143174 research outputs found
Sort by
Percutaneous coronary intervention, coronary artery bypass grafting and mortality from acute myocardial infarction in EU15+ countries, 2006–2020: a secondary analysis of the OECD database
No full textObjective Coronary revascularisation practices have evolved over the last three decades. This study sought to examine the variations in percutaneous coronary intervention (PCI) and coronary artery bypass graft (CABG) rates, alongside mortality from acute myocardial infarction (AMI) across a group of 16 high-income countries between 2006 and 2020.
Design Retrospective observational analysis using data from the Organisation for Economic Co-operation and Development (OECD) database between 2006 and 2020. Estimated annual percent change in revascularisation was analysed using Joinpoint regression model, and mortality rates were evaluated using the locally weighted scatterplot smoothing model.
Setting Publicly available data on PCI and CABG procedure rates alongside AMI mortality rate from 2006 to 2020.
Participants 16 countries from the OECD database.
Interventions Not applicable.
Main outcome measures Standardised PCI and CABG procedure rates and AMI age-standardised mortality rate (ASMR) from 2006 to 2020.
Results Over the 15 year period, 14.0 million PCI and 2.8 million CABG procedures were collectively recorded across 16 countries. PCI rates varied among nations, but from 2006 to 2020 increased in 11 of the 16 nations overall, led by Finland (+36.0%), Ireland (+34.5%) and France (+31.5%). Meanwhile, CABG rates declined in 14 out of the 16 countries, with Luxembourg (−71.3%), the UK (−62.6%) and Finland (−60.6%) experiencing the most substantial decreases. Throughout the study period, the PCI-to-CABG ratio increased, while AMI ASMR decreased consistently across all countries.
Conclusions Despite evidence supporting CABG over PCI in specific scenarios, CABG rates have declined, and PCI rates have increased. Possible factors for this trend may include patient preference and advancement in interventional techniques. The varied use of PCI among these nations, alongside a sustained decline in AMI mortality rates, may be expected given the importance of optimal medical therapy in the management of ischaemic heart disease. The results further suggest the significance of factors beyond revascularisation in driving improved outcomes.
Data availability statement
Data are available in a public, open access repository. All data are incorporated into the article and its online supplementary material. The original data is publicly available and can be accessed from: [dataset] https://data-explorer.oecd.org/
Modelling techniques to assess surgeons' cognitive workload: a systematic review
No full textBackground: Excessive intraoperative cognitive workload can impair surgical performance and threaten patient safety. Machine learning (ML) offers potential for real-time objective monitoring, yet existing studies vary widely in design, input data, and modelling strategies. This systematic review synthesises current evidence on ML-based approaches to estimate surgeon intraoperative workload.
Methods: Searches of nine databases (inception-March
2025) followed PRISMA 2020. Eligible studies applied ML to model cognitive workload in surgical or simulated
settings. Data were extracted on participants, input
modalities, labelling strategies, preprocessing, modelling approaches, and performance metrics. Quality was assessed using the Mixed Methods Appraisal Tool (MMAT).
Results: Fifteen studies were included. Electroencephalography (EEG, 60%), electrocardiography (40%), and eye-tracking (40%) were used most frequently, with 60% adopting multimodal configurations. Workload was labelled using questionnaires (40%), task difficulty (40%), or hybrid approaches (13%). Most studies (87%) addressed classification tasks, with reported accuracy ranging from 54% to 99.9%. Both classical algorithms (e.g., support vector machines, random forests) and deep learning
architecture (e.g., CNNs, LSTMs) achieved competitive
results. However, all selected studies relied on small cohorts (fewer than 30 participants) in simulated environments, with inconsistent reporting of preprocessing framework.
Conclusion: ML-based cognitive workload modelling is
feasible and shows strong performance. Adoption in clinical practice will require validated labelling frameworks, larger and more diverse cohorts, standardised preprocessing pipelines, and ecologically valid operating room datasets. Sensor practicality, comfort, and acceptability remain critical considerations for real-world deployment
Separable approximations of optimal value functions and their representation by neural networks.
No full textThe use of separable approximations is proposed to mitigate the curse of dimensionality related to the approximation of high-dimensional value functions in optimal control. The separable approximation exploits intrinsic decaying sensitivity properties of the system, where the influence of a state variable on another diminishes as their spatial, temporal, or graph-based distance grows. This property allows the efficient representation of global functions as a sum of localized contributions. A theoretical framework for constructing separable approximations in the context of optimal control is proposed by leveraging decaying sensitivity in both discrete and continuous time. Results extend prior work on decay properties of solutions to Lyapunov and Riccati equations, offering new insights into polynomial and exponential decay regimes. Connections to neural networks are explored, demonstrating how separable structures enable scalable representations of
high-dimensional value functions while preserving computational efficiency
Birational geometry of Calabi-Yau pairs and 3-dimensional Cemona transformations
No full textn this paper we develop a framework that allows one to describe the birational geometry of Calabi–Yau pairs (X, D). After establishing some general results for Calabi–Yau pairs (X, D) with mild singularities, we focus on the special case when X = P3 and D ⊂ P3 is a quartic surface. We investigate how the appearance of increasingly worse singularities in D enriches the birational geometry of the pair (P3, D), and lead to interesting subgroups of the Cremona group of P3
Impact of characterization on cross-calibration performance for multispectral sensors with SI-traceable satellite mission TRUTHS
No full textA new generation of satellites designed for low-uncertainty, SI-traceable measurements—termed
“SITSats”—marks a major advancement in Earth observation (EO) capability. These missions aim to enhance the performance and interoperability of the EO “system of systems.” Among them, the ESA Earth Watch Traceable Radiometry Underpinning Terrestrial- and Helio-Studies (TRUTHS) mission is designed
to serve as a “gold-standard” radiometric reference for cross-calibrating EO sensors in the solar reflective domain. In this work, uncertainties in cross-calibration comparisons arising from sensor characterization and design are investigated. A processing chain to prepare collocated data for uncertainty-quantified comparison is presented. This includes steps to perform spectral band adjustment and spatial resampling. Using the TRUTHS hyperspectral imaging spectrometer (HIS)
as the reference and Sentinel-2 multispectral imager (MSI) as the target, a simulation study based on high-resolution imagery assesses achievable comparison performance. A subset of uncertainty effects driven by sensor characterization is propagated through the spectral and spatial processing using a Monte Carlo approach. Sentinel-2 data are assumed at 10-m resolution, which is most sensitive to the errors considered. The results highlight the importance of sensor characterization, particularly inherent in-flight wavelength knowledge for target sensors, in such comparisons. Results from the simulation analysis give uncertainty estimates (k = 1) of 0.31% (blue), 0.50% (green), and 0.23% (red) for the combined error effects
arising from sensor characterization and geolocation uncertainty for comparisons over the Libya-4 desert pseudo-invariant calibration sites (PICS) using an instantaneous 205-m square comparison region. Results for more heterogeneous scenes, such as rainforest, still achieve uncertainties of 0.6%–1.2% for the red–green–blue (RGB) bands over a 200 × 200 m area. The uncertainty is driven largely by the spectral component—up to 1% due to the inherent Sentinel-2 wavelength knowledge of 1 nm across various representative scenes outside of the atmospheric absorption bands. While the impact of these uncertainties may decrease when considering a diverse range of scene types, they
introduce systematic errors when scenes share similar spectral characteristics. The impact of some uncertainty contributions, for example, geolocation uncertainty, is shown to be substantially reduced by aggregating samples over larger regions or over longer time periods. This analysis supports the development of low-uncertainty, ideally SITSat-enabled intercalibration
approaches needed to ensure radiometric consistency acrossmmissions for generating long-term climate data records
Sovereign AI in defence - strategic autonomy and collaborative opportunity
No full textWe present a strategic framework for achieving mission-driven and modular AI sovereignty, a model designed to offer practical guidance for governments, defence
institutions, and policymakers seeking to preserve legal authority, operational autonomy, and strategic freedom of action as artificial intelligence becomes increasingly
embedded in critical national security infrastructure. This framework builds on the principles developed through Imperial’s Trusted AI Alliance and extends our earlier
work on Sovereign AI and National AI Policy (2025) into the specific demands of the defence and security domain
How can energy-system models inform technology development? Insights for emerging energy-storage technologies
No full textEnergy-system models (ESMs) are used often to support policymakers, system operators, or investors, but
much less for offering guidance to technology developers. This paper contributes to bridging the gap between the ESM discipline and technology developers. For the example of emerging energy-storage technologies, we identify and categorise information needs of technology developers. Moreover, we discuss, which information needs can be met by an advanced analysis of ESM results, and which needs require fundamental model developments. We demonstrate the capabilities of an advanced analysis for an application study using a model that optimises investment and dispatch to assess energy-storage technology requirements in a fully decarbonised European power system. Our analysis provides insights regarding requirements and opportunities of energy-storage technologies in terms of design parameters, operational patterns, and target markets. We show that technologies with low-cost power, e.g., lithium-ion batteries (LIB), are designed with low energy-to-power (E2P) ratios, while those with low-cost energy (e.g., H2) have high E2P ratios. Concerning operational patterns, low-E2P energy-storage technologies cycle frequently (up to daily for LIB and vanadium-redox flow batteries), whereas H2 featuring a high-E2P cycles only a few times per year. Moreover, we find that requirements and cycling frequencies vary strongly between different target markets driven by their underlying electricity systems. We conclude that an advanced analysis can make a contribution to bridging the gap between ESMs and technology developers. Future work should improve the representation of technological details and develop inverse modelling approaches for technologies in very early development stages with still highly uncertain parameters
Exploring the relationship between task difficulty, head-related transfer function and spatial release from masking in a speech-on-speech experiment
No full textIt is known that individuals make use of spatial hearing cues to improve the audibility of a target signal and separate it from competing sounds. This phenomenon is known as spatial release from masking (SRM). Recent research has shown that this happens also when sources are located in the median plane, where interaural differences are limited. When assessing this within virtual conditions, it has been shown that employing individually measured head-related transfer functions (HRTFs) results in higher SRM abilities compared to using non-individual filters. In a previously published work, we found that Spanish speakers benefit from individual HRTFs when discriminating a target English speech from a single masker in the median plane. This study replicates the protocol of that previous work, varying the number of maskers and participants’ English proficiency levels to explore relationships among task difficulty and HRTF use. Results from a first experiment show that English speakers behave differently to Spanish ones; their SRM advantage is not significant. We suggest that this is due to their language proficiency, which allows them to rely on spectral glimpsing alone, that is, exploiting spectro-temporal gaps between voices rather than spectral cues introduced by spatial separation. A second experiment introduces a second speech masker, co-located with the first; by making the task more complex, participants seem to increase their reliance on spatial cues, resulting in significant effects of masker position and HRTF. This highlights a trade-off between the use of target glimpsing and spatial cues and the need for further exploration into how task difficulty influences SRM with different HRTFs
Modelling transmission thresholds and hypoendemic stability for onchocerciasis elimination
No full textThe World Health Organization (WHO) has proposed elimination of onchocerciasis transmission (EOT) in a third of endemic countries by 2030. This requires country-wide verification of EOT. Prior to the shift from morbidity control to EOT, interventions in Africa were mostly targeted at moderate- to high-transmission settings, where morbidity was most severe. Consequently, there remain numerous low transmission (hypoendemic) settings which have hitherto not received mass drug administration (MDA) with ivermectin. The WHO has prioritised the delineation of hypoendemic settings to ascertain treatment needs. However, the stability of transmission at such low levels remains poorly understood. We use the stochastic EPIONCHO-IBM transmission model to characterise the stability of transmission dynamics in hypoendemic settings and identify a range of threshold biting rates (TBRs, the annual vector biting rates below which transmission cannot be sustained). We show how TBRs are dependent on population size, inter-individual exposure heterogeneity and simulation time. In contrast with deterministic expectations, there is no fixed TBR; instead, transmission can persist between 70 and 300 bites/person/year. Using survivorship models on data generated from model simulations, we find that multiple vector biting rates can sustain hypoendemic prevalence for several decades. These findings challenge the assumption that hypoendemic foci would naturally fade out following treatment in nearby higher-endemicity regions. Our modelling suggests that, to achieve EOT, treatment should be extended to all areas where endogenous infection is identified, emphasising the need for improved diagnostic tools suitable for detecting low-prevalence infection and for strategies that allow safe treatment of communities where MDA would not be suitable
The impact of non-local parallel electron transport on plasma-impurity reaction rates in tokamak scrape-off layer plasmas
No full textPlasma-impurity reaction rates are a crucial part of modelling tokamak scrape-off layer (SOL) plasmas. To avoid calculating the full set of rates for the large number of important processes involved, a set of effective rates are typically derived which assume Maxwellian electrons. However, non-local parallel electron transport may result in non-Maxwellian electrons, particularly close to divertor targets. Here, the validity of using Maxwellian-averaged rates in this context is investigated by computing the full set of rate equations for a fixed plasma background from kinetic and fluid SOL simulations. We consider the effect of the electron distribution as well as the impact of the electron transport model on plasma profiles. Results are presented for lithium, beryllium, carbon, nitrogen, neon and argon. It is found that electron distributions with enhanced high-energy tails can result in significant modifications to the ionisation balance and radiative power loss rates from excitation, on the order of 50%–75% for the latter. Fluid electron models with Spitzer-Härm or flux-limited Spitzer-Härm thermal conductivity, combined with Maxwellian electrons for rate calculations, can increase or decrease this error, depending on the impurity species and plasma conditions. Based on these results, we also discuss some approaches to experimentally observing non-local electron transport in SOL plasmas