Abstract
Background/Objectives: Obesity is a major contributor to cardiovascular disease, yet traditional risk assessment methods may overlook behavioral and circadian influences that modulate metabolic health. Chronotype, physical activity, sleep quality, eating speed, and breakfast habits have been increasingly associated with cardiometabolic outcomes. This study aims to evaluate the associations between these behavioral factors and both anthropometric and biochemical markers of cardiovascular risk among obese candidates for bariatric surgery. Methods: A cross-sectional study was conducted in a sample of 286 obese adults (78.3% females, mean 44.3 years, SD = 10.8, mean BMI = 42.5 kg/m2, SD = 6.2) followed at a central Portuguese hospital. Chronotype (reduced Morningness-Eveningness Questionnaire), sleep quality (Pittsburgh Sleep Quality Index), physical activity (Godin-Shephard Questionnaire), eating speed, and breakfast skipping were assessed. Cardiovascular risk markers included waist-to-hip ratio (WHR), waist-to-height ratio, A Body Shape Index (ABSI), Body Roundness Index, atherogenic index of plasma (AIP), triglyceride-glucose index (TyG), and homeostatic model assessment for insulin resistance (HOMA-IR). Results: Men exhibited significantly higher WHR, ABSI, HOMA-IR, TyG, and AIP. Eveningness was associated with higher insulin (r = -0.168, p = 0.006) and HOMA-IR (r = -0.156, p = 0.011). Poor sleep quality was associated with higher body fat mass (r = 0.151, p = 0.013), total cholesterol (r = 0.169, p = 0.005) and LDL cholesterol (r = 0.132, p = 0.030). Faster eating speed was associated with a higher waist circumference (r = 0.123, p = 0.038) and skeletal muscle mass (r = 0.160, p = 0.009). Conclusions: Male sex, evening chronotype, and poor sleep quality were associated with more adverse cardiometabolic profiles in individuals with severe obesity. These findings support the integration of behavioral and circadian factors into cardiovascular risk assessment strategies.

Abstract

Abstract

Abstract
Magneto-optic surface plasmon resonances (MOSPRs) rely on the interaction of magnetic fields with surface plasmon polaritons (SPP) to modulate plasmonic bands with magnetic fields and enhance magneto-optical activity. In the present work, a study on the magnetoplasmonic behavior of Ag/Fe bilayers is carried out by VIS-NIR spectroscopy and backed with SQUID measurements, determining the thickness-dependent magnetization of thin-film samples. The MOSPR sensing properties of Ag/Fe planar bilayers are simulated using Berreman's matrix formalism, from which an optimized structure composed of 15 nm of Ag and 12.5 nm of Fe is obtained. The selected structure is fabricated and characterized for refractive index (RI) sensitivity, reaching 4946 RIU-1 and returning an effective enhancement of refractometric sensitivity after magneto-optical modulation. A new optimized and cobalt-free magnetoplasmonic Ag/Fe bilayer structure is studied, fabricated, and characterized for the first time towards refractometric sensing, to the best of our knowledge. This configuration exhibits potential for enhancing refractometric sensitivity via magneto-optical modulation, thus paving the way towards a simpler, more accessible, and safe type of RI sensor with potential applications in chemical sensors and biosensors.

Abstract
This paper presents a solution for visualizing oil spills at sea by combining satellite data with virtual choreographies. The system enables dynamic, interactive visualization of oil slicks, reflecting their shape, movement, and interaction with environmental factors like currents and wind. High resolution geospatial data supports a multiplatform experience with aerial and underwater perspectives. This approach promotes independence, interoperability, and multiplatform compatibility in environmental disaster monitoring. The results validate virtual choreographies as effective tools for immersive exploration and analysis, offering structured data narratives beyond passive visualization especially valuable for mixed reality applications.

Abstract
Deep Learning (DL) has emerged as a powerful tool for vulnerability detection, often outperforming traditional solutions. However, developing effective DL models requires large amounts of real-world data, which can be difficult to obtain in sufficient quantities. To address this challenge, DiverseVul dataset has been curated as one of the largest datasets of vulnerable and non-vulnerable C/C++ functions extracted exclusively from real-world projects. Its goal is to provide high-quality, large-scale samples for training DL models. Nevertheless, during our study several inconsistencies were identified in the raw dataset while applying pre-processing techniques, highlighting the need for a refined version. In this work, we present a refined version of DiverseVul dataset, which is used to fine-tune a large language model, LLaMA 3.2, for vulnerability detection. Experimental results show that the use of pre-processing techniques led to an improvement in performance, with the model achieving an F1-Score of 66%, a competitive result when compared to our baseline, which achieved a 47% F1-Score in software vulnerability detection.