Explainable AI (XAI) is essential for building trust in Deep Neural Networks (DNNs). SHAP (SHapley Additive exPlanations) is a well‐known XAI technique for attributing feature importance, but it struggles with exponential computational complexity as the number of features increases. Various approximation methods have been suggested, but they compromise SHAP's theoretical principles. We introduce AA‐SHAP, a novel approach that derives superpixel affinity from the explained model's internals to identify and group superpixels. AA‐SHAP constructs a relevance‐consistency affinity between superpixel interdependence, enabling much faster SHAP calculations on a reduced set of meta‐superpixels while outperforming previous methods in explanation faithfulness. Exact Shapley values are computed on the reduced meta‐superpixel game, preserving all axiomatic guarantees within the aggregated feature space. Evaluated across multiple datasets, both convolutional and transformer classification architectures show that AA‐SHAP produces more faithful attributions than competing methods while improving computational speed and maintaining SHAP's theoretical axioms. The source code is available at https://github.com/vhasic/AA‐SHAP.

The mechanisms linking chronic hyperglycemia to intestinal inflammation and epithelial dysfunction remain incompletely understood, highlighting an important gap in our understanding of diabetes-associated gastrointestinal pathology. In this study, we investigated the effects of sustained hyperglycemia on intestinal inflammation, endoplasmic reticulum (ER) stress, and autophagy in a translational porcine model of diabetes. Diabetes was induced in Yucatan mini pigs using a high-fat, high-carbohydrate/fructose diet (HFHFD) followed by streptozotocin administration. Intestinal tissues from the terminal ileum and sigmoid colon were analyzed using histological evaluation, quantitative real-time PCR, and immunohistochemistry. Histological analysis revealed structural alterations in diabetic animals, including villous degeneration, crypt depletion, goblet-cell loss, and increased inflammatory-cell infiltration. Gene expression analysis revealed significant upregulation of inflammatory mediators (NF-κB, TNF-α, IL-6, IL-1β), inflammasome components (NLRP3), and macrophage markers (CD68, CD86, CD163). In parallel, ER stress-related genes (ORMDL3, ATF6) and autophagy-associated genes (NOD2, ULK1, ATG4a) were significantly elevated in diabetic pigs. At the protein level, increased expression of ER stress markers was confirmed in both intestinal regions, while autophagy-related proteins showed less consistent changes and did not fully reflect the observed transcriptional patterns, suggesting a potential disconnect between transcriptional activation and downstream autophagy-related protein expression under diabetic conditions. Chronic hyperglycemia is associated with intestinal inflammation and disruption of cellular stress pathways, including ER stress and autophagy, in a porcine model. These findings provide mechanistic insight into how chronic hyperglycemia contributes to intestinal dysfunction through coordinated alterations in inflammatory signaling, ER stress, and autophagy pathways, identifying these processes as potential targets for therapeutic intervention in diabetes-associated gastrointestinal disease.

ABSTRACT Background: More than three decades after the Bosnian war (1992–1995), its psychological consequences persist. While early post-war mental health outcomes have been extensively studied, little is known about the long-term psychological experiences of women who lost multiple family members. Objective: This study aimed to explore the current mental health status and lived psychological consequences of women from Srebrenica and surrounding regions who experienced multiple war-related losses. Methods: Two focus group discussions were conducted with 25 women. Audio-recorded data were transcribed, translated, and analysed using qualitative content analysis to identify central themes and subthemes. Results: Five overarching themes emerged: (1) enduring war trauma, loss, and emotional consequences; (2) justice, institutional responses, and systemic failures; (3) stigma and barriers to mental health care; (4) personal and everyday coping strategies; and (5) external support systems. Participants described persistent traumatic grief, chronic psychological distress, and ongoing somatic and emotional symptoms three decades after the conflict. Narratives revealed diverse coping strategies rooted in meaning-making, work, family, spirituality, and peer support. Institutional neglect, stigma, and limited access to mental health care were identified as major obstacles to recovery, whereas NGO-based and group-based interventions were experienced as particularly valuable. Conclusions: Thirty years after the Bosnian war, women survivors from Srebrenica and surrounding regions continue to live with profound and enduring psychological consequences of trauma and loss. Their narratives highlight the long-term nature of war-related suffering, the central role of social and institutional context, and the importance of culturally embedded, relational forms of support. These findings underscore the need for sustained, survivor-centered, and trauma-informed mental health services in post-conflict settings and contribute to a deeper understanding of long-term female survivor consequences following mass violence.

AIM To determine the proportion of patients with recorded psychiatric rehospitalisation within one year and examine crude associations with antipsychotic treatment patterns and documented medication adherence. METHODS This retrospective cohort study included 87 patients with International Classification of Diseases, 10th Revision schizophrenia (F20.0-F20.9) hospitalised in Brčko District during 2014-2018 and with documented 12-month follow-up. Relapse was defined as recorded psychiatric rehospitalisation due to worsening psychotic symptoms. A separate audit described discharge treatment in 236 hospitalisation records. Exact tests and crude odds ratios (ORs) with 95% confidence intervals (CIs) were calculated. RESULTS Recorded rehospitalisation occurred in 32/87 patients (36.8%). First-generation antipsychotic exposure (OR 2.11, 95% CI 0.74-6.03; p = 0.215) and antipsychotic polypharmacy (OR 2.16, 95% CI 0.82-5.65; p = 0.164) were not significantly associated with rehospitalisation. Medication non-adherence (OR 11.33, 95% CI 3.58-35.87; p < 0.001) and documented non-cooperation (OR 10.00, 95% CI 3.16-31.62; p < 0.001) showed strong crude associations. In the audit, treatment was identifiable in 183/236 records; polypharmacy was documented in 73.8% and long-acting injectable/depot use in 32.8%. CONCLUSION Documented non-adherence and non-cooperation were strongly associated with recorded rehospitalisation. The findings support structured adherence assessment and post-discharge follow-up but do not establish causal effects of antipsychotic class or regimen complexity.

Background Studies linking self-reported symptoms to circulating protein biomarkers are increasing, partly driven by the rising number of protein biomarkers being identified. Despite research advances, current reviews that synthesise these studies are typically disease-specific. A broader, diagnosis-agnostic and multidisciplinary scoping review can uncover shared biological patterns associated with symptoms across different conditions and can generate insights that advance precision health and symptom science. This scoping review protocol builds on this need for a broad approach as it aims to consolidate current knowledge on self-reported symptom and circulating protein biomarker associations, identify potential patterns, and provide relevant insights for clinical practice and future research. Methods The protocol is aligned with the PRISMA and JBI guidelines for scoping reviews. The search strategy includes original peer-reviewed journal research articles examining associations between self-reported symptoms and protein biomarkers in blood plasma or serum. Searches will be conducted from PubMed, CINAHL, Embase, and Web of Science databases. A preliminary search retrieved more than 30,000 articles, prompting a pilot test of the Artificial Intelligence (AI)-assisted review screening tool ASReview. The pilot demonstrated time savings, refined methodological decisions, and confirmed the feasibility of proceeding with the review. Therefore, screening will be conducted using ASReview, with a total of five reviewers involved in the process. Data extraction will focus on single self-reported symptoms and circulating protein biomarker pairs that have a reported association. Analysis will involve counting and mapping these associations to identify potential patterns. The protocol was registered with the Open Science Framework (OSF) https://osf.io/bku3f. Discussion This protocol provides a structured and transparent approach for conducting a large-scale scoping review. By adhering to established guidelines, the protocol provides a more comprehensive, standardised, exact, and reproducible accumulation of knowledge, while acknowledging potential limitations. The expected results can contribute to summarise the current understanding of associations between self-reported symptoms and circulating protein biomarkers. This fosters the integration of symptom science and the omics field, specifically proteomics, to advance precision health.

The application of sewage sludge to agricultural soils is increasingly considered a sustainable option for waste disposal, nutrient recycling, and soil fertility improvement. However, sewage sludge may also introduce potentially toxic elements (PTEs), particularly lead (Pb), into soil systems, where their environmental risk depends largely on chemical speciation rather than total concentration alone. Lead in sewage sludge-amended soils poses a potential threat to food sustainability through crop uptake and subsequent entry into the human food chain. This study investigated the effects of incubation time and soil moisture on Pb fractionation in sewage sludge–amended soil under controlled experimental conditions. The sandy loam soil, amended with untreated sewage sludge at rates of 10 and 15 g kg⁻¹ soil, was incubated for 16 weeks under two moisture conditions (50% and 70% saturation percentage). Soil samples were collected after 2, 4, 8, and 16 weeks and subjected to sequential extraction to determine exchangeable (F1), carbonate-bound (F2), Fe–Mn oxide-bound (F3), organic-bound (F4), and residual (F5) Pb fractions. The application of sewage sludge significantly increased total Pb concentration compared with the unamended control, with higher application rates showing higher levels of Pb accumulation. Across all treatments, Pb was predominantly associated with the F5, followed by F3 and F4. With increasing incubation time, Pb was progressively redistributed from more labile forms, particularly the F1, into less bioavailable fractions such as F3 and F4 pools. Higher soil moisture (70%) accelerated the decline in exchangeable Pb (F1) and enhanced Pb stabilization compared with 50% moisture. In contrast, the F5 remained relatively constant throughout the incubation period. The results indicate that both ageing time and soil moisture conditions regulate Pb transformation in sludge-amended soils. Maintaining high soil moisture levels may promote the transformation of Pb into stable geochemical forms, thereby reducing its mobility and bioavailability, and further mitigating potential environmental risks.

AIM To investigate the prevalence of arterial stiffness in hypertensive patients in Bosnia and Herzegovina and to assess the effects of various antihypertensive therapies on arterial stiffness. METHODS This prospective, open-label clinical study enrolled 200 patients who met predefined inclusion criteria from a total of 217 screened between 1 November 2023 and 30 April 2024. Participants were followed for six months. Thirteen distinct antihypertensive treatment combinations were evaluated. RESULTS In 200 patients, six months of antihypertensive treatment significantly reduced systolic blood pressure from 136.06 ± 15 to 130.69 ± 12 mmHg and diastolic blood pressure from 85.35±9 to 82.69±7 mmHg. Pulse wave velocity (PWV) changed only minimally, from 8.09±1.5 m/s to 8.05±1.2 m/s (p=0.354). Augmentation index (AIx) improved from 25.17% to 23.33%. The greatest reduction in PWV was observed with angiotensin-converting enzyme inhibitors plus statins (-0.19 m/s), while the largest reduction in AIx was seen with angiotensin-converting enzyme inhibitors, calcium channel blockers, beta blockers, diuretics, and statins (-15.64%, p<0.001). Baseline systolic blood pressure was the strongest, although weak, predictor of pulse wave velocity reduction (AUC=0.627, p=0.002). CONCLUSION The study demonstrated a high prevalence of increased arterial stiffness among hypertensive patients, with higher values associated with greater cardiovascular risk. Routine assessment of arterial stiffness, alongside standard blood pressure measurement, may improve cardiovascular risk stratification and support more effective, individualized hypertension management.

Limited training data constrains deep learning models for Auditory Attention Decoding (AAD) in hearing aids (HAs). AAD uses electroencephalogram (EEG) data to decode listener's attention, enabling real-time tracking of specific sound sources. However, achieving high AAD performance with short time windows typical in HAs (<=1s) is challenging due to the scarcity of real-world speech-evoked EEG data. To address this issue, we investigate diffusion probabilistic models (DPMs) for generating synthetic speech-evoked EEG data. DPMs learn the underlying complex data structure through a denoising process and can generate realistic samples suitable for data augmentation. We evaluate the use of synthetic EEG data for augmenting datasets in locus-of-attention (LoA) classification tasks. Our experiments demonstrate that DPMs can generate realistic EEG signals and that incorporating synthetic data significantly improves AAD performance compared to models trained solely on measured EEG data (p<0.05). These results highlight the potential of diffusion-based data augmentation to mitigate training data limitations and improve the robustness of short-window AAD models in HA applications.

Fluorescence-guided surgery improves intraoperative brain tumor visualization, but currently available agents remain unreliable for several common entities, particularly lower-grade gliomas and lesions without visible 5-aminolevulinic acid (5-ALA)-induced fluorescence. We prospectively analyzed 69 surgically obtained tumor specimens from 61 patients with WHO grade 4 gliomas, WHO grade 2/3 gliomas, meningiomas, and brain metastases before and after ex vivo incubation in 4 µM water-soluble high-load hypericin-polyvinylpyrrolidone complex (HHL-PVP). Fluorescence lifetime and intensity were quantified using a dual-tap CMOS camera system with a hypericin-specific 575-615 nm bandpass filter. HHL-PVP incubation significantly increased both fluorescence intensity and lifetime in all tumor entities, including specimens without visible 5-ALA fluorescence. An independently derived combined lifetime/intensity regression model discriminated pre- from post-incubation measurements with an area under the receiver operating characteristic curve of 0.975, sensitivity of 98.6%, and specificity of 82.6%; entity-specific areas under the curve ranged from 0.933 to 1.000. These findings support robust hypericin-associated signal detection across major brain tumor entities after ex vivo incubation and provide a rationale for future in vivo evaluation of HHL-PVP in fluorescence-guided neuro-oncological surgery.

This paper presents a comparative analysis of the production and economic results of Holstein-Friesian and Simmental cows on a modern dairy farm in Bosnia and Herzegovina. The research was conducted on a sample of 30 pregnant heifers of each breed, imported from Denmark and Austria, and kept under identical farm conditions. The analysis included age at first calving, milk production in the first lactation, revenue structure, variable costs, and gross margin. Data were obtained from the farm’s electronic production database, while descriptive statistics, a two-sample t-test, and gross margin analysis were used. The results showed that Holstein-Friesian cows achieved higher milk production, with an average daily yield of 34.52 litres compared to 28.39 litres in Simmental cows. This difference was statistically significant. Holstein-Friesian cows also achieved a higher total milk yield and higher gross margin. However, the Simmental breed generated higher additional income from the sale of male and female young cattle and culled cows, confirming the economic relevance of its dual-purpose character. Feed costs were the dominant variable cost category in both breeds. 

Sugar beet (Beta vulgaris ssp. altissima var. saccharifera Alef.) is an important raw material for sugar production and is widely used in human nutrition. The use of sugar beet in pekmez production is not widely distributed on a global scale and is generally limited to specific geographical and agroecological regions, including Central Bosnia and Herzegovina. In this region, the production of sugar beet pekmez is based on traditional methods, and data regarding its production and quality remain scarce. The objective of this study was to evaluate the technological characteristics of sugar beet as a raw material, determine the nutritional and sensory properties of sugar beet pekmez, and assess the influence of agroecological growing conditions on variations in its quality. The mechanical composition, physicochemical parameters of both the raw material and pekmez, as well as the sensory properties of pekmez produced during two consecutive production seasons (S1 and S2) were investigated. Sugar beet contained 17.63°Brix soluble solids and 15.80% total sugars, while the processing recovery reached 68.76%. The pekmez yield was 4.76%. Pekmez samples produced in season S1 showed significantly higher (p˂0.05) viscosity, total phenolic content, soluble solids, total sugars, and sucrose content, as well as superior sensory quality in terms of color, taste, and consistency, whereas odor was evaluated equally for both samples. No significant differences were observed in titratable acidity and directly reducing sugar content between the two seasons. The obtained data represent the first step toward standardizing sugar beet pekmez production using traditional methods.