Tag: nevin manimala

Sci Rep. 2025 Sep 26;15(1):32954. doi: 10.1038/s41598-025-16149-4.

ABSTRACT

Cyberbullying (CB) has emerged as a growing concern among adolescents, with nearly 10% of European children affected monthly and almost half experiencing it at least once. Unlike traditional bullying, CB thrives in digital environments where anonymity and impunity are prevalent. Despite its increasing prevalence, understanding the causal mechanisms behind CB remains challenging due to the limitations of conventional statistical methods, which often rely on correlations and are prone to spurious associations. In this paper, we introduce a novel human-machine consensus framework for causal discovery, aimed at supporting social scientists in unraveling the complex dynamics of CB. We leverage recent advances in data-driven causal inference, particularly the use of Directed Acyclic Graphs (DAGs), to identify and interpret causal relationships from observational data. Our approach integrates automatic causal discovery algorithms with expert knowledge, addressing the limitations of both purely algorithmic and purely expert-driven methods, and allows for the creation of a model ensemble estimation of the causal effects. To enhance interpretability and usability, we advocate for the use of Probabilistic Graphical Causal Models (PGCMs), or Bayesian Networks, which combine probabilistic reasoning with graphical representation. This hybrid methodology not only mitigates cognitive biases and inconsistencies in expert input but also fosters transparency and critical reflection in model construction. Cyberbullying serves as a compelling case study where ethical constraints preclude experimental designs, highlighting the value of interpretable, expert-informed causal models for guiding policy and intervention strategies.

PMID:41006423 | DOI:10.1038/s41598-025-16149-4

Sci Rep. 2025 Sep 26;15(1):33098. doi: 10.1038/s41598-025-17736-1.

ABSTRACT

Artificial neuro-cognitive models can simulate human brain intelligence to enable accurate decision-making during complex agricultural operations in situ. To investigate this, twelve machine learning algorithms were employed to sequentially train 72 neurocomputing architectures using Deep Neural Network (DNN) and Artificial Neural Network (ANN) models for the neurocognitive prediction of tractive force (F_Tr). Fourteen soil-machine input variables were used, and the hyperparameters of the neuro-cognitive models were optimized through metaheuristic algorithms, targeting 50,000 neuro-perceptron epochs to minimize convergence error. The performance of the neurocomputing models was evaluated using standard accuracy metrics, including coefficient of determination (R²), root mean square error (RMSE), mean absolute error (MAE), and prediction accuracy (PA). The DNN Levenberg-Marquardt (trainlm) model with a 14-7-5-1 architecture demonstrated superior predictive performance (RMSE = 1.03e-5, R² = 1.0, MAE = 8.0e-6, and PA = 99.994), closely followed by the ANN Bayesian Regularization (trainbr) model with a 14-72-1 architecture (RMSE = 4.0e-4, R² = 0.9999, MAE = 0.0002, and PA = 99.935). Although the DNN trainlm model required slightly more epochs to reach optimal performance (55 vs. 51), it achieved faster computation (2s vs. 29s) than the ANN trainbr model. The reliability indices, i.e., a20-index (a20), scatter index (IOS), and agreement index (IOA), revealed that the DNN trainlm (14-7-5-1) and ANN trainbr (14-72-1) models are highly reliable. Notably, the ANN trainbr model attained the highest [Formula: see text] (= 240), while the DNN trainlm model was the most reliable under its configuration ([Formula: see text]= 196). Taylor’s analysis revealed no statistically significant deviation between the experimental and predicted F_Tr values for both models. Furthermore, all prediction instances (100%) for both trainbr (14-72-1) and trainlm (14-7-5-1) fell within the 95% prediction uncertainty range (PPU_95%), with a near-zero neurocognitive uncertainty index ([Formula: see text] = 0.00;0.03) and minimal logical deviation factors (d_factor=0.03;1.58), confirming Monte Carlo consistency between predicted and observed F_Tr values. The Anderson-Darling test confirmed the normality of the predictions ([Formula: see text] = 0.001), with both models satisfying the condition (P_model< P_ideal,0.05)​. Finally, the ANN model under trainbr also achieved excellent performance. The DNN trainlm (14-7-5-1) model employed a log-sig-tan-sig-purelin activation sequence, whereas the ANN trainbr (14-72-1) model used a tan-sig and purelin configuration, both suitable for in-situ prediction of F_Tr. In addition, it was observed that the Spider Wasp Optimization (SWO) algorithm enhanced the performance of the conventional DNN (trainlm) and ANN (trainbr) models in predicting agricultural traction.

PMID:41006420 | DOI:10.1038/s41598-025-17736-1

Horm Behav. 2025 Sep 25;176:105821. doi: 10.1016/j.yhbeh.2025.105821. Online ahead of print.

ABSTRACT

The care that a mother rat provides is essential for the ability of her pups to survive and thrive. Maternal care naturally varies between litters, including among animals with close genetic relatedness. There are also significant differences in behavior even among offspring reared together. Our lab and others have documented stable, naturally occurring individual differences in maternal care received by individual pups within the litter that persist throughout at least the first ten days of postnatal life. In this study, we hypothesized that within-litter variation in maternal care received constitutes a significant source of variation in offspring behavior and neurochemistry in Long-Evans rats. We analyzed measures related to maternal care behavior, offspring anxiety-like and social behaviors, and neurotransmitter levels in specific brain regions after the offspring became mothers themselves. For statistical modeling, we used the coefficient of variation (CV) to standardize and directly compare between- and within-litter variation across a range of behavioral and neurophysiological outcomes. Several variables analyzed showed greater within-litter CVs than between-litter CVs, especially for offspring behavior and levels of the monoamines dopamine, serotonin, and their primary metabolites DOPAC (3,4-dihydroxyphenylacetic acid) and 5-HIAA (5-hydroxyindoleacetic acid) in the nucleus accumbens, ventral tegmental area, medial preoptic area, hippocampus, and prefrontal cortex. Our findings suggest that within-litter variation in maternal care plays a prominent role in behavioral and physiological outcomes. This study provides a methodological advance by demonstrating that within-litter variability often exceeds between-litter variability across maternal, behavioral, and neurochemical domains, challenging a key assumption in experimental designs using littermate controls.

PMID:41004891 | DOI:10.1016/j.yhbeh.2025.105821

Appl Radiat Isot. 2025 Sep 23;226:112204. doi: 10.1016/j.apradiso.2025.112204. Online ahead of print.

ABSTRACT

There are very few radioactive environmental reference materials (RM) traceable to the International System of Units. Existing radioactive RMs for environmental samples that can be measured by mass spectrometry are even more limited and their characterisation does not always include relevant parameters such as isotopic ratios. This paper focuses on the development of two environmentally relevant candidate RMs, one liquid and one solid, which could be used for routine quality control measurements. The liquid RM was prepared by spiking seawater sampled from the North Sea, and therefore the matrix is representative of a real environmental sample, while the solid RM was prepared using a synthetic approach by spiking a mixture of silica precursors before a sol-gel reaction. The homogeneity, between-bottles and within-bottles, of both RMs was assessed using gamma-ray spectrometry and mass spectrometry. For the liquid RM, the variation among sub-samples was due mainly to the within-bottle variance, and was lower than 1 %, for all the radionuclides tested. For the solid RM, the ²⁴¹Am content measured with gamma-ray spectrometry revealed a statistically significant variation between-bottles, but was lower than 1 %. The ²³⁸U and ²³⁹Pu contents, measured by mass spectrometry, showed higher measurement variability (∼5 %), with the main contribution coming from within the bottles.

PMID:41004888 | DOI:10.1016/j.apradiso.2025.112204

Curr Res Transl Med. 2025 Sep 19;73(4):103544. doi: 10.1016/j.retram.2025.103544. Online ahead of print.

ABSTRACT

BACKGROUND: Ensuring accurate statistical reporting is critical in oncology research, where data-driven conclusions impact clinical decision-making. Despite standardized guidelines such as the Statistical Analyses and Methods in the Published Literature (SAMPL), adherence remains inconsistent. This study evaluates the performance of Gemini Advanced 2.0 Flash, an AI model, in assessing compliance with SAMPL guidelines in oncology research articles.

METHODS: A total of 100 original research articles published in four peer-reviewed oncology journals (October 2024-February 2025) were analyzed. Gemini Advanced 2.0 Flash assessed adherence to ten key SAMPL guidelines, categorizing each as “not met,” “partially met,” or “fully met.” AI evaluations were compared with independent assessments by a statistical editor, with agreement quantified using Cohen’s Kappa coefficient.

RESULTS: The overall weighted Kappa coefficient was 0.77 (95 % CI: 0.6-0.94), indicating substantial agreement between AI and manual assessment. Full agreement (Kappa = 1) was found for four guidelines, including naming statistical packages and reporting confidence intervals. High agreement was observed for specifying statistical methods (Kappa = 0.85) and confirming test assumptions (Kappa = 0.75). Moderate agreement was noted for summarizing non-normally distributed data (Kappa = 0.42) and specifying test directionality (Kappa = 0.43). The lowest agreement (Kappa = 0.37) was observed in multiple comparison adjustments due to missing justifications for post hoc tests.

CONCLUSION: AI-assisted evaluation showed substantial agreement with expert assessment, demonstrating its potential in statistical review. However, discrepancies in specific guidelines suggest human oversight remains essential for ensuring statistical rigor in oncology research. Further refinement of AI models may enhance their reliability in scientific publishing.

PMID:41004884 | DOI:10.1016/j.retram.2025.103544

Eur J Oncol Nurs. 2025 Jul 24;78:102944. doi: 10.1016/j.ejon.2025.102944. Online ahead of print.

ABSTRACT

PURPOSE: To statistically characterise both contemporaneous and temporal symptom networks to identify robust, stage-specific intervention targets.

METHODS: We conducted a prospective cohort study of patients with stage I-III breast cancer who completed the MD Anderson Symptom Inventory across each of four chemotherapy cycles (T1-T4, April 2024 to June 2025). Exploratory factor analysis (EFA) first identified co-occurring symptom clusters, which informed the construction of cycle-specific cross-sectional networks via Gaussian graphical models and bridge centrality metrics. We then built cross-lagged panel networks to estimate directed symptom effects across successive cycles and applied non-parametric bootstrapping to derive confidence intervals and assess network stability.

RESULTS: A total of 153 patients completed data collection across four chemotherapy cycles. Parallel analysis showed three relatively stable symptom clusters across the first three chemotherapy cycles (T1-T3), but consolidation into two clusters occurred at the final cycle (T4). Cross-sectional hubs influencing contemporaneous symptom interplay were identified via bridge symptoms in cycle-specific cross-sectional networks: loss of appetite (T1), sadness (T2, T3), and numbness (T4). The core temporal drivers predicting future symptom cascades were identified via cross-lagged panel networks (CLPNs): shortness of breath (T1→T2), sadness (T2→T3), and somnolence (T3→T4). The precision of edge-weight yielded narrow to moderate 95 % confidence intervals (CI) for both cross-sectional and longitudinal networks. Case-drop bootstrapping (1000 iterations) demonstrated moderate to strong stability in the rank order of centrality indices.

CONCLUSIONS: Our integrative network approach reveals a dynamic, dual-mechanistic symptom framework (cross-sectional and temporal key symptom drivers): early intervention should target somatic symptoms, mid-treatment efforts should disrupt psychological amplifiers, and late cycles require attention to neurotoxic perpetuators. This phase-specific roadmap offers precision in symptom management for patients undergoing chemotherapy.

PMID:41004877 | DOI:10.1016/j.ejon.2025.102944

Top Magn Reson Imaging. 2025 Sep 29;34(3):e0319. doi: 10.1097/RMR.0000000000000319. eCollection 2025 Oct 1.

ABSTRACT

BACKGROUND AND AIM: Magnetic resonance imaging (MRI) is essential for diagnosis but often induces anxiety, especially in claustrophobic patients, potentially affecting image quality. This study compared oxygen saturation, heart rate, and anxiety levels between claustrophobic and non-claustrophobic patients undergoing closed and open MRI in Erbil, Iraq.

MATERIAL AND METHODS: The comparative study was conducted from October 2024 to April 2025 in the Radiology Departments of Consultant Medical City and Top Med Medical Complex Centers in Erbil using purposive sampling. The questionnaire contained 3 sections: sociodemographic variables, the Claustrophobia Questionnaire, and the State-Trait Anxiety Inventory-State Subscale. Physiological measures (oxygen saturation and heart rate) were recorded at 3 timepoints: pre-, mid-, and post-MRI. Statistical analyses included one-way ANOVA, repeated measures ANOVA, post hoc tests, and both univariate and multiple linear regression, using SPSS version 26.

RESULTS: A total of 125 participants were involved in the study. The mean anxiety score was moderate, with higher levels in claustrophobic patients. Claustrophobia scores also fell within the moderate range, indicating psychological discomfort during the MRI procedure. Physiological measurements showed that claustrophobic patients, particularly those undergoing closed MRI, experienced elevated heart rates and reduced oxygen saturation compared to non-claustrophobic individuals. Statistical analysis indicated a strong positive association between anxiety and claustrophobia, with scan entry direction, age, and sex also being significant predictors of claustrophobic responses.

CONCLUSIONS: Claustrophobic patients undergoing closed MRI experience increased anxiety and physiological distress. Open MRI systems and pre-scan anxiety screening are recommended to enhance patient comfort and diagnostic outcomes.

PMID:41004856 | DOI:10.1097/RMR.0000000000000319

J Neurosurg Pediatr. 2025 Sep 26:1-9. doi: 10.3171/2025.5.PEDS24631. Online ahead of print.

ABSTRACT

OBJECTIVE: The authors’ objective was to assess the impact of telehealth on pediatric neurosurgical care access for underserved and rural populations in West Virginia. The authors explored how telehealth utilization varied over time, the socioeconomic benefits it provided to families, and its effect on visit completion rates compared with in-person appointments.

METHODS: Clinic visits from January 1, 2017, to May 31, 2023, at the sole pediatric neurosurgery clinic in West Virginia were retrospectively reviewed. The data included three types of outpatient visits: in-person, telemedicine satellite clinic, and MyChart video appointments. Initial statistical analysis focused on visit completion rates, distance traveled, and time and cost savings for families. Additional geospatial analysis used heat density mapping to recognize regional utilization patterns, and community-level socioeconomic variables were analyzed for correlation with visit type utilization.

RESULTS: Telehealth usage (telemedicine and MyChart) increased significantly during and after the COVID-19 pandemic. MyChart visits demonstrated the highest completion rates postpandemic. Telehealth visits saved families substantial travel time and cost, especially for those living more than 100 miles from the clinic. Geospatial analysis revealed that telemedicine usage was clustered in specific Appalachian regions, and in-person visits were more common among patients from economically distressed communities. Correlation analysis showed that higher poverty and unemployment rates were associated with in-person visit reliance, while telehealth adoption was lower in these populations.

CONCLUSIONS: Telehealth significantly enhances access to pediatric neurosurgical care for rural and economically disadvantaged families, reducing travel-related burdens and increasing visit adherence. However, economically distressed communities in Appalachia are less likely to use telehealth, possibly due to digital access issues or skepticism about remote care. Addressing these barriers is crucial to ensure equitable healthcare access. Further research should investigate structural and personal obstacles to telehealth uptake to improve service delivery for at-risk populations, ultimately fostering more inclusive and accessible healthcare options in remote areas.

PMID:41004850 | DOI:10.3171/2025.5.PEDS24631

J Neurosurg. 2025 Sep 26:1-7. doi: 10.3171/2025.5.JNS243160. Online ahead of print.

ABSTRACT

OBJECTIVE: Digital enhancement and visualization technologies, such as augmented reality (AR), are increasingly used in surgery. Rapid and accurate patient registration with minimal device confinements enables AR systems to increase efficiency, safety, and effectiveness, especially in urgent/emergency and/or bedside scenarios. The aim of this study was to quantitatively compare an AR headset-based neuronavigation system with a standard-of-care reference array-based neurosurgical stereotactic navigation system in a real-world setting.

METHODS: This clinical validation trial included adult patients undergoing cranial neurosurgery with stereotactic navigation at a single center from February 2024 to July 2024. Preoperative CT and MR images were acquired and used for construction of a 3D hologram model that included surface-based target fiducial markers for comparison. Preoperative images were stereotactically registered to the patient’s head using standard techniques. The registration coordinates for the fiducial markers (control) and registration time were recorded. The AR system was then deployed to create a separate stereotactic registration to the same preoperative images. A second set of registration coordinates for the fiducial markers (experimental) were acquired using the AR system, and the time for this process was also recorded. The Wilcoxon signed-rank test was used to assess differences in registration time, and a linear mixed-effects model (LMM) was used to conduct equivalence testing of coordinates between the control and experimental data.

RESULTS: Twenty patients (mean age ± SD 50.05 ± 14.38 years) were included in the trial. The mean baseline validation error of the control system was 0.73 ± 0.29 mm (range 0-1.0 mm). Using the control system as ground truth, the mean registration accuracy of the AR system was 2.16 ± 0.12 mm. LMM equivalence testing, conducted with margins of 3 mm and 2.5 mm, demonstrated statistical equivalence between the ground truth and AR system coordinates (p < 0.001 and p < 0.003, respectively). The time required for patient model registration using the AR system was a mean of 45.98 ± 15.00 seconds, which was significantly shorter compared with the control system (228.86 ± 100.06 seconds, p < 0.001).

CONCLUSIONS: The AR navigation system provided statistically similar registration accuracy and significantly faster patient model registration compared with the standard-of-care stereotactic neuronavigation system. AR navigation was accurate, fast, and had a minimal footprint, offering new opportunities to incorporate stereotaxis in low-resource, bedside, and urgent/emergency settings.

PMID:41004845 | DOI:10.3171/2025.5.JNS243160

J Comput Assist Tomogr. 2025 Sep 23. doi: 10.1097/RCT.0000000000001802. Online ahead of print.

ABSTRACT

OBJECTIVE: To assess whether few-shot prompting improves the performance of 2 popular large language models (LLMs) (ChatGPT o1 and DeepSeek-R1) in assigning Coronary Artery Disease Reporting and Data System (CAD-RADS™ 2.0) categories.

METHODS: A detailed few-shot prompt based on CAD-RADS™ 2.0 framework was developed using 20 reports from the MIMIC-IV database. Subsequently, 100 modified reports from the same database were categorized using zero-shot and few-shot prompts through the models’ user interface. Model accuracy was evaluated by comparing assignments to a reference radiologist’s classifications, including stenosis categories and modifiers. To assess reproducibility, 50 reports were reclassified using the same few-shot prompt. McNemar tests and Cohen kappa were used for statistical analysis.

RESULTS: Using zero-shot prompting, accuracy was low for both models (ChatGPT: 14%, DeepSeek: 8%), with correct assignments occurring almost exclusively in CAD-RADS 0 cases. Hallucinations occurred frequently (ChatGPT: 19%, DeepSeek: 54%). Few-shot prompting significantly improved accuracy to 98% for ChatGPT and 93% for DeepSeek (both P<0.001) and eliminated hallucinations. Kappa values for agreement between model-generated and radiologist-assigned classifications were 0.979 (0.950, 1.000) (P<0.001) for ChatGPT and 0.916 (0.859, 0.973) (P<0.001) for DeepSeek, indicating almost perfect agreement for both models without a significant difference between the models (P=0.180). Reproducibility analysis yielded kappa values of 0.957 (0.900, 1.000) (P<0.001) for ChatGPT and 0.873 [0.779, 0.967] (P<0.001) for DeepSeek, indicating almost perfect and strong agreement between repeated assignments, respectively, with no significant difference between the models (P=0.125).

CONCLUSION: Few-shot prompting substantially enhances LLMs’ accuracy in assigning CAD-RADS™ 2.0 categories, suggesting potential for clinical application and facilitating system adoption.

PMID:41004838 | DOI:10.1097/RCT.0000000000001802