Ewha Womans University Medical Center

A large body of research addresses why some natives are tolerant of outgroup members while others oppose them. For the most part, prior studies focus on macrolevel indicators (e.g., unemployment, economic globalization, immigration influx) and microlevel characteristics (e.g., age, gender, education, income). Despite the difference in the unit of analysis, they commonly emphasize ‘objective’ factors rooted in the concept of xenophobic threat. Against this backdrop, the present study shifts the analysis to a ‘subjective’ dimension, namely, perception of (in)equality of opportunity. We contend that, all else equal, individuals who perceive their society to be more unequal, i.e., lack procedural fairness, show greater resistance toward outgroup members. Based on a large-scale probability sample ( N = 42,923) of residents of Seoul, the capital of South Korea, we empirically demonstrate this argument. Specifically, Korean natives who feel a stronger sense of perceived inequality are less willing to accept, or more hostile toward, the foreign-born population. Our analysis further shows that the magnitude of this focal link is greater in the context of wealthier communities (measured by average housing price) and in the company of higher status coresidents.

This article examines fashion memes as participatory visual culture at the intersection of fashion media and platform algorithms. Drawing on convergence culture (Jenkins), everyday tactics (de Certeau), and audience agency (Fiske), alongside scholarship on algorithmic participation (Carpentier; van Dijck et al.; Abidin; Cotter), it analyses how meme creators use remix aesthetics to parody, critique, and reconfigure fashion discourse. Based on qualitative visual analysis of 22 Instagram fashion meme accounts, triangulated with secondary texts, the study codes meme-making into three strategy categories: image–text hybrids, image juxtapositions, and reconstructed visuals. Findings address three linked questions: (RQ1) memes materialise convergence by borrowing editorial and campaign frames and reworking them through meme grammars optimised for recognisability and circulation; (RQ2) they blur boundaries between amateur and professional production, producing emergent creative expertise through curation and remix rather than institutional credentials; and (RQ3) they enable accessible critique through humour while remaining shaped by platform visibility logics – where attention operates as both reward and constraint. Meme-making is shown to involve largely unpaid visibility and hope labour, generating cultural value within algorithmically curated environments characterised by datafication, commodification, and algorithmic selection. Rather than treating memes as ephemeral, the article approaches them as situated interventions in fashion communication and suggests affinities with platform-native, vernacular forms of fashion commentary.

Monocular 3D object detection is a cost-effective solution for applications like autonomous driving and robotics, but remains fundamentally ill-posed due to inherently ambiguous depth cues. Recent DETR-based methods attempt to mitigate this through global attention and auxiliary depth prediction, yet they still struggle with inaccurate depth estimates. Moreover, these methods often overlook instance-level detection difficulty, such as occlusion, distance, and truncation, leading to suboptimal detection performance. We propose MonoDLGD, a novel Difficulty-Aware Label-Guided Denoising framework that adaptively perturbs and reconstructs ground-truth labels based on detection uncertainty. Specifically, MonoDLGD applies stronger perturbations to easier instances and weaker ones into harder cases, and then reconstructs them to effectively provide explicit geometric supervision. By jointly optimizing label reconstruction and 3D object detection, MonoDLGD encourages geometry-aware representation learning and improves robustness to varying levels of object complexity. Extensive experiments on the KITTI benchmark demonstrate that MonoDLGD achieves state-of-the-art performance across all difficulty levels.

780 Background: Nectin-4 is the primary target of enfortumab vedotin (EV), an antibody-drug conjugate approved for metastatic urothelial carcinoma (mUC). Nectin-4 frequently co-expresses with HER2 (ERBB2) in urothelial tumors, suggesting overlapping molecular phenotypes. HER2 IHC adequately stratifies EV/EVP outcomes remains unclear. This study evaluates whether HER2 IHC reliably predicts response to EV versus EVP in mUC. Methods: We retrospectively abstracted all treatment lines receiving EV or EVP (2021–2024). The primary endpoint was ORR (RECIST 1.1). Secondary endpoints were PFS/OS from treatment start using a 42-day landmark Kaplan–Meier analysis. Prespecified IHC splits were 2-group (IHC 2/3 vs 0/1) and 3-group (0/1 vs 2 vs 3). Exploratory SISH-integrated schemas were: two-tier (IHC3 or IHC2/SISH+ vs IHC2/SISH− or 0/1) and three-tier (HER2-high: IHC3 or IHC2/SISH+; intermediate: IHC2/SISH−; negative: IHC0/1). Fisher’s exact test compared ORR; log-rank compared survival. Analyses were EV, EVP, and pooled. Results: Among 64 evaluable lines with HER2 IHC and response, IHC 2/3 vs 0/1 enriched for response: EVP 83.3% (20/24) vs 40.0% (4/10), p =0.034; EV 50.0% (12/24) vs 16.7% (1/6), p =0.196 (same direction, underpowered); pooled 66.7% (32/48) vs 31.3% (5/16), p =0.019. The 3-group IHC model showed a graded increase in ORR (3 > 2 > 0/1) with global significance in pooled ( p =0.037) and EVP ( p =0.018) cohorts (EV non-significant, p =0.305). Landmark PFS mirrored ORR: pooled median 7.42 vs 2.92 months (IHC2/3 vs 0/1; p =0.129); EVP 8.31 vs 4.30 ( p =0.234); EV 4.83 vs 1.35 ( p =0.308). OS was not interpretable due to insufficient events. Adding SISH did not change the clinical signal: SISH-integrated two-/three-tier schemas reproduced the same gradients and did not outperform IHC-only; defining HER2-positivity by SISH alone likewise failed to exceed the simple IHC 2/3 vs 0/1 split. Conclusions: In real-world mUC, HER2 IHC alone—independent of SISH—reliably enriches for benefit with EV-based therapy, with the clearest separation in EVP. A pragmatic IHC-only approach using 2-group or 3-group stratification is readily implementable. Prospective validation is warranted. ORR and landmark PFS by IHC 2/3 vs 0/1. Cohort IHC 2/3 ORR (k/n) IHC 0/1 ORR (k/n) Fisher p PFS, mo (2/3 vs 0/1) Log-rank p EVP 83.3% (20/24) 40.0% (4/10) 0.034 8.31 vs 4.30 0.234 EV 50.0% (12/24) 16.7% (1/6) 0.196 4.83 vs 1.35 0.308 EV+EVP pooled 66.7% (32/48) 31.3% (5/16) 0.019 7.42 vs 2.92 0.129

Severe odontogenic deep neck infections (DNIs) can rapidly extend along cervical fascial planes and progress to descending necrotizing mediastinitis (DNM) with high mortality, requiring early diagnosis and multidisciplinary management, especially in elderly patients with comorbidities. We report two elderly patients with rapidly progressive odontogenic infections: one with DNM fulfilling Estrera's criteria (Endo type IIA) and another with extensive DNI without mediastinal spread. Clinical features, serial contrast-enhanced computed tomography (CECT) findings, laboratory risk indicator for necrotizing fasciitis (LRINEC) score trends, microbiological results, surgical approaches were analyzed. Both patients presented with high LRINEC scores (≥8) and polymicrobial infections, involving multidrug-resistant organisms. Case 1 required five staged cervicothoracic debridements, negative pressure wound therapy, and delayed skin grafting. Case 2 underwent three surgeries with gland excision, hemorrhage control. Airway patency was maintained without tracheostomy through coordinated airway monitoring. Targeted antimicrobial adjustments and systemic optimization facilitated infection control. Both patients recovered fully without recurrence. These cases emphasize the importance of early CECT evaluation, cautious interpretation of LRINEC scores in patients with metabolic comorbidities, individualized surgical debridement tailored to anatomical spread, and multidisciplinary collaboration. Personalized reconstructive strategies based on defect size and tissue viability are essential for functional restoration and successful outcomes.

<title>Abstract</title> Interoceptive brain responses triggered by intestinal microbiota influence mood-related behaviors such as anxiety and depression. Although genetic mutations can alter the gut microbiome, how alterations in the gut microbiome caused by host genetic variations affect behavioral outcomes is not fully understood. We investigated how Sirtuin 3 ( <italic>Sirt3</italic> ) deficiency affects gut–brain interactions and mood-related behaviors. Gut microbiota composition and behavior were compared between <italic>Sirt3</italic> -knockout (KO) and wild-type (WT) mice. Cohousing and heterozygous littermate controls distinguished microbiome-driven from genetic effects. Transcriptomic profiling assessed region-specific gene expression in the limbic system. Metabolites in the nucleus of the solitary tract (NTS), reflecting gut–brain vagal signaling, were analyzed. The role of the vagus nerve was further tested via vagotomy, with comparative analyses of choline levels in mice and mood disorder patients. <italic>Sirt3</italic> -KO mice appeared mood-related neurobehavioral changes and alterations in synaptic plasticity-related genes in the amygdala and BNST, dependent on gut microbiome composition. Elevated plasma choline levels in both mood disorder patients and <italic>Sirt3</italic> -KO mice, together with reduced neurotransmitter-related metabolites (e.g., γ-aminobutyric acid [GABA] in the NTS), suggested that externalizing behaviors in <italic>Sirt3</italic> -KO mice are mediated by vagus nerve-dependent gut–brain axis signaling. Consistent with this, vagotomy abolished these changes, including GABA in the NTS, as well as alterations in synaptic plasticity in the amygdala and BNST. Our findings suggest the novel finding that an altered gut microbiome caused by a host genetic change, namely a <italic>Sirt3</italic> deficiency, is sensed in the NTS of the brain via the vagus nerve, leading to externalizing behaviors.

Non-von Neumann architectures overcome the memory-compute separation of von Neumann systems by distributing computation and memory locally, thereby reducing data-transfer bottlenecks and power consumption. These features are particularly advantageous for reinforcement learning (RL) workloads that rely on frequent value-function updates across large state-action spaces. When combined with event-driven spiking neural networks (SNNs), non-von Neumann architectures can further improve overall computational efficiency by leveraging the sparse nature of spike-based processing. In this study, we propose a hardware-feasible SNN-based non-von Neumann architecture that performs Q-learning, one of the most widely known reinforcement learning algorithms. The proposed architecture maps states and actions to individual neurons using one-hot encoding and locally stores each state-action pair's Q-value in the corresponding synapse. To enable each synapse to update its local Q-value based on the next state maximum Q stored in other synapses, a neuron group connected through a lateral inhibition structure is employed to produce the maximum Q, which is then globally transmitted to all synapses. A delay circuit is also added to align the next-state and current-state values to ensure temporally consistent updates. Each synapse locally generates a learning selection signal and combines it with the globally transmitted signals to update only the target synapse. The proposed architecture was validated through simulations on the Cart-pole benchmark, showing stable learning performance under low-bit precision and achieving comparable accuracy to software-based Q-learning with sufficient bit precision.

Abstract Measurement is important for the scientific programmes of addictive behaviours. In the present study, we investigated the measurement of gaming disorder, which recently became the first formal mental disorder owing to technology use (International Classification of Diseases 11). In particular, we tested the gambling hypothesis: whether survey studies on gaming disorder are contaminated by a linguistic conflation of gambling and gaming behaviours. The 500 most cited studies were analysed. Both confirmatory and exploratory analyses supported the gambling hypothesis. Of all 13 studies for which authors shared their survey materials, only one had excluded gambling (95% credible interval (CrI) [1.8%, 33.9%]; Bayes factor (BF) compared to 50% prevalence = 45). Based on an exploratory review of 358 full texts, approximately 30% of the studies had excluded gambling (95% CrI [23.54%, 32.80%]; BF compared to 50% prevalence = 2.8 × 10¹4). We further surveyed English (EN; n = 312) and Slovak (SK; n = 319) speakers about their views on gaming and gambling, which revealed terminological confusions in both languages. For instance, about half of the participants (EN = 48.7%/SK = 59.1%) interpreted ‘gambling’ to be ‘gaming’ as well. The majority of gaming disorder survey research appears to be contaminated by a systematic measurement error stemming from conceptual confusion. This problem is linguistic and does not concern all languages similarly. To minimize the impact of the error, we recommend researchers clarify survey instructions and seek alternatives to survey data.

Rho guanine nucleotide exchange factor (Rgnef/p190RhoGEF), a RhoA-specific guanine nucleotide exchange factor, has been implicated in cancer and amyotrophic lateral sclerosis, but little is known about its role in bone. Here we investigate the roles of Rgnef in bone metabolism using Rgnef-deficient and overexpressing mice. Compared with littermate wildtype mice, Rgnef-deficient mice had increased bone mass owing to lower osteolysis and higher osteogenesis, and Rgnef-overexpressing transgenic mice had the opposite bone phenotype. Rgnef deficiency inhibited osteoclast formation and resorptive function and promoted osteoblast differentiation and mineralization, whereas Rgnef overexpression had the reverse effect. Mechanistically, Rgnef promotes osteoclastogenesis by enhancing the activity of nuclear factor kappa B (NF-κB), mitogen-activated protein kinases and AKT through the activation of RhoA and Rac1 and attenuates osteoblastogenesis through the RhoA/Rac1-mediated NF-κB activation. Moreover, Rgnef-deficient mice were protected from bone loss caused by lipopolysaccharide-induced inflammation or ovariectomy. Thus, Rgnef is a crucial regulator of bone metabolism and could serve as a potential new target for treating bone diseases.

Engineering fibrous scaffolds with anisotropic architecture, tunable mechanics, and hydrated stability is essential for constructing physiologically relevant muscle tissue models. Here, we introduce gelatin–chitosan (GC) composite scaffolds fabricated via a tensile spinning process, enabling the formation of continuously aligned fibers without toxic solvents. Systematic modulation of chitosan content governed precursor viscoelasticity, which translated into controlled fiber alignment, pore characteristics (20–26%), hydrated-state resilience, and stiffness spanning 3–87 kPa. Spectroscopic and surface-potential analyses revealed compositional regulation of intermolecular interactions and surface-accessible functional groups, establishing a direct link between the GC formulation and interfacial chemical properties. Intermediate formulations exhibited an optimal convergence of structural and interfacial attributes─moderate stiffness, sustained diffusivity, and accessible oxygen- and nitrogen-bearing functionalities─conditions that supported high myoblast adhesion, viability, and directional cytoskeletal alignment. Notably, GC4.0 scaffolds promoted anisotropic myotube formation along the fiber axis, underscoring the requirement for co-optimized fiber alignment, hydrated mechanical integrity, and surface chemistry in functional myogenesis. By integrating compositional control with architectural and interfacial outcomes, this work establishes a unified formulation–structure–function relationship for gelatin–chitosan scaffolds. Rather than introducing a new material class, the findings reveal how incremental variation in the chitosan content within a toxic solvent-free spinning platform yields structurally aligned, compositionally responsive fibrous scaffolds capable of directing muscle cell organization, offering a refined design framework for next-generation muscle tissue engineering.