23343 research outputs found
Sort by
Hybrid deep learning framework for brain tumor type, grade, and segmentation in MRI images
No full textBrain tumours are a critical global health challenge, accounting for 85-90% of all primary central nervous system tumours, with an estimated 308,102 new cases globally in 2020. These tumours often lead to severe physical and cognitive impacts on patients, but early and accurate diagnosis is essential for improving survival rates through better treatment planning. Current diagnostic methods rely heavily on radiologist expertise, making the process time-consuming, expensive, and resource-intensive, placing strain on healthcare systems. In this thesis, we aim to address this diagnostic bottleneck by developing a unified hybrid deep learning framework namely Brain Tumor Diagnostics and Analysis (TDA) that aims to produce accurate diagnostic output for several vital processes within Brain Tumour Diagnostics including classification, grading, and segmentation in MRI images. A novel ensemble dual-modality approach for 3D brain tumor segmentation is proposed which combines dual MRI modalities to enhance segmentation accuracy. The ensemble model, integrating the best-performing dual-modality combinations, achieved a Dice Coefficient of 97.73% and a Mean IoU of 60.08% on the BraTS2020 dataset, significantly outperforming single-modality and dual-modality models. A parallel residual convolutional network (PRCnet) is proposed for tumour classification which leverages advanced techniques such as parallel layers with varied filter sizes and dropout layers to achieve an accuracy of 94.77% on dataset A and 97.1% on dataset B respectively. The framework further addresses key challenges in medical imaging through the development of novel data augmentation method namely orientated Combination MRI (OCMRI) to further enhance performance by mitigating class imbalance and data scarcity. OCMRI, through the controlled fusion of MRI images based on mean squared error thresholds, improve PRCnet classification accuracy. Additionally, the Brain Tumor Diagnostics and Analysis (TDA) framework includes an interactive web-based interface to support clinical interaction, enabling feedback-driven refinement through dynamic validation mechanisms. Also, the inclusion of explainability techniques such as Grad-CAM enhances clinical trust and interpretability. Overall, this thesis contributes an end-to-end, adaptable, and explainable AI-based diagnostic solution for brain tumor analysis, aiming to support radiologists and improve clinical decision-making processes
An exploration into the experiences of newly qualified midwives during their transition to practice in the UK: A systematic review
No full textAim: The aim of this systematic review was to explore the experiences of newly qualified midwives (NQMs) during their transition to practice in the UK. Background: It has been recognised that the transition from student to NQM is a time of significant change that has implications for staff recruitment and retention. Thus, the rationale for this review was that understanding the experiences of NQMs at this crucial time in their career is essential for developing systems of support to improve maternity care and no previous reviews were found on this topic based in the UK. Methods: A systematic search was conducted using the SPIDER – sample, phenomenon of interest, design, evaluation, research type – search strategy. Studies were critically appraised using the Consolidated Criteria for Reporting Qualitative Research Checklist and the Critical Appraisal Skills Programme Qualitative Studies Checklist. Relevant findings were extracted using thematic analysis in which text was coded, descriptive themes developed, and analytical themes generated. Results: The search returned eleven qualifying qualitative studies and findings were synthesised into four analytical themes – a journey of confidence; in at the deep end: varying support; the rollercoaster of transition; and the disparity between expectations and reality. Conclusion: NQMs experienced various challenges during the transition to practice such as lack of supernumerary time and unstructured rotations which were exacerbated by poor staffing and a heavy workload. A lack of support from colleagues and experiences of bullying increased stress and anxiety. However, structured preceptorship and support from midwifery colleagues facilitated a more positive experience
Using Lyα Absorption to Measure the Intensity and Variability of z ∼ 2.4 Ultraviolet Background Light
Full text linkWe present measurements of z ∼ 2.4 ultraviolet (UV) background light using Lyα absorption from galaxies at z ∼ 2–3 in the Hobby–Eberly Telescope Dark Energy Experiment (HETDEX) database. Thanks to the wide area of this survey, we also measure the variability of this light across the sky. The data suggest an asymmetric geometry where integrated UV light from background galaxies is absorbed by H i within the halo of a foreground galaxy, in a configuration similar to damped Lyα systems. Using stacking analyses of over 400,000 HETDEX LAE spectra, we argue that this background absorption is detectable in our data. We also argue that the absorption signal becomes negative due to HETDEX's sky-subtraction procedure. The amount that the absorption is oversubtracted is representative of the z ∼ 2.4 UV contribution to the overall extragalactic background light (EBL) at Lyα. Using this method, we determine an average intensity (in νJν units) of 12.9 ± 3.7 nW m−2 sr−1 at a median observed wavelength of 4134 Å, or a rest-frame UV background intensity of 508 ± 145 nW m−2 sr−1 at z ∼ 2.4. We find that this flux varies significantly depending on the density of galaxies in the field of observation. Our estimates are consistent with direct measurements of the overall EBL
Preface
No full textMargaret Thatcher remains the pre-eminent political figure of British post-war history. She is to this period what Winston Churchill was to the Second World War and David Lloyd George was to the First World War and its immediate aftermath. Thatcher, the politician, remains germane and worthy of study and discussion as much for what she represents in the public mind, as for the concrete achievements and failures of her administrations
Political connections and investment inefficiency: a machine learning approach
Full text linkThis study examined the impact of political connections on corporate investment inefficiency, focusing on the governance role of former politicians serving on the board of commissioners within a two-tier board structure. Various advanced machine learning algorithms were employed to analyse data from publicly listed Indonesian firms (a laboratory in this context), the results of which indicate that the presence of former politicians on a company’s board of commissioners significantly mitigates real investment inefficiency concerns. The Random Forest algorithm emerged as the best estimator, demonstrating the lowest root relative squared error level, closely followed by the Bootstrap aggregation (Bagging) technique. These findings significantly advance empirical aspects of the literature on political connections, highlighting the potential of former politicians on a two-tier board to effectively mitigate investment inefficiency. This study provides robust evidence of the advantages that former politicians bring to supervisory boards in Indonesia. The results provide valuable insights for regulators and capital market authorities and, moreover, have the potential to reshape the academic discourse on political connections and investment inefficiency
Diffusion of digital transformation initiatives in multi-tier supply chains: The double agency role of Tier-1 suppliers
Full text linkGrowing global supply chain complexity has generated interest in multi-tier management, yet most research focuses on focal firm and Tier–1 ties, overlooking Tier-2 dynamics. A specific gap exists in investigating Digital Transformation Initiatives (DTI) within triadic relationships among the focal firm and its Tier-1 and Tier-2 suppliers. This study addresses this gap by utilizing a multiple-case study approach across diverse sectors, drawing on 30 interviews, to examine the diffusion of DTI across multi-tier supply chains, with particular attention to the Double Agency Role (DAR) of Tier-1 suppliers. Using social capital theory and agency theory, our findings emphasize the key role of social capital in the DAR of Tier-1 suppliers. The importance of relational and cognitive capital varies as we shift from downstream to upstream segments of the supply chain. Relational capital is crucial in the downstream segment (focal firm–Tier-1 supplier), while cognitive capital grows in significance as we move upstream (Tier-1–Tier-2 suppliers). Additionally, we create a matrix to illustrate the relationships among social capital, the DAR, and the effectiveness of DTI diffusion. This study contributes to the digital transformation literature by expanding its focus from dyadic to triadic relationships. It also offers practical implications for supply chain managers, providing actionable insights to enhance the diffusion of DTI
Study on a novel solar light concentration and transmission system for applications in light-driven air dehumidification systems
Full text linkThe paper presents a systematic computer simulation work on a first-it-kind of solar light concentration and transmission system for applications in light-driven air dehumidification systems. This system consists of a double Fresnel lenses solar light concentration unit with sun tracking function and a tubular light transmission unit, it can track, collect, concentrate, collimate, and transmit solar light to site for air dehumidification. Numerical models were developed, and simulation work was carried out for performance investigation of the system. The simulation results showed the transmission efficiency is mainly determined by tube internal reflectivity and number of the light reflections inside the tube, the latter is affected by the tube geometry and the light incidence angle, the high number of the reflections lead to decreased transmission efficiency. A case study for a typical summer day at Hull (UK) showed that for a system with Ø1m/Ø0.5 m of double Fresnel lenses light concentrator and a Ø0.45 m/0.61 m of straight tube connected with a 90◦ bend tube and under average daily sun light of 393 W/m2, the concentrated light can reach 1647.5 W/m2, the output light from the system is 988.4 W/ m2 which is 4.2 times higher than the sun light of 393 W/m2 without concentration. In addition, the practical viability of the system validated by preliminary experimental work and discussions of practical feasibility of the system in building integration as well as mechanical design considerations, cost, maintenance and durability are also presented in this paper
P212 Screening for liver disease in alcohol use disorder – is it worthwhile?
Full text linkIntroduction Alcohol Use Disorder (AUD) is a major risk factor for alcohol-related liver disease (ArLD). Screening for ArLD in AUD may offer an opportunity for earlier detection or intervention. We sought to determine the burden of liver disease in a cohort of patients with suspected AUD referred for assessment to our hospital Alcohol Care Team (ACT).Method Universal screening for AUD is performed for all admissions to our hospital using electronic entry of FAST (modified AUDIT) score alongside initial observations. ACT referrals are generated automatically on a virtual dashboard for FAST5+ (i.e. possible alcohol dependence), or if a person is already known to ACT. Referrals can also be made where staff suspect AUD regardless of FAST. Following review, Enhanced Liver Fibrosis (ELF) test is advised for those with suspected AUD followed by Transient Elastography (TE) for those with non-reassuring result. We performed a case note review to identify the proportion of those with AUD who ultimately required 2-stage fibrosis testing, severity of ArLD identified, and reason for admission.Results All unique admissions referred to ACT from 1/11/22 to 30/6/23 were reviewed (n=857). 112 (13.1%) required 2-stage fibrosis assessment and were further analysed. The cohort was 63.4% male, mean age 56 (23-82) - representative of our AUD population. Mean FAST score was 9.4 (SD 5.1), indicating high levels of alcohol dependence. There was no direct correlation between FAST score and ArLD stage by ELF/TE, suggesting alcohol severity does not reliably predict ArLD severity. Mean ELF was 10.1 (7.7–17.2), and was >9.8 in 51.8% (n=58), demonstrating a high level of suspected advanced ArLD in the cohort. Mean TE was 11.3 (2.2 – 75 kPa). 38.4% (n=43) had abnormal TE (>7 kPa), with 24 (21.4%) suggestive of advanced fibrosis (>11 kPa). 62% of the group had normal TE, emphasising the opportunity ACT can offer for ArLD prevention, even in severe AUD. The 20.1% (n=23) admitted primarily due to alcohol were younger (mean age 52.8 vs 56) with higher FAST (11 vs 9.4), but there was no significant difference in ArLD severity versus rest of cohort. 53.6% (n=60) were admitted for non-alcohol/gastroenterology reasons and were identified at risk of ArLD only as a result of universal screening. Most frequent reasons for admission were falls (n=17; 15.8%) and abdominal pain (n=13; 11.6%) but we identified 19 admission triggers overall, including elective surgery.Conclusion We identified high rates of advanced ArLD in our AUD cohort, showing the potential value of screening in this group to give opportunities to modify ArLD trajectory via specialist AUD/ArLD management. As AUD severity did not directly correlate with ArLD extent, screening should be considered for all presenting with AUD. Universal screening identified many people who would not otherwise have had ArLD screening or AUD support, and broader adoption should be encouraged
Fast mapping in hominids
Full text linkFast mapping is essential when children acquire language, but whether the required cognition is uniquely human or shared with animals is debated. Although documented in dogs and cats, both species have a history of domestication of social cognition, so that it remains unclear whether fast mapping is naturally present in non-domesticated animals. Here, we used an eye-tracking paradigm to test three species of hominids – gorillas, orangutans and humans – in their ability to rapidly learn to associate novel sounds with objects in their everyday noisy environment. The task was difficult for all participants, but while adult humans showed evidence of fast mapping, we could not detect any sign of learning in the other hominids. These species differences could have trivial causes, such as problems with attention or motivation, but it is also possible that fast mapping requires a preexisting lexicon before becoming an effective learning mechanism, or that it has simply evolved after the shared ancestor of all great apes
Loneliness trajectories and dementia risk: Insights from the HUNT cohort study
Full text linkINTRODUCTION: Loneliness is postulated to be a risk factor for dementia. However, the findings are inconsistent, and long-term studies on this association remain scarce. METHODS: In all, 9389 participants self-reported loneliness in the Trøndelag Health Study (HUNT) in HUNT1 (1984–1986), HUNT2 (1995–1997), and/or HUNT3 (2006–2008) and underwent cognitive assessment in HUNT4 (2017–2019) at age 70 years or older. Logistic regression was employed to analyze the association between the course of loneliness and dementia, with those never lonely as a reference. RESULTS: In the fully adjusted model, the odds ratio (OR) for persistent loneliness was 1.47 (95% confidence interval [CI] 1.10, 1.95). This attenuated when adjusting for depression (OR 1.28, 95% CI 0.95, 1.72). DISCUSSION: Persistent loneliness from midlife into older age, as well as becoming lonely, were associated with increased odds of dementia, whereas transient loneliness in midlife was not. These findings underscore the importance of reducing loneliness. Clinical trial registration:: The study was registered with ClinicalTrials.gov (NCT04786561) and is available online. Highlights: Persistent and incident loneliness was associated with a higher risk of dementia. Transient loneliness was not associated with a higher risk of dementia. Loneliness 11 years before to the cognitive assessment was associated with dementia. Reducing the sense of loneliness might reduce or delay the onset of dementia