Háskólinn í Reykjavík

Varanleg URI fyrir þennan undirflokkhttps://hdl.handle.net/20.500.11815/5667

Skoða

Niðurstöður 1 - 20 af 686

Understanding funding-related bias in gambling research : insights from unhealthy commodities industries and medical/pharmaceutical research–a scoping review
(2026-08-09) Dapo-Famodu, Adetoun; Young, Matthew M.; Tuico, Kyle; Dymond, Simon; Ludvig, Elliot A.; Roberts, Amanda; Rogers, Jim; Sharman, Steve; Stark, Sasha; Tunney, Richard J.; Department of Psychology
Despite the risk of funding-related bias (RoFRB) introduced by industry funding in gambling research, empirical investigations remain limited. To inform understanding of RoFRB in gambling research, this scoping review synthesized 52 peer-reviewed studies (2019–2024) spanning gambling, other unhealthy commodity sectors, and medical research–fields with similar commercial dynamics to gambling and more extensive examination of RoFRB. Findings indicate that industry influence manifests through identifiable patterns and mechanisms, at both micro (individual study) and macro (evidence base and research ecosystem) levels, with varying detectability. While industry funding is associated with influence, the extent of funder involvement in the research process appears to shape the relationship between funding source and RoFRB. Compared to other sectors, gambling studies show inconsistent uses of the term ‘industry funding,’ narrower explorations of industry influence, and more limited methodological approaches to assessing RoFRB. We propose a two-dimensional framework mapping influence by level and detectability, along with a typology of funders to improve definitional consistency. To improve the identification and mitigation of funder influence in gambling research, stronger disclosure standards, adoption of open science practices, and development of a dedicated RoFRB assessment tool are recommended.
The potential for groundwater heat pump systems for mitigating urban heat islands : aquifer suitability evaluation in Canada’s major cities
(2026-09-17) Lee, Victoria; Rivard, Christine; Raymond, Jasmin; Parent, Michel; Lavoie, Denis; Comeau, Félix Antoine; Newson, Juliet
Urban heat islands (UHI) are a growing issue in the context of global warming and pose significant threats to the public’s health and well-being. The heat rejected into the city air from conventional air conditioners is a significant contributor to UHI. Geothermal energy is an alternative for mitigating this issue by redirecting this heat into the subsurface. This study assesses the suitability for open-loop geothermal technology of 10 selected Canadian cities that commonly experience UHI effects during summer, focusing on groundwater heat pump (GWHP) systems. Utilizing six key criteria related to hydrogeology and geochemistry, an evaluation tool in the form of an index is developed to perform a preliminary assessment of aquifer suitability. Results indicate that Vancouver, Winnipeg, and Kitchener/Waterloo/Cambridge are cities with highly favorable conditions for implementing GWHP systems. The other evaluated cities (Toronto, Montreal, Calgary, Edmonton, Ottawa/Gatineau, Quebec City, and Saskatoon) show moderate potential. While these results are contingent on the hydrogeological data available for these major cities, which are often limited, this study, nonetheless, yields promising results for several Canadian cities and has allowed to verify the usefulness of the developed index for conducting systematic preliminary assessments of site suitability.
The Strengths and Difficulties Questionnaire (SDQ) parent-report : bifactor structure and method effects in a community-based sample
(2026-04-19) Thorlacius, Örnólfur; Valborgarson, Auðun
The Strengths and Difficulties Questionnaire (SDQ) is a widely used screening measure for child mental health, yet findings regarding its latent structure have been inconsistent. The present study examined the structural validity of the SDQ parent-report in a large community sample of mothers of Icelandic children aged 6–12 years (N = 2332). Competing measurement models were evaluated, including the original five-factor structure, second-order and bifactor models, estimated with and without a method factor capturing covariance among positively worded items. Model comparisons indicated that a bifactor model including two broad general factors—Externalizing and Internalizing problems—together with a Prosocial Behavior factor and an orthogonal wording-related method factor provided the most interpretable representation. The Externalizing (ωH = 0.69; ECV = 0.65) and Internalizing (ωH = 0.58; ECV = 0.55) factors accounted for a substantial proportion of reliable variance among the symptom-based items, whereas the specific SDQ factors, particularly Conduct Problems and Peer Problems, contributed relatively limited variance beyond these broader factors. The wording method factor accounted for a notable share of covariance among positively worded items, particularly those from the Prosocial Behavior domain, consistent with structured wording-related variance. Sensitivity analyses using alternative specifications, including exploratory structural equation modeling and a random-intercept factor approach, yielded converging results. The findings highlight the influence of wording-related covariance and skewed and restricted response distributions, underscoring the importance of accounting for these sources of variance when interpreting SDQ scores. Broad Internalizing and Externalizing scores may therefore provide more stable indicators of child mental health difficulties in non-clinical populations.
Thoracoabdominal Asynchrony in Healthy Children
(2026-10) Dereksson, Kristjan; Lund, Sigrun Helga; Markström, Agneta; Sigurdardottir, Sigridur; Arnardottir, Erna Sif; Clausen, Michael V.; Sigurdardottir, Sigurveig T.; Helgason, Tryggvi; Eklund, Erik A.; Department of Engineering
Measuring asynchrony between the thorax and abdomen during sleep provides information about inspiratory effort and upper airway obstruction in children, aiding in the diagnosis of obstructive sleep apnea (OSA). This study aimed to define a normative range for thoracoabdominal asynchrony (TAA) in healthy children by studying polysomnographies of 113 children aged 10–13 years. The percentage of total sleep time (TST) spent with elevated TAA, > 40°, was measured, as well as obstructive apnea-hypopnea index (OAHI), sleep position, sleep stage and snoring. Subjects without measurable OSA (N = 94) were defined as healthy subjects and they had elevated TAA for an average of 3.3% of their TST while subjects with OSA (N = 19) had elevated TAA for 10.7% of their TST (p < 0.00001). The cut-off point for the normal prevalence of elevated TAA was estimated at 9.6% of TST, which can be rounded up to 10% for ease of use. TAA prevalence rose in REM-sleep and in the supine position and was related to increased snoring. Children with obesity had higher OAHI, more prevalent TAA and more snoring. No gender differences were observed. Parent-reported PSQ scores showed limited concordance with objective measures. TAA assessment, available from standard RIP-bands placed around chest and abdomen, may aid in the interpretation of sleep studies in children, both home sleep apnea tests and polysomnographies. With further research, they might help guide clinical decision making in paediatric sleep medicine.
Technology Matters : From innovation to implementation – Nationwide self-referral for digital mental healthcare is possible
(2026-08-18) Halldorsson, Brynjar; Kristjansdottir, Hafrun; Islind, Anna Sigridur; Department of Psychology; Department of Sport Science; Department of Computer Science
Digital mental health interventions for children are developing rapidly, yet a persistent gap remains between demonstrating that they work and making them available to families in routine care. This article describes Iceland's national rollout of the Online Support and Intervention for Child Anxiety (OSI) programme, a government-funded digital mental health intervention for parents of children aged 5–12 with common anxiety difficulties, adapted for Icelandic primary care. Its broader importance, we argue, rests not only on the intervention content but also on the service model through which it was delivered. We describe three features: a deliberately simple self-referral pathway that retains clinical oversight, delivery embedded within existing primary care services and national rollout achieved through collaboration across seven health institutions. We argue that scaling digital mental healthcare demands substantial organisational, technical and governance infrastructure and is rarely simpler than face-to-face provision. While Iceland's context may not be directly replicable in other countries, the implementation principles it illustrates may be informative for clinicians, service leaders and policymakers internationally. Improving access depends on implementation as much as innovation.
Relapse in disordered gambling : a systematic review from a biopsychosocial perspective
(2026-05-07) Torrance, Jamie; Quigley, Martyn; Heath, Conor; Cousins, Alecia; Dighton, Glen; Wright, Simon; Dymond, Simon; Newall, Philip; Department of Psychology
Background: Relapse is common among individuals undergoing treatment for gambling disorder (GD). Understanding the biological, psychological, and social (biopsychosocial (BPS)) factors that influence relapse in GD is important to guide clinical practice and improve treatment efficacy. However, evidence synthesis in this area is warranted to provide clarity and cohesion to the literature. Method: A systematic review of empirical research (2015–2026) was conducted. Included studies focused on factors that either predicted relapse in GD, or those that offered protection against relapse. Findings were synthesized according to the BPS model (PROSPERO ID: CRD42024521445). Results: Twenty-six studies were included in the review. Overall, studies lacked heterogeneity, with most conducted in Spain (n = 19) and deriving from the same participant groups (n = 14). Biological factors such as younger age were associated with higher rates of GD relapse. Psychological predictors encompassed alcohol/substance use, personality traits (higher harm avoidance and lower self-directedness), cognitive factors (cognitive distortions, low cognitive flexibility, impulsivity), and poor psychopathological state. Social factors such as marital status (being single, divorced, or separated), lower education levels, involvement in crime, and exposure to gambling advertising also contributed to relapse risk. Relatedly, familial support, engaging in meaningful activities, and attendance at self-help meetings were protective against relapse. Conclusions: Overall, a broad range of BPS factors influence GD relapse. Despite the lack of heterogeneity in the included studies, these findings may contribute to better therapeutic understanding and practice. Expanding this research area is crucial for improving long-term recovery outcomes in GD.
(Mis)Comprehension and (Mistaken) Attractiveness of Financial Gambling Inducements among UK Bettors
(2026-03-25) Torrance, Jamie; Wright, Simon; Newall, Philip; Crawford, Tomos; Quigley, Martyn; Dymond, Simon; Department of Psychology
Financial gambling inducements such as ‘free’ bets and welcome bonuses involve complex wagering requirements that many consumers miscomprehend. UK regulations cap these wagering requirements at x10 (i.e., users must wager x10 the bonus before withdrawal) but don’t require worked examples to aid comprehension. This study assessed UK bettors’ comprehension of inducement wagering requirements, whether comprehension varied by gambling severity, whether perceived attractiveness differed when worked examples were displayed, and bettors’ broader perceptions of inducements. A between-subjects, mixed-methods online experiment randomised UK bettors (n = 585) to view a welcome bonus inducement either with or without (control) a worked example detailing its wagering requirement. Participants completed measures of comprehension, perceived attractiveness, gambling severity, and qualitative questioning. Participants in the control condition significantly underestimated the inducement’s wagering requirements (p<.001, r=.94), with a median estimate of £500 versus the correct value of £750. Underestimation did not differ by gambling severity. The inducement was rated significantly less attractive (p<.001, η²=.18) when the worked example was displayed (M = 2.39, SD = 1.46) compared to the control (M = 3.75, SD = 1.48). However, this effect was moderated by gambling severity (B=-0.112, p=.005), with the reduction in perceived attractiveness associated with the worked example becoming smaller as gambling severity scores increased. Qualitative analysis identified three overarching perceptions of inducements: as manipulative, economically worthless, and requiring better regulation. UK bettors significantly underestimate inducements’ wagering requirements, while worked examples significantly reduce their mistaken attractiveness. These findings demonstrate how informed choice can be enhanced via worked examples.
Lost in translation? Not anymore : a Mokken scale analysis of the revised Icelandic translation of the PHQ-9
(2026-01-20) Jóns, Laufey Ósk; Kristófersdóttir, Kristín Hulda; Vésteinsdóttir, Vaka; Kristjánsdóttir, Hafrún; Karlsson, Thorlakur; Thorsdottir, Fanney; Department of Sport Science; Department of Psychology
The Patient Health Questionnaire-9 (PHQ-9) is among the most widely used self-report instruments for assessing depressive symptoms. Although previously translated into Icelandic and widely adopted in clinical and research contexts, recent studies identified notable translation issues that may have affected its psychometric quality. This study examined a revised Icelandic version of the PHQ-9, designed to improve clarity and fidelity to the original English version. A general population sample (n = 139) completed the revised version in an online survey. Mokken scale analysis was conducted to evaluate scalability, monotonicity, local independence, and invariant item ordering. Results indicated that the revised version demonstrated stronger psychometric properties compared to the original translation. The overall scalability coefficient suggested a strong scale (H =.548) according to Mokken’s criteria; however, the confidence interval (95% CI [.466,.630]) crossed.50, indicating that strong scalability cannot be guaranteed at the population level. All items surpassed the conventional.30 benchmark, with most achieving Hi ≥.50. No significant violations of monotonicity were observed, and only minor indications of negative local dependence emerged. While partial evidence for invariant item ordering was found, it remained limited. These findings suggest that the revised Icelandic PHQ-9 is a more internally consistent and interpretable assessment of depression severity, broadly comparable to findings reported in other language versions. The scale appears suitable for use in research settings and for screening purposes in Iceland, although further validation in clinical populations is needed.
Intolerance of uncertainty and its impact on university students’ confidence, awareness and experience with employer selection tests
(2026-07-14) Quigley, Martyn; Kamenova, Zdravka; Blackmore, Maya; Dymond, Simon; Bradley, Alex; Department of Psychology
Intolerance of Uncertainty (IU) is a disposition typified by negative reactions to uncertainty and may influence how university students engage with career development opportunities. IU can be ‘Prospective’ or ‘Inhibitory’. Prospective IU is characterised by a ‘desire for predictability’, whilst Inhibitory IU is characterised by inaction due to ‘uncertainty paralysis’. In this study, we surveyed 568 students to examine whether Inhibitory and Prospective IU (as measured by the IUS-12) were associated with students’ attendance at careers’ events and their awareness of, and confidence and experience with, commonly used graduate employer selection tests. Results revealed that higher inhibitory IU scores were associated with lower awareness, confidence and experience with employer selection tests but not attendance levels. Prospective IU scores were positively associated with confidence and experience levels. These novel findings demonstrate that IU is a factor impacting on students’ employability. Recognising the role of inhibitory IU and cultivating a tolerance to uncertainty may help to improve students’ employability.
Gender bias in lending-based crowdfunding : the case of technology development projects
(2026-08-23) Eriksson, Victoria; Ek, Kristina; Lazarczyk Carlson, Ewa; Söderholm, Patrik; Department of Business and Economics
The purpose of the paper is to investigate whether lending crowdfunding offers equal opportunities for female and male entrepreneurs in the context of a technology demonstration project. The analysis recognises the potential role of signalling, i.e. suggesting that any success in attracting funding may not only be contingent on gender per se, but also on how signals from women and men concerning the various project attributes will be interpreted. We design a split-sample discrete choice experiment that, apart from financial factors, consider two quality signals: private co-funding and downstream actor collaboration. Data are collected through a survey of over 2000 Swedish individuals and analysed using mixed logit models. Overall, the results indicate little gender bias against women in the case of a (characteristically) male-dominated technological development project. We find very limited evidence suggesting that any of the two project quality signals are perceived differently based on the project owner’s gender. Still, the results show that male lenders place a larger premium on actor collaboration when the project owner is female. On the whole, our findings indicate that lending crowdfunding could potentially help democratize the access to capital and reduce any existing funding obstacles for women.
Defining collectivism as duties in relationships rather than generalized warmth improves predictive validity in 100 cultures
(2026-09-15) The Responsibilism Collaboration Team; Department of Business and Economics
Experts widely agree that East Asian cultures are more collectivistic than Western cultures. Yet 76 survey studies found no reliable difference. Our explanation is that the most-used surveys are contaminated with ‘warm fuzzy’ items, which describe warmth and prosociality towards everyone near and far. We argue that collectivism is better understood as ‘responsibilism’, which emphasizes responsibilities in close relationships. A survey found that many cultural psychologists expect collectivistic cultures to agree with warm fuzzy items. Then we developed a responsibilism scale and tested 12,247 people in 100 cultures. Results showed large cultural differences (d = 3.98 [2.50, 5.43]) that were previously obscured by flawed measurement. Responsibilism predicted differences across societies (such as family ties and corruption) and individuals (social class and politics) that most earlier scales missed. This study offers a solution to previous puzzling findings by distinguishing warm fuzzy prosociality from responsibilism and introducing a scale with stronger predictive validity.
Cortical Network Overconnectivity Relates to Sensory, Cognitive, and Social Dimensions in Young Children With Autism Spectrum Disorder
(2026-09) Rodríguez-Herreros, Borja; Mheich, Ahmad; Osório, Joana Maria Almeida; Richetin, Sonia; Junod, Vincent; Arnold, Lorène; Aeschbach, Victoria; Adamou, Lilia Nora; Gschwend, Katherina; Mendes, Laura; Romascano, David; Yu, Paola; Gygax, Marine Jequier; Maillard, Anne M.; Murray, Micah M.; Hassan, Mahmoud; Chabane, Nadia
The heterogeneity in both the neurobiological mechanisms and the phenotypic presentations of autism spectrum disorder (ASD) poses a major challenge to clinical and translational research. Alterations in functional connectivity (FC) have been associated with ASD, yet it remains unclear whether and how divergent brain network properties may account for individual differences across ASD-related symptomatology and behaviors. We applied source-level reconstruction to rest-like non-task-related high-density EEG data in a cohort of 104 young children (38 with ASD) to identify global and local alterations of cortical network connectivity. We subsequently used regularized canonical correlation analysis (rCCA) to characterize specific FC patterns linked to variation in cognitive, social and sensory dimensions derived from standard clinical instruments. We found increased low-frequency FC in frontotemporal cross-hemispheric networks and lateral-occipital regions of young ASD children versus healthy peers. RCCA revealed three distinct FC patterns in recurrent ASD-related networks, each contributing to predict individual differences in cognitive, social and sensory features. These linked FC-behavior dimensions may shed light on atypical brain network topology associated with specific phenotypic manifestations of ASD, which might implicate unique underlying neurobiological mechanisms.
Consumer search patterns in online retailing
(2026-06-01) Dreneva, Anna; Orquin, Jacob L.; Department of Business and Economics
Online retailers routinely rely on sorting, filters, and personalized assortments, but the effectiveness of these design choices depends on how consumers search and decide in digital environments. This research develops a unified and behaviorally grounded framework that conceptualizes consumer search as a context-dependent screening process. Using a novel approach that integrates eye-tracking, mouse-scroll behavior, subjective product preferences, and choice data, we analyze sequences of search events across two studies. The findings show that consumers do not follow a single search strategy but instead adapt their search flexibly to product value and context. Search unfolds as an adaptive screening process in which consumers scan many low-value products, examine a smaller set of higher-value options, and compare only a few top candidates before choosing. Moreover, consumers search less when the overall value of seen products is higher. This framework integrates previously disconnected research streams on decision processes, stopping rules, and consideration set formation, and offers actionable guidance for design and decision-support tools in online retailing.
Broad-Range Inverse Design of Antennas Using Deep Neural Networks with Material Parameter Embedding and Rapid Fine-Tuning
(2026-09-01) Koziel, Slawomir; Pietrenko-Dabrowska, Anna; Department of Engineering
Contemporary antenna design increasingly depends on numerical optimization to manage multiple design variables, objectives, and constraints. In many applications, such as metasurfaces, antenna arrays, or redesign over wide operating ranges, global optimization is no longer optional but essential. However, conventional bioinspired algorithms are notoriously inefficient, while machine learning-based methods struggle to construct sufficiently accurate metamodels. This communication presents a new broad-range optimization strategy based on inverse data-driven surrogates implemented with deep convolutional neural networks (CNNs). A custom embedding process is applied to allow constructing models covering diverse material properties (e.g., substrate permittivity). The surrogate maps antenna frequency responses to geometry parameters. Training data are generated in a dimensionally reduced space that captures the most influential directions in the design space, enabling accurate model construction with a limited number of samples. In the first phase, the inverse model produces an initial design directly from the performance specifications (e.g., desired operating bandwidth) mapped into a target frequency response, which is subsequently refined using an accelerated gradient-based optimizer. The presented approach is verified with two planar antennas optimized for the best reflection response and gain. The outcomes indicate that the inverse surrogate can be applied across a broad spectrum of frequencies, geometries, and substrate materials at a low cost of fewer than 30 electromagnetic (EM) analyses on average. Benchmarking demonstrates the benefits of our method and its practical edge over a broad spectrum of bioinspired, machine learning (ML), and surrogate-assisted population-based approaches.
Associations between strength and power and throwing performance in high-level male and female javelin throwers
(2026-05-18) Köykkä, Miika; Ihalainen, Johanna K.; Saari, Anssi; Waller, Benjamin; Cronin, Neil J.
Javelin throwing presents some unique characteristics compared to other overarm throwing sports and imposes considerable physical demands on the throwers. As these demands have not been adequately quantified in the literature, this study investigated the associations between strength, power, and javelin throwing performance. Thirteen male (age: 24 ± 3 years, personal best: 77.0 ± 5.7 m, javelin throwing as the primary sport: 8 ± 3 years) and fourteen female (24 ± 5 years, 55.6 ± 5.7 m, 8 ± 4 years) javelin throwers were tested for javelin throwing performance, unilateral and bilateral vertical jumps, standing overhead medicine ball throw, and isometric maximal strength in various upper and lower body exercises. In males, throwing performance was associated moderately to strongly with bilateral and unilateral drop and depth jump measures, and with forearm supination strength of the throwing arm (R = .56 to .72, p < .05). In females, throwing performance was associated moderately to strongly with ankle plantarflexion and hip abduction strength in each leg, as well as with grip and shoulder internal rotation strength of the throwing arm (R = .55 to .73, p < .05). The results show clear sex-specific differences in important strength and power measures for javelin throwing. In males, better throwers possess greater lower body power whereas better female throwers display greater upper and lower body maximal strength.
A Randomized Controlled Trial of an iCBT Program to Reduce Infertility-Related Stress : A Novel Digital Approach
(2026-06-23) Haflidadottir, Sigurbirna; Helgadottir, Fjola Dogg; Sighvatsson, Magnus Blondahl
The global burden of infertility and its associated psychological distress is substantial, yet many individuals face barriers to accessing psychological support. A scalable, low-cost online intervention presents a promising solution. A randomized controlled trial (RCT) was conducted (experimental vs. waitlist control group) to evaluate the efficacy of a fully automated cognitive behavioral therapy (CBT) program in reducing infertility-related stress among women attempting to conceive. Data were analyzed using linear mixed models on an intention-to-treat basis. The study enrolled 63 Icelandic women experiencing infertility-related stress. The intervention, Overcome Fertility Stress, is an online program designed to identify and modify maladaptive behaviors and thought patterns contributing to infertility-related distress. It delivers personalized treatment through prewritten text, audio, and video responses based on user input. Infertility-related stress was measured using the 46-item Fertility Problem Inventory (FPI), which assesses a global stress score and five domains of infertility-related stress. Participants in the experimental group demonstrated a significant reduction in infertility-related stress (mean reduction score = 22.85, p < 0.05), whereas the control group showed no significant changes (p > 0.05). The within-group effect was large, but controlled between-group effects and confidence intervals should be prioritized when interpreting treatment benefit. These findings provide preliminary controlled evidence that a fully automated CBT program may reduce infertility-related stress, although the underpowered sample and low adherence require cautious interpretation. This RCT, alongside prior feasibility studies, suggests that Overcome Fertility Stress may be a promising low-intensity intervention for infertility-related stress. No significant effects were observed on secondary outcomes measuring depression, anxiety, stress, or worry, suggesting the program is specific to infertility-related stress rather than a broader intervention for emotional disorder symptoms. Larger, adequately powered trials with stronger engagement strategies are needed. Future research should focus on expanding access and integrating digital psychological support into standard fertility care.
An Icelandic pangenome reference
(2026-08-19) Holley, Guillaume; Eggertsson, Hannes P.; Kristmundsdottir, Snaedis; Beyter, Doruk; Skuladottir, Astros Th; Moore, Kristjan H.S.; Olason, Pall I.; Gylfason, Arnaldur; Magnusson, Olafur T.; Oddsson, Asmundur; Stefansson, Hreinn; Helgason, Agnar; Masson, Gisli; Sulem, Patrick; Gudbjartsson, Daniel F.; Stefansson, Kari; Halldorsson, Bjarni V.; Department of Engineering
Reference bias is an issue that affects most genomic studies analysing short reads mapped to a reference genome1,2. It can be mitigated by mapping to multiple haplotypes represented in a pangenome3, 4–5. Here we introduce two new methods to address reference bias: Emblask for pangenome construction and Weaver for mapping to pangenomes at scale. Emblask is a hybrid long- and short-read haplotype-resolved dual assembly pipeline for parent–offspring trio data. Using Emblask, we assembled 698 Icelandic haplotypes and added them to the Human Pangenome Reference Consortium (HPRC) pangenome4 to construct an Icelandic pangenome reference (HPRC-ICE) including 51.41 million small variants. We mapped the short reads of 57,630 Icelanders to HPRC-ICE with Weaver and called 98.96 million variants, representing a 6.17% increase over mapping to a linear reference. We uncovered new variants in low-mappability regions, including a pathogenic single nucleotide polymorphism (SNP) in GBA1 that associates with early onset Parkinson’s disease and a missense SNP in CBS that is pathogenic for homocystinuria. We replicated the GBA1 association in the UK Biobank6 with a targeted remapping of 429,193 British and Irish participants.
Robust statistical feature extraction and deep learning surrogates for highly accurate guided wave-based determination of global corrosion parameters in metallic plates
(2027-01) Koziel, Slawomir; Zima, Beata; Pietrenko-Dabrowska, Anna; Department of Engineering
Accurate evaluation of corrosion in metal plates is essential across many industries, particularly in shipbuilding. Among available techniques, non-destructive testing (NDT) is especially valuable, as it enables the assessment of material degradation without compromising structural integrity. Within this domain, guided ultrasonic waves have attracted increasing attention in recent years due to their potential for estimating statistical corrosion parameters, such as mean thickness loss and its standard deviation. However, practical application of these methods remains challenging because of the complex interaction between wave propagation and spatially varying thickness profiles. This study presents an advanced methodology for precise estimation of corrosion parameters using deep learning-based inverse models. The corrosion morphology is modeled as a stochastic field, while guided wave responses, obtained through numerical simulations and recorded by a sensor network, are postprocessed to extract statistical descriptors. In particular, the mean vector and covariance matrix of sensor signals, including wave packet location, width, and amplitude, are computed. To ensure robustness, feature scaling and elliptic envelope-based covariance estimation are employed, effectively mitigating the influence of outliers, which are especially pronounced at higher levels of thickness loss and surface roughness. The inverse model establishes a mapping between extracted feature vectors and the corresponding corrosion parameters, enabling their direct prediction. This framework addresses the intrinsic non-uniqueness of the problem, where identical statistical descriptors may arise from infinitely many realizations of the stochastic corrosion field. Extensive numerical investigations demonstrate that reliable feature extraction, particularly robust handling of outliers, combined with the ability of deep neural networks to capture nonlinear relationships, is critical for accurate parameter estimation. Prediction performance is further improved by aggregating responses from multiple independent realizations of the stochastic corrosion field. Overall, the results demonstrate that the proposed framework, combining robust statistical feature extraction, covariance-based feature representation and nonlinear inverse modeling, provides improved prediction accuracy compared with the previously developed feature-based linear regression model and the conventional time-of-flight approach under the considered numerical simulation conditions.
Design-ready farfield pattern modeling of broadband antennas using multi-branch attention-based recurrent convolutional decoder neural network
(2027-01) Koziel, Slawomir; Sahu, Kaustab C.; Pietrenko-Dabrowska, Anna; Department of Engineering
Behavioral modeling is becoming increasingly important in antenna development, particularly in accelerating simulation-based design processes such as optimization. However, the construction of accurate data-driven models remains a challenging task, typically limited to electrical responses such as reflection coefficients or integral performance metrics (e.g., gain). In contrast, the behavioral modeling of radiation patterns—although crucial from a design perspective—has been relatively scarce in the existing literature. This research introduces a new surrogate modeling framework for broadband antennas that leverages customized deep learning techniques. The proposed approach enables accurate and comprehensive modeling of both electrical responses (e.g., reflection coefficient, gain) and far-field radiation patterns across a broad range of frequencies, elevation/azimuth angles, and geometric parameters. As a result, the developed surrogates are highly versatile and applicable for addressing a wide spectrum of design and optimization tasks. For accurate representation of radiation patterns, a hybrid deep learning architecture is developed that integrates multi-branch recurrent processing with cross-attention–based feature fusion. Six independent RNN branches (each combining FC, LSTM, and GRU layers) extract complementary representations of the geometric input space, which are subsequently fused through a cascade of attention modules. The fused embedding is decoded using a transposed-convolutional refinement network to reconstruct the full radiation pattern. Extensive validation studies, performed using two ultra-wideband antenna designs, confirm the excellent predictive accuracy and robustness of the proposed surrogates. The results also demonstrate their advantages over multiple benchmark methods, including conventional regression models and various deep learning schemes. Furthermore, the practical utility of the framework is demonstrated through real-world design tasks, including impedance matching enhancement, size reduction, omnidirectional radiation, and gain stability. Overall, the presented methodology represents a significant advancement toward comprehensive multi-response antenna modeling and substantially extends the capabilities of behavioral metamodeling in contemporary antenna design practice.
Optimizing SiGe–SiO2 Visible–Short‐Wave Infrared Photoresponse by Modulating Interplay Between Strain and Defects Through Annealing
(2024-08) Sultan, Muhammad Taha; Stavarache, Ionel; Manolescu, Andrei; Arnalds, Unnar Bjarni; Teodorescu, Valentin Serban; Svavarsson, Halldor Gudfinnur; Ingvarsson, Snorri; Ciurea, Magdalena Lidia; Verkfræðideild

Skoða

Nýlegt