{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,8,14]],"date-time":"2026-08-14T06:26:56Z","timestamp":1786688816208,"version":"build-2736575974"},"reference-count":66,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2026,8,14]],"date-time":"2026-08-14T00:00:00Z","timestamp":1786665600000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":["frontiersin.org"],"crossmark-restriction":true},"short-container-title":["Front. Digit. Health"],"abstract":"<jats:p>\n                    Linguistic-based anxiety screening has become a widely adopted approach for detecting anxiety in social media, with pre-trained language models (PLMs) now forming the state-of-the-art foundation for early mental health detection. However, most existing PLMs are not explicitly optimized for anxiety detection (AD), limiting their adaptability to fine-grained psycholinguistic cues. To address this gap, we introduce\n                    <jats:italic>SOFIA<\/jats:italic>\n                    (\n                    <jats:italic>\n                      <jats:underline>A<\/jats:underline>\n                      nxious?\n                      <jats:underline>SOF<\/jats:underline>\n                      ten-\n                      <jats:underline>I<\/jats:underline>\n                      t-Out\n                    <\/jats:italic>\n                    ), a novel prompt-based reasoning framework that enhances PLM capability for binary AD without full fine-tuning.\n                    <jats:italic>SOFIA<\/jats:italic>\n                    integrates multiple prompt template strategies with a dual-end truncation method that preserves both initial contextual cues and concluding emotional markers, enabling richer and more stable representations of conversational anxiety signals. Additionally, a knowledge-enhanced verbalizer incorporates structured domain knowledge through soft, manual, and automatic verbalizers, improving reasoning within frozen PLMs. We evaluate\n                    <jats:italic>SOFIA<\/jats:italic>\n                    across three self-reported anxiety corpora, such as Guo, SMHD, and Kim, using multiple 100M\u2212300M-parameter PLMs. Experimental results demonstrate consistent improvements over traditional baselines, including F1 score gains of 5.37%, 5.40%, and 7.92% for MentalBERT on the respective datasets when using soft templates of length 20 with knowledge base-driven verbalizers. These findings highlight\n                    <jats:italic>SOFIA<\/jats:italic>\n                    's effectiveness as a scalable, low-resource, and knowledge-informed framework for digital mental health screening applications.\n                  <\/jats:p>","DOI":"10.3389\/fdgth.2026.1899582","type":"journal-article","created":{"date-parts":[[2026,8,14]],"date-time":"2026-08-14T05:31:50Z","timestamp":1786685510000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":0,"title":["SOFIA: a knowledge-infused prompt-based reasoning framework for anxiety screening in social media"],"prefix":"10.3389","volume":"8","author":[{"given":"Ashala","family":"Senanayake","sequence":"first","affiliation":[{"name":"Tamai Investment Education Inc","place":["Kyoto, Japan"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Prasan","family":"Yapa","sequence":"additional","affiliation":[{"name":"Luxembourg Centre for Systems Biomedicine (LCSB), University of Luxembourg","place":["Esch-sur-Alzette, Luxembourg"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Sidath R.","family":"Liyanage","sequence":"additional","affiliation":[{"name":"Faculty of Computing & Technology, University of Kelaniya","place":["Kelaniya, Sri Lanka"]}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1965","published-online":{"date-parts":[[2026,8,14]]},"reference":[{"key":"B1","doi-asserted-by":"publisher","first-page":"22170","DOI":"10.1007\/s12144-024-05980-z","article-title":"Social media: a digital social mirror for identity development during adolescence","volume":"43","author":"P\u00e9rez-Torres","year":"2024","journal-title":"Curr Psychol"},{"key":"B2","first-page":"1","article-title":"Personalized state anxiety detection: an empirical study with linguistic biomarkers and a machine learning pipeline","author":"Wang","year":"2023"},{"key":"B3","doi-asserted-by":"publisher","first-page":"1423","DOI":"10.2147\/PRBM.S365012","article-title":"State of emotional health disorders of undergraduate students in the United Arab Emirates: a cross-sectional survey","volume":"15","author":"Al Marzouqi","year":"2022","journal-title":"Psychol Res Behav Manag"},{"key":"B4","doi-asserted-by":"publisher","first-page":"1087528","DOI":"10.3389\/fpsyg.2022.1087528","article-title":"Predictors and prevalence of stress, anxiety, depression, and PTSD among university students during the second wave of the COVID-19 pandemic in Turkey","volume":"13","author":"Aslan","year":"2023","journal-title":"Front Psychol"},{"key":"B5","doi-asserted-by":"publisher","first-page":"777","DOI":"10.1080\/j.1440-1614.2006.01883.x","article-title":"Web-based survey of depression, anxiety and stress in first-year tertiary education students in Hong Kong","volume":"40","author":"Wong","year":"2006","journal-title":"Aust N Z J Psychiatry"},{"key":"B6","doi-asserted-by":"publisher","first-page":"779039","DOI":"10.3389\/fdgth.2022.779039","article-title":"Toward linguistic recognition of generalized anxiety disorder","volume":"4","author":"Rook","year":"2022","journal-title":"Front Digit Health"},{"key":"B7","doi-asserted-by":"publisher","first-page":"115579","DOI":"10.1016\/j.psychres.2023.115579","article-title":"Identifying self-disclosed anxiety on Twitter: a natural language processing approach","volume":"330","author":"Zarate","year":"2023","journal-title":"Psychiatry Res"},{"key":"B8","doi-asserted-by":"publisher","first-page":"2","DOI":"10.48550\/arXiv.2502.13278","article-title":"Performance evaluation of sentiment analysis on text and emoji data using end-to-end, transfer learning, distributed and explainable ai models","volume":"13","author":"Velampalli","year":"2022","journal-title":"J Adv Inf Technol"},{"key":"B9","doi-asserted-by":"publisher","first-page":"109","DOI":"10.12720\/jait.11.3.109-118","article-title":"Quantifying the natural sentiment strength of polar term senses using semantic gloss information and degree adverbs","volume":"11","author":"Darwich","year":"2020","journal-title":"J Adv Inf Technol"},{"key":"B10","first-page":"58","article-title":"Detecting anxiety through reddit","author":"Shen","year":"2017"},{"key":"B11","doi-asserted-by":"publisher","first-page":"720","DOI":"10.1080\/02699931.2011.602048","article-title":"Linguistic correlates of social anxiety disorder","volume":"26","author":"Hofmann","year":"2012","journal-title":"Cogn Emot"},{"key":"B12","first-page":"92","article-title":"Expressions of anxiety in political texts","author":"Rheault","year":"2016"},{"key":"B13","first-page":"69","article-title":"Hierarchical neural model with attention mechanisms for the classification of social media text related to mental health","author":"Ive","year":"2018"},{"key":"B14","first-page":"7184","article-title":"MentalBERT: publicly available pretrained language models for mental healthcare","author":"Ji","year":"2022"},{"key":"B15","first-page":"1077","article-title":"PsychBERT: a mental health language model for social media mental health behavioral analysis","author":"Vajre","year":"2021"},{"key":"B16","first-page":"15305","article-title":"DisorBERT: a double domain adaptation model for detecting signs of mental disorders in social media","author":"Arag\u00f3n","year":"2023"},{"key":"B17","first-page":"3816","article-title":"Making Pre-trained language models better few-shot learners","author":"Gao","year":"2021"},{"key":"B18","first-page":"3045","article-title":"The power of scale for parameter-efficient prompt tuning","author":"Lester","year":"2021"},{"key":"B19","first-page":"100","article-title":"What makes good in-context examples for GPT-3?","author":"Liu","year":"2022"},{"key":"B20","doi-asserted-by":"publisher","first-page":"1430","DOI":"10.1109\/TCSS.2023.3260080","article-title":"A prompt-based topic-modeling method for depression detection on low-resource data","volume":"11","author":"Guo","year":"2023","journal-title":"IEEE Trans Comput Soc Syst"},{"key":"B21","first-page":"156","article-title":"How does in-context learning help prompt tuning?","author":"Sun","year":"2024"},{"key":"B22","doi-asserted-by":"publisher","first-page":"43","DOI":"10.1038\/s41746-020-0233-7","article-title":"Methods in predictive techniques for mental health status on social media: a critical review","volume":"3","author":"Chancellor","year":"2020","journal-title":"npj Digit Med"},{"key":"B23","first-page":"71","article-title":"Mental health discourse on reddit: self-disclosure, social support, and anonymity","author":"De Choudhury","year":"2014"},{"key":"B24","doi-asserted-by":"publisher","first-page":"e38911","DOI":"10.2196\/38911","article-title":"Text topics and treatment response in internet-delivered cognitive behavioral therapy for generalized anxiety disorder: text mining study","volume":"24","author":"Myll\u00e4ri","year":"2022","journal-title":"J Med Internet Res"},{"key":"B25","doi-asserted-by":"publisher","first-page":"875","DOI":"10.1037\/0022-006X.69.6.875","article-title":"A multidimensional meta-analysis of treatments for depression, panic, and generalized anxiety disorder: an empirical examination of the status of empirically supported therapies","volume":"69","author":"Westen","year":"2001","journal-title":"J Consult Clin Psychol"},{"key":"B26","doi-asserted-by":"publisher","first-page":"130449","DOI":"10.1109\/ACCESS.2021.3112102","article-title":"Enhanced word embedding variations for the detection of substance abuse and mental health issues on social media writings","volume":"9","author":"Ram\u00edrez-Cifuentes","year":"2021","journal-title":"IEEE Access"},{"key":"B27","doi-asserted-by":"publisher","first-page":"38","DOI":"10.1007\/s44163-023-00090-4","article-title":"Review and content analysis of textual expressions as a marker for depressive and anxiety disorders (DAD) detection using machine learning","volume":"3","author":"Sharma","year":"2023","journal-title":"Discov Artif Intell"},{"key":"B28","doi-asserted-by":"publisher","first-page":"1309","DOI":"10.32604\/cmc.2020.010172","article-title":"Review of text classification methods on deep learning","volume":"63","author":"Wu","year":"2020","journal-title":"Comput Mater Contin"},{"key":"B29","doi-asserted-by":"publisher","first-page":"1819","DOI":"10.1007\/s11831-022-09863-z","article-title":"Mental health analysis in social media posts: a survey","volume":"30","author":"Garg","year":"2023","journal-title":"Arch Comput Methods Eng"},{"key":"B30","article-title":"Attentive multi-stage learning for early risk detection of signs of anorexia and self-harm on social Media","author":"Ragheb","year":"2019"},{"key":"B31","first-page":"7685","article-title":"A time-aware transformer-based model for suicide ideation detection on social media","author":"Sawhney","year":"2020"},{"key":"B32","first-page":"5998","article-title":"Attention is all you need","author":"Vaswani","year":"2017"},{"key":"B33","first-page":"72","article-title":"Publicly available clinical BERT embeddings","author":"Alsentzer","year":"2019"},{"key":"B34","first-page":"4489","article-title":"MentaLLaMA: interpretable mental health analysis on social media with large language models","author":"Yang","year":"2024"},{"key":"B35","doi-asserted-by":"publisher","first-page":"1410947","DOI":"10.3389\/fdgth.2024.1410947","article-title":"Prompt engineering for digital mental health: a short review","volume":"6","author":"Priyadarshana","year":"2024","journal-title":"Front Digit Health"},{"key":"B36","doi-asserted-by":"publisher","first-page":"15","DOI":"10.1109\/MIS.2023.3254179","article-title":"Will affective computing emerge from foundation models and general artificial intelligence? A first evaluation of ChatGPT","volume":"38","author":"Amin","year":"2023","journal-title":"IEEE Intell Syst"},{"key":"B37","doi-asserted-by":"publisher","first-page":"4193","DOI":"10.1109\/TCSS.2025.3569400","article-title":"Dynamenta: dynamic prompt engineering and weighted transformer architecture for mental health classification using social Media data","volume":"12","author":"Kumar","year":"2025","journal-title":"IEEE Trans Comput Soc Syst"},{"key":"B38","doi-asserted-by":"publisher","first-page":"1646974","DOI":"10.3389\/fpsyt.2025.1646974","article-title":"Evaluation of large language models on mental health: from knowledge test to illness diagnosis","volume":"16","author":"Xu","year":"2025","journal-title":"Front Psychiatry"},{"key":"B39","first-page":"360","article-title":"Depression detection on malay dialects using GPT-3","author":"Hayati","year":"2022"},{"key":"B40","article-title":"Selective annotation makes language models better few-shot learners","author":"Hongjin","year":"2022"},{"key":"B41","first-page":"645","article-title":"Reasoning before responding: integrating commonsense-based causality explanation for empathetic response generation","author":"Fu","year":"2023"},{"key":"B42","doi-asserted-by":"publisher","first-page":"e49415","DOI":"10.2196\/49415","article-title":"Promoting personalized reminiscence among cognitively intact older adults through an AI-driven interactive multimodal photo album: development and usability study","volume":"7","author":"Wang","year":"2024","journal-title":"JMIR Aging"},{"key":"B43","first-page":"6797","article-title":"Improving few-shot domain transfer for named entity disambiguation with pattern exploitation","author":"Blair","year":"2022"},{"key":"B44","first-page":"18481","article-title":"Dialogue for prompting: a policy-gradient-based discrete prompt generation for few-shot learning","author":"Li","year":"2024"},{"key":"B45","doi-asserted-by":"publisher","first-page":"2151","DOI":"10.1093\/jamia\/ocae090","article-title":"The first step is the hardest: pitfalls of representing and tokenizing temporal data for large language models","volume":"31","author":"Spathis","year":"2024","journal-title":"Journal of the American Medical Infor. Assoc"},{"key":"B46","first-page":"1418","article-title":"Utilizing large language models for psychological assessment: enhancing suicide risk detection through social media analysis","author":"Xu","year":"2024"},{"key":"B47","author":"Fu","year":"2023"},{"key":"B48","first-page":"6056","article-title":"Towards interpretable mental health analysis with large language models","author":"Yang","year":"2023"},{"key":"B49","first-page":"8446","article-title":"Improving the generalizability of depression detection by leveraging clinical questionnaires","author":"Nguyen","year":"2022"},{"key":"B50","first-page":"4057","article-title":"Towards intention understanding in suicidal risk assessment with natural language processing","author":"Ji","year":"2022"},{"key":"B51","first-page":"11834","article-title":"HELP ME THINK: a simple prompting strategy for non-experts to create customized content with models","author":"Mishra","year":"2023"},{"key":"B52","first-page":"168","article-title":"AnxiBERT-Prompt engineering for social Media anxiety detection","author":"Senanayake","year":"2024"},{"key":"B53","first-page":"105","article-title":"Openprompt: an open-source framework for prompt-learning","author":"Ding","year":"2022"},{"key":"B54","first-page":"4921","article-title":"WARP: word-level adversarial ReProgramming","author":"Hambardzumyan","year":"2021"},{"key":"B55","first-page":"16261","article-title":"Worrywords: norms of anxiety association for over 44k English words","author":"Mohammad","year":"2024"},{"key":"B56","first-page":"8","article-title":"Emotion-based modeling of mental disorders on social media","author":"Guo","year":"2021"},{"key":"B57","first-page":"1485","article-title":"SMHD: a large-scale resource for exploring online language usage for multiple mental health conditions","author":"Cohan","year":"2018"},{"key":"B58","doi-asserted-by":"publisher","first-page":"11846","DOI":"10.1038\/s41598-020-68764-y","article-title":"A deep learning model for detecting mental illness from user content on social media","volume":"10","author":"Kim","year":"2020","journal-title":"Sci Rep"},{"key":"B59","first-page":"1","article-title":"kNN approach to unbalanced data distributions: a case study involving information extraction","author":"Mani","year":"2003"},{"key":"B60","first-page":"4171","article-title":"Bert: pre-training of deep bidirectional transformers for language understanding","author":"Devlin","year":"2019"},{"key":"B61","article-title":"DistilBERT, a distilled version of BERT: smaller, faster, cheaper and lighter","author":"Sanh","year":"2019"},{"key":"B62","author":"Liu","year":"2019"},{"key":"B63","article-title":"A baseline for detecting misclassified and out-of-distribution examples in neural networks","author":"Hendrycks","year":"2022"},{"key":"B64","article-title":"Adam: a method for stochastic optimization","author":"Kingma","year":"2014"},{"key":"B65","doi-asserted-by":"publisher","first-page":"1564828","DOI":"10.3389\/frai.2025.1564828","article-title":"ExDoRA: enhancing the transferability of large language models for depression detection using free-text explanations","volume":"8","author":"Priyadarshana","year":"2025","journal-title":"Front Artif Intell"},{"key":"B66","doi-asserted-by":"publisher","first-page":"1898","DOI":"10.1109\/TAFFC.2022.3145634","article-title":"Cost-sensitive boosting pruning trees for depression detection on Twitter","volume":"14","author":"Tong","year":"2022","journal-title":"IEEE transactions on Affective Computing"}],"container-title":["Frontiers in Digital Health"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fdgth.2026.1899582\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,8,14]],"date-time":"2026-08-14T05:31:51Z","timestamp":1786685511000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fdgth.2026.1899582\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,8,14]]},"references-count":66,"alternative-id":["10.3389\/fdgth.2026.1899582"],"URL":"https:\/\/doi.org\/10.3389\/fdgth.2026.1899582","relation":{},"ISSN":["2673-253X"],"issn-type":[{"value":"2673-253X","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,8,14]]},"article-number":"1899582"}}