[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"detail-sidebar-cat-1-en-105":3,"doc-seo-246544-105":53,"doc-detail-246544-en":126},{"code":4,"msg":5,"data":6},0,"success",[7,14,19,24,29,34,39,44,49],{"id":8,"doc_module":9,"doc_module_name":10,"category_name":11,"show_sort_weight":12,"slug":13},11,1,"Template","Presentations",90,"presentations",{"id":15,"doc_module":9,"doc_module_name":10,"category_name":16,"show_sort_weight":17,"slug":18},12,"Resumes",80,"resumes",{"id":20,"doc_module":9,"doc_module_name":10,"category_name":21,"show_sort_weight":22,"slug":23},14,"Invoices",70,"invoices",{"id":25,"doc_module":9,"doc_module_name":10,"category_name":26,"show_sort_weight":27,"slug":28},15,"Posters",60,"posters",{"id":30,"doc_module":9,"doc_module_name":10,"category_name":31,"show_sort_weight":32,"slug":33},16,"Social Media",50,"social-media",{"id":35,"doc_module":9,"doc_module_name":10,"category_name":36,"show_sort_weight":37,"slug":38},17,"Forms",40,"forms",{"id":40,"doc_module":9,"doc_module_name":10,"category_name":41,"show_sort_weight":42,"slug":43},18,"Letters",30,"letters",{"id":45,"doc_module":9,"doc_module_name":10,"category_name":46,"show_sort_weight":47,"slug":48},21,"Paper Templates",5,"papers-templates",{"id":50,"doc_module":9,"doc_module_name":10,"category_name":51,"show_sort_weight":4,"slug":52},158,"General","general-158",{"code":4,"msg":54,"data":55},"ok",{"site_id":56,"language":57,"slug":58,"title":59,"keywords":60,"description":61,"schema_data":62,"social_meta":119,"head_meta":121,"extra_data":123,"updated_unix":125},105,"en","multistream-modelling-for-mental-health-modelling-linguistic-and-temporal-contexts-with-mutual-and-self-excitation-in-social-media","Multistream Modelling for Mental Health: Modelling Linguistic and Temporal Contexts with Mutual and Self-Excitation in Social Media","","We present MHRoBERT (Multistream HEAT over Recurrence over BERT), a hierarchical transformer architecture for longitudinal mental health monitoring that models self- and mutual-excitation patterns in linguistic and temporal signals across multivariate event streams. To enable this, an LLM-based annotation extracts three complementary streams from social media posts: emotional states, personal life events, and mental health symptoms. Multi-task learning with these automatically generated labels yields substantial, consistent gains across evaluated architectures. LLM baselines using the stream annotations improve macro F1 by 12.6% over text-only prompting, and learned parameters provide interpretable temporal interaction patterns.",{"@graph":63,"@context":118},[64,80,101],{"@type":65,"itemListElement":66},"BreadcrumbList",[67,71,74,77],{"item":68,"name":69,"@type":70,"position":9},"https://docshare.wps.com","Home","ListItem",{"item":72,"name":10,"@type":70,"position":73},"https://docshare.wps.com/template/",2,{"item":75,"name":51,"@type":70,"position":76},"https://docshare.wps.com/template/general/",3,{"item":78,"name":59,"@type":70,"position":79},"https://docshare.wps.com/template/multistream-modelling-for-mental-health-modelling-linguistic-and-temporal-contexts-with-mutual-and-self-excitation-in-social-media/246544/",4,{"url":78,"name":59,"@type":81,"image":82,"author":87,"headline":59,"publisher":90,"fileFormat":93,"inLanguage":57,"description":61,"dateModified":94,"datePublished":95,"encodingFormat":93,"isAccessibleForFree":96,"interactionStatistic":97},"DigitalDocument",{"url":83,"@type":84,"width":85,"height":86},"https://docshare.wps.com/thumbnails/multistream-modelling-for-mental-health-modelling-linguistic-and-temporal-contexts-with-mutual-and-self-excitation-in-social-media/246544.png","ImageObject",442,249,{"name":88,"@type":89},"Elsa","Person",{"url":68,"name":91,"@type":92},"DocShare","Organization","application/pdf","2026-09-21","2026-09-12",true,{"@type":98,"interactionType":99,"userInteractionCount":73},"InteractionCounter",{"@type":100},"ViewAction",{"@type":102,"mainEntity":103},"FAQPage",[104,110,114],{"name":105,"@type":106,"acceptedAnswer":107},"What is MHRoBERT and what problem does it address?","Question",{"text":108,"@type":109},"MHRoBERT is a hierarchical transformer for longitudinal mental health monitoring. It models self- and mutual-excitation patterns across multiple linguistic and temporal event streams extracted from social media posts.","Answer",{"name":111,"@type":106,"acceptedAnswer":112},"How are the three information streams for each social media post obtained?",{"text":113,"@type":109},"The model uses an LLM-based annotation to extract three streams: emotional states, personal life events, and mental health symptoms.",{"name":115,"@type":106,"acceptedAnswer":116},"What performance benefit does multistream multi-task learning provide?",{"text":117,"@type":109},"Multi-task learning with the automatically generated stream labels produces substantial, consistent improvements across all evaluated architectures, including simpler models.","https://schema.org",{"og:url":78,"og:type":120,"og:title":59,"og:site_name":91,"og:description":61},"article",{"robots":122,"canonical":78},"index,follow",{"doc_id":124,"site_id":56},246544,1789962547,{"code":4,"msg":5,"data":127},{"doc_id":124,"user_id":128,"nickname":88,"user_avatar":129,"doc_module":9,"category_id":50,"category_name":51,"doc_title":59,"doc_description":61,"doc_content":130,"file_id":131,"file_url":132,"file_type":133,"file_size":134,"view_count":76,"is_deleted":4,"is_public":9,"is_downloadable":9,"audit_status":9,"page_count":135,"language":136,"language_code":57,"site_id":56,"html_lang":57,"table_of_contents":137,"faqs":138,"seo_title":139,"seo_description":61,"update_tm":140,"read_time":141},137455077381,"https://ap-avatar.wpscdn.com/davatar_994ba38a5ba835b3df7d355c54d3ed8d","Multistream Modelling for Mental Health: Modelling Linguistic and Temporal Contexts with Mutual and Self-Excitation in Social Media  \nAnthony Hills1 , Talia Tseriotou1 , Mahmud Elahi Akhter1 , Junyu Mao5 , Iqra Ali1 , Xenia Miscouridou3,4 , Maria Liakata1,2  \n1 Queen Mary University of London, 2The Alan Turing Institute, 3University of Cyprus,  \n4Imperial College London, 5University of Southampton  \n{a. r .hills,t.tseriotou,[m.liakata}@qmul.ac.uk](m.liakata}@qmul.ac.uk)  \nAbstract  \nWe present MHRoBERT (Multistream HEAT over Recurrence over BERT), a hierarchical transformer architecture for longitudinal mental health monitoring that models self-and mutual excitation patterns in linguistic and temporal data across multivariate event streams relating to an individual’s mental health. To supply the model with complementary perspectives on each post, we apply a Large Language Model (LLM) based annotation to extract three streams from social media posts: emotional states, personal life events, and mental health symptoms. A central finding is that multi-task learning with these automatically-generated stream labels provides substantial, consistent improvements across all model architectures evaluated. Multistream information further consistently benefits simpler models not explicitly designed to exploit it: LLM baselines incorporating stream annotations improve macro F1 by 12.6% over text-only prompting. These results have direct implications on Moments of Change detection: multistream auxiliary supervision yields consistent, substantial gains regardless of architecture, suggesting it is a simple and portable strategy that future systems can readily adopt with minimal architectural changes. MHRoBERT additionally produces interpretable learned parameters across streams, revealing temporal interaction patterns between mental health indicators.  \n1 Introduction  \nMental health conditions manifest through multiple, interconnected signals – from emotional expressions and behavioural changes to physical symptoms and life events. Despite that, traditional approaches to analysing and modeling social media for mental health monitoring examine individual behavioural signals in isolation, without modelling the cross-stream dependencies and temporal interactions between them (Bao et al., 2024) . Network theory of psychopathology posits that mental health  \nconditions emerge from systems of mutually reinforcing symptoms that interact causally over time (Borsboom and Cramer, 2013 ; Borsboom, 2017), while subsequent work demonstrates that emotions form dynamic, mutually influencing networks overtime (Bringmann et al., 2013) and that the interplay between positive and negative affect carries clinically meaningful information about mental health trajectories (Wichers et al., 2012) . On social media, prior work has largely focused on individual behavioural signals without modelling how different streams co-evolve and mutually interact over time (Yazdavar et al., 2020 ; Garg, 2023), leaving important cross-stream temporal dynamics unaddressed.  \nIn this work, we leverage the stream taxonomy inspired by Mao et al. (2025) to obtain three complementary perspectives on each post – emotional states, personal life events, and mental health symptoms – via LLM-based annotation, and investigate how these streams can be integrated into temporal modelling architectures for Moments of Change (MoC) in mood detection (Tsakalidis et al., 2022b,a; Tseriotou et al., 2025) . A widely applicable finding is that multi-task learning with these automatically-generated stream labels provides substantial, consistent performance improvements across all evaluated model architectures. Multistream information widely benefits diverse architectures, with LLM baselines showing substantial improvements when provided with stream annotations. These results suggest that incorporating multistream auxiliary supervision is a broadly applicable strategy for improving MoC detection. Buildin","cbCaibLozUz9IRG5","https://ap.wps.com/l/cbCaibLozUz9IRG5","pdf",593800,20,"English","# Abstract\n# 1 Introduction\n# 2 Related Work","[{\"question\":\"What is MHRoBERT and what problem does it address?\",\"answer\":\"MHRoBERT is a hierarchical transformer for longitudinal mental health monitoring. It models self- and mutual-excitation patterns across multiple linguistic and temporal event streams extracted from social media posts.\"},{\"question\":\"How are the three information streams for each social media post obtained?\",\"answer\":\"The model uses an LLM-based annotation to extract three streams: emotional states, personal life events, and mental health symptoms.\"},{\"question\":\"What performance benefit does multistream multi-task learning provide?\",\"answer\":\"Multi-task learning with the automatically generated stream labels produces substantial, consistent improvements across all evaluated architectures, including simpler models.\"}]","Multistream Modelling for Mental Health: Modelling Linguistic and Temporal Contexts with Mutual and Self-Excitation in Social Media | PDF",1789236365,7]