ASL MRI is derived from the difference between the spin labeled image and the spin untagged control image. Limited by the T1 decay of arterial blood, ASL MRI has an intrinsic low signal-to-noise-ratio. Solving this problem is challenging because the ground truth is often unknown, and it is difficult to preserve textures when suppressing heavy noise. In this paper we propose an unsupervised Locally Adaptive regularization with Collaborative data Selection (LACS) scheme, which exploits the high affinity between the paired label and control (L/C) images to select highly correlated contents to form the low-rank matrices. The low-rank regularization applied to such matrices could be better adapted to local structures compared with slice-level global models and more robust against noise compared with voxel-level local models. Further, we used the log-determinant of covariant matrices as the non-convex surrogate of the low-rank penalty instead of the widely used convex surrogates. We demonstrated that the adopted surrogate essentially exploits near-optimal sparsity in the underlying principal component analysis (PCA) domain without explicit training. Apparently, LACS does not rely on any ground-truth training data. When tested on a real-world ASL MRI dataset, LACS significantly improved the quality of ASL perfusion maps using just one pair of L/C images, compared with the standard pipeline that requires multiple L/C pairs. The proposed scheme could set a new benchmark for ASL MRI denoising.
Language models (LMs) call for a theoretical rethinking grounded within linguistics itself. Rather than signalling the "end of linguistics" or merely encouraging interdisciplinarity, LMs function as an empirical testing ground for formal linguistic concepts. They prompt a renewed examination of form-meaning mapping, theoretical autonomy, and the conditions under which computational systems can genuinely inform linguistic explanation.
The convergence of financial technology (fintech), sport event sponsorship, and destination marketing constitutes an underexplored but increasingly consequential research paradigm in tourism studies. This paper presents a conceptual analysis of how commercial banks-particularly regional and development-oriented financial institutions operating in emerging economy contexts-utilize fintech-based marketing strategies through sport event sponsorship to simultaneously advance destination branding and inclusive tourism. The study grounds its inquiry in the Uzbekistan and Central Asian context, where rapidly developing digital payment infrastructure intersects with ambitious tourism growth targets and significant financial exclusion among potential visitors. Drawing on four complementary theoretical lenses-Stakeholder Theory, service-dominant logic (S-D logic), the Technology Acceptance Model (TAM), and Sen's Capability Approach-the paper develops an integrative conceptual framework identifying three causal pathways through which fintech innovation, including mobile payment systems, blockchain, artificial intelligence, and alternative lending products, can transform sport event sponsorship into a vehicle for destination competitiveness and social inclusion. The paper also critically engages with the digital divide literature to acknowledge conditions under which fintech sponsorship may deepen rather than reduce exclusion. Seven theoretically grounded propositions are derived, along with four key moderating factors. The study makes conceptual contributions to tourism marketing literature by clarifying the triadic relationship among financial institutions, sport properties, and destination management organizations, and identifies priority directions for empirical research in emerging market contexts.
BackgroundMild cognitive impairment is a prodromal stage of dementia, and early identification is crucial for prognosis.ObjectiveThis study aims to create and validate a machine learning model for diagnosing mild cognitive impairment (MCI) using eye movement and gait analysis data.MethodsTo facilitate model training and internal validation, a cohort of 235 patients was recruited from the Memory Clinic at Xi'an NO.3 Hospital between August 2024 and November 2025. In addition, data from 71 patients were randomly selected to form an independent test set. Feature selection was conducted using the Least Absolute Shrinkage and Selection Operator (LASSO) and multivariable logistic regression. Subsequently, various machine learning classifiers were compared. Model performance was assessed using metrics such as the area under the receiver operating characteristic curve (AUC) and decision curve analysis. To evaluate model interpretability, SHapley Additive exPlanations (SHAP) were employed.ResultsThe study involved 235 participants, divided into mild cognitive impairment (MCI) (n = 130) and healthy control (HC) (n = 105) groups. The final prediction model used four features: gait speed during a dual-task test, ground reaction force in a single-task test, antisaccade task accuracy, and noise rate in a saccade-to-pursuit task. The Gaussian Naive Bayes (GNB) classifier showed excellent performance with an AUC of 0.952 (95% CI: 0.923-0.981) in the validation group and 0.944 (95% CI: 0.912-0.967) in the test set.ConclusionsThe GNB model, combining eye movement and gait parameters, enables early MCI detection with high accuracy and practical clinical use.
In-stent restenosis (ISR) remains a significant complication after percutaneous coronary intervention. Identifying patients at high risk for ISR from visually patent post-procedural angiograms is challenging. We developed and validated deep learning models using routine post-procedural digital subtraction angiography (DSA) to predict future ISR risk, emphasizing the role of mask-guided spatial attention. We retrospectively enrolled 237 patients with 1-year angiographic follow-up across three centers. Immediate post-procedural DSA frames were used as inputs, with 1-year outcomes as ground truth. Deep learning architectures were trained using full-image and mask-guided strategies. On the independent external test set, the mask-guided DenseNet-121 achieved the highest predictive performance (AUC: 0.885, AUPRC: 0.912). Crucially, models trained on full images exhibited severe shortcut learning, erroneously focusing on task-irrelevant background noise. The mask-guided strategy successfully corrected this, demonstrating excellent calibration and robust clinical net benefit. Grad-CAM visualization confirmed precise attention on stented vessel regions associated with future restenosis. In conclusion, mask-guided deep learning analysis of post-procedural DSA provides an accurate, interpretable, and automated prognostic tool. By overcoming shortcut learning, this approach effectively stratifies future ISR risk, holding potential to guide personalized post-procedural surveillance and optimize secondary prevention.Trial registration: This study was registered with the Chinese Clinical Trial Registry (ChiCTR) under the registration number ChiCTR2500115474 on December 26, 2025.
Mathematical models linking zygote size to survival until maturity are central to evolutionary ecology, yet many existing formulations remain fragmented and conceptually disconnected. These models fall broadly into two categories: phenomenological models, which lack explicit biological justification, and those grounded in survival analysis, which we, in contrast to phenomenological models, call mechanistic models. Although the latter are theoretically preferable, many commonly used mechanistic functions lack a common basis despite being better integrated than phenomenological models, and some are known to contain inconsistent dimensional assumptions. Here, we use a general framework that links zygote survival functions to growth dynamics and size-dependent mortality to show that many existing zygote survival models are corollaries of allometric growth models and generalize readily to account for allometric mortality, revealing previously hidden connections among models. The unification presented here provides principled grounds for re-evaluating the parameters and predictions of existing models.
Depression is a common mental health disorder which frequently co-occurs with increased body mass index or increased waist circumference (hereafter 'overweight'), causing heightened cardiovascular risk. Unhealthy lifestyle behaviours underlie both conditions. The Multimodal Lifestyle Intervention (MLI) LEEF integrates physical activity, nutrition, and behavioural strategies, tailored to motivational challenges common in depression, offering a dual focus on mental and physical health. However, unpublished process data revealed very limited referral to MLI‑LEEF from both primary and secondary care, and consequently low initiation rates, signalling clear implementation challenges. This study addresses this gap by translating prioritised implementation determinants into conceptually and empirically grounded implementation strategies and examining what works, how, and under which conditions. This quasi‑experimental, explanatory sequential mixed‑methods study will implement MLI‑LEEF across ten general practices and three secondary mental health care outpatient clinics in the Northern Netherlands. Implementation strategies will be developed and tailored using Causal Pathway Diagramming and applied over a six‑month period. Quantitative data will be collected before, during, and after implementation to assess proximal implementation outcomes (the immediate, observable effects of an implementation strategy), distal implementation outcomes (adoption and sustainability), service‑level penetration, and patient outcomes. Following the implementation period, semi‑structured interviews will be conducted with referrers from each participating organisation. Site‑specific topic guides, informed by each organisation's implementation plan, causal pathway diagrams, and quantitative implementation outcomes, will probe how strategies generate change (mechanisms), how they influence targeted determinants, and which contextual conditions (preconditions, moderators) shape implementation strategy functioning. This study will respond to fieldwide calls in implementation science for rigorous, transparent, and context-sensitive approaches to developing and evaluating implementation strategies. By examining how strategies shape referral and initiation in routine general practice and specialist mental health services, and by clarifying the mechanisms and contextual conditions under which they are effective, the study will generate actionable insight into why implementation succeeds or falters in real-world care. These insights will provide essential groundwork for strengthening the reach and adoption of MLI-LEEF and will offer transferable guidance for embedding multimodal lifestyle interventions into everyday care, while advancing generalizable knowledge of how implementation strategies produce change. 10.17605/OSF.IO/XJHCB.
Tree growth is a critical carbon sink for tropical forests, yet how light and climate jointly shape long-term individual-level growth remains poorly understood. Using 34 years of data from 2881 trees (10 species: six evergreen, four deciduous) in a Costa Rican wet forest, we examined how interannual climate variation and individual light availability influence tree growth in diameter at breast height, height and above-ground biomass across phenological types and size classes. Light availability consistently emerged as the primary growth driver, with deciduous species more responsive than evergreen species; however, light effects diminished with increasing tree size. Higher mean daily maximum temperatures enhanced radial growth in evergreen trees and height growth in deciduous trees, whereas elevated minimum temperatures suppressed radial growth in evergreen trees. Increased wet-season precipitation (WSP) reduced radial growth in both phenological types and height growth in evergreen trees, whereas higher dry-season precipitation decreased radial growth in deciduous trees. Notably, larger trees exhibited weaker positive radial growth responses to maximum temperatures but stronger negative responses to WSP. These findings underscore the dominance of light over climate in regulating tropical tree growth and highlight the necessity of incorporating tree sizes and functional groups into ecological modelling and forest management.
Freshwater ecosystem services in peri-urban catchments are increasingly threatened by land-use intensification, infrastructure deficiencies, and climate variability, with significant implications for environmental management and community well-being. This study examines ecosystem services in the Kat River Catchment, Eastern Cape, South Africa, using a multidimensional participatory mapping approach to assess service types, perceived importance, perceived spatial and temporal changes in availability and quality. A total of 54 stakeholders, representing communities within the catchment (the Kat River Catchment Forum), commercial farmers, subsistence farmers, and non-farming households participated in this study. There were 23 participants in a participatory mapping workshop and 31 semi-structured interview respondents. Provisioning services such as water, reeds, wood, sand, fish, and medicinal plants were consistently rated as highly important to local livelihoods, while cultural services including spiritual and recreational uses were rated as moderately important. Most ecosystem services were perceived to have deteriorated since the year 2000, particularly water quality and availability, reeds used for craft production, and fish for food and recreation. Participants attributed these declines to climate variability, agricultural intensification, failing wastewater treatment infrastructure, and inadequate solid waste management. The findings demonstrate how participatory approaches can generate locally grounded evidence of ecosystem service degradation and identify governance and infrastructure gaps. Integrating local ecological knowledge into catchment management, strengthening wastewater treatment performance, and improving pollution control are critical for enhancing ecosystem service sustainability and supporting adaptive environmental management in peri-urban river systems.
Positron Emission Tomography (PET) diagnostic precision is often compromised by low spatial resolution. Deep learning restoration models tend to sacrifice quantitative accuracy for visual sharpness, and most are trained on a single fixed degradation profile, limiting generalization across scanners. This paper presents a metabolically faithful 3D restoration framework pairing a volumetric extension of SwinFIR with two innovations: (1) a composite metabolic-aware loss enforcing structural, distributional, and frequency-domain agreement with the ground truth, and (2) a stochastic degradation augmentation strategy that randomizes point spread function parameters, voxel sampling, and counting noise during training, exposing the model to a distribution of simulated scanner-like degradations rather than a single fixed simulation. Evaluated on NeuroEXPLORER data, the proposed method outperforms baselines with a Structural Similarity Index Measure (SSIM) of 0.843, Peak Signal to Noise Ratio (PSNR) of 27.08 dB, and Normalized Root Mean Squared Error (NRMSE) of 0.117, while maintaining metabolic fidelity (Concordance Correlation Coefficient (CCC) 0.948, Wasserstein distance 0.018). Ablation experiments confirm that stochastic degradation augmentation improves robustness over fixed-profile training. The framework recovers anatomical detail with only small but measurable regional SUVR biases (≤3.3%) in selected cortical regions, which are symmetric across diagnostic groups.
The commentary argues the authors employ misdirection and strawmanning to cast others as polarized extremes and themselves as the reasonable centrists. We argue that these patterns of misrepresentation ultimately damage any consensus and middle ground they claim to hope to reach.