Search Results (178)

The deployment of Vision-Language Models (VLMs) in Satellite IoT scenarios is critical for real-time disaster assessment but is often hindered by the substantial memory and compute requirements of state-of-the-art models. While parameter-efficient fine-tuning (PEFT) enables adaptation, with minimal computational overhead, standard supervised methods often fail to ensure robustness and reliability on resource-constrained edge devices. To address this, we propose EdgeV-SE, a self-reflective fine-tuning framework that significantly enhances the performance of VLM without introducing any inference-time overhead. Our framework incorporates an uncertainty-aware self-reflection mechanism with asymmetric dual pathways: a generative linguistic pathway and an auxiliary discriminative visual pathway. By estimating uncertainty from the linguistic pathway using a log-likelihood margin between class verbalizers, EdgeV-SE identifies ambiguous samples and refines its decision boundaries via consistency regularization and cross-pathway mutual learning. Experimental results on hurricane damage assessment demonstrate that our approach improves image classification accuracy, enhances image–text semantic alignment, and achieves superior caption quality. Notably, our work achieves these gains while maintaining practical deployment on a commercial off-the-shelf edge device such as NVIDIA Jetson Orin Nano, preserving the inference latency and memory footprint. Overall, our work contributes a unified self-reflective fine-tuning framework that improves robustness, calibration, and deployability of VLMs on edge devices. Full article

Background/Objectives: This study aims to explore both the psychosocial outcomes of hearing aid use and the factors that differentiate users from non-users among older adults with sensorineural hearing loss (SNHL) in Romania. Methods: We conducted a cross-sectional, comparative study with follow-up, including 201 patients aged between 49 and 92 years (mean age 70.76 ± 11.86 years), diagnosed with moderate to severe SNHL, evaluated between 1 November 2023, and 30 November 2024, at the Municipal Clinical Hospital Orăștie, Romania. Audiological assessment involved pure-tone audiometry and speech testing. Outcome measures included the Word Recognition Score (WRS), International Outcome Inventory for Hearing Aids (IOI-HA), Hearing Handicap Inventory for Adults (HHIA), Tinnitus Handicap Inventory (THI), and the Self-Esteem Scale (SES). Results: Of the 201 patients, 105 (52.2%) accepted hearing aid (HA) fitting and 96 (47.8%) declined. No significant differences were found in age (p = 0.565) or sex (p = 0.476) between groups. HA users reported significantly lower perceived handicap (HHIA: 46.48 ± 24.83 vs. 77.74 ± 28.02, p = 0.015) and higher self-esteem scores (SES: 35.68 ± 4.88 vs. 23.03 ± 4.90, p < 0.001), while tinnitus-related distress (THI) did not differ significantly (p = 0.785). Word recognition scores improved significantly post-fitting across all degrees of hearing loss: moderate (48.52% → 86.13%), moderately severe (47.47% → 85.31%), and severe (47.55% → 85.46%), all p < 0.001. Conclusions: Hearing aid use in older adults with SNHL was associated with significant improvements in speech perception and reduced perceived hearing handicap. These benefits were consistent across all severity levels and were independent of unilateral or bilateral device use. The difference in self-esteem observed between users and non-users may reflect pre-existing psychological factors influencing HA adoption, underlining the importance of personalized counseling in hearing rehabilitation. Full article

COVID-19 led to substantial changes in early literacy instruction. Although emerging evidence documents its effects on children’s reading achievement, much less is known about how these changes relate to young children’s reading motivation and school-related well-being. This study compared two cohorts of second graders (N = 287) from the same four low-SES schools, all assessed at the beginning of second grade. A pre-COVID-19 cohort, whose first-grade instruction was delivered entirely face-to-face, was compared with a during-COVID-19 cohort whose first-grade reading instruction took place amid extended distance learning with intermittent, restricted in-person schooling. Cohorts were compared on reading motivation, school-related well-being (covitality), and reading achievement (word reading, vocabulary, and reading comprehension). Multivariate analyses of variance (MANOVAs) were conducted to compare cohorts (pre- vs. during COVID-19) and reader groups (typical vs. poor readers), with gender, class, and school entered as control variables. In motivation, self-concept was higher during COVID-19, and typical readers reported higher motivation than their peers. In well-being, covitality was higher during COVID-19 at the total score and across gratitude, optimism, zest, and persistence; a cohort by group interaction for persistence indicated higher scores for typical readers during COVID-19. In achievement, phonological decoding and orthographic word identification were lower during COVID-19; typical readers scored higher than poor readers on all achievement outcomes. Together, these findings suggest that the educational setting shapes motivation and well-being alongside achievement, and that distance learning is not uniformly detrimental, as it coincided with higher covitality and reading self-concept at school reentry. Full article

Background/Objectives: Scaling is central to non-surgical periodontal therapy; however, it is often perceived as uncomfortable, particularly in periodontitis. Psychological distress may amplify pain and reduce adherence. This comparative pilot study assessed whether ultrasonic technology influences patient-reported pain and psychoemotional status while quantifying the contribution of the periodontal condition and baseline psychological factors. Methods: A monocentric split-mouth design enrolled 42 adults (21 with stage I–III, grade B periodontitis; 21 periodontally healthy). Maxillary scaling was performed with Device nr.1 and mandibular scaling was performed with Device nr.2, and no anesthesia was used. Pain was measured immediately post-procedure using the Short-Form McGill Pain Questionnaire (SF-MPQ; sensory and affective subscales). Psychological status was assessed pre- and post-session with the Kessler Psychological Distress Scale (K10) and the Rosenberg Self-Esteem Scale (RSES). Construct validity was examined via exploratory factor analysis. The Mann–Whitney U, Wilcoxon signed-rank, Spearman’s ρ, and Cliff’s δ were applied (α = 0.05). Results: The overall pain was low. Between devices, sensory pain did not differ, whereas affective pain was modestly lower with the “No Pain” device (p = 0.017). Periodontitis was the dominant determinant of pain: higher sensory (U = 509.00, p = 0.0004; δ = 0.42) and affective scores (U = 290.00, p < 0.0001; δ = 0.67) occurred irrespective of device, while device-related effects were negligible (sensory δ = −0.03) to small (affective δ = 0.27). Somatic distress correlated with affective pain (ρ = 0.25, p = 0.023) and was borderline for sensory pain (ρ = 0.21, p = 0.060); emotional distress showed no significant associations. During the session, K10 scores decreased and RSES values increased, indicating immediate psychoemotional benefits. Conclusions: Pain perception during scaling is shaped primarily by periodontal status and psychological distress rather than by ultrasonic technology per se. Although the electronic module to dynamically adjust the power of the instrument technology may attenuate the affective component, standardized atraumatic techniques and routine psychosocial screening are likely to yield greater gains. The observed short-term improvements in distress and self-esteem support integrating patient-reported outcomes into individualized, patient-centered periodontal care. Full article

This study examined the impact of a shift schedule change on firefighter sleep and health outcomes (n = 24). Firefighters from a U.S. department transitioned from a 24 h on, 48 h off (24/48) schedule to a 48 h on, 96 h off (48/96) schedule. Wrist actigraphy and self-reported health outcomes were assessed at three time points: baseline (24/48), 3 months post-transition, and 6 months post-transition. Objective sleep measures included total sleep time (TST), sleep efficiency (SE), sleep onset latency (SOL), and wake after sleep onset (WASO). Self-reported health outcomes included the Insomnia Severity Index (ISI), Beck Depression Inventory–II (BDI-II), Beck Anxiety Inventory (BAI), Multidimensional Assessment of Fatigue (MAF), and the Alcohol Use Disorders Identification Test (AUDIT). Linear mixed-effects models (LMMs) with random intercepts were used to evaluate changes over time, adjusting for age, years of service, and individual night-time call volume. Results showed significant improvements in TST, SE, SOL, and WASO at the 3-month follow-up, which were sustained but did not further increase at 6 months. ISI and BDI-II scores also improved, while BAI, MAF, and AUDIT remained stable. These findings suggest that the 48/96 schedule may provide short-term improvements in sleep and psychological health for firefighters in low call-volume settings. Additional research is needed in higher-volume departments and over longer timeframes. Full article

Inactivity and increases in psychological and behavioral symptoms are common for people with dementia, and current assessment relies on proxy-rated tools. We investigate the feasibility and adherence of the use of sensor technology by exploring the relationship between daytime activity and sleep quality. For a total of 42 day–night data pairs in nursing home residents with dementia (N = 11), Garmin Vivoactive5 and Somnofy monitored continuous physical activity levels, sleep efficiency (SE), sleep score, sleep regularity index (SRI), and wake after sleep onset (WASO). Using the Spearman coefficient, we explored correlations between digital and proxy-rated tools (Personal Self Maintenance Scale (PSMS) and Neuropsychiatric Inventory–Nursing Home version (NPI-NH)) and the relationships between the digital biomarkers (SE, SRI, WASO, sleep score, physical activity). Participants (mean age 84 years) had moderate to severe degrees of dementia. Daytime activity levels correlated to sleep quality parameters WASO (−0.34, p = 0.03), and SRI (0.43, p = 0.01), and traditional sleep measures were associated with digital biomarkers (WASO/NPI-NH-K, p = 0.03). We found a relationship between daytime activity and sleep quality; however, the bidirectional relationship remains ambiguous and should be further investigated. The use of sensing technologies for people with dementia residing in a nursing home is feasible, although not without limitations, and has the potential to identify subtle changes, improving clinical assessment and the corresponding care recommendations. Full article

Research has shown that language ability can vary enormously depending on variables such as musical ability, musical training, and second and/or foreign language experience. In this study, we simulated initial foreign language learning conditions in which learners must recognize and match unfamiliar language input. We recruited 500 participants with different levels of foreign language experience, different levels of musical training and different socio-economic backgrounds. Their auditory phonological pattern recognition ability, short-term memory (STM) capacity, musical ability, musical self-estimation, educational status, and socio-economic status (SES) were assessed. Both overall and group-specific analyses were conducted to investigate the impact of these variables. For the group-specific analysis, participants were assigned to four groups based on the presence or absence of musical training and extensive foreign language experience. For the overall analysis, regression models were applied to the entire sample to examine the combined effects of all variables. Group-specific analyses revealed that both musical training and extensive foreign language experience contributed to individual differences in the ability to recognize phonological patterns in unintelligible auditory stimuli. A key finding was that musical training appeared to have a stronger influence on auditory phonological pattern recognition than extensive foreign language experience, particularly in the early stages of language learning. This suggests that musical training may exert a greater impact on initial phonetic acquisition processes than extensive foreign language proficiency, especially when the language stimuli are relatively poor in linguistic content. The overall analysis revealed that musical variables, short-term memory capacity, socioeconomic status, and educational status all contributed to individual differences in auditory phonological pattern recognition. Notably, the most significant finding of the overall analysis was the association between SES and auditory phonological pattern recognition in unfamiliar speech—a result that challenges the notion of aptitude measures as stable and environment-independent and highlights the potential influence of environmental factors on this capacity. Full article

Background: Higher socioeconomic status (SES) is generally associated with lower engagement in health-risk behaviors, in part due to increased access to health information, preventive resources, and supportive environments. However, emerging evidence suggests that this protective pattern may not extend uniformly to all forms of substance use, including adolescent e-cigarette use, and may vary by gender. For instance, some studies have found higher rates of e-cigarette use among adolescents from higher SES backgrounds. Aim: This study examined whether the associations between family SES and tobacco use differ between girls and boys. We also explored whether these associations vary by age group. Methods: A cross-sectional survey was conducted among students (age 14–20) attending public high schools in Baltimore City. Family SES was assessed using three indicators: parental education, parental employment, and household income. Tobacco use was measured using self-reported past use of e-cigarettes and conventional cigarettes. Demographic covariates included age, sex, race/ethnicity, and household composition. Separate logistic regression models were estimated for each tobacco use outcome, adjusting for covariates. To examine subgroup differences, analyses were stratified by gender and age. Results: Higher parental education was associated with lower odds of e-cigarette use, but no SES indicators were significantly associated with conventional cigarette use. Subgroup analyses showed that the protective association of parental education against e-cigarette use was evident among girls but not boys and among older but not younger adolescents. Conclusions: These findings differ from previous studies that reported a positive association between SES and adolescent e-cigarette use. In this predominantly low-income, urban sample, higher parental education appeared to be protective for girls but not for boys. These results suggest that SES influences on tobacco use may be context- and subgroup-specific. Further research is needed to better understand how sociodemographic and contextual factors shape adolescent tobacco use behaviors. Full article

Virtual reality (VR) is often regarded as the “ultimate empathy machine” because of its ability to immerse users in alternative perspectives and environments beyond physical reality. In this study, 105 participants (average age 22.43 ± 5.31 years, range 19–45, 75% female) with diverse educational and professional backgrounds experienced three-dimensional 360° VR videos featuring actors expressing different emotions. Despite the availability of established methodologies in both research and clinical domains, there remains a lack of a universally accepted “gold standard” for empathy assessment. The primary objective was to explore the relationship between the empathy levels of the participants and the changes in their physiological responses. Empathy levels were self-reported using questionnaires, while physiological attributes were recorded through various sensors. The main outcomes of the study are machine learning (ML) models capable of predicting state empathy levels and trait empathy scores during VR video exposure. The Random Forest (RF) regressor achieved the best performance for trait empathy prediction, with a mean absolute percentage error (MAPE) of 9.1%, and a standard error of the mean (SEM) of 0.32% across folds. For classifying state empathy, the RF classifier achieved the highest balanced accuracy of 67%, and a standard error of the proportion (SE) of 1.90% across folds. This study contributes to empathy research by introducing an objective and efficient method for predicting empathy levels using physiological signals, demonstrating the potential of ML models to complement self-reports. Moreover, by providing a novel dataset of VR empathy-eliciting videos, the work offers valuable resources for future research and clinical applications. Additionally, predictive models were developed to detect non-empathic arousal (78% balanced accuracy ± 0.63% SE) and to distinguish empathic vs. non-empathic arousal (79% balanced accuracy ± 0.41% SE). Furthermore, statistical tests explored the influence of narrative context, as well as empathy differences toward different genders and emotions. We also make available a set of carefully designed and recorded VR videos specifically created to evoke empathy while minimizing biases and subjective perspectives. Full article

The incorporation of educational robotics (ER) into classroom learning has emerged as a significant goal in contemporary education, with instructors assuming a pivotal role. Recent research has shown the influence of teachers’ perceptions of ER and their self-efficacy on the learning process, while the primary goal in these inquiries is to the development of appropriate scales that guarantee correct measurements. Serving this goal, the present study presents the TRi-Robotics scale and its psychometric properties, which assesses teachers’ readiness to integrate ER into their classrooms. TRi-Robotics is a novel multidimensional tool that integrates self-efficacy, commitment, and affective conditions, validated through both CFA and network psychometrics. The proposed 14-item scale is three-dimensional and includes self-efficacy (SE), commitment (C), and affective conditions (AC). The validation procedure included the customary Exploratory and Confirmatory Factor Analysis, applied to a sample of 817 in-service teachers. Reliability analysis showed satisfactory internal consistency, while measurement invariance for gender was sustained. Furthermore, network psychometrics was applied via Exploratory Graph Analysis (EGA), which supported the proposed structure and its dimensionality and measurement invariance. The TRi-Robotics scale proved a valid instrument with satisfactory psychometric properties, and it is a significant asset to implement in educational and psychological research for testing further research hypotheses. Full article

Background/Objectives: This study aimed to assess the effectiveness and clinical relevance of intraprocedural invasive measurements—specifically intraprocedural mean pressure gradients (IC MPGs) and diastolic delta (DD)—in comparison with echocardiography for evaluating transcatheter heart valve (THV) performance across different prosthesis types. Particular attention was paid to comparing outcomes between balloon-expandable (BE) and self-expandable (SE) valves, with further stratification by aortic annulus size. Methods: A retrospective analysis was performed on 926 patients who underwent transcatheter aortic valve replacements (TAVRs) between 2012 and 2021. Patients were categorized into BE (n = 301) and SE (n = 625) valve groups. Intraprocedural MPG was measured immediately before and after valve deployment. Postprocedural echocardiographic MPG (EC MPG) and the degree of aortic regurgitation were assessed within five days after implantation. Aortic annuli were classified as small (≤23 mm) or large (≥24 mm). Results: After implantation, EC MPG was consistently higher than IC MPG, with only a weak correlation observed between the two modalities. SE valves were generally associated with lower EC MPG than BE valves. DD was higher in the BE group; however, no significant correlation was found between DD and echocardiographically assessed aortic regurgitation. Conclusions: Intraprocedural invasive measurements offer a reliable and immediate assessment of prosthesis function during TAVR but tend to underestimate gradients compared to echocardiography. Newer SE valves show performance comparable to BE valves, particularly in small annuli, supporting their use in challenging anatomies. DD appears to lack a diagnostic value for postprocedural aortic regurgitation. Full article

Background/Objectives: It is unclear whether enamel margins should be beveled in direct resin-based restorations. This study evaluated the influence of enamel beveling on the marginal quality of mesio-occluso-distal (mod) cavity boxes. Methods: Seventy-five caries-free human molars were divided into three groups. Mod-cavities with the entire margin in the enamel were prepared ± proximal bevel (n = 25). Twenty-five beveled mod-cavities served as control. Each group was restored with five material combinations: micro hybrid composite with etch-and-rinse (ER) or self-etch (SE) adhesive, compomer with ER or SE, and low-shrinkage composite with ER. A complex filling technique was used in the control. After artificial aging (1000 thermal cycles, 5/55 °C), the percentage of continuous margins (PCM) of the proximal boxes was analyzed by scanning electron microscopy using epoxy replicas (×300), and the marginal seal was assessed by light microscopy after dye penetration (×64). Statistical analysis was performed using Kruskal–Wallis and Mann–Whitney U tests (p < 0.05). Results: No significant differences in PCM were found between the restorative systems (p = 0.075). The composite with low shrinkage showed the highest mean PCM of all groups (78.1%). Significant differences in marginal seal were observed between the restoratives after bevel preparation (p < 0.05). Beveling significantly improved the PCM only for the hybrid composite (p < 0.05), whereby the effect on marginal seal was less pronounced. Conclusions: Clinicians should be aware that beveling mod-cavities does not necessarily improve the marginal quality of direct resin-based restorations bonded with well-established adhesives and may be more beneficial for traditional hybrid composites. Full article

Ternary chalcogenide-based sulfide materials with distorted morphologies such as BaZrS₃, CaZrS₃, and SrZrS₃, have recently gained much attention in optoelectronics and photovoltaics due to their high structural and thermal stability and compatibility with low-cost, earth-abundant synthesis routes. However, their relatively large bandgaps often limit their suitability for near-infrared (NIR) photodetectors. Here, we conducted a comprehensive investigation of SrHfSe₃, a ternary chalcogenide with an orthorhombic crystal structure and distinctive needle-like morphology, as a promising candidate for NIR photodetection. SrHfSe₃ exhibits a direct bandgap of 1.02 eV, placing it well within the NIR range. Its robust structure, high temperature stability, phase stability and natural abundance make it a compelling material for next-generation, self-powered NIR photodetectors. An in-depth analysis of the SrHfSe₃-based photodetector was performed using SCAPS-1D simulations, focusing on key performance metrics such as J–V behavior, photoresponsivity, and specific detectivity. Device optimization was achieved by thoroughly altering each layer thickness, doping concentrations, and defect densities. Additionally, the influence of interface defects, absorber bandgap, and operating temperature was assessed to enhance the photoresponse. Under optimal conditions, the device achieved a short-circuit current density (Jsc) of 45.88 mA/cm², an open-circuit voltage (Voc) of 0.7152 V, a peak photoresponsivity of 0.85 AW⁻¹, and a detectivity of 2.26 × 10¹⁴ Jones at 1100 nm. A broad spectral response spanning 700–1200 nm confirms its efficacy in the NIR region. These results position SrHfSe₃ as a strong contender for future NIR photodetectors and provide a foundation for experimental validation in advanced optoelectronic applications. Full article

Background/Objectives: Schizophrenia has been associated with increased inflammatory and metabolic disturbances. Perceived family support potentially affects inflammatory and metabolic biomarkers. The aim of this study was to determine the interrelations between family support, C-reactive protein (CRP) and Body Mass Index (BMI) in a sample of outpatients with schizophrenia. Importantly, this study sought to elucidate the effect of perceived family support on inflammatory processes among patients with schizophrenia. Methods: In this cross-sectional correlation study, 206 outpatients with schizophrenia in clinical remission completed a standardized self-report questionnaire that assessed family support (Family Support Scale—FSS). Sociodemographic, clinical and laboratory data were also recorded. Results: Among the participants, 49.5% had detectable CRP values (≥0.11 mg/dL), whereas 14.6% had positive CRP levels (>0.6 mg/dL). There was a significant difference in CRP levels among the different BMI groups (normal weight/overweight vs. obese). For obese patients, the crude odds ratios (ORs) for detectable and positive CRP values were 1.980 (95% confidence interval (CI) [1.056, 3.713]) and 27.818 (95% CI [6.300, 122.838]), respectively. Significant positive correlations were observed among CRP, BMI and illness duration, while scores on the FSS were negatively associated with these variables. The results of binary logistic regression analysis indicated that both BMI and family support were significant factors in determining the likelihood of having positive CRP levels, with each unit increase in the BMI associated with a 17% (95% CI [0.025, 0.337]) increase in the odds, and with each unit increase in family support leading to an 8.6% (95% CI [0.018, 0.15]) decrease. A moderation analysis revealed that the association between family support and the probability of having positive CRP levels depends on the BMI value, but only for obese patients did the protective effect of family support significantly decrease the magnitude of the risk of having positive CRP (b = −0.1972, SE = 0.053, OR = 0.821, p = 0.000, 95% CI [−0.3010, −0.0934]). Conclusions: The effect of perceived family support on inflammatory responses becomes evident in cases where beyond metabolic complications, inflammatory processes have already been established. Increased perceived family support seems to protect against inflammation and, notably, the association between low perceived family support and increased inflammation is even stronger. Establishing the role of family involvement during the treatment of patients with schizophrenia through inflammatory processes is a novelty of this study, emphasizing the need to incorporate family therapy into psychiatric treatment plans. However, primary interventions are considered necessary for patients with schizophrenia in order to maintain their BMI within normal limits and avoid the subsequent nosological sequelae. Full article

Background/Objectives: Anorexia nervosa (AN) is a severe disorder with limited treatment efficacy. This interim analysis aimed to assess the preliminary efficacy and safety of transcranial direct current stimulation (tDCS) in reducing core AN symptoms, stress, depression, low self-esteem, and BMI in adolescent females, to determine the rationale for continuing the study. Methods: A single-center, randomized, double-blind, placebo-controlled trial included 20 adolescent females with AN assigned to an active tDCS group (n = 10) or a sham group (n = 10). The intervention involved 30 sessions over three weeks, targeting the dorsolateral prefrontal cortex. Outcomes were assessed at baseline, post-treatment, and follow-up using the Eating Attitudes Test (EAT-26) for eating disorder symptoms, the Perceived Stress Scale (PSS-10) for stress, the Beck Depression Inventory (BDI) for depression, the Rosenberg Self-Esteem Scale (SES) for self-esteem, and body mass index (BMI) measurements. Safety and tolerability were assessed using the tDCS Side Effects Questionnaire. Results: Eating disorder symptoms significantly decreased in the active tDCS group at study end (p = 0.003) and follow-up (p = 0.02), while no significant changes were observed in the sham group. Although BMI increased more in the active group (13.78%) than in the sham group (7.31%), this difference was not statistically significant (p = 0.10). Conclusions: Adverse effects were mild and transient, with no serious safety concerns reported. Based on the results of this interim analysis, the study will proceed due to promising efficacy outcomes and good treatment tolerability. Full article