Search Results (2,031)

Social media platforms host abundant and timely descriptions of medication experiences that can complement traditional pharmacovigilance systems. Yet the linguistic informality of these data presents a major challenge for mapping adverse drug event (ADE) expressions to standardized medical terminology. In this study, we developed BERT-based language models to classify ADE mentions from social media into MedDRA System Organ Classes (SOCs). Using the SMM4H and CADEC corpora, as well as their combination, we performed 20 iterations of 20% holdout validation for 3-, 6-, 22-, and 25-SOC classification tasks with a selected fixed training configuration (learning rate, batch size, and training epochs) based on training-loss convergence. The models achieved accuracies ranging from 75% to 94%, demonstrating strong performance for SOC-level classification of noisy and informal ADE expressions under the evaluated settings. These results are based on a controlled mention-level evaluation using deduplicated adverse drug event strings and do not establish document-level or real-world deployment generalization. This work provides a systematic evaluation of BERT-based models for SOC-level classification of ADEs and demonstrates consistent performance within the evaluated datasets and label granularities. While direct comparison with prior studies is limited by differences in datasets and evaluation protocols, the results demonstrate that transformer-based models can effectively classify ADEs into SOCs. These findings support the use of transformer-based normalization for SOC-level aggregation of user-reported adverse events and their integration into large-scale social media pharmacovigilance pipelines as a downstream component under controlled conditions. Full article

Background/Objectives: Nowadays, social media is increasingly utilized in the field of orthodontics for information sharing and promotion, yet its influence on patients’ willingness to initiate orthodontic treatment remains insufficiently defined. This systematic review aims to synthesize the available evidence on the impact of orthodontic-related social media content on patients’ willingness to seek orthodontic treatment. Methods: An extensive literature search was performed across five electronic databases up to August 2025, complemented by manual screening of reference lists. Randomized and non-randomized studies evaluating orthodontic-related social media exposure and reported treatment-related willingness or motivation outcomes were considered for inclusion. Results: A total of 1243 records were identified, and eight studies met the inclusion criteria, including six cross-sectional studies, one randomized controlled trial, and one qualitative study. Given the diversity of study designs and assessment methods, the results were synthesized narratively. Visually oriented orthodontic-related social media posts, particularly outcome-focused imagery such as before–after photographs, were more frequently associated with increased willingness to seek orthodontic treatment compared with technical content. Gender-related differences were reported, with female participants appearing more responsive to orthodontic-related social media exposure. Across the included studies, Instagram was identified as the platform exerting the strongest influence. Conclusions: The findings of this systematic review indicate that visually oriented orthodontic-related social media content, particularly outcome-focused imagery such as before–after photographs, shows more consistent associations with willingness to seek orthodontic treatment, alongside gender-related differences and platform-specific effects. Full article

Australia’s ban on social media for under-16s, introduced in December 2025, made it the first country worldwide to implement a nationwide prohibition on major platforms for adolescents. This narrative literature review compares Australia’s age-based restriction with international approaches to protecting young people from online risks. The review synthesized 26 academic studies and 15 grey literature sources (policy documents, legislation, and official reports published between 2015 and 2025). It employed Bronfenbrenner’s bioecological systems theory to examine effects across family, platform, institutional, and broader socio-legal contexts. Three key themes emerged: (A) Empirical findings on age-threshold policies remain inconclusive and context-dependent. While unregulated use relates to psychological vulnerabilities, structured and intentional engagement can promote social connection, identity exploration, and support access, especially for marginalized youth. (B) Global responses vary, favoring alternatives like parental consent, platform duty-of-care obligations, and screen-time control measures. (C) Balanced, sustainable harm reduction depends on combining parental involvement, platform accountability, and digital literacy education. Overall, while Australia’s precautionary approach addresses legitimate developmental and public health concerns, its effectiveness seems limited by enforcement challenges, risks of digital exclusion, and potential human rights issues. Bronfenbrenner’s framework underscores the need for coordinated governance across interconnected systems to lessen online harm. Full article

The contemporary media landscape has sustained a substantial transformation with the rise of AI-driven algorithmic platforms that enable activist organizations to produce and disseminate their own forms of political communication and campaigns. This study examines the YouTube channel of Code Pink, a prominent U.S.-based anti-war and social justice organization, to explore how activist media practices intersect with contemporary forms of journalism. Over a one-month period, video transcripts from the organization’s YouTube channel were analyzed using NVivo 15, employing a hybrid qualitative approach that combined inductive and deductive coding. Deductive codes were informed by sustained observation of the channel over one year (short and long videos on YouTube, TikTok, and X), supplemented by engagement with relevant news coverage, while inductive coding followed grounded theory principles, allowing themes to emerge directly from the transcripts. Large Language Models (LLMs) were employed as exploratory analytic tools to support AI-assisted qualitative analysis, complementing manual coding processes. The analysis focuses on how Code Pink frames political events and U.S. foreign policy through confrontational interviews, protest documentation, and the dissemination of commentary to online audiences. Findings suggest that the organization’s video content operates simultaneously as political activism, protest performance, and quasi-journalistic reporting. Activists frequently adopt journalistic techniques—including interviewing political figures, providing on-the-ground commentary, and framing narratives around public accountability—while also advancing explicit ideological positions that challenge dominant media narratives. The study highlights how platform-based activist media blurs the boundaries between journalism, advocacy, and political performance, contributing to the construction of alternative public narratives in the digital age. Full article

Attention has become a central resource of contemporary political communication, yet existing accounts do not fully explain how visibility acquires social credibility and political force under platformised conditions. This article addresses that problem through the ritual model of communication and media rituals. It develops a theory-building framework linking attention, recognition, legitimacy, and participation within a platformised ritual circuit. Methodologically, it proceeds through conceptual synthesis and illustrative analytical reconstruction rather than causal testing. It reconstructs three public episode types centred on witnessing, conflict, and commemoration, using public artefacts, trace-based evidence, platform affordances, and reporting. The analysis argues that attention-based politics is a ritualised struggle over socially recognised salience. Visibility becomes politically consequential when publicly ratified through legible participation and when recognition traces are narrativised as claims to legitimacy. The article proposes a provisional comparative vocabulary for distinguishing dominant configurations of online political media rituals across concentrated witnessing, cyclical antagonism, and prolonged commemorative alignment. It concludes that platforms do not simply amplify visibility or host participation. They organise the recurring social forms through which visibility becomes usable in legitimacy claims. Full article

Facial expressions transmit information about internal states, both during social interaction and in response to shared stimuli such as films. When individuals view the same content, synchrony in their expressions reflects shared information processing, and the degree to which their expressions correlate indicates how similarly their perceptual and affective systems are responding to the common input. This makes interindividual expression synchrony a potential marker of engagement and subjective experience. However, the acquisition and analysis of facial data pose both ethical and technical challenges to researchers. ‘Trace’ is a research media player implemented in PsychoPy’s online platform Pavlovia, which captures anonymised facial landmark coordinates through a webcam, without the ethical and technical constraints of capturing and storing video images of participants. Nonetheless, its usefulness is currently limited due to the lack of available preprocessing and analysis tools. This paper describes the functionality of TraceLAB, a MATLAB-based toolbox designed for the preprocessing of Trace data: specifically, the formatting, aligning, and filtering of data. In addition, TraceLAB implements some novel analysis techniques to allow researchers to quantify interindividual synchrony of expressions (through correlated component analysis) and head movements (through Surrogate Synchrony), which may be interpreted as measures of shared information processing. These techniques are demonstrated here on both simulated and real datasets. Full article

Social media engagement drives both individual behavior and content dissemination, yet traditional analytics often reduce interactions to simple counts, obscuring the complex structures underlying user activity. In the highly competitive digital landscape, understanding how users interact with content is crucial for businesses aiming to optimize social media campaigns and maximize return on investment (ROI). Traditional engagement metrics, such as likes and shares, fail to capture the underlying structure and dynamics of user behavior. This study investigates the latent patterns of engagement by combining topological data analysis (TDA) with behavioral clustering across 100,000 posts on multiple platforms. Using persistent homology and k-nearest neighbour graphs, we reveal a primary bifurcation between Active (validation-focused) and Passive (consumption/propagation) users, nested four-strain substructures, and over 650 significant $H_1$ loops indicating recurring feedback cycles. Active users exhibit strong cluster cohesion and high engagement rates, while Passive users contribute broadly to content diffusion with slightly higher loop counts, highlighting distinct functional roles in social media dynamics. These findings provide a principled framework for targeting content, reinforcing feedback loops, and leveraging hub posts to amplify engagement. By linking topological structure to behavioral patterns, this work advances both the theoretical understanding of digital interaction and the practical design of more effective social media campaigns. Full article

This study examines how user-generated content (UGC) is incorporated and negotiated within platformized public service media (PSM) journalism, using the German content network funk as a case study. Based on a qualitative content analysis of selected formats and their social media posts, the study shows that participatory affordances offered by social media platforms (SMPs) are present but rarely foregrounded as central elements of storytelling. Instead, UGC is typically used as illustrative material or selectively embedded within editorial narratives. The analysis investigates how UGC is solicited, incorporated, and visually integrated into editorial storytelling across different formats. The findings identify three recurring patterns of UGC integration that illustrate how audience participation is negotiated within everyday editorial production: (1) illustrative UGC integration, (2) community-oriented UGC integration, and (3) minimalist UGC integration. Overall, the study highlights how platformized PSM journalism integrates UGC in ways that remain strongly editorially moderated rather than fully participatory, demonstrating how participation is enabled, constrained, and strategically applied within platform infrastructures. Full article

Disinformation is a global challenge that affects areas such as politics, health, economics, and science and is spread rapidly by social media (SM), suggesting the necessity of advancing educational proposals to strengthen critical thinking (CT) and digital literacy (DL). This quantitative, non-experimental, descriptive study identified the self-perception that master’s students (n = 72; at three Spanish universities; October 2024–March 2025) have regarding their DL, along with their CT, in post-truth and fake news settings. A self-administered online questionnaire (53 items) was conducted, covering aspects such as the responsible use of information and platforms, algorithmic perceptions, actions taken to verify this information, and concepts of CT, post-truth, and fake news. The results show that Instagram (97%) and WhatsApp (96%) predominated, with a notable proportion of users (86%) reporting that algorithms influenced them ‘highly’ or ‘moderately’. Despite being aware of disinformation they find on social media (65%) as well as its close link to hate speech (90% who ‘strongly’ or ‘somewhat’ agreed), this knowledge does not fully translate into taking measures to counter it. In fact 61% of respondents report sharing news on at least some occasions, while only 25% are able to identify a professional fact-checking organization. In conclusion, these findings suggest the merit of assessing the prevalence of skills such as Critical Thinking (CT) and strategies like fact-checking among students in other postgraduate education systems. Such assessments could inform the potential promotion of media and digital literacy as cross-curricular skills in education. This approach would help bridge the gap between theoretical knowledge and the active verification needed to counter disinformation. Full article

Short-form video content has gained importance as a popular form of digital media due to the rising popularity of social media platforms and the decreasing attention spans of consumers. However, a major obstacle to popularity detection in short-form content is the heterogeneous nature of the data, encompassing textual, visual, and metadata components. To tackle this challenge, we propose FUSEPOP, a robust multi-modal architecture. The proposed framework utilizes ResNet-50 for visual feature extraction and XLM-RoBERTa for encoding multilingual textual information. FUSEPOP employs a mutual information-based modality weighting mechanism with logarithmic smoothing and a 0.7 weight ceiling to balance contributions from each input stream. Furthermore, FUSEPOP implements a robust stacked generalization strategy trained via stratified 5-fold cross-validation. This approach utilizes a logistic regression meta-learner to dynamically synthesize predictions from random forest, XGBoost, and a neural network-based classifier. Experimental results show that this architecture significantly outperforms benchmark models, achieving an accuracy of 0.980 and an average F1-score of 0.964 on the feature configuration selected for this study, and remains competitive on a literature-aligned alternative configuration. These findings confirm that the proposed model successfully detects popularity on short-form social media content. Full article

The COVID-19 pandemic highlighted substantial variability in public health information environments, yet the relationship between information source, perceived credibility, and behavioral response remains incompletely understood. This study evaluated how information sources influence COVID-19-related knowledge and behaviors and whether targeted educational interventions modify these relationships. We conducted a prospective survey-based study (July–December 2021) among adults recruited from outpatient clinics in Michigan (N = 209). Participants completed pre- and post-intervention surveys assessing information sources, perceived reliability, knowledge, and behaviors, and were randomized to receive either a video or infographic. Social media was the most frequently reported source (n = 95) but had lower perceived reliability (mean 2.97/5), whereas healthcare workers (HCWs) were rated most reliable (mean 4.26/5) despite lower utilization (n = 60). Use of HCWs, print media, and websites was associated with higher baseline knowledge, while television and radio were associated with lower knowledge of vaccine side effects (p = 0.011 and p = 0.003). Educational interventions improved knowledge and attitudes, with differential effects across source groups, while infographic-based interventions were more effective among social media users (p = 0.034). Information sources and perceived credibility significantly shape health knowledge and behavior, highlighting the need for communication strategies that integrate trusted messengers, high-reach platforms, and health literacy to improve public health outcomes. Full article

The literature emphasizes the importance of perceived social media marketing activities (SMMAs) in shaping various brand-related outcomes. However, their importance in brand attachment formation remains underexplored. Grounded in the Stimulus–Organism–Response (S-O-R) framework and Attachment Theory, this study examines the relationship between SMMAs and brand attachment, and the impact of brand attachment on brand loyalty and consumer engagement with brand social media (CEBSM). A questionnaire survey was conducted with 502 consumers of outdoor and sports brands in Algeria. Data analysis was performed using Partial Least Squares Structural Equation Modeling (PLS-SEM). The findings unveil that SMMAs, including interactivity, informativeness, personalization, trendiness, and WOM, are positively associated with brand attachment. Furthermore, brand attachment is significantly associated with both brand loyalty and CEBSM. This study makes several theoretical contributions by being among the early studies to examine the individual effects of social media marketing dimensions, the role of SMMA in brand attachment formation, and brand-related outcomes alongside in-platform outcomes. This study offers recommendations to guide community managers and brand managers in clarifying the roles and capabilities of social media marketing in evoking and reinforcing brand attachment. Full article

Social media (SM) are virtual platforms for interpersonal interactions and play a significant role in the well-being of adolescents. It is therefore essential to understand how they use SM and how this use relates to interpersonal difficulties and social anxiety. The primary aim of this study was to examine SM use profiles together with interpersonal difficulties and social anxiety in a sample of adolescents. The sample comprised 304 adolescents aged 16–18 years, selected through probabilistic sampling: 170 (55.9%) were female and 134 (44.1%) were male. Four SM use profiles were identified, according to sex, time spent on SM, and number of accounts. Differences in interpersonal difficulties and social anxiety were observed across profiles, particularly among adolescents who showed intensive SM-centred use of the internet. Differences between males and females were also evident. The findings and their implications for the development of programmes aimed at encouraging healthy SM use among adolescents with a special focus on interpersonal well-being are discussed. Full article

This study examines how the Science of Reading is represented in Twitter discourse and compares these representations to contemporary models of reading development. Although the Science of Reading is frequently positioned as an equity-oriented reform, little is known about how related ideas circulate in public discourse, particularly across social media platforms that increasingly shape teacher learning, policymaking, and public opinion. This content analysis study analyzed a sample of 14,165 tweets containing the hashtag #scienceofreading from 2020–2021 and 2022–2023. It explores two primary questions investigating (1) the extent to which essential literacy skills (e.g., phonological awareness, phonics, comprehension, vocabulary) are referenced in tweets or linked content and (2) the extent to which specific subgroup classifications identified by the National Assessment of Educational Progress (e.g., Black, Hispanic, students with disabilities, low-income, and other populations) are mentioned in the same sample of discourse on Twitter. Findings demonstrate that online discourse on Twitter (now X) includes more references to decoding-related skills such as phonological awareness and phonics, with far fewer mentions of language-related skills such as comprehension or vocabulary. Mentions of subgroups were minimal, while references to students with disabilities with explicit mention of dyslexia occurred at four times the frequency of race- or income-related subgroups. These distributions contrast with persistent national achievement disparities and suggest that contemporary Science of Reading discourse is more strongly oriented toward decoding-related skills than toward equity-focused concerns. Implications for teacher preparation, policy enactment, and critical media literacy are discussed. Full article

Trust decay in social media is a serious threat to user experience and platform ecology. To solve this problem, this paper focuses on triadic closure in the infrastructure of social networks and explores its mechanism in trust decay prevention. Based on the systematic comparison of the ER random graph, the BA scale-free network, a forest fire model, and complete graph approaches, two core metrics, the trust decay risk index and trust resilience index, are proposed in this paper. Combined with structural indices such as the clustering coefficient, the average path length, and the triangular closure number and its growth rate, the quantitative relationship between network structure evolution and trust decay risk is established. It is found that the forest fire model exhibits optimal trust resilience in structure due to its power-law growth characteristics of high clustering, short path length and triangular closure; the dynamic mechanism of trust decay under different network growth modes is significantly different. The validity of the theoretical framework is further supported by the verification of Sina Weibo attention relationship network data. The analysis framework of network growth evolution based on fusion triangle closure and the risk and resilience indicators defined in this paper provides a computable theoretical tool for understanding and predicting trust evolution in social media from the perspective of network structure. Full article