Ahmed G et al. · Jul 1, 2026
Precision livestock farming (PLF) leverages activity sensors to monitor behaviours like grazing, resting and walking, yet class imbalance in datasets often leads to underrepresentation of minority behaviours such as 'escaping' and 'being mounted.' This study proposes a novel framework combining long short-term memory (LSTM) networks with the synthetic minority oversampling technique (SMOTE) to address this challenge. Unlike existing methods that use complex SMOTE variants such as DeepSMOTE or latent space augmentations, which add computational complexity and overhead, our approach integrates simple SMOTE with non-overlapping windowed segmentation, preserving sequential patterns during synthetic data generation while augmenting minority classes. The LSTM architecture captures temporal dependencies in the balanced dataset, enabling robust behaviour recognition. Evaluated on a composite accelerometer dataset derived from three distinct cows, the framework generalises across breeds, overcoming limitations of breed-specific models. It achieves state-of-the-art performance with 97.24% accuracy, 97.56% precision, 97.24% recall and a 97.29% F1-score, significantly improving detection of rare behaviours without compromising majority class precision. By unifying data from multiple cows, the model ensures robustness to behavioural variability, enhancing scalability for diverse farming environments. The simplicity of using basic SMOTE reduces computational overhead, making the solution practical for real-world deployment. This work bridges classical data balancing techniques with modern deep learning, offering a resource-efficient blueprint for handling imbalanced time-series data in agricultural AI. The results advance precision livestock farming by improving the reliability of automated behaviour monitoring, directly contributing to enhanced animal welfare and farm productivity through accessible, breed-agnostic AI tools.
Veterinary
Kenea AM et al. · Jul 1, 2026
Background Natural mineral water (hora) is a mineral source widely available in Ethiopia. However, its specific impacts on sheep growth performance and carcass characteristics remain understudied. Objective To evaluate the effect of hora supplementation on nutrient intake, growth performance and carcass characteristics of sheep. Methods Twenty Horro breed sheep with an initial body weight (IBW) of 17.27 ± 0.71 kg were assigned to five treatment groups in a randomized complete block design (RCBD), blocked on the basis of their IBW. The treatments included control group (CON) fed basal diet only, and four treatment groups (H-200, H-300, H-400, and H-500) receiving 200, 300, 400, and 500 mL, of hora per day respectively. All sheep were fed grass hay and water ad libitum. Throughout the feeding trial data on dry matter (DM) and nutrient intake, body weight changes and various carcass traits were measured. Results Hora supplementation significantly improved (p 0.05). Total edible offal components showed a significant quadratic pattern (p = 0.001), whereas total non-edible offal showed a significant linear trend (p = 0.002). Conclusions Supplementation with 400 mL/day of hora is the optimal dose for significantly improving sheep growth performance, nutrient intake, and carcass characteristics. Overall, although hora provides clear benefits, its effects are dose-dependent, with potential diminishing returns beyond the 400 mL/day level.
Agricultural and Biological Sciences
do Carmo Júnior NM et al. · Jul 1, 2026
Purpose To determine the trajectories of benzodiazepine (BZD) and Z-drug use among older adults (n = 3590) followed in a network of geriatric outpatient clinics in Brazil. Methods This longitudinal study evaluated BZD and Z-drug use over a 24-month follow-up period. The proportion of use at baseline and at the end of the study period was compared using the McNemar test. Two trajectories were considered: started using and stopped using. Comparisons between trajectories and independent variables were performed using Pearson's chi-square test. Variables with p Results A small but significant reduction in overall BZD or Z-drug use was observed (17.4% to 16.0%; p = 0.007), with a marked decrease in isolated Z-drug use (6.1% to 4.0%; p Conclusions Clinical vulnerability and psychiatric conditions were associated with BZD or Z-drug use and initiation, while pharmacist consultations were associated with both initiation and discontinuation trajectories.
Psychology
Kim S et al. · Jul 1, 2026
Education for sustainable development (ESD) is increasingly recognized as a developmental imperative in a world facing ecological, social, and political challenges. While many approaches to ESD emphasize knowledge and competencies, emerging research suggests that sustainable action in youth requires the integration of cognitive, emotional, and identity-based developmental processes. We propose that Digital Civic Learning (DCL) serves as a developmentally appropriate educational approach that can scaffold this process in early adolescence. Using a sample of 149 4th and 5th graders in Mid-Western United States, we examined how civic competencies, namely civic knowledge, skills, and dispositions, relate to civic purpose and further investigated the mediating role of moral self-concept. Our findings underscore the importance of nurturing moral self-concept so that enhanced civic competencies translate into a form of civic purpose. SUMMARY: Conceptual innovation: It introduces Digital Civic Learning as a design-based, developmentally aligned model that integrates civic learning with sustainable development goals, aiming for transformative impact in education. Methodological advancement: It develops and applies a novel civic competencies coding scheme tailored for a nuanced analysis of students' civic knowledge, skills, and dispositions. Theoretical contribution: It positions moral self-concept as a developmental link connecting civic learning with a sustainability-oriented civic purpose in early adolescence, a critical period for identity formation. Empirical rigor: It employs performance-based assessments to evaluate students' civic competencies demonstrated in authentic digital classroom settings, emphasizing real-world application of civic reasoning within sustainability education.
Environmental Science
Nguyen HV. · Jul 1, 2026
Issue addressed Underage alcohol use has been linked to risks of physical, mental, and social harms to young people. Despite the known risks, research shows that parents may choose to supply alcohol to their children on occasions for various reasons. This has prompted several health promotional campaigns aimed at parents to discourage the practice of parental supply of alcohol, but there has been little evidence of how the messages are received by the target audiences, i.e., parents of adolescents. Methods Grounded in the existing literature on alcohol-related harms, social dimensions and communication of risks, the current paper conducted a qualitative analysis of interviews with parents of adolescents to understand their interpretation of risks in a series of Australian health promotion campaigns that addressed underage drinking. Result The study demonstrated how target audiences brought in their own lived experiences and social worldviews to interpret and internalise messaging about risks in ways that are nuanced and situational. Conclusions The findings demonstrated how parents' lived experiences and worldviews influenced their interpretation and alignment with the health promotion messages about parental supply of alcohol. While the ways the parents negotiated with the health promotion messages may not be scientifically-grounded, it was not always due to unawareness of risks but based upon strategies and assessment of risks in situational contexts. SO WHAT?: Understanding of how lived experiences inform interpretation of health promotion campaigns has implications for more effective alcohol-related risk communication aimed at behaviour change to reduce alcohol-related harms among young people.
Medicine
Fotou E et al. · Jul 1, 2026
Brain tumor therapy remains limited by the blood-brain barrier (BBB), which restricts drug access. BBB-penetrating peptides offer a promising strategy for delivering therapeutic and diagnostic payloads. Angiopep-2 is a well-established vector, yet novel radioconjugates based on this vector remain of interest. We report the synthesis and evaluation of DOTA-Angiopep-2 for radiolabeling with Lutetium-177 ( 177 Lu) and Terbium-161 ( 161 Tb). Notably, 177 Lu serves as a β- and γ-emitter, whereas 161 Tb is an Auger and β-emitter; both are utilized in therapy and SPECT imaging. Peptides were synthesized via solid-phase peptide synthesis. Cytotoxicity assays in T98 glioblastoma cells showed that Angiopep-2 is well-tolerated, maintaining ~100% viability at 20 μM and a moderate decline up to 100 μM. Radiolabeling achieved yields > 95% with excellent radiochemical stability at room temperature for up to 10 days and moderate stability in the presence of human serum. Biodistribution in healthy CFW mice showed a brain-associated radioactivity of 0.24% ± 0.05% IA/g at 5 min p.i. and a 12-fold increase in the brain-to-blood ratio (0.028-0.339) by 60 min p.i. These results support DOTA-Angiopep-2 as a versatile platform for radionuclide delivery and a potential candidate for future glioma-targeted studies. Further studies in tumor-bearing models are ongoing to evaluate therapeutic efficacy and translational potential.
Materials Science
Chintolo E et al. · Jul 1, 2026
Background Little is known about myocardial perturbations in African children hospitalised with severe anaemia. Methods An observational study nested within a clinical trial of blood transfusion was conducted on the paediatric ward in Blantyre, Malawi. Children were ages 2 months-12 years hospitalized with uncomplicated severe anaemia (haemoglobin 4-6 g/dl). By randomisation, 13 children received 30 ml/kg whole blood, 13 received 20 ml/kg whole blood and 26 had no immediate transfusion (usual care). We measured standard parameters of cardiac function using ultrasonic cardiac output monitoring (USCOM) at enrolment, 8 and 24 hours and discharge. Results Fifty-two children, median age 39 months (interquartile range [IQR] 25-58) and median haemoglobin 5.1 g/dl (IQR 4.8-5.6) were studied. Severe tachycardia and tachypnoea over time corrected faster in the transfused arms than the controls. At enrolment, the stroke volume index was within the normal range and 26/52 (50%) had a cardiac output index (COI) >97.5% the standard centile. The COI decreased in all arms by discharge but was greatest in the transfusion arms (p=0.05 for 20 ml/kg and p=0.009 for 30 ml/kg). A higher volume or receipt of whole blood did not worsen cardiac function. No child required diuretics. Conclusions The data generated by this small but granular study of haemodynamic and cardiac function provide reassuring physiological evidence showing the safety of higher doses of blood transfusion than currently recommended. It also supports the findings of a secondary analysis of the Transfusion and Treatment of Severe Anaemia in African Children trial indicating that whole blood transfusions are safe. These data support the new evidence-based paediatric transfusion algorithm for anaemic African children and its recommendation for safe use.
Medicine
Ricotti A et al. · Jul 1, 2026
Background Language development in early childhood varies considerably, making early detection of Developmental Language Disorders (DLDs) challenging despite their high prevalence and long-term effects on learning and mental health. In Italy, no culturally adapted, easy-to-use screening tools are currently available in primary care. To address this gap, a screening tool was developed to support the early identification of children aged 24-72 months at risk of DLD and other clinically relevant language difficulties. Aims To evaluate the psychometric properties and accuracy of the Comunicazione e Linguaggio in Ambulatorio Pediatrico (CLAP), a brief age-specific screening tool designed for use in Italian paediatric outpatient settings. Methods and procedures In this pilot validation study, children were recruited by primary care paediatricians during routine well-child visits and stratified into four age groups: 24-30, 36-42, 48-54, and 60-72 months. After administration of the CLAP screening tool, each child underwent a blinded speech-language pathologist (SLP) assessment. Psychometric evaluation included internal consistency, item-total correlations, confirmatory factor analysis, and item response theory indices (discrimination and difficulty). Diagnostic accuracy was assessed using ROC curves, area under the curve (AUC), sensitivity, specificity, and optimal cut-offs. Analyses were conducted separately for each age group. Outcomes and results Fifty children were enrolled in each age group; overall, 24% of the sample fell into the pathological subgroup after the blinded SLP assessment. Internal consistency was acceptable in the 24-30-month (KR-20 = 0.695) and 36-42-month (KR-20 = 0.777) groups, but lower in older children. Factor analyses supported a mainly unidimensional structure in the younger groups. Item response theory showed good discrimination and informativeness for several items. ROC analyses indicated excellent diagnostic accuracy in the 24-30-month group (AUC = 0.93; sensitivity = 92%; specificity = 87%), fair accuracy in the 36-42- and 48-54-month groups (AUC = 0.75 and 0.74), and poor performance in the 60-72-month group (AUC = 0.46). Conclusion and implications The CLAP demonstrates promising psychometric properties and good-to-fair accuracy as a brief screening tool for identifying children aged 24-54 months at risk of clinically relevant language difficulties, including those who may need further assessment for DLD. Its age-specific design, quick administration, and non-invasive nature support its potential integration into routine primary care. For older children, an age-specific revision or an alternative tool might be required. A larger validation study is currently in progress. What this paper adds What is already known on this subject Developmental Language Disorder (DLD) is common in early childhood; however, early identification remains difficult due to variable developmental pathways and the absence of validated screening tools in primary care. Currently, no brief, culturally adapted instrument is available for routine use in Italian paediatric settings. What does this study add to existing knowledge This study demonstrates that the CLAP tool has promising psychometric properties, with good accuracy in the youngest age group and fair accuracy up to 54 months for identifying children at risk of clinically relevant language difficulties, including those who may later meet criteria for DLD. It provides the first evidence supporting an age-specific, feasible screening option that can be integrated into Italian primary care, while also identifying areas requiring revision for older pre-schoolers. What are the potential or actual clinical implications of this work? CLAP can assist paediatricians in the early detection of clinically relevant language difficulties in children during routine well-child visits. Its adoption could help standardize early language screening in Italy, leading to earlier referral for further diagnostic assessment and appropriate speech-language evaluation.
Psychology
Can M et al. · Jul 1, 2026
Background Terbium-161 (Tb-161) emits gamma rays and beta radiation, enabling both therapeutic and imaging applications. However, the multiple gamma emissions of 161 Tb can affect image quality by increasing the scattering rate during SPECT imaging. To improve image quality, appropriate scatter correction methods, such as Dual-Energy Window (DEW) and Triple-Energy Window (TEW) need to be optimized. Although these methods are used in clinical practice, studies investigating the efficacy of spectral analysis approaches for next-generation radionuclides with multiple gamma emissions, such as 161 Tb, are limited. Purpose This study aims to evaluate the effects of DEW and TEW scatter correction methods on image quality and quantitative accuracy in 161 Tb SPECT imaging compared to uncorrected images. Methods Three distinct reference geometries were utilized to determine the gamma camera calibration factor (CF) and to evaluate the image quality parameters. Five image protocols were used, each with different combinations of main photopeak and scatter energy windows. Image quality parameters (CF, contrast-to-noise ratio (CNR), signal-to-noise ratio (SNR), spatial resolution) and activity bias (%) were compared using LifeX software in both uncorrected and corrected SPECT images. Results The CF calculated in images acquired with an energy window of 48.9 keV was found to be higher than that calculated in an energy window of 74.6 keV. The best CNR was calculated for images acquired in air with a photopeak of 48.9 keV ± 10% and a scatter window of 6%. In scattered media, it was observed in images obtained with a photopeak of 74.6 keV ± 10% and a scatter window of 6%. When scatter correction was applied, the images' SNR values decreased slightly, ranging from 1.188 to 1.724 across all media. Scatter correction techniques reduced the FWHM values for all protocols except the air medium, thereby improving spatial resolution. Activity bias results showed an overestimation in uncorrected protocols including the 48.9 keV peak. Conversely, utilizing the 74.6 keV peak with TEW correction improved quantitative accuracy to a 3.7%-4.2% absolute bias. Conclusion Our results show that both DEW and TEW correction approaches improve spatial resolution and increase CNR by reducing scattering contributions and background noise. However, as expected with the subtraction of scattered effects, these enhancements are accompanied by a slight decrease in SNR. The TEW method performed better than the DEW method in terms of quantitative accuracy under scattering conditions. In SPECT imaging using therapeutic amounts of Tb-161, high image quality can be achieved with an energy window of 74.6 keV ± 10%.
Medicine
Han Y et al. · Jul 1, 2026
Background Accurate segmentation of brain metastases (BM) is essential for diagnosis, stereotactic radiosurgery planning, and longitudinal assessment. However, manual contouring is time-intensive, limiting clinical scalability, and exhibits substantial inter-observer variability. This variability complicates objective assessment of automated segmentation methods and challenges interpretation of model performance. Purpose To address these limitations, we developed TUM-SAM, a hybrid foundation-model framework for fully automated BM segmentation, and introduced a bias-controlled, blinded multi-rater evaluation paradigm to determine whether AI-based BM segmentation has reached expert-level performance and whether AI-generated contours are preferred by human experts under unbiased assessment. Methods TUM-SAM integrates nnU-Net-based lesion detection with a tumor-adapted Med-SAM segmentation model to enable prompt-free, fully automated segmentation. Training used 301 patients (2548 lesions), and external evaluation used an independent cohort of 105 patients (397 lesions). Segmentation accuracy was benchmarked against DeepMedic and nnU-Net using Dice similarity coefficient (DSC) and 95th-percentile Hausdorff distance (HD95). Two physicians contoured all external cases, and a third physician contoured a 20-patient subset for a blinded, tumor-level, multi-rater preference study. Pairwise contour preferences were analyzed using a Bradley-Terry probabilistic model to obtain bias-adjusted estimates of relative contour quality while accounting for rater-specific tendencies and case difficulty. Results In the external cohort, TUM-SAM achieved a lesion-wise detection sensitivity of 0.94 and outperformed DeepMedic and nnU-Net across all tumor sizes, with a mean DSC of 0.84 and HD95 of 1.9 mm (nnU-Net/DeepMedic: DSC 3.3 mm). Across voxel-wise evaluation, TUM-SAM's geometric performance fell within the range of inter-observer variability among physicians and was sensitive to reference construction. In contrast, in the blinded rater study, experts preferred TUM-SAM-generated contours over individual physician contours in 81-87% of raw comparisons; Bradley-Terry analysis yielded conservative, bias-corrected win probabilities of 55-56%, indicating consistent preference after adjustment for rater and case difficulty. Conclusion Using a bias-controlled, blinded multi-rater evaluation framework, TUM-SAM demonstrates brain metastasis segmentation quality that is consistently preferred by expert physicians, highlighting the limitations of agreement-based voxel-wise metrics under inter-observer variability. These findings underscore the dependence of conventional evaluation on reference definition and support preference-based assessment as a complementary approach for evaluating AI segmentation quality in BM MRI.
Medicine