Mental health challenges and cyberbullying are increasingly prevalent in digital spaces, necessitating scalable and interpretable detection systems. This paper introduces a unified multiclass classification framework for detecting ten distinct mental health and cyberbullying categories from social media data. We curate datasets from Twitter and Reddit, implementing a rigorous "split-then-balance" pipeline to train on balanced data while evaluating on a realistic, held-out imbalanced test set. We conducted a comprehensive evaluation comparing traditional lexical models, hybrid approaches, and several end-to-end fine-tuned transformers. Our results demonstrate that end-to-end fine-tuning is critical for performance, with the domain-adapted MentalBERT emerging as the top model, achieving an accuracy of 0.92 and a Macro F1 score of 0.76, surpassing both its generic counterpart and a zero-shot LLM baseline. Grounded in a comprehensive ethical analysis, we frame the system as a human-in-the-loop screening aid, not a diagnostic tool. To support this, we introduce a hybrid SHAPLLM explainability framework and present a prototype dashboard ("Social Media Screener") designed to integrate model predictions and their explanations into a practical workflow for moderators. Our work provides a robust baseline, highlighting future needs for multi-label, clinically-validated datasets at the critical intersection of online safety and computational mental health.
Difficult decision-making problems abound in various disciplines and domains. The proliferation of generative techniques, especially large language models (LLMs), has excited interest in using them for decision support. However, LLMs cannot yet resolve missingness in their training data, leading to hallucinations. Retrieval-Augmented Generation (RAG) enhances LLMs by incorporating external information retrieval, reducing hallucinations and improving accuracy. Yet, RAG and related methods are only partial solutions, as they may lack access to all necessary sources or key missing information. Even everyday issues often challenge LLMs' abilities. Submitting longer prompts with context and examples is one approach to address knowledge gaps, but designing effective prompts is non-trivial and may not capture complex mental models of domain experts. For tasks with missing critical information, LLMs are insufficient, as are many existing systems poorly represented in available documents. This paper explores how LLMs can make decision-making more efficient, using a running example of evaluating whether to respond to a call for proposals. We propose a technology based on optimized human-machine dialogue and monotone Boolean and k-valued functions to discover a computationally tractable personal expert mental model (EMM) of decision-making. Our EMM algorithm for LLM prompt engineering has four steps: (1) factor identification, (2) hierarchical structuring of factors, (3) generating a generalized expert mental model specification, and (4) generating a detailed generalized expert mental model from that specification.
Koustuv Saha, Yoshee Jain, Violeta J. Rodriguez
et al.
The ubiquity and widespread use of digital and online technologies have transformed mental health support, with online mental health communities (OMHCs) providing safe spaces for peer support. More recently, generative AI and large language models (LLMs) have introduced new possibilities for scalable, around-the-clock mental health assistance that could potentially augment and supplement the capabilities of OMHCs. Although genAI shows promise in delivering immediate and personalized responses, its effectiveness in replicating the nuanced, experience-based support of human peers remains an open question. In this study, we harnessed 24,114 posts and 138,758 online community (OC) responses from 55 OMHCs on Reddit. We prompted several state-of-the-art LLMs (GPT-4-Turbo, Llama-3, and Mistral-7B) with these posts, and compared their responses to human-written (OC) responses based on a variety of linguistic measures across psycholinguistics and lexico-semantics. Our findings revealed that AI responses are more verbose, readable, and analytically structured, but lack linguistic diversity and personal narratives inherent in human--human interactions. Through a qualitative examination, we found validation as well as complementary insights into the nature of AI responses, such as its neutral stance and the absence of seeking back-and-forth clarifications. We discuss the ethical and practical implications of integrating generative AI into OMHCs, advocating for frameworks that balance AI's scalability and timeliness with the irreplaceable authenticity, social interactiveness, and expertise of human connections that form the ethos of online support communities.
Mental health care-seeking among marginalized young adults has received limited attention in CSCW research. Through in-depth interviews and visual elicitation methods with 18 diverse U.S. participants, our study reveals how marginalized identities shape mental health care-seeking journeys, often characterized by low aspirations and passive care-seeking influenced by lived experiences of marginalization. However, we found the transformative function of "care encounters" - serendipitous interactions with mental health resources that occur when individuals are not actively seeking support. These encounters serve as critical turning points, catalyzing shifts in aspiration and enabling more proactive care-seeking behaviors. Our analysis identifies both the infrastructural conditions that enable transformative care encounters and the aspiration breakdowns that impede care-seeking processes. This work makes conceptual contributions by supplementing traditional motivation-based care-seeking models with a reconceptualization of "care encounters" that accounts for the infrastructural and serendipitous nature of mental health access. We advance understanding of how marginalized identity uniquely influences care-seeking behaviors while providing actionable design implications for embedding technology-mediated "care encounters" into socio-technical interventions that can better support mental health care access for vulnerable populations.
We propose a clinical decision support system (CDSS) for mental health diagnosis that combines the strengths of large language models (LLMs) and constraint logic programming (CLP). Having a CDSS is important because of the high complexity of diagnostic manuals used by mental health professionals and the danger of diagnostic errors. Our CDSS is a software tool that uses an LLM to translate diagnostic manuals to a logic program and solves the program using an off-the-shelf CLP engine to query a patient's diagnosis based on the encoded rules and provided data. By giving domain experts the opportunity to inspect the LLM-generated logic program, and making modifications when needed, our CDSS ensures that the diagnosis is not only accurate but also interpretable. We experimentally compare it with two baseline approaches of using LLMs: diagnosing patients using the LLM-only approach, and using the LLM-generated logic program but without expert inspection. The results show that, while LLMs are extremely useful in generating candidate logic programs, these programs still require expert inspection and modification to guarantee faithfulness to the official diagnostic manuals. Additionally, ethical concerns arise from the direct use of patient data in LLMs, underscoring the need for a safer hybrid approach like our proposed method.
Large Language Models (LLMs) are increasingly deployed in mental health contexts, from structured therapeutic support tools to informal chat-based well-being assistants. While these systems increase accessibility, scalability, and personalization, their integration into mental health care brings privacy and safety challenges that have not been well-examined. Unlike traditional clinical interactions, LLM-mediated therapy often lacks a clear structure for what information is collected, how it is processed, and how it is stored or reused. Users without clinical guidance may over-disclose personal information, which is sometimes irrelevant to their presenting concern, due to misplaced trust, lack of awareness of data risks, or the conversational design of the system. This overexposure raises privacy concerns and also increases the potential for LLM bias, misinterpretation, and long-term data misuse. We propose a framework embedding Artificial Intelligence (AI) literacy interventions directly into mental health conversational systems, and outline a study plan to evaluate their impact on disclosure safety, trust, and user experience.
Artificial intelligence and social computing rely on hundreds of thousands of content reviewers to classify high volumes of harmful and forbidden content. Many workers report long-term, potentially irreversible psychological harm. This work is similar to activities that cause psychological harm to other kinds of helping professionals even after small doses of exposure. Yet researchers struggle to measure the mental health of content reviewers well enough to inform diagnoses, evaluate workplace improvements, hold employers accountable, or advance scientific understanding. This systematic review summarizes psychological measures from other professions and relates them to the experiences of content reviewers. After identifying 1,673 potential papers, we reviewed 143 that validate measures in related occupations. We summarize the uses of psychological measurement for content reviewing, differences between clinical and research measures, and 12 measures that are adaptable to content reviewing. We find serious gaps in measurement validity in regions where content review labor is common. Overall, we argue for reliable measures of content reviewer mental health that match the nature of the work and are culturally-relevant.
Waganesh A. Zeleke, Mengistu Dagnew, Yemataw Wondie
et al.
<b>Background:</b> Internal displacement is a global crisis, with Ethiopia being among the most affected countries due to conflict, violence, and natural disasters. Internally displaced people (IDPs) face multifaceted trauma at the individual, family, and community levels, exacerbating mental health issues such as PTSD and depression. Despite ongoing interventions, many programs lack cultural and contextual adaptations that are suited to Ethiopia’s diverse communities. <b>Aims:</b> This study aimed to explore the cultural and contextual factors influencing trauma-informed interventions for IDPs in Ethiopia and develop a framework for culturally responsive mental health support. <b>Methods</b>: Utilizing Community-Based Participatory Action Research (CBPAR) and interpretative phenomenological research design, data were collected from 42 stakeholders through Focus Group Discussions and in-depth individual interviews, and subsequently analyzed using thematic analysis to identify patterns and themes. <b>Results:</b> Key findings highlighted the ongoing trauma faced by IDPs, the importance of demographic characteristics (e.g., gender and education), and the role of cultural stereotypes in shaping trauma perceptions. Traditional community rituals such as coffee ceremonies have been identified as vital for healing. Effective interventions require cultural alignment, respect for religious values, and integration into community activities. <b>Conclusions:</b> This study underscores the need for culturally and contextually responsive trauma-informed intervention. Incorporating community rituals and engaging local leaders enhances intervention acceptance and effectiveness. The findings provide a framework to address mental health needs while fostering resilience among internally displaced Ethiopian populations.
Prior research on young adults' mental health help-seeking mostly focuses on one particular resource such as a mobile app or digital platform, paying less attention to their lived experiences interacting with the ecosystem of resources. We conducted in-depth interviews with 18 participants about their help-seeking and non-help-seeking experiences. Guided by Social Ecological Theory, we proposed a Socio-technical Ecosystem Framework for mental health care, consisting of four levels of resources, including technological-, interpersonal-, community-, and societal level resources. Using this framework, we identified two types of support systems for help-seeking, single-resource support system and multi-resource support system. These resources support young adults' help-seeking via three mechanisms, \textit{care-giving}, \textit{care-mediating}, and \textit{care-outreaching}, forming various pathways to care. We then pointed out the barriers to resource use at each level and the general challenges in finding a support system. Our findings contributed to a conceptual framework to categorize mental health care. It also serves as a practical framework to identify challenges in the pathways to care and discover design implications.
Glassy polymer melts such as the plastics used in pipes, structural materials, and medical devices are ubiquitous in daily life. They accumulate damage over time due to their use, which limits their functionalities and demands periodic replacement. The resulting economic and social burden could be mitigated by the design of self-healing mechanisms that expand the lifespan of materials. However, the characteristic low molecular mobility in glassy polymer melts intrinsically limits the design of self-healing behavior. We demonstrate through numerical simulations that controlled oscillatory deformations enhance the local molecular mobility of glassy polymers without compromising their structural or mechanical stability. We apply this principle to increase the molecular mobility around the surface of a crack, inducing fracture repair and recovering the mechanical properties of the pristine material. Our findings establish a general physical mechanism of self-healing in glasses that may inspire the design and processing of new glassy materials.
The COVID-19 pandemic has intensified the urgency for effective and accessible mental health interventions in people's daily lives. Mobile Health (mHealth) solutions, such as AI Chatbots and Mindfulness Apps, have gained traction as they expand beyond traditional clinical settings to support daily life. However, the effectiveness of current mHealth solutions is impeded by the lack of context-awareness, personalization, and modularity to foster their reusability. This paper introduces CAREForMe, a contextual multi-armed bandit (CMAB) recommendation framework for mental health. Designed with context-awareness, personalization, and modularity at its core, CAREForMe harnesses mobile sensing and integrates online learning algorithms with user clustering capability to deliver timely, personalized recommendations. With its modular design, CAREForMe serves as both a customizable recommendation framework to guide future research, and a collaborative platform to facilitate interdisciplinary contributions in mHealth research. We showcase CAREForMe's versatility through its implementation across various platforms (e.g., Discord, Telegram) and its customization to diverse recommendation features.
Rebecca Roppolo, William Aldridge, Christina DiSalvo
et al.
Background External implementation support (EIS) can aid implementation and scale-up efforts, but less has been reported about the experience of those receiving EIS, such as the feasibility and usability of participating in the support process. Method From November 2016 to April 2022, data were collected from the support participants across 13 regions in North Carolina and South Carolina implementing the Triple P system of interventions and the regional support team members who provided EIS to these partners. The experience of participating in EIS was assessed using measures of acceptability, appropriateness, accessibility, quality of delivery, feasibility, likelihood and actual use of support materials received, degree of collaboration, and frequency of contact. Mann–Whitney U tests or Kruskal–Wallis tests were conducted to explore differences in these measures across a variety of regional characteristics and contexts. Results Support participants generally found EIS to be accessible, acceptable, appropriate, feasible, and delivered with high quality across different states, regions, and over the course of the support relationship. Support was generally provided 1–2 times per month and collaboration between regional support teams and regional Triple P partners was rated highly significant differences between support participant experiences were generally limited to ratings of support accessibility, engagement with data collection processes, and number of monthly contacts. Conclusions This pattern of findings suggests that EIS as provided by regional support teams is feasible for support participants across a diversity of contexts. Additional research on EIS would help refine the field and illuminate promising practices and mechanisms of change to accelerate successful and sustainable implementation.
Abstract: Cultural ideals are repeatedly coded into hidden messages through drums, sampling, and signifying, which is all embodied through various dance styles. This transformation brings new meaning to political, social, historical, and cultural issues. The policing of the black moving body has become an international symbol of struggle, pain, oppression and injustice. How strong is a symbol? To be seen is to be remembered. When will we forget what has been learned? When will we receive what our ancestors have earned? The purpose of this research is to deepen an understanding of the sub- group or ethnic group known as African- Americans in the new world, also known as, the Americas. This paper begins with an introduction to phenomena such as present day stereotypes, social constructs and mandates on what is considered by Brenda Dixon Gottschild to be the "black dancing body" in America. The discussion to follow deals with how policing the black dancing and moving body occurs throughout various interlinked systems in America. The black female and male forms are constantly violated by lack of access to education, diagnosis of illness and reinforced stereotypes of aggression. An embodied exploration of the Pan- African dance technique known as Umfundalai (pronounced ma-foon-da-la) provides a deeper understanding of protest within the arts. This writer will show therapeutic values inherent in the stylized movement vocabulary of people of the African Diaspora and the utilization of their culture as a viable resource for healing in an acute care psychiatric hospital.
Extended reality (XR) technology has the incredible potential to revolutionize mental health treatment and support, bringing a whole new dimension to the field. Through the use of immersive virtual and augmented reality experiences, individuals can enter entirely new worlds and realities that provide a safe and controlled space for therapy and self-exploration. Whether it's stepping into a calming natural environment, practicing social interactions or confronting past traumas in a controlled environment, extended reality offers endless possibilities. Engaging these virtual realities, individuals can gain a deeper understanding of themselves and their emotions, learn coping strategies, and practice important life skills in a way that is both engaging and effective. The wonders of extended reality for mental health are truly awe-inspiring and offer a powerful tool for improving the well-being of individuals around the world. However, we should remember, everything has its disadvantages, and XR is no different. While XR is a revolution, the human brain is very complex, fragile and unique (like with fingerprints, no two people have the same brain anatomy), leading to varying conditions, results, experiences and consequences. This article presents insights and information on how immersive interactive digital experiences can shape our minds and behaviors. Research to date suggests that XR experiences can change regions of the brain responsible for attention and visuospatial skills.
Today, the impact of the environment as one of the most important factors in mental health and human behavioral patterns, as well as the influence of nature and green spaces on reducing daily psychological stress and enhancing mental health, has been proven. Researchers believe that nature and green spaces can effectively reduce daily psychological stress and enhance mental health by helping to cope with anxiety. The main goal of this research is to incorporate healing components into landscape design and urban environments, equipping these landscapes with healing qualities and providing appropriate solutions to create healing urban landscapes. Given the current time frame and the existence of social crises, economic instability in society, and the resulting psychological damage, conducting this research is timely. This research, conducted to analyze concepts and indicators in Paknejad Boulevard in Tehran, is situated within the scope of applied research in terms of objectives and subject matter. In terms of research methodology, it is a descriptive-analytical study with a content analysis approach, leaning towards qualitative methods. "Initially, some key indicators were extracted through library studies, and then field studies such as behavioral mapping were conducted using research tools. Subsequently, a questionnaire was designed based on the healing components extracted from relevant sources. To assess validity, the questionnaire was randomly distributed to users. The responses were analyzed using the statistical analysis software SPSS, and the data outputs were presented in the form of charts and frequency tables. The results of this research indicate that influential healing indicators in urban landscape design include elements such as the presence of water, interaction with nature, sensory stimulation, environmental tranquility, a sense of security in space, colors, pleasant scents, and soothing sounds. Ultimately, strategic actions and appropriate solutions were proposed by analyzing influential healing principles and indicators essential for designing a healing urban landscape.
Moeen Ali Naqvi, Sehrish Malik, Merve Astekin
et al.
With the growing adoption of self-adaptive systems in various domains, there is an increasing need for strategies to assess their correct behavior. In particular self-healing systems, which aim to provide resilience and fault-tolerance, often deal with unanticipated failures in critical and highly dynamic environments. Their reactive and complex behavior makes it challenging to assess if these systems execute according to the desired goals. Recently, several studies have expressed concern about the lack of systematic evaluation methods for self-healing behavior. In this paper, we propose CHESS, an approach for the systematic evaluation of self-adaptive and self-healing systems that builds on chaos engineering. Chaos engineering is a methodology for subjecting a system to unexpected conditions and scenarios. It has shown great promise in helping developers build resilient microservice architectures and cyber-physical systems. CHESS turns this idea around by using chaos engineering to evaluate how well a self-healing system can withstand such perturbations. We investigate the viability of this approach through an exploratory study on a self-healing smart office environment. The study helps us explore the promises and limitations of the approach, as well as identify directions where additional work is needed. We conclude with a summary of lessons learned.
Donghoon Shin, Subeen Park, Esther Hehsun Kim
et al.
Social support in online mental health communities (OMHCs) is an effective and accessible way of managing mental wellbeing. In this process, sharing emotional supports is considered crucial to the thriving social supports in OMHCs, yet often difficult for both seekers and providers. To support empathetic interactions, we design an AI-infused workflow that allows users to write emotional supporting messages to other users' posts based on the elicitation of the seeker's emotion and contextual keywords from writing. Based on a preliminary user study (N = 10), we identified that the system helped seekers to clarify emotion and describe text concretely while writing a post. Providers could also learn how to react empathetically to the post. Based on these results, we suggest design implications for our proposed system.
Women agency defined as the ability to conceive of purposeful plan and to carry out action consistent with such a plan can play an important role in determining health status. Using data from female respondents conducted in a survey in the city of Dhaka in Bangladesh, this paper explores how women agency relates to their physical and mental health. The findings indicate women with high agency to experience significantly lesser mental distress on average. Counterintuitively, these women are more likely to report poor physical health. As an explanation, we propose purposeful action among women with high agency as a potential reason, wherein they conceive purpose in the future and formulate action that is feasible today. Hence, these women prefer to report illness and get the required treatment to ensure better future health. This illuminates our understanding of sustainable development and emphasises the critical role of women agency for sustainable human development.
M. Elizabeth Lewis Hall, Jason McMartin, Crystal L. Park
et al.
Current measures of religious coping are generally etic in nature, measuring constructs across religions. Emic variables (i.e., those specific to particular religions) are often left out, which limits our ability to assess religious/spiritual coping during times of stress and adversity. Here we provide findings from three studies we conducted to develop and test an emic Christian meaning-making coping method: identifying with Christ in his suffering. We ground this construct in Christian theology, the psychology of religious/spiritual coping literature, and existing qualitative research. In the first study, we developed items and tested the items for clarity and generalizability to diverse Christian groups using expert review and cognitive interviewing with participants from five distinct Christian groups. In the second study, we conducted exploratory factor analysis using data from MTurk (N = 335), which revealed a two-factor structure consistent with our theoretical formulation. In the third study, we established factor stability and construct validity using data from Prolific (N = 355). Because we conceptualize identification with Christ in his suffering as a form of meaning-making coping, in this third study we also explored the relationship of the measure to well-being using incremental validity analyses. We found that identification with Christ in suffering predicted measures of well-being above and beyond the variance explained by etic religious coping measures. Collectively, these results highlight the value of emic religious measures of coping with suffering.