Search Results (2,883)

Urban streets are critical public spaces that support residents’ daily psychological recovery, and their landscape quality is directly related to pedestrians’ physical and mental well-being. In the rapid urbanization process, numerous urban streets have exhibited problems such as excessive building density, cluttered visual interfaces, a lack of natural elements, and an absence of regional characteristics, leading to a continuous decline in the psychological restorative capacity of street spaces and failure to meet residents’ demands for a healthy urban environment. Existing research mostly employs qualitative assessment methods to evaluate walking experiences and psychological restoration levels of street environments, lacking high-precision, pixel-level quantification of street landscape elements and rarely incorporating regional cultural elements into the analytical framework of restorative environments. This study takes Foshan, a famous historical and cultural city in China, as the research object, and selects five typical streets in the main urban area, including comprehensive streets, living streets, landscape streets, commercial streets, and historical–cultural streets, to construct a technical route of “data collection–element quantification–model construction–effect analysis.” Leveraging the pre-trained Mask2Former semantic segmentation model and pedestrian-perspective street-view images (SVIs), combined with field research, the study quantifies 22 street landscape elements across five dimensions: environment, transportation, social interaction, facilities, and culture. Through PCA principal component analysis and K-means clustering, 20 typical photos were objectively sampled, and public psychological evaluations were conducted using the Perceived Restorativeness Scale (PRS). A stepwise multiple linear regression model was then employed to construct an exploratory explanatory model for street psychological restoration, identifying key influencing factors and their effect intensities. The results indicate the following: (1) Environmental and cultural elements are the core characteristics associated with pedestrians’ psychological restoration, whereas transportation, social, and facility elements are correlated only with certain restoration dimensions and show no significant association with the overall psychological restoration level. (2) Among the 22 element indicators, the Green View Index showed the strongest positive association with psychological restoration (β = 0.681, p < 0.001); historical memory markers and the Blue View Index also exhibited significant positive associations. (3) By integrating the elements associated with pedestrians’ psychological restoration and their association strengths, an exploratory explanatory model of the psychological restoration benefits of urban street landscapes was constructed, with an adjusted coefficient of determination of 69.8%, accounting for 69.8% of the variation in street psychological restoration levels. The findings establish an exploratory analytical framework and furnish empirical evidence for healthy city planning and street renewal in similar historical and cultural cities. Full article

Global olive oil consumption continues to grow, and Spain, particularly Andalusia, holds the largest cultivation area within the EU, covering 1.68 million hectares, which expanded by 7.7% between 2013 and 2023. However, precise data on the planting systems used are lacking; this information is essential for assessing crop intensification and its environmental, economic, and social implications. This study applies Deep Learning (DL) techniques to classify the planting system of olive parcels previously identified in a national cadastral database, using high-resolution RGB imagery acquired from a nationwide public aerial orthophotography program. Five DL algorithms (VGG19, InceptionV3, MobileNet, ResNet50, and Xception) were compared across different planting-system classification schemes and spatial resolutions, using a dataset of 8000 images (2000 per class) divided into training (70%), validation (20%), and testing (10%) subsets. The best results were obtained with a three-class scheme (traditional, intensive, and super-intensive) at 0.5 m/pixel resolution, reaching precision values of 99.00% with MobileNet and 98.66% with VGG19. These results support extending the approach to larger areas for regional-scale monitoring of planting systems, including other Mediterranean olive-growing regions, and provide a basis for further research on the water use and labor demand associated with each planting system. Full article

Polymetallic nodules constitute a strategically important deep sea mineral resource enriched in nickel, cobalt, copper, and other critical metals. Accurate prediction of their spatial distribution is crucial for resource assessment and sustainable exploitation. However, sparse sampling, high survey costs, and complex nonlinear interactions among environmental factors make this task difficult. To address these challenges, this study applies a quantile regression forest to analyze the data from 233 box-core stations across a 23,000 km² area in the eastern Pacific, integrating bathymetry and backscatter intensity. Cross-validation verifies model performance. The trained model predicts nodule abundance across an adjacent 3000 km² study area and validates it against 12 independent stations that were not involved in the training. This external validation gives a root mean square error of 4.11 kg/m². The quantile regression forest model also estimates uncertainty to quantify prediction reliability. The resulting uncertainty map distinguishes high-confidence zones from areas with elevated uncertainty. In 46% of the study area, normalized uncertainty falls below 0.50, indicating higher reliability. The remaining 54% shows higher uncertainty and requires additional sampling. This method combining independent spatial validation and uncertainty visualization provides a transparent tool for deep sea mineral resource assessment where data are sparse. Full article

Background: Exercise intolerance is a common symptom in patients with chronic heart failure (CHF) and is often associated with reduced pulmonary function. Breathing exercise, including inspiratory muscle training (IMT) has been shown to alleviate respiratory symptoms in patients with pulmonary and cardiac disease. The purpose of this systematic review is to assess the impact of IMT on maximal inspiratory pressure (Pimax), diaphragm thickness, functional vital capacity (FVC), dyspnea and maximum oxygen uptake (VO_2peak) in patients with CHF. Methods: Database research was conducted (PubMed, EMBASE, Cochrane Library of Controlled Trials) by using an appropriate algorithm to indicate the IMT effect on pulmonary function studies. Eligibility criteria included pulmonary function measurements following IMT in patients with CHF. A continuous random effect model (PROSPERO-CRD420261360295) was used to calculate standard mean differences (Std.MD) in all outcomes (between an experimental and control group). Results: A total of 395 participants (13 randomized controlled trials) were identified. More studies indicate improvement in Pimax after IMT (Pimax: Std.MD: 1.07; 95%CI, 0.72–1.42; p < 0.00001). Functional vital capacity (Std.MD, 0.41; 95%CI, 0.03–0.79; p = 0.03) and dyspnea (Std.MD, 1.11; 95%CI, 0.57–1.66; p < 0.0001) were also improved. Moreover, VO_2peak assessed by cardiopulmonary exercise testing before and after IMT showed statistical significance (VO_2peak: Std.MD: 0.86; 95%CI, 0.06–1.66; p = 0.04) in trained patients with CHF compared to controls. Conclusions: The meta-analysis indicates that isolated IMT performed either as short- to medium-term intervention (4 to 8 weeks) with moderate to high intensity (40 to 90% Pimax) or long-term program (8 to 12 weeks) with low intensity (30% Pimax) promotes pulmonary function and metabolic responses. More studies are required to examine the impact of IMT on diaphragm thickness in patients with chronic heart failure. Full article

Short-term financial risk in intelligent trading systems is reflected not only in prices, trading volumes, and textual sentiment but also in infrastructure operating states, including server workload, device power consumption, network latency, and packet loss rate. We propose HSF-LLMNet, a hardware sensor and large language model semantic fusion network for jointly modeling external information shocks and infrastructure responses. A large language model extracts event category, sentiment polarity, risk intensity, and semantic uncertainty from financial texts. Reliability-aware temporal modeling handles sensor missingness, drift, and abnormal noise, while asynchronous soft alignment, bidirectional cross-attention, and reliability-aware gated fusion integrate irregular textual events with continuous hardware signals. The model jointly predicts market direction, realized volatility, and three-level risk over the subsequent 30 min. Experiments were conducted on eight Chinese A-share indices: the SSE Composite Index (000001.SH), SSE 50 Index (000016.SH), CSI 300 Index (000300.SH), STAR 50 Index (000688.SH), CSI 500 Index (000905.SH), CSI 1000 Index (000852.SH), Shenzhen Component Index (399001.SZ), and ChiNext Index (399006.SZ). The common observation period for market, textual, and hardware data extended from 1 March 2024 to 30 June 2025. After data cleaning, timestamp matching, and multimodal temporal alignment, 169,208 aligned asset–time prediction windows were retained for the 30 min forecasting task. Realized volatility was defined as the square root of the sum of squared one-minute log returns over the future 30 min interval. The three-level risk label was constructed from future realized volatility, absolute 30 min return, and liquidity stress, with all thresholds estimated exclusively from the training portion of each fold. A sample was labeled high risk when at least two of the three indicators exceeded their 85th-percentile thresholds or when any indicator exceeded its 95th-percentile threshold. It was labeled medium risk when, after excluding high-risk samples, at least two indicators exceeded their 60th-percentile thresholds or any indicator exceeded its 85th-percentile threshold; all remaining samples were labeled low risk. Results showed that HSF-LLMNet achieved an accuracy of 78.62%, a precision of 78.14%, a recall of 77.83%, an F1-score of 77.98%, an area under the receiver operating characteristic curve of 84.91%, and a Matthews correlation coefficient of 57.36% for directional prediction. For realized-volatility regression, the MAE, RMSE, MAPE, and $R^2$ were 0.0089, 0.0135, 9.21%, and 0.812, respectively. For high-risk-event warning, the mean effective warning time, defined as the interval between the first valid alarm and the corresponding event, was 15.37 min; the false-alarm rate and missed-alarm rate were 6.82% and 8.14%, respectively. Ablation experiments showed performance reductions after removing semantic encoding, sensor-reliability estimation, asynchronous alignment, bidirectional cross-attention, gated fusion, or multi-task learning. These results indicate that textual events and infrastructure operating states provide complementary information for quantitative risk analytics and fintech applications. Full article

This study assessed the feasibility of a technology-supported high-intensity interval training (HIIT) program for occupational drivers. A multiphase feasibility study incorporating participant feedback was conducted in Flanders, Belgium, including consultation, pilot, and intervention phases. Occupational drivers were randomized to a HIIT group, an active control group (mindfulness program), or a passive control group (no intervention). The 16-week, equipment-free HIIT program consisted of two 15-min sessions per week, delivered via a technology-supported mobile application, preceded by a 2-week baseline. Primary feasibility outcomes included program completion, session adherence, intervention fidelity, adverse events, and participant experiences, while changes in health-related outcomes were assessed as secondary exploratory outcomes using validated questionnaires. During the consultation (n = 11) and pilot (n = 2) phases, participants reported positive attitudes toward the intervention. Key barriers, including scheduling constraints (n = 5) and bicycle transportation (n = 4), informed refinements such as an equipment-free, technology-supported protocol and flexible scheduling. Of the 23 randomized participants across all groups, 16 participants initiated their assigned study protocol, and 10 completed all assessment points and were included in the final analysis. Of the nine participants allocated to the HIIT group, four initiated the intervention (44.4%), and three completed the program (33.3% of those allocated; 75% of those who initiated). Among the three completers, mean session attendance was 80.2%. Session fidelity was well maintained, but exercise-intensity fidelity could not be objectively confirmed. No exercise-related adverse events were reported. Participant feedback indicated generally positive usability but mixed acceptability of the technology-supported HIIT program among the three participants who completed the intervention. Exploratory individual-level changes in health-related outcomes were heterogeneous, with lower mean scores for depression, anxiety, stress, and fatigue. A technology-supported HIIT program was feasible among the small subgroup of occupational drivers who initiated it, with high session attendance. However, considerable pre-initiation attrition and low retention limit its feasibility for the broader target population. Observed health-related outcomes were heterogeneous and should be interpreted descriptively, warranting evaluation in a fully powered randomized controlled trial. Full article

Manual fruit thinning is labor-intensive and inefficient, making the development of intelligent visual detection systems a crucial approach for improving the quality and production efficiency of the pear industry. However, in natural orchard environments, young pear fruits are small in size with slender fruit stalks, and their texture and color characteristics are highly similar to those of tender branches. Furthermore, variations in illumination and occlusions caused by branches and leaves make it difficult for existing detection models to simultaneously and accurately identify fruits and fruit stalks, limiting their application in automated thinning equipment. In this study, Yuluxiang pear was selected as the research object, and image data were collected under diverse field conditions, including forward-lighting, backlighting, close-range shooting, long-range shooting and fruit overlapping. A dedicated dataset containing 3057 images was established. Based on the RT-DETR-R18 network, a lightweight and high-precision fruit–stalk synchronous detection model was proposed. Specifically, the backbone network was reconstructed by integrating GCConv with C2f modules to enhance global feature extraction for slender fruit stalks. The bottleneck structure was optimized using GCConvC3 to reduce feature degradation under occlusion conditions, and an additional 4× down-sampling P2 detection head was introduced to improve the detection capability for small targets. To fully validate the model performance and stability, three types of experiments were conducted in this study: ablation experiments, repeated experiments with different random seeds, and comparative experiments. Ablation experiments verified the cumulative performance improvements brought by the introduced modules. Repeated experiments with different random seeds were performed to explore training randomness-induced performance fluctuations, and the results demonstrated that the proposed model maintains stable overall detection accuracy with minor metric fluctuations. Comparative experiments demonstrated that the proposed model achieved a compact parameter size of only 15.97 M, with a precision of 95.0% for young pear fruit detection and an mAP50 of 83.0% for fruit stalk detection, outperforming all comparative models in overall mAP50. The training convergence curves and Grad-CAM++ visualization results further confirmed the stable optimization process and enhanced feature attention capability of the proposed model. By achieving a favorable balance between detection accuracy and model lightweightness, this approach provides effective technical support for the development of intelligent fruit thinning equipment and vision-based systems for smart pear orchards. Full article

Rootstock–scion interactions are recognised as key determinants of grapevine physiology and secondary metabolism, yet their role in modulating the phenolic profile and sensory typicity of cv. Xinomavro, a Greek red variety characterised by low anthocyanin content and high acidity, remains insufficiently documented. This study evaluated the effect of five rootstocks (110R, 161-49C, 3309C, 101-14MGt and Vitis riparia) on Xinomavro grape composition and wine quality. The vines were cultivated under identical vineyard management in an experimental vineyard in Amyntaio and the wines were produced under similar vinification conditions. Grape physicochemical parameters and their phenolic composition were determined, while wines were analysed for basic oenological parameters, phenolic composition and individual anthocyanins by HPLC. Sensory evaluation of wines was conducted by a trained panel. Rootstocks significantly affected anthocyanin accumulation, extractability and tannin distribution in grapes, and modulated wine phenolic composition, colour intensity and aromatic profile. Overall, rootstock selection emerges as an important viticultural tool to regulate phenolic composition, colour expression and sensory typicity of Xinomavro wines under evolving climatic conditions. Full article

As the scaling of silicon-based devices approaches physical limits, two-dimensional transition-metal dichalcogenide field-effect transistors (2D TMD FETs) have emerged as promising candidates for logic devices in the post-Moore era. However, their design optimization relies heavily on computationally intensive TCAD simulations, thereby limiting efficient exploration of multidimensional parameter spaces. This paper proposes an efficient automated design framework for 2D TMD FETs under small-sample conditions and validates it using a monolayer MoS₂ FET as a case study. The framework integrates device design, physics-based simulation, performance prediction, and inverse design, establishing a bidirectional mapping between device parameters and electrical performance. Target-driven closed-loop optimization is achieved through TCAD-based feedback validation. Results demonstrate that, using a dataset comprising 300 TCAD samples, the forward model achieves an average coefficient of determination (${\mathrm{R}}^2$) of 0.9503. TCAD revalidation of the inverse-designed devices yields an average mean absolute error (MAE) of 0.0464 and an average mean absolute percentage error (MAPE) of 5.46% for performance metrics. Regarding computational efficiency, while a single TCAD simulation takes approximately 25 to 50 min, the trained model performs inference in under 50 ms, achieving a speedup of at least $3\times {10}^4$ during the inference phase. Accounting for the generation of the 300 TCAD samples and the training of both forward and inverse models, the framework’s one-time computational cost ranges from 160.27 to 285.27 h. Once the cumulative number of design tasks exceeds approximately 342 to 385, the total computational cost falls below that of direct TCAD simulation, with the computational advantage becoming increasingly significant as the number of tasks grows. Consequently, this method is highly suitable for large-scale parameter sweeps, device screening, and multi-objective, high-frequency design iterations. It drastically reduces repetitive TCAD calls, offering a scalable solution for the efficient, automated design of 2D TMD FETs. Full article

Metabolic syndrome (MetS) is a clinical condition defined by the coexistence of interconnected cardiometabolic risk factors, including central obesity, dyslipidemia, elevated blood pressure, and impaired glucose regulation, which collectively increase the risk of type 2 diabetes mellitus and cardiovascular disease. Beyond these clinical diagnostic features, MetS is characterized by complex pathophysiological alterations involving systemic dysregulation of metabolic signaling across adipose tissue, skeletal muscle, liver, vascular endothelium, and the immune system. Key molecular alterations include impaired insulin receptor substrate (IRS)–Akt signaling, chronic nuclear factor kappa B (NF-κB) activation, mitochondrial dysfunction, and oxidative stress. Physical exercise is recognized as a pleiotropic biomedical intervention capable of restoring metabolic homeostasis through coordinated modulation of intracellular signaling pathways and inter-organ communication. Exercise activates AMP-activated protein kinase (AMPK), enhances peroxisome proliferator-activated receptor gamma coactivator-1 alpha (PGC-1α)-mediated mitochondrial biogenesis, and stimulates nuclear factor erythroid 2-related factor 2 (Nrf2)-dependent antioxidant responses. These adaptations improve glucose uptake, enhance fatty acid oxidation, and reduce ectopic lipid accumulation across metabolically active tissues. At the systemic level, skeletal muscle functions as an endocrine organ by releasing myokines such as irisin, interleukin-6 (IL-6), and fibroblast growth factor 21 (FGF21), which contribute to metabolic regulation across the liver, adipose tissue, and vasculature. These exercise-induced signals promote immune modulation, reduce pro-inflammatory cytokine production, and improve endothelial function. Different exercise modalities including aerobic, resistance, and high-intensity interval training (HIIT) activate both common and modality-specific molecular pathways, supporting individualized exercise strategies. Collectively, exercise targets the multi-organ pathophysiology of MetS and provides a mechanistic foundation for precision exercise medicine in cardiometabolic disease management. Full article

Background: Physical exercise is a vital component of cancer rehabilitation, with demonstrated improvements in cancer health-related outcomes, relapse rates, mortality and overall survival, yet global participation remains low. In Singapore, uptake of community cancer rehabilitation is limited despite high prevalence of treatment-related impairments. Methods: A cross-sectional study was conducted among adults (≥21 years) enrolled in the Singapore Cancer Society Rehabilitation Center between December 2021 and March 2023. Clinical data was collected from medical records and assessments included the Distress Thermometer (DT), Brief Illness Perception Questionnaire (Brief IPQ), and modified Bandura’s Exercise Self-Efficacy (ESE) scale. Adequate exercise was defined as ≥150 min/week of moderate intensity aerobic exercise and ≥2 days/week of resistance training. Regression analyses were performed to identify factors associated with meeting recommended moderate-intensity aerobic exercise levels. Results: The median age of our study cohort was 60.0 (IQR: 51.0–66.3) years. Of 132 analyzed participants, only 29.5% met recommended aerobic exercise levels and 9.1% met resistance training level recommendations. The median duration of moderate intensity aerobic exercise per week was 85.0 (30.0–180.0) minutes. The three most common cancer diagnoses amongst the participants were breast (53.8%), gastrointestinal (11.4%) and gynecological (7.6%) cancers. Clinically significant distress (DT ≥ 5) was present in 40.9% of patients. Multivariate regression analyses revealed that having lower ESE was the only significant psychosocial factor associated with a lower likelihood of meeting recommended aerobic exercise activity levels (OR = 0.856; 95% CI = 0.787–0.931; p < 0.001). Conclusions: Our study found a low prevalence of meeting recommended physical activity levels amongst our cohort of Asian cancer survivors enrolled in a community rehabilitation program in Singapore. Reinforcing the need for future longitudinal studies involving Asian cancer survivors to explore additional factors associated with physical inactivity and whether behavioral interventions targeting ESE will improve physical activity participation. Full article

Certified seed purity is a prerequisite for sustaining the agronomic performance and commercial value of Khao Dawk Mali 105 (KDML-105), Thailand’s premium aromatic rice; however, conventional inspection methods are destructive, labour-intensive, and poorly suited to high-throughput operations. This study developed a non-destructive purity inspection system integrating five-band MSI acquired using a MicaSense RedEdge-MX sensor with three machine learning classifiers: Extra Trees (ET), Random Forest (RF), and Support Vector Machine (SVM). The system was evaluated systematically across four illumination levels (360.90–13,188.46 lx) to discriminate KDML-105 from three morphologically similar contaminating varieties: Chainat-1, RD-6, and RD-15. Two-way ANOVA confirmed that classifier type was the dominant performance determinant (η² = 0.847). Under optimal illumination (L4, 13,188.46 lx), ET achieved the highest accuracy (88.8%), recall (86.6%), and F1-score (88.4%) with a training time of 0.098 s. External validation confirmed model generalisability: KDML-105 seeds were identified with 90.6% accuracy and 95.6% recall, while Chainat-1 and RD-6 yielded accuracies of 85.1% and 90.6%, respectively; RD-15 remained challenging (62.9%; 67.4%) owing to spectral proximity to KDML-105. These findings establish MSI combined with ET classification as a viable, cost-effective approach for automated seed purity screening in certified rice production. Full article

Although deep learning methods have shown high performance in chest X-ray classification, high accuracy alone does not guarantee reliable reasoning. A model may still exhibit pathological behavior, such as unstable evidence usage under harmless input changes, inconsistent reasoning across augmented views, excessive dependence on surrounding frame information, and appearance bias dominance, where prediction relies too heavily on intensity while neglecting texture and shape. In this paper, we propose a bio-inspired pathology-aware, factor-aware framework for explainable and reliable chest X-ray classification, inspired by biological vision principles such as figure–ground separation, selective attention, and balanced use of complementary visual cues. During training, the method regularizes appearance bias dominance through evidence-guided counterfactual perturbations that mimic cue-suppression analysis in biological perception, thereby revealing and penalizing excessive factor dependence. During testing, it evaluates model behavior using four criteria: reasoning stability, augmentation inconsistency, appearance bias dominance, and framing sensitivity. This combination enables the framework to go beyond conventional inference-time explanation by both correcting pathological behavior during training and exposing it during evaluation. From a biomimetic perspective, the framework encourages the model to separate relevant foreground anatomy from surrounding background and to avoid over-reliance on a single dominant cue. The proposed approach improves interpretability and reliability without modifying the backbone architecture or increasing model size or inference-time cost. The proposed training process improved the F1-score of DenseNet-121 from 0.899 to 0.931, while also producing more stable and balanced reasoning. Full article

Background: The rapid integration of Artificial Intelligence (AI), specifically Clinical Decision Support Systems (CDSS), into healthcare offers substantial efficiency but introduces critical ethical and legal challenges, particularly the perpetuation of systemic bias against older adults (“Digital Ageism”). While technological advances may improve care, they can violate fundamental bioethical principles when models are trained on unrepresentative data. Aim: This article argues that traditional clinical risk-management models are structurally insufficient to address opaque algorithmic bias and presents a conceptual, multidimensional governance framework designed to prevent the codification of human ageism into AI infrastructure. Methods: Drawing on systemic failures observed during the COVID-19 pandemic, the normative model integrates legal and governance standards aligned with the EU AI Act, Explainable AI (XAI) tools, and a three-phase implementation protocol. Results: To illustrate potential application without overburdening medical staff, the article introduces a theoretical Targeted Escalation Protocol and an Autonomous High-Load Safety Mode. The latter applies deterministic hardcoded constraints to contain age-dominant outputs during acute surges while preserving attending-clinician authority. The framework is explored through an Intensive Care Unit (ICU) thought experiment. Conclusions: The framework provides a structured roadmap for policymakers, ethicists, and healthcare administrators to move from reactive defensive medicine toward proactive ethical safety, safeguarding the dignity of the aging population while aiming to mitigate institutional legal exposure. Full article

Background/Objectives: Brain-derived neurotrophic factor (BDNF) and antioxidant networks mediated by Selenoprotein P (SEPP1) are core drivers of structural neuroplasticity and blood–brain barrier integrity, yet their co-regulatory behavior during subacute stroke neurorehabilitation remains poorly understood. This pilot study quantified concurrent changes in serum BDNF and SEPP1 during rehabilitation. Methods: Sixteen subacute post-stroke patients with mild stroke severity were assigned to an intensive multi-modal neurorehabilitation protocol incorporating high-intensity interval training (Treated, n = 7) or standard care (Control, n = 9); the biomarker sampling window averaged 73.4 days. Fasting venous blood was collected at baseline and post-intervention and analyzed by ELISA. Results: BDNF changes (ΔBDNF) differed significantly between arms (U = 57.0, p = 0.0081): the Treated cohort showed a uniform decrease (mean Δ: −0.240 ± 0.103 ng/mL), while Controls showed stabilization or a slight increase (mean Δ: +0.118 ± 0.339 ng/mL). ΔBDNF and ΔSEPP1 were significantly, positively correlated across the cohort (ρ = 0.596, p = 0.015). Conclusions: The BDNF decline in the Treated group is consistent with the “central sink” hypothesis, but may more plausibly reflect stress/cortisol-mediated suppression induced by the high training intensity, contrasting with increases typically reported after moderate-intensity subacute-phase exercise. The collinear coupling with SEPP1 suggests a link between neurotrophic synthesis and antioxidant buffering during post-stroke tissue remodeling, though this correlational finding does not by itself establish a single, coordinated mechanism. Full article