21827 research outputs found
Sort by
Human-Centered User Interface Design for Explainable AI in Chest Radiology: A Multi-Phase Co-Design Approach
Full text linkThe AI-powered computer vision is transforming medical imaging by improving analysis, diagnosis, and treatment. However, the opaque nature of deep learning often limits its adoption in critical clinical settings, where interpretability and trust are paramount. Explainable AI (XAI) aims at mitigating those limitations by creating visual and interpretable explanations of model decisions. However, poor user interface (UI) design frequently hampers the usability and clinical integration of these explanations. This paper addresses the critical need for human-centered UI design in XAI systems for chest radiology, a vital diagnostic domain for diseases such as pneumonia, lung cancer, and pulmonary embolism. Two deep learning–based XAI systems were developed for detecting pneumonia from chest X-rays and diagnosing COVID-19 from chest CT scans using post-hoc explanation methods, Gradient-weighted Class Activation Mapping (Grad-CAM) and Local Interpretable Model-Agnostic Explanations (LIME). Building on these systems, we introduce a novel multi-phase Human-Centered Design (HCD) methodology that actively involves radiologists and clinicians through participatory co-design, iterative prototyping, and multidisciplinary evaluation workshop. This process identified fifteen preliminary UI features tailored to clinical needs and led to a prototype XAI interface. Empirical evaluation from a multidisciplinary workshop revealed that radiologists preferred diagnostic prediction displays combining the original and AI-annotated images shown side-by-side or with adjustable overlays accompanied by explanatory text tailored to various audiences, including radiologists, clinicians, and patients. Participants agreed that highlighting the confidence scores of AI outputs aligned with clinical reasoning enhances perceived trust, diagnostic efficiency, and willingness to adopt XAI in daily practice. Our study demonstrates that collaborative multidisciplinary co-design is essential for bridging technical innovation and clinical utility, offering valuable insights for the future development of effective, trustworthy XAI user interfaces in medical imaging
CPFTransGAN: A Cross Perception Fusion Transformer-based Generative Adversarial Network for Head and Neck Cancer Dose Prediction in Radiotherapy
No full textRadiation therapy is one of the primary treatment modalities for head and neck (H&N) cancer in clinical practice, aiming to deliver sufficient dose to Planning Target Volume (PTV) while protecting surrounding Organs at Risk (OAR) from or minimizing exposure to radiation. Quantitative dose prediction of various tissues and organs is a prerequisite for implementing intelligent precision radiotherapy. In order to improve dose prediction accuracy, we propose a generative adversarial network CPFTrans- GAN based on Cross Perception Fusion Transformer (CPF Transformer). Specifically, we design a CPF Transformer module through deeply integrating CNN and Transformer. Using the CPF Transformer as basic unit, we constructed a generator with four-stage encoding-decoding structure called CPFTransGenerator. An adaptive weight loss is used to train the discriminator to alleviate the issues of imbalance training in adversarial learning. To further improve the prediction accuracy, a multiscale cross-window encoding network is designed, which can constrain the differences between predicted dose and the reference one at different granularity levels by calculating feature losses between them at different scales. The proposed method is evaluated on two public head and neck cancer datasets and a local clinical dataset. Extensive experiments demonstrate the superior performance of our method compared with the state-of-the-art ones
Evaluating the influence of grass distribution patterns on runoff and sediment yield dynamics: A flow path length perspective
No full textVegetation distribution patterns exert a first-order control on hillslope hydrology and erosion; however, the mechanisms by which spatial heterogeneity in vegetation regulates runoff generation and sediment yield remain inadequately understood. This knowledge gap constrains the development of physically based erosion models and effective soil conservation strategies. To elucidate how heterogeneous vegetation distributions govern hillslope hydrological connectivity and associated runoff–erosion processes, rainfall–runoff plot experiments were conducted under five vegetation distribution patterns—vertical strips (VS), horizontal strips (HS), X-shaped strips (XS), chessboard uniform distribution (CD), and random patchy distribution (RP)—with a bare slope (BS) serving as the control. Hydrological connectivity was quantified using relative flow path length (RFL), allowing systematic assessment of its influence on overland flow hydraulics, runoff and sediment yield. Results show that key hydrodynamic parameters respond nonlinearly to RFL and are well described by quadratic relationships (adjusted R² > 0.70). Mean flow velocity (v), stream power (ω), and unit energy (E) initially increased and subsequently declined with increasing RFL, reaching extreme values at RFL = 1. Under rainfall intensities of 60–120 mm·h⁻¹ , v, ω, and E increased by 100–114 %, 54–79 %, and 18–38 %, respectively. In contrast, flow resistance (f) and shear stress (τ) exhibited inverse responses, decreasing by 78–85 % and 11–29 % under the same conditions. Erosion rate (ER) also displayed a pronounced nonlinear response to RFL: as RFL increased from 0.513 to 1, ER rose by 65–118 %, with the sensitivity of ER to RFL diminishing at higher rainfall intensities. Building on these relationships, an erosion rate model coupling stream power ω and RFL was developed and validated using multi-source datasets. The model exhibited strong predictive skill and robustness, with adjusted R² and Nash–Sutcliffe efficiency (NSE) values exceeding 0.75, and substantially outperformed the Water Erosion Prediction Project (WEPP) model (adjusted R² = 0.316; NSE = −0.283). Overall, this study establishes a clear mechanistic link between vegetation-induced heterogeneity in hillslope hydrological connectivity and erosion dynamics, providing new insights for improving erosion modeling and designing vegetation-based soil conservation measures
Metformin: Antidiabetic actions from cells to tissues
Full text linkMetformin (dimethyl biguanide) is a primary pharmacotherapy to treat hyperglycemia in type 2 diabetes. It counters the effects of insulin resistance, improves glucose homeostasis, assists weight control and avoids overt hypoglycemia via reduced hepatic gluconeogenesis, increased splanchnic glucose turnover and greater peripheral glucose utilization. The underlying cellular actions of metformin differ between tissues and drug exposures. High concentrations of metformin (e.g. millimolar in the intestine) can interrupt the mitochondrial respiratory chain at complex 1, increase cytosolic NADH (favouring pyruvate conversion to lactate), decrease ATP synthesis, raise cytosolic AMP and activate AMP-activated protein kinase (AMPK). Lesser concentrations of metformin in liver can interrupt the respiratory chain at complex 4, which inhibits mitochondrial glycerol-3-phosphate dehydrogenase and impedes the mitochondrial glycerophosphate shuttle. Low concentrations of metformin (e.g., ∼10 μM) can activate AMPK by a lysosomal pathway without interrupting oxidative metabolism. While AMPK implements many of the metabolic effects of metformin, other contributing mechanisms include separate effects on metabolic pathways (e.g. inhibiting fructose-1,6-bisphosphatase) and signalling intermediates (e.g. inhibiting phosphatases) to reinforce the actions of insulin. Thus, the antidiabetic effects of metformin reflect diverse concentration-dependent cellular actions on nutrient metabolism and energetics in different tissues. The breadth of cellular actions of metformin encourages investigation of potential opportunities to assist in the management of cardiovascular, inflammatory, neoplastic and neurodegenerative disorders
Cybersecurity as a Dynamic Capability: How Micro and Small Social Enterprises Build Digital Resilience
Full text linkDespite the increasing digitalisation of organisational processes, the cybersecurity capability of micro and small social enterprises remains substantially underexamined within Information Systems research. These organisations occupy a critical yet vulnerable position in the digital ecosystem, handling sensitive beneficiary data while operating with informal structures, limited technical expertise, and mission-driven resource priorities. Existing cybersecurity maturity models assume formal governance and stable resources, providing limited insight into how capability emerges in such contexts. Addressing this gap, this study adopts a qualitative, inductive approach based on 23 semi-structured interviews with owners and managers of UK social enterprises to investigate how cybersecurity capability is developed and enacted in practice. Drawing on the Dynamic Capabilities View, the analysis reveals that cybersecurity capability in social enterprises is constituted through three interrelated and iterative dimensions: technical (readiness, prior exposure, and data sensitivity), organisational (informal coordination, training, and partnership-based support), and psychological (risk perceptions, ethical responsibility, and mission-driven motivation). The findings advance theory by showing that capability development does not follow linear maturity stages but emerges through experiential learning, social capital mobilisation, and values-aligned adaptation. The study contributes an empirically grounded Cybersecurity Capability Framework that explains how resource-constrained, mission-driven organisations sense threats, seize available resources, and reconfigure practices to maintain digital resilience. Practical implications highlight how managers, policymakers, and support organisations can strengthen cybersecurity capability by leveraging collaborative networks, informal learning mechanisms, and mission-aligned security practices. This work extends IS scholarship by illuminating an overlooked organisational form and by reconceptualising cybersecurity capability as a dynamic, context-dependent socio-technical process
Where angels fear to tread: FDI into sanctioned locations
Full text linkMultinational enterprises (MNEs) are increasingly challenged by the strategic implications of economic sanctions, which are imposed in response to geopolitical instability, international conflict, and violations of international norms. In this paper, we propose that superior resources and capabilities enhance the ownership advantages of MNEs, enabling them to pursue foreign direct investment (FDI) in sanctioned locations. We also build on institutional theory to examine contextual conditions and find that effective home country institutions deter investment to sanctioned locations and decrease the magnitude of the moderating effect of firm resources and experience. Moreover, being in a sanctioned location leads firms to invest more to other sanctioned locations because of the resulting specific ownership advantages. We test our conjectures on a large panel dataset and find support for our arguments. In post hoc analysis, we also examine the impact of sanctions on locational choice, highlighting that they have a deterrent effect. Our results have important implications for managers and policy makers in terms of international management and institutional dynamics
ADHD and adherence to antihypertensive medication treatment: a multinational cohort study
Full text linkBackground: Adherence to antihypertensive medication, alongside lifestyle modifications, is fundamental to managing hypertension and reducing the risk of cardiovascular disease. Attention-deficit/hyperactivity disorder (ADHD) is a common neurodevelopmental disorder associated with a range of cardiovascular diseases, including hypertension. ADHD medication has also been associated with hypertension. However, the influence of ADHD and ADHD medication on discontinuation and adherence to antihypertensive treatments is unknown. Methods: We conducted a multinational cohort study using electronic health databases from seven countries, which included adults who initiated antihypertensive medication between 2010 and 2020. ADHD was identified by a diagnosis of ADHD or dispensation of ADHD medications. The outcomes were (1) time to the first discontinuation of antihypertensive medication and (2) poor adherence, defined as the proportion of days covered (PDC) below 80% during 1-, 2-, and 5-year follow-up periods. We used Cox proportional hazards models and logistic regression to estimate associations, adjusting for age, sex, and calendar year of antihypertensive medication initiation. We pooled results from different countries via random-effects meta-analysis. Results: We identified 12,174,321 adults who initiated antihypertensive medication during the study period, including 320,691 (2.6%) with ADHD. In the pooled analysis across all countries, ADHD was associated with an increased rate of discontinuation in 5-year follow-up of antihypertensive medication (hazard ratio [HR] 1.14; 95% CI, 1.02–1.27). In age-stratified analyses, ADHD was associated with a higher rate of antihypertensive medication discontinuation in middle-aged (HR, 1.11; 95% CI, 1.01–1.23) and older adults (HR, 1.14; 95% CI, 1.01–1.29), but not in young adults. Individuals with ADHD also had higher odds of poor adherence across 1 year after treatment initiation (odds ratio [OR] 1.45, 95% CI 1.26–1.67) to 5 years (OR 1.64, 95% CI 1.34–2.00). Among those with ADHD, use of ADHD medications was associated with lower odds of poor adherence (1 year OR 0.66, 95% CI 0.60–0.73; 5 years OR 0.58, 95% CI 0.46–0.72). Conclusions: Adults with ADHD are more likely to discontinue antihypertensive treatment and exhibit poor medication adherence. However, ADHD medication use appears to be associated with better adherence among individuals with ADHD
Type 1 diabetes, ageing and frailty: an underexplored intersection
Full text linkOver recent decades, the life expectancy of individuals with type 1 diabetes has steadily improved due to advances in therapies that enhance metabolic control alongside better prevention and management of complications. However, this extended survival brings new challenges. Type 1 diabetes, through sustained hyperglycaemia and recurrent hypoglycaemia, may act as an accelerator of ageing, predisposing individuals to the development of geriatric syndromes such as frailty. Frailty, defined as a state of reduced physiological reserve that heightens susceptibility to stressors and impairs the ability to restore homeostasis after acute events, has emerged as a recognised complication of diabetes and has been associated with several adverse outcomes including increased risks of hypoglycaemia, hospitalisation, disability, institutionalisation and death. The putative pathophysiology of frailty in type 1 diabetes is complex and multifactorial. It reflects the direct effects of chronic exposure to hyperglycaemia and consequent micro- and macrovascular complications, superimposed on age- and diabetes-related hormonal changes. Additional contributors include sarcopenia, cognitive decline and other comorbidities. Currently, most of the literature on diabetes and frailty focuses on type 2 diabetes, while the relationship with type 1 diabetes and the impact on outcomes remain to be fully elucidated. In this review we discuss the growing evidence on the link between frailty and type 1 diabetes, explore its underlying pathophysiological mechanisms, discuss assessment and treatment strategies, and highlight the key knowledge gaps and suggest future research directions in this evolving field
Responsible leadership and front-line employees’ career commitment: the roles of status striving and perceptions of developmental HRM practices
Full text linkAdopting a responsible leadership (RL) lens and drawing on social learning and contextual approach, we investigate the link between first-line managers’ RL and front-line employee career commitment, the mediating role of status striving, and the contingent role of developmental human resource management (DHRM) practices. To investigate these relationships, we utilized both experimental and field-based research designs with samples of employees for service organizations. In Study 1, utilizing an experimental design, we found support for the direct effects of RL on career commitment via status striving. In Study 2, using a multi-wave research design, we replicated and extended the Study 1 findings and found that status striving mediated the RL-career commitment relationship. We also found that perceptions of DHRM practices moderated the indirect relationship between RL and career commitment via status striving. Our study findings highlight the value of studying leadership dynamics in the context of front-line employee career commitment, and that career research can benefit from taking a broader stakeholder view of leadership
A hybrid multi-agent and system dynamics approach for risk-informed selection of third-party logistics providers in supply chains
No full textIn recent years, global supply chains have faced significant challenges from uncertainties and disruptions, highlighting the urgent need for resilient and flexible decision-making mechanisms. In order to address these challenges, this study integrates Agent-Based Modelling (ABM) and System Dynamics (SD) with the multi-criteria decision-making (MCDM) method, providing a novel hybrid approach to third-party logistics (3PL) providers’ selection under multiple risk factors. This model captures the impact of risks on weighted criteria and decision-making behaviours within dynamic, multi-agent environments. A case study in the ceramics industry validates the model, identifying critical risk factors influencing supply chain decisions and proposing optimal strategies for stakeholders. Key contributions include a comprehensive methodology for risk-informed decision-making, a validated system architecture for hybrid models, and practical recommendations for resilient supply chain management. Despite some limitations, the findings demonstrate the potential of the model to enhance decision-making in uncertain environments, offering a valuable tool for practitioners
Data sources:
Access Repository Dashboard
Do you manage Aston Publications Explorer? Access insider analytics, issue reports and manage access to outputs from your repository in the CORE Repository Dashboard!