Digital Representation of Emotions through Social Media: A Systematic Review on Tracking Emotions

Digital Representation of Emotions through Social Media: A Systematic Review on Tracking Emotions | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Systematic Review Digital Representation of Emotions through Social Media: A Systematic Review on Tracking Emotions Sarath CJ, Aparna Pandey This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-9047352/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 5 You are reading this latest preprint version Abstract The present systematic review aims to synthesise the existing evidence on tracking the digital representation of emotions through social media platforms, thereby providing scope for identifying users' emotional vulnerability and emotional psychopathology. Emotions are central to human experience, and their expression has undergone a major shift with the rise of digital media platforms. Social media currently serve as a primary medium of communication and emotional exchange. The PRISMA, 2020 guidelines were followed in this systematic review. The review was conducted on four databases, including Scopus, Web of Science, PubMed, and APA Psycnet. The final list consists of 28 articles, specifically focusing on emotional representation through social media, published in English and selected based on the inclusion and exclusion criteria. The findings show an acceleration in interdisciplinary research on emotional representation and digital media, with 90% of studies published after 2022. The most widely researched social media platform from the reviewed research articles was Twitter, with predominantly textual analysis using linguistic and semantic markers. A strong association was found from the reviewed articles on the recognition of the emotional patterns and the early detection of mood disorders and suicidal ideation. The review indicates the function of social media as an emotional expression outlet and a diagnostic mirror of the users' affective and psychological processes. emotional representation emotional distress social media digital representation Figures Figure 1 Figure 2 Figure 3 Figure 4 Introduction Emotions were found to be central to human lived experiences. The explanations given to emotions from the early researchers were specifically in association with the physiological reactions or arousal. James-Lange, in the late 19th century, proposed and stated that emotions follow the physiological responses (Lang, 1994 ). Furthermore, Cannon in the early 20th century, stated the limitations of the James-Lange theory and suggested that emotions and physiological response pathways proceed in parallel without one following the other (Dror, 2013 ). Schatcher and Singer introduced the dimension of cognitive labelling, leading to specific emotions (Cherry, 2025 ). However, the recent advancement in the study and concept of emotions has given the newest and most diversified dimensions and explanations to emotions. The Modular or Discrete category approach and Dimensional approach of emotions are the two broad paradigms that dominate the theoretical landscape of emotions. The Modular approach to emotions views emotions as distinct modules or specific categories, focusing on emotions as states that take into account their triggers and features, such as anger, happiness, fear, and others. On the other hand, the dimensional approach considers emotions as varying along a continuous axis, such as valence (positive vs negative), arousal (high vs low), predominantly treating emotion as a process (Harmon-Jones et al., 2017 ). However, this paper adopts a more integrative and functional definition of emotion proposed by APA, which explains emotions as a complex reaction pattern that includes experiential, behavioural, and physiological components through which one tries to deal with a significant triggering event of life. The definitions of emotion provided by each theoretical perspective predominantly explain the internal manifestation and formation of emotion; likewise, it is essential to understand and identify the patterns and ways through which emotions are represented externally. Emotion representation could be defined as the expression, communication, and sharing of emotions externally, through verbal or nonverbal means. The verbal emotional communication includes spoken or written forms of emotional representation, and the non-verbal emotional representation includes facial expression, body language, voice tone, and artistic and symbolic representation. With the emergence and advancement of digital communication, emotional representation also shifted towards digital forms, such as text, images, symbols, and so on. In this advanced digital era, social media acts as a primary medium for emotional representation (Kapoor et al., 2018 ). Social media could be defined as a digital communication platform that facilitates interaction, content dissemination, and relationship-building among individuals and groups. The social media platforms support the exchange of information, emotions, and social meanings through various modes such as texts, images, emojis, hashtags, audio and video. It enables continuous social interaction across personal, professional, and socio-cultural contexts (Kapoor et al., 2018 ). Social media is found to be a central mode of communication in the present generation, ranging from text messages to face-to-face through video interaction (Kaplan & Haenlein, 2010 ). It provides an open platform for users to build connections, express themselves, and engage in emotional exchange through multiple modalities. Self-presentation and self-disclosure are the mechanisms that facilitate social media's functioning at the heart of social media, providing opportunities for users to curate, share, and reveal personal information online (Kaplan & Haenlein, 2010 ). The social media platform features enable and facilitate the subtle ways of emotional actions and emotional expressions to a large extent (Steinert & Dennis, 2022 ). Currently, emotional representation has become a part of social media culture. They, along with the mitigation of private emotions, also facilitate social interaction. We generally prefer to broadcast or present positive emotional expression in public, and negative emotions, on the other hand, are mostly presented ironically, satirically or through an indirect mode, providing a collective outlet (Steinert & Dennis, 2022 ). Being a central and significant part in human life, emotions were identified to be foundational. Hence, it is important to understand how emotions are represented in the recent advanced digital era. Understanding emotions as an internal process of the emotional representation is significant in developing knowledge of emotions from a psychological perspective with a communication lens, which is further expanded with the rise of digital media. Digital emotional representation is a significant new layer of recognising and tracking the patterns of emotions and their variations. Research indicates the crucial role of social media platform design in shaping emotional expression and interactions (Steinert & Dennis, 2022 ). The study further introduces emotional affordances, which refer to the features in social media as that facilitate specific emotional expressions, such as like and dislike buttons, reaction emojis, comments tab, share function, and automated suggestions (Steinert & Dennis, 2022 ). The notion of emotional architecture facilitated by social media platforms suggests that social media platforms are not merely neutral conduits for communication or emotional expression, but an active environment that shapes the emotional flow of users (Jorgenson, 2019). Digital media enables in the transformation of emotional experiences and emotional expression patterns through hashtags, viral campaigns, and symbolic outrage, through which emotions can rapidly circulate and inspire coordinated actions (Manikonda et al., 2018 ). The emotional architecture of social media facilitates both the expression and amplification of emotions. The emotional affordances and architecture introduced through the digital media creates the emotional condition in addition to their mediation role in interaction and emotional expression. Understanding the dual nature of social media by decoding the patterns of emotional expression and amplification among users could provide a wider scope for human-computer interaction through the emotional spectrum. Rationale The focus of the recent research was found increasingly in the area of emotional expression through social media; even then, the field remains widely fragmented. Studies in this area have limited scope with confined integration of theoretical framework. Further, the methodological rigour remains inadequate (Schreiner et al., 2019 ). Hence, the present systematic review aims to provide a structured and transparent synthesis of existing research papers on emotional representation through social media contexts, to address the gaps mentioned above. The two key questions covered in this systematic review are; How do individuals represent different states and processes of emotions through varying modes of social media communication forms? What are the characteristics and platform-specific variations based on specific emotional representation? The current study seeks to deepen the understanding of the psychosocial dynamics of emotional representation in digital/online environments. It further provides direction for interdisciplinary research, thereby guiding digital well-being interventions and support systems for the design of emotionally supportive social media platforms and its healthy usage. Methodology A systematic review design was followed as per the PRISMA guidelines (2020) for synthesising and reviewing research articles. It aimed at finding evidence from authentic research articles in exploring the psychosocial dynamics of emotional representation through social media platforms. Four databases, including Scopus, PubMed, Web of Science, and APA PsycNET, were reviewed in this systematic review. Inclusion & Exclusion criteria Peer-reviewed research articles published in the English language were considered in this systematic review, which investigated emotional representation from posts or comments shared through social media platforms. Eligible studies examining emotional expression or perception across diverse socio-demographic groups ranging from adolescence to adulthood, considering all forms of users, involving active, passive or observational users. Adolescence is a critical period, characterised by identity formation, emotion regulation development and heightened peer influence. Social media plays a major role in emotion representation and social comparison of this group (Steinberg, 2014 ). The adults group generally demonstrates a stable and mature emotion regulation and variations in their motivations for social media usage, which also helps in making an understanding across developmental stages. The quantitative, qualitative, or computational studies, were considered to capture a comprehensive range of research evidence on the same. Case studies, secondary literature, or studies focusing on offline media were excluded. Literature published in languages other than English were also excluded. Search Strategy A strategic search was conducted across four databases (SCOPUS, APA PsycNET, PubMed, and Web of Science) from October 2024 to November 2024 with keywords such as "social media", "emotional representation", "emotional states", and "emotion". Boolean operators and database-specific filters were utilised to refine search results by date (last 10 years) and language (English). The selection process comprised independent screening by reviewers based on title and abstract, followed by a full-text review for eligibility. The final selection of studies was made based on their relevance to the representation of emotions through social media, judged by independent reviewers. Any disagreements were resolved by reaching a consensus. Data Extraction Data collection involved the extraction of the following information from eligible studies by reviewers: general information (title, authors, year of publication), study characteristics (research design, social media context (platform(s), type of data - text, images, videos), type of emotions, representation characteristics (linguistic, symbolic, semantic features), and findings (main outcomes, implications, and research gaps). The filtration and extraction process lasted from November 2024 to June 2025. A narrative synthesis was performed to categorise findings based on the type of emotions, SMPs, representation style (e.g., textual vs. visual) and outcomes. Results The systematic review followed the PRISMA 2020 guidelines across four databases, including Scopus, PubMed, Web of Science, and APA PsycNET. The initial search resulted in a total of 421 articles. After the removal of 19 duplicates and 7 non-research articles from the list, a total of 395 research articles were screened for title and abstract. A total of 325 articles were removed after title and abstract filtration, and 70 articles were sorted for full-text retrieval, with 66 articles having full text accessed for eligibility. The primary reason for the exclusion of articles after reviewing the full text is due to the lack of clear focus on emotional representation (n = 34), and lack of clear distinct eligibility (n = 4). The final list consists of 28 articles which met all the inclusion and exclusion criteria. The detailed flow chart in the study selection is given in Fig. 1.1 , the PRISMA flow chart. The characteristics and summary of the 28 included studies are given in Table 1.1 . Studies were spread across different nations, but a major focus was found with Asian countries (from China (31%), in collaboration with India (17%) & Saudi Arabia (10%), and the USA (24%)) as mentioned in Table 1.1 . The chronological distribution of studies spikes after 2022, which accounts for 90% of studies as given in Table 1.1 . The most frequently studied social media, according to these studies, was Twitter (25%), followed by Reddit (10%). The majority of the studies focused on analysing textual data through linguistic and semantic analysis (95%). Some of the studies used emojis and symbolic analysis (5%). The majority of participants across the 28 reviewed studies fall into adulthood, belonging to 19 to 64 years of age group (84.3%), and adolescence 13 to 18 years of age group (11.2%), and the remaining in late adulthood. The review and synthesis of research articles indicates that a higher range of emotional representations were expressed and segregated through the verbal mode, more specifically through textual contents which were determined predominantly using linguistic (41%), semantic (39%) and lexical (19%) markers given in Fig. 1.3 . Linguistic markers were frequently used in emotional words and phrases, narrative self-disclosure, pronoun usage, informal language and slang. These linguistic patterns potentially represent immediate emotional responses, sustained emotional distress, and so on. The semantic markers were found to capture the underlying emotional meaning and intent beyond surface-level words and meaning. It indicates emotional states and thematic meaning based on subtle emotions such as anxiety, uncertainty or vulnerability could be inferred, and the emotional shifts over time were also being interpreted. This supports the contextual determination of emotions. Lexical markers used in emotional segregation are based on word frequency, emotional polarity, and affective lexicons. Even though the lexical markers were less likely to be used, but it remains essential in emotional valence and polarity with strength of emotion determination. The representation of emotions through social media was observed with the combination of explicit linguistic expression and implicit semantic meaning, supported by the lexical indicators. The specific emotional representation and characteristics were found to highly vary based on the social media platform on communication affordances, user roles, and content domains of the major themes identified and shown in Fig. 1.4 . The microblogging platforms such as Twitter, facilitate short, immediate emotional expression and high emotional intensity. It covers the themes, including public health, misinformation and breaking news. The influencers and celebrity-driven platforms such as Instagram indicate emotional amplification for social reach, users mirror disclosed emotions, and high emotional engagements. The community or support-oriented social media platforms show a more elaborate and reflective emotional representation, encouraging empathy and understanding facilitation. Support oriented social media features a more reflective emotional representation, encouraging empathy and facilitating understanding among users. In the support-based community online forums, highly arousing emotional expressions are less likely to dominate visibility, as they focus mainly on encouraging empathetic validation and coping. (Berger, & Milkman, 2012 ; Barak et al., 2008 ). Table 1.1 Study characteristics Sl. No Title Year Country Author(s) Methodology Major findings 1 Happiness, stress, a bit of vulgarity, and lots of discursive conversation: A pilot study examining nursing students' tweets about nursing education posted to Twitter 2015 Canada Richard G. Booth Tweets about nursing courses, classes and clinical work were analysed using thematic analysis with five thematic clusters Informal social media expression and need for digital professionalism among nursing students 2 Language-agnostic deep learning framework for automatic monitoring of population-level mental health from social networks 2022 Thailand Thanapon Noraset, Krittin Chatrinan, Tanisa Tawichsri, Tipajin Thaipisutikul, Suppawong Tuarob Proposes a language-agnostic deep learning framework as a feasible alternative to traditional methods for gathering mental health data across diverse populations. Effective low-resource language framework for social media–based mental health monitoring 3 Deep parallel contextual analysis framework-based emotion prediction in community wellness communications on social media 2024 China Feng Liu, Kun Hou, Yang Dong Introduces the DPCAF framework, which uses dual word embedding techniques to better capture semantic information in short texts for wellness communications. Deep contextual framework improving emotion prediction in short social media texts 4 A Deep Learning Framework for News Readers’ Emotio Prediction Based on Features From News Articles and Pseudo Comments 2023 China Xu Mou, Qinke Peng, Zhao Sun, Ying Wang, Xintong Li, Muhammad Fiaz Bashir Proposes a block emotion attention network (BEAN) that merges article content with "pseudo-comments" to predict reader emotions, even when actual reader feedback is missing. A hybrid framework enhancing emotion prediction in news articles through comment integration 5 Differing Content and Language Based on Poster-Patient Relationships on the Chinese Social Media Platform Weibo: Text Classification, Sentiment Analysis, and Topic Modelling of Posts on Breast Cancer 2024 Japan & New Zealand Zhouqing Zhang, Kongmeng Liew, Roeline Kuijer, Wan Jou She, Shuntaro Yada, Shoko Wakamiya, Eiji Aramaki Weibo posts on breast cancer fine-tuned through two step with Chinese BERT classification to identify poster- patient relationships determined using sentiment analysis using linguistic inquiry and word count Influence of poster–patient relationship on emotional expression in health-related social media posts 6 Emotion Topology: Extracting Fundamental Components of Emotions from Text Using Word Embeddings 2024 Poland Hubert Plisiecki, Adam Sobieszek Go Emotions dataset of Reddit of communities with a minimum 10k comments with emotion taxonomy developed through manual annotations. Word embedding approach for mapping valence–arousal dimensions in textual emotion analysis 7 Increased Online Aggression During COVID-19 Lockdowns: Two-Stage Study of Deep Text Mining and Difference-in-Differences Analysis 2022 Taiwan Jerome Tze-Hou Hsu, Richard Tzong-Han Tsa A two-stage study combining Deep Text Mining (training a BERT model to detect anger, offensive language, and hate speech) with a Difference-in-Differences (DID) econometric analysis to establish causal links between lockdowns and aggression. Rise in aggressive social behaviours during COVID-19 lockdown and its policy implications 8 Emotional Expression on Social Media Support Forums for Substance Cessation: Observational Study of Text-Based Reddit Posts 2023 USA Genevieve Yang, Sarah G. King, Hung-Mo Lin, Rita Z. Goldstein Reddit posts across 394 forums, quantified and categorised emotion word frequencies in substance cessation of alcohol, nicotin, and cannabis. Online communities as valuable sources for understanding emotional recovery experiences 9 Exploring Public Emotions on Obesity During the COVID-19 Pandemic Using Sentiment Analysis and Topic Modelling:Cross-Sectional Study 2024 Switzerland, UK, Pakistan Jorge César Correia, Sarmad Shaharyar Ahmad, Ahmed Waqas, Hafsa Meraj, Zoltan Pataky Analyzes 53,414 tweets using the XLM-RoBERTa-base model for sentiment analysis and the BERTopic library for topic modelling. Negative public sentiment toward obesity highlighting need for informed health communication and policy 10 Guide for the Application of the Data Augmentation Approach on Sets of Texts in Spanish for Sentiment and Emotion Analysis 2023 Chile Rodrigo Gutiérrez Benítez, Alejandra Segura Navarrete, Christian Vidal-Castro, Claudia Martínez-Araneda Evaluates Machine Learning and Deep Learning performance using data augmentation (DA) techniques like Easy Data Augmentation (EDA), back-translation (BT), and SentiGAN to expand small Spanish-language datasets. Data augmentation techniques enhancing sentiment and emotion analysis in Spanish social media texts 11 EMFSA: Emoji-Based Multifeature Fusion Sentiment Analysis 2024 China Hongmei Tang, Wenzhong Tang, Dixiongxiao Zhu, Shuai Wang, Yanyang Wang, Lihong Wang Proposes a multifeature fusion model (EMFSA) that integrates emoji, topic, and text features. It uses a sentiment- and emoji-masked language model (Senti_MLM) and a cross-attention mechanism to improve accuracy in short social media texts. Multifeature fusion model leveraging emojis to improve sentiment analysis in short social media texts 12 Emotion-Based Reinforcement Attention Network for Depression Detection on Social Media: Algorithm Development and Validation 2022 China Bin Cui, Jian Wang, Hongfei Lin, Yijia Zhang, Liang Yang, Bo Xu Proposes an Emotion-Based Reinforcement Attention Network (ERAN) that uses a TextCNN for deep emotional feature extraction and Reinforcement Learning (RL) to select specific "depression indicator" posts from a user's history. Deep learning model enhancing depression detection through emotional semantic feature extraction on social media 13 An Optimised Deep Learning Approach for Suicide Detection through Arabic Tweets 2022 Saudi Arabia, & Egypt Nadiah A. Baghdadi, Amer Malki, Hossam Magdy Balaha, Yousry AbdulAzeem, Mahmoud Badawy, Mostafa Elhosseini Proposes a framework for binary classification (Normal vs. Suicide) of Arabic tweets using Bidirectional Encoder Representations from Transformers (BERT) and Universal Sentence Encoder (USE) models after rigorous Arabic text preprocessing (lemmatization and stemming). Social media as a tool for tracking depression and mental health trend 14 Computational Linguistics-Based Text Emotion Analysis Using Enhanced Beetle Antenna Search with Deep Learning During the COVID-19 Pandemic 2023 Saudi Arabia, & India Youseef Alotaibi, Arun Mozhi Selvi Sundarapandi, Subhashini P, Surendran Rajendran Computational linguistics based mood analysis using enhanced beetle antenna search with deep learning (CLSA- EBASSDL using BERT word embedding classifier using attention based BiLSTM network. Application of computational intelligence and nature-inspired algorithms for real-world optimisation problems 15 Machine Learning and Natural Language Processing to Assess the Emotional Impact of Influencers’ Mental Health Content on Instagram 2024 Spain Noemi Merayo, Alba Ayuso-Lanchares, Clara González-Sanguino Created a labelled dataset of influencer’s Instagram responses based on mental health posts categorized by emotion laden and applied machine learning algorithms Emotional impact of mental health disclosures on social media beyond psychopathology detection 16 A hybrid deep learning approach for depression prediction from user tweets using feature-rich CNN and bi-directional LSTM 2022 India Harnain Kour, Manoj K. Gupta Development of a hybrid deep learning model using a feature-rich Convolutional Neural Network (CNN) and a Bi-directional Long Short-Term Memory (Bi-LSTM) network to classify depressive versus non-depressive textual narratives from Twitter data. Predicting users’ mental health conditions through Twitter data analysis 17 Research on the detection model of mental illness of online forum users based on a convolutional network 2023 China, & India Yuliang Guo, Zheng Zhang, Xuejun Xu Hierarchical user post feature representation models including Single- Gated LeakReLU- CNN and Multi- Gated LeakReLU- CNN to extract emotional features from user posts and to identify mental illness by analysing online forums. Improved models for accurate extraction of key emotional features from social media posts 18 Emotion recognition for human–computer interaction using high-level descriptors 2024 India, & Ethiopia Chaitanya Singla, Sukhdev Singh, Preeti Sharma, Nitin Mittal, Fikreselam Gared Construction and preprocessing of a labelled speech corpus from diverse social media sources, followed by the application of Deep Learning techniques, specifically Convolutional Neural Networks (CNN), for Speech Emotion Recognition (SER). Enhanced emotion recognition accuracy in Punjabi speech using an advanced SER approach 19 Meaningful messaging: Sentiment in elite social media communication with the public on the COVID-19 pandemic 2021 USA Janet M. Box-Steffensmeier, Laura Moses Analysis of the tone and sentiment in elite social media messaging to determine its effect on information spread and public reaction during the pandemic. Influence of partisanship and emotional tone on crisis communication and message engagement 20 Detecting Substance Use Disorder Using Social Media Data and the Dark Web: Time- and Knowledge-Aware Study 2024 USA, India, & Thailand Usha Lokala, Orchid Chetia Phukan, Triyasha Ghosh Dastidar, Francois Lamy, Raminta Daniulaityte, Amit Sheth Utilization of state-of-the-art deep learning models to generate sentiment and emotion from social media posts and crypto market listings to understand users' perceptions and the relationship between substance misuse and mental health. Detection of Substance Use Disorder through social media and Dark Web data analysis 21 User-based Hierarchical Network of Sina Weibo Emotion Analysis 2023 China Qian Chen, Xiao Sun, Jiamin Wang, Meng Wang User based hierarchical network with combination of multi head attention and convolutional neural network to jointly analyse individual Weibo texts and related posts ober time, capturing contextual emotional information for improved five category emotion classification. User-level contextual information improves accuracy in emotional classification, 22 Twitter Perspectives on Cochlear Implantation: Sentiment and Thematic Analysis 2023 USA Joel S. Feier; Kenny Nguyen; Janet S. Choi All English language tweets mentioning ‘Cochlear implantation’, from 2019–2021 collected using custom Python script, analysed for sentiment with VADER tool, and based on positive, negative, liked tweets using thematic analysis. Understanding public positive and negative perceptions of cochlear implantation 23 “Twitter is Really Therapeutic at Times”: Examination of Black Men’s Twitter Conversations Following Hip-Hop Artist Kid Cudi’s Depression Disclosure 2021 USA Diane B. Francis The study was conducted in the United States and used thematic analysis to examine a sample of 1,482 tweets from the hashtag #YouGoodMan to identify recurring patterns in Black men’s mental health conversations following Kid Cudi's depression disclosure. Celebrity disclosure facilitates emotional expression online, and depression disclosure through an influencer 24 Topics and Sentiments of Public Concerns Regarding COVID-19 Vaccines: Social Media Trend Analysis 2021 USA Monselise et al. Tweets related to COVID-19 vaccines were collected over 60 days and analysed using topic modelling and sentiment analysis to identify key topics and emotions. Fear as a leading reaction, news events dictate spikes 25 Social media’s dark secrets: A propagation, lexical and psycholinguistic oriented deep learning approach for fake news proliferation 2024 China, & Saudi Arabia Kanwal Ahmed; Muhammad Asghar Khan; Ijazul Haq; Alanoud Al Mazroa; Syam M.S.; Nisreen Innab; Masoud Alajmi; Hend Khalid Alkahtani A deep learning framework using Graph Convolutional Neural Networks (GCNN) with multi-head attention, incorporating BERT embeddings and psycholinguistic attributes (sentiment, personality, etc.) across user interaction and semantic propagation graphs. 1. Disinformers are generally newer users, tweet more frequently, and are more active at night (higher insomnia index). 2. Users with lower popularity (low F2F ratio) are more likely to spread fake news. 26 Sentiment, we-talk and engagement on social media: insights from Twitter data mining on the US presidential elections 2020 2022 Germany Linus Hagemann and Olga Abramova A data mining approach using a dataset of over three million tweets from the 2020 US presidential elections, applying dual process theory to test how affective cues (sentiment) and cognitive cues (insight, causation) impact audience engagement. Emotional and cognitive cues influence social media engagement, and negative bias attracting users 27 Sentiment Classification of Anxiety-Related Texts in Social Media via Fusing Linguistic and Semantic Features 2024 China Jianghong Zhu, Zhenwen Zhang, Zhihua Guo, and Zepeng Li Fuses linguistic and semantic features by using the SC-LIWC dictionary for linguistic extraction and a TextCNN-based model for deep semantic analysis. It uses a novel recursive feature selection algorithm on Sina Weibo data. Anxiety-related posts are a combination of negative and meaningful highly supporting the incorporation of the fusion model. 28 Revealing the spatial co-occurrence patterns of multi-emotions from social media data 2023 China & USA Dongyang Wang a, Yandong Wang a,*, Xiaokang Fu a,b, Mingxuan Dou a, Shihai Dong a, Duocai Zhang c Used BERT based model to classify social media posts into six emotions, applied K- means clustering to analyse special emotion co-occurrence patterns, and employed a Multi Model Logit Model to examine their relationship with land use characteristics. Significant special co-occurrence patterns of conflicting or consistent emotions in urban space on social media Themes Emotional Vulnerability Detection The theme of Emotional vulnerability covers 36% of the identified themes from reviewed papers. The studies indicate the detection in the segments of emotional vulnerability, including sadness, fear, aggression and other negative emotions. Sentiment analysis, emotion classification, and lexicon-based or deep-learning models were used to segment social media texts. It helps in the early identification of emotional distress and its vulnerability on the development of mood disorders. It further indicates the understanding of platform-specific emotional risk detection and the necessary action. Emotional Support Sequences The theme of Emotional support sequences covers 23% of identified themes from reviewed papers, with the dimensions of positive emotions, including happiness, joy, satisfaction and other positive emotions. Emotion recognition models were used in analysing and identifying these segments. It helps in understanding well-being, resilience, and social support patterns. The digital well-being initiatives and potentiality were discussed with positive social media engagement strategies. The positive emotions were identified using polarity-based sentiment models and multi-class emotion classifiers. The algorithm and thematic classification based on the emotional support theme help in identifying segments in fostering wellbeing and protective factors against vulnerability. Emotion Valence Measurement The theme of Emotion valence measurement covers 16% of the identified themes from the reviewed papers. Research depicting the classification of emotions based on the range of positivity or negativity using emotional polarity categorization or sentiment classification using supervised machine learning models. It helps in the simplification of large-scale emotional representation patterns and prevalence through social media comparing emotional polarity across time, topic, or population. Technological Application The theme of Technological application covers a total of 9% of identified themes from research papers. The specific ML algorithms learning models of LSTM, CNN, Transformer-based, and NLP pipelines were predominantly trained and used in the emotion recognition of social media textual contents. It improves the accuracy and scalability of emotion detection. It further enables in the real time emotion monitoring on social media. The ethical usage of AI has been highlighted based on the training and model performance. Mental Health Monitoring The theme of Mental health monitoring covered a total of 7% from identified themes from reviewed papers. It indicates the monitoring of mental health through emotion and sentiment patterns for mental health using computational models. Computational models have been used in emotion detection and sentiment analysis. This do not just assess individual’s emotion representation, but is also helpful in monitoring mental health at population level. It promotes interdisciplinary researches that the incorporate computer science with psychology for the early diagnosis of emotional dysfunctions, including depression, and anxiety using predictive computational models. Emotional Communication Strategies The theme of emotional communication strategy covered a total of 5% of identified themes from reviewed papers. It indicates analysis of language, slang and interaction patterns to understand emotional expression. It shows verbal emotional expressions communicated differently across different platforms. It further helps in improving the human-computer interaction and development of empathetic systems, and designing better online communication tools. The linguistic structures, discourse patterns and interaction patterns features indicate examination of differences in emotional expression across platforms and cultures. Stress Detection The theme of stress detection covered a total of 4% of themes from the identified reviewed papers. It indicates the analysis of stress through linguistic cues, sentiment shifts, and temporal emotion patterns. The studies identify the stress triggers related to work, crisis, or social events. It further indicates the potential of the early warning systems during emergencies or pandemics. Stress is detected with related keywords, sentiment shifts, and temporal emotion changes using classification ML models trained on stress labelled dataset. The common procedures inferred from the research papers based on machine learning and supervised models are represented in a structured and systematic order as follows. The technique begins with the data acquisition through APIs and web scraping tools such as Scrapy. A comprehensive data pre-processing is followed, which includes data cleaning, artefact removal and normalisation. Data augmentation techniques such as random word swapping, synonym replacement, and back translation are applied. The textual data is then translated to numerical representations using feature extraction methods ranging from traditional bag of words to advanced word2vec, Glove, and BERT. The emotion classification or sentiment range was performed using a deep learning framework, including CNN and Bi-LSTM architectures. In addition, it incorporates topic modelling using LDA to uncover latent topicshemes, causal and statistical analysis through Difference in Difference or emotional topology. The performance is further evaluated using standard metrics of precision, recall, and F1 score to ensure accuracy and reliability. Discussion The present study aims to synthesise the rapidly growing body of research examining how emotions are digitally represented, detected, and interpreted in social media by youth population. The systematic review indicates that the interdisciplinary dimension of social media and emotion is widely accelerating, as shown by the growing number of research in recent years, as shown in Fig. 1.2 . The study specifically focused on identifying how individuals express diverse emotional states and process across different social media modalities and further analysing the platform-specific variations in emotional representation. The systematic review process resulted in a final list of 28 peer-reviewed research articles. The research shows the convergence of psychology, computer science, communication studies, and data science with natural language processing by leveraging the analysis of linguistics and semantics more sequentially. The models, to show high performance in areas including predicting mental health conditions such as mood disorders and for recognising emotional states more specifically. The review further shows evidence for the methodological advancement moving beyond textual data to visual or auditory multi lingual interpretation. This multicultural expansion is significant in the development of a robust detection system. A consistent pattern was detected from the 28 reviewed research articles showing the development of the digital environment into a significant ecosystem, in which emotions are expressed and interpreted. The findings indicate that the ML-based approaches of CNN, Bi LSTM, and Reinforcement attention mechanism demonstrate significant capacities in decoding patterns of emotional representation through texts, visual contents or symbols. These techniques further expand the reach and expansion of studies of emotional representations through digital media (Liu et al., 2024 ). The research evidence shows that emotional recognition is not limited to linguistic or semantic markers, but extends to semantic, syntactic and symbolic cues, such as emojis, hashtags, reaction symbols (Tang et al., 2023 ). Digital communication channels serve as a significant source for the multi-model structure of emotional representation (Steinert & Dennis, 2022 ). Social medias were found not just a platform to express emotions, but research shows evidence of them playing an active role in shaping the pattern of emotions based on media culture The text-based emotion analysis was found to be predominant and widely researched as the majority of research articles from the review uses textual content. Linguistic and semantic markers were predominantly used for analysing the text for detecting emotional patterns (Pliseki & Sobieszek, 2024 ). The emerging recent models adopt a hybrid and multilingual approach that expands beyond the English language, addressing cultural and linguistic variation in emotional expression (Gutierrez et al., 2024). A strong intersection of emotional representation with mental health was also identified from this review. Emotional language, sentiment polarity and emotional valence were found to reliably indicate the symptoms of mood disorders such as depression or anxiety (Kour & Gupta, 2022 ). The ethical usage of this user-generated social media data can act as a mirror of the user’s emotional process and state. Even then, the technical usage remains a challenge, especially regarding informed consent and user privacy (Steinert & Dennis, 2022 ). The thematic analysis of the 28 finally selected articles shows three dominant trends in emotional representation research. The first is regarding the contingent effect or emotional amplification effect is prevalent more specifically during social crisis situations, such as the COVID-19 pandemic, where solidarity became widespread (Hsu, Tsa, 2022 ). The second trend depicts emotional personalisation driven by the user's algorithmic polarisation reinforces the user's unique affective contents by influencing public discourse and promoting wellbeing (Box- Steffensmeier & Moses, 2021). The third trend is regarding the interdisciplinary researches with the combination of computer science, psychology and language to interpret the complex interaction of emotions through the digital ecosystem (Manikanda et al., 2018). The current literature shows a high skewness towards text-based analysis while non-verbal indicators of emotions, such as emojis, symbols, or pictures, are less explored. The real-time emotional cues were found to be relatively less explored, which limits the dynamic affective shifts in digital interactions. Bridging this gap would enhance the clinical utility and interpretability. Current systematic review synthesises the methodological sequences and theoretical insights based on emotions represented through the social media ecosystem. A strong convergence was found towards data-driven, text-centric approaches. The incorporation in the advancement in machine learning and deep learning. Reviewed studies indicated the beginning of large-scale data acquisition through APIs and web scraping tools such as Scrapy. Data acquisition at the preliminary stage in collecting data from social media platforms through Application Programming Interface (API), that allows structured access to social media platforms and using Web Scrapping tools, extracts data from public webpages where APIs are limited. This is further followed by data pre-processing, including noise removal, normalisation, tokenisation, and lemmatisation, to improve model robustness (Manikonda et al., 2018 ; Singla et al., 2024 ). The third step of data augmentation was facilitated using techniques such as word swapping or back translation, which were highly adopted for data sparsity. Social media raw data cannot be processed by machine learning models, hence, the traditional Bag- of- Words or advanced Word2Vec, Glove or BERT, converts words into densevector representations capturing semantic meaning, contextual similarity or Emotional nuance. (Kour & Gupta, 2022 ). The deep learning frameworks were predominantly used for the classification of emotions and sentiment of the data using CNNs and LSTMs, often in hybrid or parallel sequences (Pliseki & Sobieszek, 2024 ). Topic modelling techniques of Latent Dirichlet Allocation were used to synthesise the thematic structures within emotional discourse (Correia et al., 2024 ). The interpretability is further enhanced using approaches including difference in differences approaches. Further, the performance of the model was evaluated using standard matrices based on accuracy, precision, recall and F1 score (Baghdadi et al., 2022 ). The findings indicates that social media acts as both an emotional outlet and a diagnostic mirror as an emotional ecosystem. Social media was found to act as an emotional outlet and a diagnostic mirror, as verified from this review. Social media platforms provide and facilitate affective expression, and it simultaneously expose users to heightened emotional distress. Hence, the integration of social media platforms for emotional recognition and health must balance between empathetic responsiveness and ethical responsibility. Future research could focus more into building cross-platform emotional models capable of recognising and tracking emotional injuries. Conclusion This systematic review shows evidence for the transformative potential of social media platforms in decoding emotional patterns and expressions through digital media. The findings show the potential of the human-centred computational algorithms that not only recognise emotions, but also modify based on the emotional pattern and preference of the users. It can further help in the early recognition of mental health and emotional conditions and to improve digital empathy with the development of emotionally intelligent algorithms. By maintaining the technical advancement and ethical boundaries, the digital emotional space not only limits in providing a space for emotional expression, but to a compassionate and empathetic technological understanding. Declarations Author Contribution Sarath CJ wrote the first draft of the manuscript. Both authors reviewed and edited the manuscript. References Alotaibi Y, Sundarapandi AMS, Subhashini P, Rajendran S. Computational linguistics-based text emotion analysis using enhanced beetle antenna search with deep learning during COVID-19 pandemic. PeerJ Comput Sci. 2023. https://doi.org/10.7717/peerj-cs.1714 . 9, Article e1714. APA (American Psychological Association). (n.d.). Emotion. In APA Dictionary of Psychology. Retrieved October 21. 2025, from https://dictionary.apa.org/emotion Baghdadi NA, Malki A, Balaha HM, AbdulAzeem Y, Badawy M, Elhosseini M. An optimised deep learning approach for suicide detection through Arabic tweets. PeerJ Comput Sci. 2022. https://doi.org/10.7717/peerj- cs.1070 . 8, Article e1070. Barak A, Boniel-Nissim M, Suler J. Fostering empowerment in online support groups. Comput Hum Behav. 2008;24(5):1867–83. https://doi.org/10.1016/j.chb.2008.02.004 . Berger J, Milkman KL. What makes online content viral? J Mark Res. 2012;49(2):192–205. https://doi.org/10.1509/jmr.10.0353 . Booth RG. Happiness, stress, a bit of vulgarity, and lots of discursive conversation: A pilot study examining nursing students' tweets about nursing education posted to Twitter. Nurse Educ Today. 2015;35:322–7. https://pubmed.ncbi.nlm.nih.gov/25467718/ . Box-Steffensmeier JM, Moses L. Meaningful messaging: Sentiment in elite social media communication with the public on the COVID-19 pandemic. Sci Adv. 2021;7(29):eabg2898. https://pubmed.ncbi.nlm.nih.gov/34261655/ . Cannon WB. The James-Lange theory of emotions: A critical examination and an alternative theory. Am J Psychol. 1927;39(1/4):106–24. https://emotion.wisc.edu/wp- content/uploads/sites/1353/2020/11/Cannon_1927AmJPsych.pdf . Cherry K. (2025). The Schachter–Singer two-factor theory of emotions. *Verywell Mind* https://share.google/pQekNpesIJYQFhHw1 Correia JC, Ahmad SS, Waqas A, Meraj H, Pataky Z. Exploring public emotions on obesity during the COVID-19 pandemic using sentiment analysis and topic modelling: Cross-sectional study. J Med Internet Res. 2024;26(1):e52142. https://doi.org/10.2196/52142 . Cui B, Wang J, Lin H, Zhang Y, Yang L, Xu B. Emotion-based reinforcement attention network for depression detection on social media: Algorithm development and validation. JMIR Med Inf. 2022;10(8). https://doi.org/10.2196/37818 . Article e37818. Dror OE. The Cannon–Bard thalamic theory of emotions: A brief genealogy and reappraisal. *Emotion Review*. 2013. https://doi.org/10.1177/1754073913494898 . Guo Y, Zhang Z, Xu X. Research on the detection model of mental illness of online forum users based on a convolutional network. BMC Psychol. 2023;11(1). https://doi.org/10.1186/s40359-023-01460-4 . Article 424. Gutiérrez Benítez R, Segura Navarrete A, Vidal-Castro C, Martínez-Araneda C. Guide for the application of the data augmentation approach on sets of texts in Spanish for sentiment and emotion analysis. PLoS ONE. 2024;19(9):e0310707. https://doi.org/10.1371/journal.pone.0310707 . Harmon-Jones E, Harmon-Jones C, Summerell E. On the importance of both dimensional and discrete models of emotion. Behav Sci. 2017;7(4). https://doi.org/10.3390/bs7040066 . Article 66. Hsu JT-H, Tsa RT-H. Increased online aggression during COVID-19 lockdowns: A two-stage deep text mining study. Front Psychol. 2022;13:928485. https://www.jmir.org/2022/8/e38776/ . James W. (1884). What is an emotion? Mind, 9, 188–205. https://www.jstor.org/stable/2246769?seq=1 Kaplan AM, Haenlein M. Users of the world, unite! The challenges and opportunities of social media. Bus Horiz. 2010;53(1):59–68. https://doi.org/10.1016/j.bushor.2009.09.003 . Kapoor K, Tamilmani K, Rana NP, Patil P, Dwivedi YK, Nerur S. Advances in social media research: Past, present and future. Inform Syst Front. 2018;20(3):531–58. https://doi.org/10.1007/s10796-017-9810-y . https://link.springer.com/article/10.1007/s10796-017-9810-y . Kour H, Gupta MK. A hybrid deep learning approach for depression prediction from user tweets using CNN and Bi-LSTM. Expert Syst Appl. 2022;203:117394. https://pubmed.ncbi.nlm.nih.gov/35317471/ . Lang PJ. The varieties of emotional experience: A meditation of James–Lange theory. *Psychology Review*. 1994. https://doi.org/10.1037/0033-295X.101.2.211 . Liu F, Hou K, Dong Y. Deep parallel contextual analysis framework based emotion prediction in community wellness communications on social media. Heliyon. 2024;10(11):e31626. https://doi.org/10.1016/j.heliyon.2024.e31626 . Lokala U, Phukan OC, Dastidar TG, Lamy F, Daniulaityte R, Sheth A. Detecting substance use disorder using social media data and the dark web: Time- and knowledge-aware study. JMIRx Med. 2024;5(1). https://doi.org/10.2196/48519 . Article e48519. Manikonda L, Beigi G, Liu H, Kambhampati S. (2018). Twitter for sparking movements: Using hashtags to drive collective action. Proceedings of the International AAAI Conference on Web and Social Media, 12(1), 105–114. https://www.researchgate.net/publication/323931993_Twitter_for_Sparking_a_Move ment_Reddit_for_Sharing_the_Moment_metoo_through_the_Lens_of_Social_Media. Merayo N, Ayuso-Lanchares A, González-Sanguino C. Machine learning and natural language processing to assess the emotional impact of influencers' mental health content on Instagram. PeerJ Comput Sci. 2024. https://doi.org/10.7717/peerj-cs.2251 . 10, Article e2251. Mou X, Peng Q, Sun Z, Wang Y, Li X, Bashir MF. A deep learning framework for news readers' emotion prediction based on features from news article and pseudo comments. IEEE Trans Cybernetics. 2023;53(4):2186–99. https://doi.org/10.1109/TCYB.2021.3112578 . Pliseki H, Sobieszek A. (2024). Emotion topology: Extracting fundamental components of emotions from text using word embeddings. Front Psychol, 15, Article 3410084. Schachter S, Singer JE. Cognitive, social, and physiological determinants of emotional state. Psychol Rev. 1962;69(5):379–99. https://doi.org/10.1037/h0046234 . https://psycnet.apa.org/record/1963-06064-001 . Singla C, Singh S, Sharma P, Mittal N, Gared F. Emotion recognition for human-computer interaction using high-level descriptors. Sci Rep. 2024;14., Article 59294. https://doi.org/10.1038/s41598-024-59294-y . Schreiner M, Fischer T, Riedl R. Impact of content type and communication channel on emotional engagement in social media marketing. J Interact Mark. 2019;47:38–55. 10.1007/s10660-019-09353-8 . https://link.springer.com/article/ . Steinberg L. (2014). Age of opportunity: Lessons from the new science of adolescenc Houghton Mifflin Harcourt. https://www.researchgate.net/publication/277621882_Laurence_Steinberg_Age_of_O pportunity_Lessons_from_the_New_Science_of_Adolescence. Steinert S, Dennis MJ. (2022). Emotions and Digital Well-Being: on Social Media’s Emotional Affordances. Philosophy & Technology, 35(2), Article 36. 10.1007/s13347-022-00530-6 , https://link.springer.com/article/10.1007/s13347-022-00530-6 Steinert S, Dennis M. Emotional affordances: How digital technologies shape emotion and emotion regulation. Emot Rev. 2022;14(3):181–93. https://pubmed.ncbi.nlm.nih.gov/35450167/ . Tang H, Tang W, Zhu D, Wang S, Wang K, Wang J. EMFSA: Emoji-based multifeature fusion sentiment analysis. PLoS ONE. 2023;18(9):e0310715. https://doi.org/10.1371/journal.pone.0310715 . Wahl-Jorgensen K. Emotion and journalism: Theory, practice, and culture. Polity; 2018. Yang G, King SG, Lin H, Goldstein RZ. Emotional expression on social media support forums for substance cessation: Observational study of text-based Reddit posts. J Med Internet Res. 2023;25:e45267. https://doi.org/10.2196/45267 . Zhang Z, Liew K, Kuijer R, She WJ, Yada S, Wakamiya S, Aramaki E. Differing content and language based on poster-patient relationships on the Chinese social media platform Weibo: Text classification, sentiment analysis, and topic modeling of posts on breast cancer. JMIR Cancer. 2024;10:e51332. https://doi.org/10.2196/51332 . Additional Declarations No competing interests reported. Cite Share Download PDF Status: Under Review Version 1 posted Reviewers agreed at journal 16 Apr, 2026 Reviewers invited by journal 20 Mar, 2026 Editor assigned by journal 16 Mar, 2026 Submission checks completed at journal 16 Mar, 2026 First submitted to journal 06 Mar, 2026 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-9047352","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Systematic Review","associatedPublications":[],"authors":[{"id":609394369,"identity":"3cc58e1c-6a0f-4657-a813-8ae79e74d5d7","order_by":0,"name":"Sarath CJ","email":"","orcid":"","institution":"Indian Institute of Technology Bhubaneswar","correspondingAuthor":false,"prefix":"","firstName":"Sarath","middleName":"","lastName":"CJ","suffix":""},{"id":609394370,"identity":"2e32e651-a6cc-4dd1-abc3-09c32a722c84","order_by":1,"name":"Aparna Pandey","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABCUlEQVRIiWNgGAWjYHCDBAaGDwYSDEAEBjxEaWGcYSAhQZoWZqAyuBacgH9287PPBTV35M3Zk49J2xRY1JlLtz9g+FHDIGOOQ4vEnWPGs2cce2a4s+dZmnQO0GGWc84YMPYcY+CxbMCh50aCMTMP22HGDTdyzMBaDG7kMDDwNjDwGBzArkP+RvpnZp5/h+033Mj/Jm0B1pL+gPEvHi1AM42ZedsOJwJtYZNmAGtJMGDGZ4vhnTPFzLx9z5I3nHlmbNljICG5E+iXwzLHJHBqkbvdvpmZ59sd2w3Hkx/e+PGnjh8YYg8fvqmxscelhQESC2BZFniMHICJE9LC/AG3slEwCkbBKBjJAABwulqPzBHF4wAAAABJRU5ErkJggg==","orcid":"","institution":"Indian Institute of Technology Bhubaneswar","correspondingAuthor":true,"prefix":"","firstName":"Aparna","middleName":"","lastName":"Pandey","suffix":""}],"badges":[],"createdAt":"2026-03-06 07:25:31","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-9047352/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-9047352/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":105417548,"identity":"2d9eb744-a7f6-4d5b-9d1b-507f2dbf8e43","added_by":"auto","created_at":"2026-03-25 19:33:49","extension":"jpeg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":85409,"visible":true,"origin":"","legend":"\u003cp\u003eFig 1.1: PRISMA Flow chart\u003c/p\u003e","description":"","filename":"floatimage1.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-9047352/v1/b972d8b2be83cc865ff07985.jpeg"},{"id":105566189,"identity":"d5514b39-8537-4c3a-bf5e-2a6a6810ead7","added_by":"auto","created_at":"2026-03-27 12:55:38","extension":"jpg","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":139378,"visible":true,"origin":"","legend":"\u003cp\u003eFig 1.2: Social media platform focused on the study distribution\u003c/p\u003e","description":"","filename":"Picture1.jpg","url":"https://assets-eu.researchsquare.com/files/rs-9047352/v1/c24b4c91fbf7266a4615b7c2.jpg"},{"id":105417547,"identity":"54e3ae4e-324b-4824-b202-bdd67487cb33","added_by":"auto","created_at":"2026-03-25 19:33:49","extension":"jpg","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":112186,"visible":true,"origin":"","legend":"\u003cp\u003eFig 1.3: Number of studies using Specific Markers\u003c/p\u003e","description":"","filename":"Picture3.jpg","url":"https://assets-eu.researchsquare.com/files/rs-9047352/v1/e148948e810f41ed994e9ce1.jpg"},{"id":105417549,"identity":"dccc85ba-560b-4b52-a9e6-7c6cd1ec7866","added_by":"auto","created_at":"2026-03-25 19:33:49","extension":"jpg","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":138933,"visible":true,"origin":"","legend":"\u003cp\u003eFig 1.4: Major themes distribution\u003c/p\u003e","description":"","filename":"Picture4.jpg","url":"https://assets-eu.researchsquare.com/files/rs-9047352/v1/6a918136fc0e3856c1f27485.jpg"},{"id":105569575,"identity":"9b4311d0-dc8e-4bf9-bb1e-074a4a3da429","added_by":"auto","created_at":"2026-03-27 13:12:47","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1300437,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9047352/v1/3960422e-8c63-4fb1-b468-f8b66b34db39.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Digital Representation of Emotions through Social Media: A Systematic Review on Tracking Emotions","fulltext":[{"header":"Introduction","content":"\u003cp\u003eEmotions were found to be central to human lived experiences. The explanations given to emotions from the early researchers were specifically in association with the physiological reactions or arousal. James-Lange, in the late 19th century, proposed and stated that emotions follow the physiological responses (Lang, \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e1994\u003c/span\u003e). Furthermore, Cannon in the early 20th century, stated the limitations of the James-Lange theory and suggested that emotions and physiological response pathways proceed in parallel without one following the other (Dror, \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2013\u003c/span\u003e). Schatcher and Singer introduced the dimension of cognitive labelling, leading to specific emotions (Cherry, \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e2025\u003c/span\u003e). However, the recent advancement in the study and concept of emotions has given the newest and most diversified dimensions and explanations to emotions.\u003c/p\u003e \u003cp\u003eThe Modular or Discrete category approach and Dimensional approach of emotions are the two broad paradigms that dominate the theoretical landscape of emotions. The Modular approach to emotions views emotions as distinct modules or specific categories, focusing on emotions as states that take into account their triggers and features, such as anger, happiness, fear, and others. On the other hand, the dimensional approach considers emotions as varying along a continuous axis, such as valence (positive vs negative), arousal (high vs low), predominantly treating emotion as a process (Harmon-Jones et al., \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e2017\u003c/span\u003e). However, this paper adopts a more integrative and functional definition of emotion proposed by APA, which explains emotions as a complex reaction pattern that includes experiential, behavioural, and physiological components through which one tries to deal with a significant triggering event of life.\u003c/p\u003e \u003cp\u003eThe definitions of emotion provided by each theoretical perspective predominantly explain the internal manifestation and formation of emotion; likewise, it is essential to understand and identify the patterns and ways through which emotions are represented externally. Emotion representation could be defined as the expression, communication, and sharing of emotions externally, through verbal or nonverbal means. The verbal emotional communication includes spoken or written forms of emotional representation, and the non-verbal emotional representation includes facial expression, body language, voice tone, and artistic and symbolic representation. With the emergence and advancement of digital communication, emotional representation also shifted towards digital forms, such as text, images, symbols, and so on. In this advanced digital era, social media acts as a primary medium for emotional representation (Kapoor et al., \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e2018\u003c/span\u003e). Social media could be defined as a digital communication platform that facilitates interaction, content dissemination, and relationship-building among individuals and groups. The social media platforms support the exchange of information, emotions, and social meanings through various modes such as texts, images, emojis, hashtags, audio and video. It enables continuous social interaction across personal, professional, and socio-cultural contexts (Kapoor et al., \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e2018\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eSocial media is found to be a central mode of communication in the present generation, ranging from text messages to face-to-face through video interaction (Kaplan \u0026amp; Haenlein, \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e2010\u003c/span\u003e). It provides an open platform for users to build connections, express themselves, and engage in emotional exchange through multiple modalities. Self-presentation and self-disclosure are the mechanisms that facilitate social media's functioning at the heart of social media, providing opportunities for users to curate, share, and reveal personal information online (Kaplan \u0026amp; Haenlein, \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e2010\u003c/span\u003e). The social media platform features enable and facilitate the subtle ways of emotional actions and emotional expressions to a large extent (Steinert \u0026amp; Dennis, \u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). Currently, emotional representation has become a part of social media culture. They, along with the mitigation of private emotions, also facilitate social interaction. We generally prefer to broadcast or present positive emotional expression in public, and negative emotions, on the other hand, are mostly presented ironically, satirically or through an indirect mode, providing a collective outlet (Steinert \u0026amp; Dennis, \u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e2022\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eBeing a central and significant part in human life, emotions were identified to be foundational. Hence, it is important to understand how emotions are represented in the recent advanced digital era. Understanding emotions as an internal process of the emotional representation is significant in developing knowledge of emotions from a psychological perspective with a communication lens, which is further expanded with the rise of digital media. Digital emotional representation is a significant new layer of recognising and tracking the patterns of emotions and their variations. Research indicates the crucial role of social media platform design in shaping emotional expression and interactions (Steinert \u0026amp; Dennis, \u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). The study further introduces emotional affordances, which refer to the features in social media as that facilitate specific emotional expressions, such as like and dislike buttons, reaction emojis, comments tab, share function, and automated suggestions (Steinert \u0026amp; Dennis, \u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). The notion of emotional architecture facilitated by social media platforms suggests that social media platforms are not merely neutral conduits for communication or emotional expression, but an active environment that shapes the emotional flow of users (Jorgenson, 2019). Digital media enables in the transformation of emotional experiences and emotional expression patterns through hashtags, viral campaigns, and symbolic outrage, through which emotions can rapidly circulate and inspire coordinated actions (Manikonda et al., \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e2018\u003c/span\u003e). The emotional architecture of social media facilitates both the expression and amplification of emotions. The emotional affordances and architecture introduced through the digital media creates the emotional condition in addition to their mediation role in interaction and emotional expression. Understanding the dual nature of social media by decoding the patterns of emotional expression and amplification among users could provide a wider scope for human-computer interaction through the emotional spectrum.\u003c/p\u003e\n\u003ch3\u003eRationale\u003c/h3\u003e\n\u003cp\u003eThe focus of the recent research was found increasingly in the area of emotional expression through social media; even then, the field remains widely fragmented. Studies in this area have limited scope with confined integration of theoretical framework. Further, the methodological rigour remains inadequate (Schreiner et al., \u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e2019\u003c/span\u003e). Hence, the present systematic review aims to provide a structured and transparent synthesis of existing research papers on emotional representation through social media contexts, to address the gaps mentioned above. The two key questions covered in this systematic review are;\u003c/p\u003e \u003cp\u003e \u003col\u003e \u003cspan\u003e \u003cli\u003e \u003cp\u003eHow do individuals represent different states and processes of emotions through varying modes of social media communication forms?\u003c/p\u003e \u003c/li\u003e \u003c/span\u003e \u003cspan\u003e \u003cli\u003e \u003cp\u003eWhat are the characteristics and platform-specific variations based on specific emotional representation?\u003c/p\u003e \u003c/li\u003e \u003c/span\u003e \u003c/ol\u003e \u003c/p\u003e \u003cp\u003eThe current study seeks to deepen the understanding of the psychosocial dynamics of emotional representation in digital/online environments. It further provides direction for interdisciplinary research, thereby guiding digital well-being interventions and support systems for the design of emotionally supportive social media platforms and its healthy usage.\u003c/p\u003e"},{"header":"Methodology","content":" \u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003cp\u003eA systematic review design was followed as per the PRISMA guidelines (2020) for synthesising and reviewing research articles. It aimed at finding evidence from authentic research articles in exploring the psychosocial dynamics of emotional representation through social media platforms. Four databases, including Scopus, PubMed, Web of Science, and APA PsycNET, were reviewed in this systematic review.\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eInclusion \u0026 Exclusion criteria\u003c/h3\u003e\n\u003cp\u003ePeer-reviewed research articles published in the English language were considered in this systematic review, which investigated emotional representation from posts or comments shared through social media platforms. Eligible studies examining emotional expression or perception across diverse socio-demographic groups ranging from adolescence to adulthood, considering all forms of users, involving active, passive or observational users. Adolescence is a critical period, characterised by identity formation, emotion regulation development and heightened peer influence. Social media plays a major role in emotion representation and social comparison of this group (Steinberg, \u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e2014\u003c/span\u003e). The adults group generally demonstrates a stable and mature emotion regulation and variations in their motivations for social media usage, which also helps in making an understanding across developmental stages. The quantitative, qualitative, or computational studies, were considered to capture a comprehensive range of research evidence on the same. Case studies, secondary literature, or studies focusing on offline media were excluded. Literature published in languages other than English were also excluded.\u003c/p\u003e \u003cp\u003eSearch Strategy\u003c/p\u003e \u003cp\u003eA strategic search was conducted across four databases (SCOPUS, APA PsycNET, PubMed, and Web of Science) from October 2024 to November 2024 with keywords such as \"social media\", \"emotional representation\", \"emotional states\", and \"emotion\". Boolean operators and database-specific filters were utilised to refine search results by date (last 10 years) and language (English). The selection process comprised independent screening by reviewers based on title and abstract, followed by a full-text review for eligibility. The final selection of studies was made based on their relevance to the representation of emotions through social media, judged by independent reviewers. Any disagreements were resolved by reaching a consensus.\u003c/p\u003e \u003cp\u003eData Extraction\u003c/p\u003e \u003cp\u003eData collection involved the extraction of the following information from eligible studies by reviewers: general information (title, authors, year of publication), study characteristics (research design, social media context (platform(s), type of data - text, images, videos), type of emotions, representation characteristics (linguistic, symbolic, semantic features), and findings (main outcomes, implications, and research gaps). The filtration and extraction process lasted from November 2024 to June 2025. A narrative synthesis was performed to categorise findings based on the type of emotions, SMPs, representation style (e.g., textual vs. visual) and outcomes.\u003c/p\u003e"},{"header":"Results","content":"\u003cp\u003eThe systematic review followed the PRISMA 2020 guidelines across four databases, including Scopus, PubMed, Web of Science, and APA PsycNET. The initial search resulted in a total of 421 articles. After the removal of 19 duplicates and 7 non-research articles from the list, a total of 395 research articles were screened for title and abstract. A total of 325 articles were removed after title and abstract filtration, and 70 articles were sorted for full-text retrieval, with 66 articles having full text accessed for eligibility. The primary reason for the exclusion of articles after reviewing the full text is due to the lack of clear focus on emotional representation (n\u0026thinsp;=\u0026thinsp;34), and lack of clear distinct eligibility (n\u0026thinsp;=\u0026thinsp;4). The final list consists of 28 articles which met all the inclusion and exclusion criteria. The detailed flow chart in the study selection is given in Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1.1\u003c/span\u003e, the PRISMA flow chart.\u003c/p\u003e \u003cp\u003eThe characteristics and summary of the 28 included studies are given in Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1.1\u003c/span\u003e. Studies were spread across different nations, but a major focus was found with Asian countries (from China (31%), in collaboration with India (17%) \u0026amp; Saudi Arabia (10%), and the USA (24%)) as mentioned in Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1.1\u003c/span\u003e. The chronological distribution of studies spikes after 2022, which accounts for 90% of studies as given in Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1.1\u003c/span\u003e. The most frequently studied social media, according to these studies, was Twitter (25%), followed by Reddit (10%). The majority of the studies focused on analysing textual data through linguistic and semantic analysis (95%). Some of the studies used emojis and symbolic analysis (5%). The majority of participants across the\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e28 reviewed studies fall into adulthood, belonging to 19 to 64 years of age group (84.3%), and adolescence 13 to 18 years of age group (11.2%), and the remaining in late adulthood.\u003c/p\u003e \u003cp\u003eThe review and synthesis of research articles indicates that a higher range of emotional representations were expressed and segregated through the verbal mode, more specifically through textual contents which were determined predominantly using linguistic (41%), semantic (39%) and lexical (19%) markers given in Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e1.3\u003c/span\u003e. Linguistic markers were frequently used in emotional words and phrases, narrative self-disclosure, pronoun usage, informal language and slang. These linguistic patterns potentially represent immediate emotional responses, sustained emotional distress, and so on. The semantic markers were found to capture the underlying emotional meaning and intent beyond surface-level words and meaning. It indicates emotional states and thematic meaning based on subtle emotions such as anxiety, uncertainty or vulnerability could be inferred, and the emotional shifts over time were also being interpreted. This supports the contextual determination of emotions. Lexical markers used in emotional segregation are based on word frequency, emotional polarity, and affective lexicons. Even though the lexical markers were less likely to be used, but it remains essential in emotional valence and polarity with strength of emotion determination. The representation of emotions through social media was observed with the combination of explicit linguistic expression and implicit semantic meaning, supported by the lexical indicators.\u003c/p\u003e \u003cp\u003eThe specific emotional representation and characteristics were found to highly vary based on the social media platform on communication affordances, user roles, and content domains of the major themes identified and shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e1.4\u003c/span\u003e. The microblogging platforms such as Twitter, facilitate short, immediate emotional expression and high emotional intensity. It covers the themes, including public health, misinformation and breaking news. The influencers and celebrity-driven platforms such as Instagram indicate emotional amplification for social reach, users mirror disclosed emotions, and high emotional engagements. The community or support-oriented social media platforms show a more elaborate and reflective emotional representation, encouraging empathy and understanding facilitation. Support oriented social media features a more reflective emotional representation, encouraging empathy and facilitating understanding among users. In the support-based community online forums, highly arousing emotional expressions are less likely to dominate visibility, as they focus mainly on encouraging empathetic validation and coping. (Berger, \u0026amp; Milkman, \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e2012\u003c/span\u003e; Barak et al., \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e2008\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1.1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eStudy characteristics\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"7\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSl. No\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eTitle\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eYear\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eCountry\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eAuthor(s)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eMethodology\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003eMajor findings\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eHappiness, stress, a bit of vulgarity, and lots of discursive conversation: A pilot study examining nursing students' tweets about nursing education posted to Twitter\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e2015\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eCanada\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eRichard G. Booth\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eTweets about nursing courses, classes and clinical work were analysed using thematic analysis with five thematic clusters\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eInformal social media expression and need for digital professionalism among nursing students\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eLanguage-agnostic deep learning framework for automatic monitoring of population-level mental health from social networks\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e2022\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eThailand\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eThanapon Noraset, Krittin Chatrinan, Tanisa Tawichsri, Tipajin Thaipisutikul, Suppawong Tuarob\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eProposes a language-agnostic deep learning framework as a feasible alternative to traditional methods for gathering mental health data across diverse populations.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eEffective low-resource language framework for social media\u0026ndash;based mental health monitoring\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eDeep parallel contextual analysis framework-based emotion prediction in community wellness communications on social media\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e2024\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eChina\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eFeng Liu, Kun Hou, Yang Dong\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eIntroduces the DPCAF framework, which uses dual word embedding techniques to better capture semantic information in short texts for wellness communications.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eDeep contextual framework improving emotion prediction in short social media texts\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eA Deep Learning Framework for News Readers\u0026rsquo; Emotio Prediction Based on Features From News Articles and Pseudo Comments\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e2023\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eChina\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eXu Mou, Qinke Peng, Zhao Sun, Ying Wang, Xintong Li, Muhammad Fiaz Bashir\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eProposes a block emotion attention network (BEAN) that merges article content with \"pseudo-comments\" to predict reader emotions, even when actual reader feedback is missing.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eA hybrid framework enhancing emotion prediction in news articles through comment integration\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eDiffering Content and Language Based on Poster-Patient Relationships on the Chinese Social Media Platform Weibo: Text Classification, Sentiment Analysis, and Topic Modelling of Posts on Breast Cancer\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e2024\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eJapan \u0026amp; New Zealand\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eZhouqing Zhang, Kongmeng Liew, Roeline Kuijer, Wan Jou She, Shuntaro Yada, Shoko Wakamiya, Eiji Aramaki\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eWeibo posts on breast cancer fine-tuned through two step with Chinese BERT classification to identify poster- patient relationships determined using sentiment analysis using linguistic inquiry and word count\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eInfluence of poster\u0026ndash;patient relationship on emotional expression in health-related social media posts\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eEmotion Topology: Extracting Fundamental Components of Emotions from Text Using Word Embeddings\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e2024\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003ePoland\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eHubert Plisiecki, Adam Sobieszek\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eGo Emotions dataset of Reddit of communities with a minimum 10k comments with emotion taxonomy developed through manual annotations.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eWord embedding approach for mapping valence\u0026ndash;arousal dimensions in textual emotion analysis\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e7\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eIncreased Online Aggression During COVID-19 Lockdowns: Two-Stage Study of Deep Text Mining and Difference-in-Differences Analysis\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e2022\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eTaiwan\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eJerome Tze-Hou Hsu, Richard Tzong-Han Tsa\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eA two-stage study combining Deep Text Mining (training a BERT model to detect anger, offensive language, and hate speech) with a Difference-in-Differences (DID) econometric analysis to establish causal links between lockdowns and aggression.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eRise in aggressive social behaviours during COVID-19 lockdown and its policy implications\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e8\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eEmotional Expression on Social Media Support Forums for Substance Cessation: Observational Study of Text-Based Reddit Posts\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e2023\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eUSA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eGenevieve Yang, Sarah G. King, Hung-Mo Lin, Rita Z. Goldstein\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eReddit posts across 394 forums, quantified and categorised emotion word frequencies in substance cessation of alcohol, nicotin, and cannabis.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eOnline communities as valuable sources for understanding emotional recovery experiences\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e9\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eExploring Public Emotions on Obesity During the COVID-19 Pandemic Using Sentiment Analysis and Topic Modelling:Cross-Sectional Study\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e2024\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eSwitzerland, UK, Pakistan\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eJorge C\u0026eacute;sar Correia, Sarmad Shaharyar Ahmad, Ahmed Waqas, Hafsa Meraj, Zoltan Pataky\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eAnalyzes 53,414 tweets using the XLM-RoBERTa-base model for sentiment analysis and the BERTopic library for topic modelling.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eNegative public sentiment toward obesity highlighting need for informed health communication and policy\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e10\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eGuide for the Application of the Data Augmentation Approach on Sets of Texts in Spanish for Sentiment and Emotion Analysis\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e2023\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eChile\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eRodrigo Guti\u0026eacute;rrez Ben\u0026iacute;tez, Alejandra Segura Navarrete, Christian Vidal-Castro, Claudia Mart\u0026iacute;nez-Araneda\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eEvaluates Machine Learning and Deep Learning performance using data augmentation (DA) techniques like Easy Data Augmentation (EDA), back-translation (BT), and SentiGAN to expand small Spanish-language datasets.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eData augmentation techniques enhancing sentiment and emotion analysis in Spanish social media texts\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e11\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eEMFSA: Emoji-Based Multifeature Fusion Sentiment Analysis\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e2024\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eChina\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eHongmei Tang, Wenzhong Tang, Dixiongxiao Zhu, Shuai Wang, Yanyang Wang, Lihong Wang\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eProposes a multifeature fusion model (EMFSA) that integrates emoji, topic, and text features. It uses a sentiment- and emoji-masked language model (Senti_MLM) and a cross-attention mechanism to improve accuracy in short social media texts.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eMultifeature fusion model leveraging emojis to improve sentiment analysis in short social media texts\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e12\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eEmotion-Based Reinforcement Attention Network for Depression Detection on Social Media: Algorithm Development and Validation\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e2022\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eChina\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eBin Cui, Jian Wang, Hongfei Lin, Yijia Zhang, Liang Yang, Bo Xu\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eProposes an Emotion-Based Reinforcement Attention Network (ERAN) that uses a TextCNN for deep emotional feature extraction and Reinforcement Learning (RL) to select specific \"depression indicator\" posts from a user's history.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eDeep learning model enhancing depression detection through emotional semantic feature extraction on social media\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e13\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eAn Optimised Deep Learning Approach for Suicide Detection through Arabic Tweets\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e2022\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eSaudi Arabia, \u0026amp; Egypt\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eNadiah A. Baghdadi, Amer Malki, Hossam Magdy Balaha, Yousry AbdulAzeem, Mahmoud Badawy, Mostafa Elhosseini\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eProposes a framework for binary classification (Normal vs. Suicide) of Arabic tweets using Bidirectional Encoder Representations from Transformers (BERT) and Universal Sentence Encoder (USE) models after rigorous Arabic text preprocessing (lemmatization and stemming).\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eSocial media as a tool for tracking depression and mental health trend\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e14\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eComputational Linguistics-Based Text Emotion Analysis Using Enhanced Beetle Antenna Search with Deep Learning During the COVID-19 Pandemic\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e2023\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eSaudi Arabia, \u0026amp; India\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eYouseef Alotaibi, Arun Mozhi Selvi Sundarapandi, Subhashini P, Surendran Rajendran\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eComputational linguistics based mood analysis using enhanced beetle antenna search with deep learning (CLSA- EBASSDL using BERT word embedding classifier using attention based BiLSTM network.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eApplication of computational intelligence and nature-inspired algorithms for real-world optimisation problems\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e15\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eMachine Learning and Natural Language Processing to Assess the Emotional Impact of Influencers\u0026rsquo; Mental Health Content on Instagram\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e2024\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eSpain\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eNoemi Merayo, Alba Ayuso-Lanchares, Clara Gonz\u0026aacute;lez-Sanguino\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eCreated a labelled dataset of influencer\u0026rsquo;s Instagram responses based on mental health posts categorized by emotion laden and applied machine learning algorithms\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eEmotional impact of mental health disclosures on social media beyond psychopathology detection\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e16\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eA hybrid deep learning approach for depression prediction from user tweets using feature-rich CNN and bi-directional LSTM\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e2022\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eIndia\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eHarnain Kour, Manoj K. Gupta\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eDevelopment of a hybrid deep learning model using a feature-rich Convolutional Neural Network (CNN) and a Bi-directional Long Short-Term Memory (Bi-LSTM) network to classify depressive versus non-depressive textual narratives from Twitter data.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003ePredicting users\u0026rsquo; mental health conditions through Twitter data analysis\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e17\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eResearch on the detection model of mental illness of online forum users based on a convolutional network\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e2023\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eChina, \u0026amp; India\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eYuliang Guo, Zheng Zhang, Xuejun Xu\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eHierarchical user post feature representation models including Single- Gated LeakReLU- CNN and Multi- Gated LeakReLU- CNN to extract emotional features from user posts and to identify mental illness by analysing online forums.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eImproved models for accurate extraction of key emotional features from social media posts\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e18\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eEmotion recognition for human\u0026ndash;computer interaction using high-level descriptors\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e2024\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eIndia, \u0026amp; Ethiopia\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eChaitanya Singla, Sukhdev Singh, Preeti Sharma, Nitin Mittal, Fikreselam Gared\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eConstruction and preprocessing of a labelled speech corpus from diverse social media sources, followed by the application of Deep Learning techniques, specifically Convolutional Neural Networks (CNN), for Speech Emotion Recognition (SER).\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eEnhanced emotion recognition accuracy in Punjabi speech using an advanced SER approach\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e19\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eMeaningful messaging: Sentiment in elite social media communication with the public on the COVID-19 pandemic\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e2021\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eUSA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eJanet M. Box-Steffensmeier, Laura Moses\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eAnalysis of the tone and sentiment in elite social media messaging to determine its effect on information spread and public reaction during the pandemic.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eInfluence of partisanship and emotional tone on crisis communication and message engagement\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e20\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eDetecting Substance Use Disorder Using Social Media Data and the Dark Web: Time- and Knowledge-Aware Study\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e2024\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eUSA, India, \u0026amp; Thailand\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eUsha Lokala, Orchid Chetia Phukan, Triyasha Ghosh Dastidar, Francois Lamy, Raminta Daniulaityte, Amit Sheth\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eUtilization of state-of-the-art deep learning models to generate sentiment and emotion from social media posts and crypto market listings to understand users' perceptions and the relationship between substance misuse and mental health.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eDetection of Substance Use Disorder through social media and Dark Web data analysis\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e21\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eUser-based Hierarchical Network of Sina Weibo Emotion Analysis\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e2023\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eChina\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eQian Chen, Xiao Sun, Jiamin Wang, Meng Wang\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eUser based hierarchical network with combination of multi head attention and convolutional neural network to jointly analyse individual Weibo texts and related posts ober time, capturing contextual emotional information for improved five category emotion classification.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eUser-level contextual information improves accuracy in emotional classification,\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e22\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eTwitter Perspectives on Cochlear Implantation: Sentiment and Thematic Analysis\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e2023\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eUSA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eJoel S. Feier; Kenny Nguyen; Janet S. Choi\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eAll English language tweets mentioning \u0026lsquo;Cochlear implantation\u0026rsquo;, from 2019\u0026ndash;2021 collected using custom Python script, analysed for sentiment with VADER tool, and based on positive, negative, liked tweets using thematic analysis.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eUnderstanding public positive and negative perceptions of cochlear implantation\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e23\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u0026ldquo;Twitter is Really Therapeutic at Times\u0026rdquo;: Examination of Black Men\u0026rsquo;s Twitter Conversations Following Hip-Hop Artist Kid Cudi\u0026rsquo;s Depression Disclosure\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e2021\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eUSA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eDiane B. Francis\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eThe study was conducted in the United States and used thematic analysis to examine a sample of 1,482 tweets from the hashtag #YouGoodMan to identify recurring patterns in Black men\u0026rsquo;s mental health conversations following Kid Cudi's depression disclosure.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eCelebrity disclosure facilitates emotional expression online, and depression disclosure through an influencer\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e24\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eTopics and Sentiments of Public Concerns Regarding COVID-19 Vaccines: Social Media Trend Analysis\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e2021\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eUSA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eMonselise et al.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eTweets related to COVID-19 vaccines were collected over 60 days and analysed using topic modelling and sentiment analysis to identify key topics and emotions.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eFear as a leading reaction, news events dictate spikes\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e25\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSocial media\u0026rsquo;s dark secrets: A propagation, lexical and psycholinguistic oriented deep learning approach for fake news proliferation\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e2024\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eChina, \u0026amp; Saudi Arabia\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eKanwal Ahmed; Muhammad Asghar Khan; Ijazul Haq; Alanoud Al Mazroa; Syam M.S.; Nisreen Innab; Masoud Alajmi; Hend Khalid Alkahtani\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eA deep learning framework using Graph Convolutional Neural Networks (GCNN) with multi-head attention, incorporating BERT embeddings and psycholinguistic attributes (sentiment, personality, etc.) across user interaction and semantic propagation graphs.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e1. Disinformers are generally newer users, tweet more frequently, and are more active at night (higher insomnia index).\u003c/p\u003e \u003cp\u003e2. Users with lower popularity (low F2F ratio) are more likely to spread fake news.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e26\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSentiment, we-talk and engagement on social media: insights from Twitter data mining on the US presidential elections 2020\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e2022\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGermany\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eLinus Hagemann and Olga Abramova\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eA data mining approach using a dataset of over three million tweets from the 2020 US presidential elections, applying dual process theory to test how affective cues (sentiment) and cognitive cues (insight, causation) impact audience engagement.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eEmotional and cognitive cues influence social media engagement, and negative bias attracting users\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e27\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSentiment Classification of Anxiety-Related Texts in Social Media via Fusing Linguistic and Semantic Features\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e2024\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eChina\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eJianghong Zhu, Zhenwen Zhang, Zhihua Guo, and Zepeng Li\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eFuses linguistic and semantic features by using the SC-LIWC dictionary for linguistic extraction and a TextCNN-based model for deep semantic analysis. It uses a novel recursive feature selection algorithm on Sina Weibo data.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eAnxiety-related posts are a combination of negative and meaningful highly supporting the incorporation of the fusion model.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e28\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eRevealing the spatial co-occurrence patterns of multi-emotions from social media data\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e2023\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eChina \u0026amp; USA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eDongyang Wang a, Yandong Wang a,*, Xiaokang Fu a,b, Mingxuan Dou a, Shihai Dong a, Duocai Zhang c\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eUsed BERT based model to classify social media posts into six emotions, applied K- means clustering to analyse special emotion co-occurrence patterns, and employed a Multi Model Logit Model to examine their relationship with land use characteristics.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eSignificant special co-occurrence patterns of conflicting or consistent emotions in urban space on social media\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e\n\u003ch3\u003eThemes\u003c/h3\u003e\n\u003cdiv id=\"Sec7\" class=\"Section2\"\u003e \u003ch2\u003eEmotional Vulnerability Detection\u003c/h2\u003e \u003cp\u003eThe theme of Emotional vulnerability covers 36% of the identified themes from reviewed papers. The studies indicate the detection in the segments of emotional vulnerability, including sadness, fear, aggression and other negative emotions. Sentiment analysis, emotion classification, and lexicon-based or deep-learning models were used to segment social media texts. It helps in the early identification of emotional distress and its vulnerability on the development of mood disorders. It further indicates the understanding of platform-specific emotional risk detection and the necessary action.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003eEmotional Support Sequences\u003c/h2\u003e \u003cp\u003eThe theme of Emotional support sequences covers 23% of identified themes from reviewed papers, with the dimensions of positive emotions, including happiness, joy, satisfaction and other positive emotions. Emotion recognition models were used in analysing and identifying these segments. It helps in understanding well-being, resilience, and social support patterns. The digital well-being initiatives and potentiality were discussed with positive social media engagement strategies. The positive emotions were identified using polarity-based sentiment models and multi-class emotion classifiers. The algorithm and thematic classification based on the emotional support theme help in identifying segments in fostering wellbeing and protective factors against vulnerability.\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eEmotion Valence Measurement\u003c/h3\u003e\n\u003cp\u003eThe theme of Emotion valence measurement covers 16% of the identified themes from the reviewed papers. Research depicting the classification of emotions based on the range of positivity or negativity using emotional polarity categorization or sentiment classification using supervised machine learning models. It helps in the simplification of large-scale emotional representation patterns and prevalence through social media comparing emotional polarity across time, topic, or population.\u003c/p\u003e\n\u003ch3\u003eTechnological Application\u003c/h3\u003e\n\u003cp\u003eThe theme of Technological application covers a total of 9% of identified themes from research papers. The specific ML algorithms learning models of LSTM, CNN, Transformer-based, and NLP pipelines were predominantly trained and used in the emotion recognition of social media textual contents. It improves the accuracy and scalability of emotion detection. It further enables in the real time emotion monitoring on social media. The ethical usage of AI has been highlighted based on the training and model performance.\u003c/p\u003e \u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003eMental Health Monitoring\u003c/h2\u003e \u003cp\u003eThe theme of Mental health monitoring covered a total of 7% from identified themes from reviewed papers. It indicates the monitoring of mental health through emotion and sentiment patterns for mental health using computational models. Computational models have been used in emotion detection and sentiment analysis. This do not just assess individual\u0026rsquo;s emotion representation, but is also helpful in monitoring mental health at population level. It promotes interdisciplinary researches that the incorporate computer science with psychology for the early diagnosis of emotional dysfunctions, including depression, and anxiety using predictive computational models.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003eEmotional Communication Strategies\u003c/h2\u003e \u003cp\u003eThe theme of emotional communication strategy covered a total of 5% of identified themes from reviewed papers. It indicates analysis of language, slang and interaction patterns to understand emotional expression. It shows verbal emotional expressions communicated differently across different platforms. It further helps in improving the human-computer interaction and development of empathetic systems, and designing better online communication tools. The linguistic structures, discourse patterns and interaction patterns features indicate examination of differences in emotional expression across platforms and cultures.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec13\" class=\"Section2\"\u003e \u003ch2\u003eStress Detection\u003c/h2\u003e \u003cp\u003eThe theme of stress detection covered a total of 4% of themes from the identified reviewed papers. It indicates the analysis of stress through linguistic cues, sentiment shifts, and temporal emotion patterns. The studies identify the stress triggers related to work, crisis, or social events. It further indicates the potential of the early warning systems during emergencies or pandemics. Stress is detected with related keywords, sentiment shifts, and temporal emotion changes using classification ML models trained on stress labelled dataset.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eThe common procedures inferred from the research papers based on machine learning and supervised models are represented in a structured and systematic order as follows. The technique begins with the data acquisition through APIs and web scraping tools such as Scrapy. A comprehensive data pre-processing is followed, which includes data cleaning, artefact removal and normalisation. Data augmentation techniques such as random word swapping, synonym replacement, and back translation are applied. The textual data is then translated to numerical representations using feature extraction methods ranging from traditional bag of words to advanced word2vec, Glove, and BERT. The emotion classification or sentiment range was performed using a deep learning framework, including CNN and Bi-LSTM architectures. In addition, it incorporates topic modelling using LDA to uncover latent topicshemes, causal and statistical analysis through Difference in Difference or emotional topology. The performance is further evaluated using standard metrics of precision, recall, and F1 score to ensure accuracy and reliability.\u003c/p\u003e \u003c/div\u003e"},{"header":"Discussion","content":"\u003cp\u003eThe present study aims to synthesise the rapidly growing body of research examining how emotions are digitally represented, detected, and interpreted in social media by youth population. The systematic review indicates that the interdisciplinary dimension of social media and emotion is widely accelerating, as shown by the growing number of research in recent years, as shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e1.2\u003c/span\u003e. The study specifically focused on identifying how individuals express diverse emotional states and process across different social media modalities and further analysing the platform-specific variations in emotional representation. The systematic review process resulted in a final list of 28 peer-reviewed research articles. The research shows the convergence of psychology, computer science, communication studies, and data science with natural language processing by leveraging the analysis of linguistics and semantics more sequentially. The models, to show high performance in areas including predicting mental health conditions such as mood disorders and for recognising emotional states more specifically. The review further shows evidence for the methodological advancement moving beyond textual data to visual or auditory multi lingual interpretation. This multicultural expansion is significant in the development of a robust detection system.\u003c/p\u003e \u003cp\u003eA consistent pattern was detected from the 28 reviewed research articles showing the development of the digital environment into a significant ecosystem, in which emotions are expressed and interpreted. The findings indicate that the ML-based approaches of CNN, Bi LSTM, and Reinforcement attention mechanism demonstrate significant capacities in decoding patterns of emotional representation through texts, visual contents or symbols. These techniques further expand the reach and expansion of studies of emotional representations through digital media (Liu et al., \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). The research evidence shows that emotional recognition is not limited to linguistic or semantic markers, but extends to semantic, syntactic and symbolic cues, such as emojis, hashtags, reaction symbols (Tang et al., \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). Digital communication channels serve as a significant source for the multi-model structure of emotional representation (Steinert \u0026amp; Dennis, \u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). Social medias were found not just a platform to express emotions, but research shows evidence of them playing an active role in shaping the pattern of emotions based on media culture\u003c/p\u003e \u003cp\u003eThe text-based emotion analysis was found to be predominant and widely researched as the majority of research articles from the review uses textual content. Linguistic and semantic markers were predominantly used for analysing the text for detecting emotional patterns (Pliseki \u0026amp; Sobieszek, \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). The emerging recent models adopt a hybrid and multilingual approach that expands beyond the English language, addressing cultural and linguistic variation in emotional expression (Gutierrez et al., 2024). A strong intersection of emotional representation with mental health was also identified from this review. Emotional language, sentiment polarity and emotional valence were found to reliably indicate the symptoms of mood disorders such as depression or anxiety (Kour \u0026amp; Gupta, \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). The ethical usage of this user-generated social media data can act as a mirror of the user\u0026rsquo;s emotional process and state. Even then, the technical usage remains a challenge, especially regarding informed consent and user privacy (Steinert \u0026amp; Dennis, \u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e2022\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eThe thematic analysis of the 28 finally selected articles shows three dominant trends in emotional representation research. The first is regarding the contingent effect or emotional amplification effect is prevalent more specifically during social crisis situations, such as the COVID-19 pandemic, where solidarity became widespread (Hsu, Tsa, \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). The second trend depicts emotional personalisation driven by the user's algorithmic polarisation reinforces the user's unique affective contents by influencing public discourse and promoting wellbeing (Box- Steffensmeier \u0026amp; Moses, 2021). The third trend is regarding the interdisciplinary researches with the combination of computer science, psychology and language to interpret the complex interaction of emotions through the digital ecosystem (Manikanda et al., 2018).\u003c/p\u003e \u003cp\u003eThe current literature shows a high skewness towards text-based analysis while non-verbal indicators of emotions, such as emojis, symbols, or pictures, are less explored. The real-time emotional cues were found to be relatively less explored, which limits the dynamic affective shifts in digital interactions. Bridging this gap would enhance the clinical utility and interpretability. Current systematic review synthesises the methodological sequences and theoretical insights based on emotions represented through the social media ecosystem. A strong convergence was found towards data-driven, text-centric approaches. The incorporation in the advancement in machine learning and deep learning. Reviewed studies indicated the beginning of large-scale data acquisition through APIs and web scraping tools such as Scrapy. Data acquisition at the preliminary stage in collecting data from social media platforms through Application Programming Interface (API), that allows structured access to social media platforms and using Web Scrapping tools, extracts data from public webpages where APIs are limited. This is further followed by data pre-processing, including noise removal, normalisation, tokenisation, and lemmatisation, to improve model robustness (Manikonda et al., \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e2018\u003c/span\u003e; Singla et al., \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). The third step of data augmentation was facilitated using techniques such as word swapping or back translation, which were highly adopted for data sparsity. Social media raw data cannot be processed by machine learning models, hence, the traditional Bag- of- Words or advanced Word2Vec, Glove or BERT, converts words into densevector representations capturing semantic meaning, contextual similarity or Emotional nuance. (Kour \u0026amp; Gupta, \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). The deep learning frameworks were predominantly used for the classification of emotions and sentiment of the data using CNNs and LSTMs, often in hybrid or parallel sequences (Pliseki \u0026amp; Sobieszek, \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). Topic modelling techniques of Latent Dirichlet Allocation were used to synthesise the thematic structures within emotional discourse (Correia et al., \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). The interpretability is further enhanced using approaches including difference in differences approaches. Further, the performance of the model was evaluated using standard matrices based on accuracy, precision, recall and F1 score (Baghdadi et al., \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). The findings indicates that social media acts as both an emotional outlet and a diagnostic mirror as an emotional ecosystem.\u003c/p\u003e \u003cp\u003eSocial media was found to act as an emotional outlet and a diagnostic mirror, as verified from this review. Social media platforms provide and facilitate affective expression, and it simultaneously expose users to heightened emotional distress. Hence, the integration of social media platforms for emotional recognition and health must balance between empathetic responsiveness and ethical responsibility. Future research could focus more into building cross-platform emotional models capable of recognising and tracking emotional injuries.\u003c/p\u003e"},{"header":"Conclusion","content":"\u003cp\u003eThis systematic review shows evidence for the transformative potential of social media platforms in decoding emotional patterns and expressions through digital media. The findings show the potential of the human-centred computational algorithms that not only recognise emotions, but also modify based on the emotional pattern and preference of the users. It can further help in the early recognition of mental health and emotional conditions and to improve digital empathy with the development of emotionally intelligent algorithms. By maintaining the technical advancement and ethical boundaries, the digital emotional space not only limits in providing a space for emotional expression, but to a compassionate and empathetic technological understanding.\u003c/p\u003e"},{"header":"Declarations","content":"\u003ch2\u003eAuthor Contribution\u003c/h2\u003e\u003cp\u003eSarath CJ wrote the first draft of the manuscript. Both authors reviewed and edited the manuscript.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eAlotaibi Y, Sundarapandi AMS, Subhashini P, Rajendran S. Computational linguistics-based text emotion analysis using enhanced beetle antenna search with deep learning during COVID-19 pandemic. PeerJ Comput Sci. 2023. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.7717/peerj-cs.1714\u003c/span\u003e\u003cspan address=\"10.7717/peerj-cs.1714\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. 9, Article e1714.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAPA (American Psychological Association). (n.d.). Emotion. In APA Dictionary of Psychology. Retrieved October 21. 2025, from \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://dictionary.apa.org/emotion\u003c/span\u003e\u003cspan address=\"https://dictionary.apa.org/emotion\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBaghdadi NA, Malki A, Balaha HM, AbdulAzeem Y, Badawy M, Elhosseini M. An optimised deep learning approach for suicide detection through Arabic tweets. PeerJ Comput Sci. 2022. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.7717/peerj- cs.1070\u003c/span\u003e\u003cspan address=\"10.7717/peerj- cs.1070\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. 8, Article e1070.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBarak A, Boniel-Nissim M, Suler J. Fostering empowerment in online support groups. Comput Hum Behav. 2008;24(5):1867\u0026ndash;83. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.chb.2008.02.004\u003c/span\u003e\u003cspan address=\"10.1016/j.chb.2008.02.004\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBerger J, Milkman KL. What makes online content viral? J Mark Res. 2012;49(2):192\u0026ndash;205. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1509/jmr.10.0353\u003c/span\u003e\u003cspan address=\"10.1509/jmr.10.0353\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBooth RG. Happiness, stress, a bit of vulgarity, and lots of discursive conversation: A pilot study examining nursing students' tweets about nursing education posted to Twitter. Nurse Educ Today. 2015;35:322\u0026ndash;7. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://pubmed.ncbi.nlm.nih.gov/25467718/\u003c/span\u003e\u003cspan address=\"https://pubmed.ncbi.nlm.nih.gov/25467718/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBox-Steffensmeier JM, Moses L. Meaningful messaging: Sentiment in elite social media communication with the public on the COVID-19 pandemic. Sci Adv. 2021;7(29):eabg2898. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://pubmed.ncbi.nlm.nih.gov/34261655/\u003c/span\u003e\u003cspan address=\"https://pubmed.ncbi.nlm.nih.gov/34261655/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCannon WB. The James-Lange theory of emotions: A critical examination and an alternative theory. Am J Psychol. 1927;39(1/4):106\u0026ndash;24. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://emotion.wisc.edu/wp- content/uploads/sites/1353/2020/11/Cannon_1927AmJPsych.pdf\u003c/span\u003e\u003cspan address=\"https://emotion.wisc.edu/wp- content/uploads/sites/1353/2020/11/Cannon_1927AmJPsych.pdf\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCherry K. (2025). The Schachter\u0026ndash;Singer two-factor theory of emotions. *Verywell Mind* \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://share.google/pQekNpesIJYQFhHw1\u003c/span\u003e\u003cspan address=\"https://share.google/pQekNpesIJYQFhHw1\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCorreia JC, Ahmad SS, Waqas A, Meraj H, Pataky Z. Exploring public emotions on obesity during the COVID-19 pandemic using sentiment analysis and topic modelling: Cross-sectional study. J Med Internet Res. 2024;26(1):e52142. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.2196/52142\u003c/span\u003e\u003cspan address=\"10.2196/52142\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCui B, Wang J, Lin H, Zhang Y, Yang L, Xu B. Emotion-based reinforcement attention network for depression detection on social media: Algorithm development and validation. JMIR Med Inf. 2022;10(8). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.2196/37818\u003c/span\u003e\u003cspan address=\"10.2196/37818\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Article e37818.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDror OE. The Cannon\u0026ndash;Bard thalamic theory of emotions: A brief genealogy and reappraisal. *Emotion Review*. 2013. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1177/1754073913494898\u003c/span\u003e\u003cspan address=\"10.1177/1754073913494898\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGuo Y, Zhang Z, Xu X. Research on the detection model of mental illness of online forum users based on a convolutional network. BMC Psychol. 2023;11(1). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1186/s40359-023-01460-4\u003c/span\u003e\u003cspan address=\"10.1186/s40359-023-01460-4\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Article 424.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGuti\u0026eacute;rrez Ben\u0026iacute;tez R, Segura Navarrete A, Vidal-Castro C, Mart\u0026iacute;nez-Araneda C. Guide for the application of the data augmentation approach on sets of texts in Spanish for sentiment and emotion analysis. PLoS ONE. 2024;19(9):e0310707. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1371/journal.pone.0310707\u003c/span\u003e\u003cspan address=\"10.1371/journal.pone.0310707\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHarmon-Jones E, Harmon-Jones C, Summerell E. On the importance of both dimensional and discrete models of emotion. Behav Sci. 2017;7(4). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3390/bs7040066\u003c/span\u003e\u003cspan address=\"10.3390/bs7040066\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Article 66.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHsu JT-H, Tsa RT-H. Increased online aggression during COVID-19 lockdowns: A two-stage deep text mining study. Front Psychol. 2022;13:928485. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.jmir.org/2022/8/e38776/\u003c/span\u003e\u003cspan address=\"https://www.jmir.org/2022/8/e38776/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJames W. (1884). What is an emotion? Mind, 9, 188\u0026ndash;205. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.jstor.org/stable/2246769?seq=1\u003c/span\u003e\u003cspan address=\"https://www.jstor.org/stable/2246769?seq=1\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKaplan AM, Haenlein M. Users of the world, unite! The challenges and opportunities of social media. Bus Horiz. 2010;53(1):59\u0026ndash;68. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.bushor.2009.09.003\u003c/span\u003e\u003cspan address=\"10.1016/j.bushor.2009.09.003\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKapoor K, Tamilmani K, Rana NP, Patil P, Dwivedi YK, Nerur S. Advances in social media research: Past, present and future. Inform Syst Front. 2018;20(3):531\u0026ndash;58. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s10796-017-9810-y\u003c/span\u003e\u003cspan address=\"10.1007/s10796-017-9810-y\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://link.springer.com/article/10.1007/s10796-017-9810-y\u003c/span\u003e\u003cspan address=\"https://link.springer.com/article/10.1007/s10796-017-9810-y\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKour H, Gupta MK. A hybrid deep learning approach for depression prediction from user tweets using CNN and Bi-LSTM. Expert Syst Appl. 2022;203:117394. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://pubmed.ncbi.nlm.nih.gov/35317471/\u003c/span\u003e\u003cspan address=\"https://pubmed.ncbi.nlm.nih.gov/35317471/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLang PJ. The varieties of emotional experience: A meditation of James\u0026ndash;Lange theory. *Psychology Review*. 1994. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1037/0033-295X.101.2.211\u003c/span\u003e\u003cspan address=\"10.1037/0033-295X.101.2.211\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLiu F, Hou K, Dong Y. Deep parallel contextual analysis framework based emotion prediction in community wellness communications on social media. Heliyon. 2024;10(11):e31626. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.heliyon.2024.e31626\u003c/span\u003e\u003cspan address=\"10.1016/j.heliyon.2024.e31626\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLokala U, Phukan OC, Dastidar TG, Lamy F, Daniulaityte R, Sheth A. Detecting substance use disorder using social media data and the dark web: Time- and knowledge-aware study. JMIRx Med. 2024;5(1). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.2196/48519\u003c/span\u003e\u003cspan address=\"10.2196/48519\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Article e48519.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eManikonda L, Beigi G, Liu H, Kambhampati S. (2018). Twitter for sparking movements: Using hashtags to drive collective action. Proceedings of the International AAAI Conference on Web and Social Media, 12(1), 105\u0026ndash;114. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.researchgate.net/publication/323931993_Twitter_for_Sparking_a_Move\u003c/span\u003e\u003cspan address=\"https://www.researchgate.net/publication/323931993_Twitter_for_Sparking_a_Move\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e ment_Reddit_for_Sharing_the_Moment_metoo_through_the_Lens_of_Social_Media.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMerayo N, Ayuso-Lanchares A, Gonz\u0026aacute;lez-Sanguino C. Machine learning and natural language processing to assess the emotional impact of influencers' mental health content on Instagram. PeerJ Comput Sci. 2024. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.7717/peerj-cs.2251\u003c/span\u003e\u003cspan address=\"10.7717/peerj-cs.2251\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. 10, Article e2251.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMou X, Peng Q, Sun Z, Wang Y, Li X, Bashir MF. A deep learning framework for news readers' emotion prediction based on features from news article and pseudo comments. IEEE Trans Cybernetics. 2023;53(4):2186\u0026ndash;99. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1109/TCYB.2021.3112578\u003c/span\u003e\u003cspan address=\"10.1109/TCYB.2021.3112578\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePliseki H, Sobieszek A. (2024). Emotion topology: Extracting fundamental components of emotions from text using word embeddings. Front Psychol, 15, Article 3410084.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSchachter S, Singer JE. Cognitive, social, and physiological determinants of emotional state. Psychol Rev. 1962;69(5):379\u0026ndash;99. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1037/h0046234\u003c/span\u003e\u003cspan address=\"10.1037/h0046234\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://psycnet.apa.org/record/1963-06064-001\u003c/span\u003e\u003cspan address=\"https://psycnet.apa.org/record/1963-06064-001\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSingla C, Singh S, Sharma P, Mittal N, Gared F. Emotion recognition for human-computer interaction using high-level descriptors. Sci Rep. 2024;14., Article 59294. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1038/s41598-024-59294-y\u003c/span\u003e\u003cspan address=\"10.1038/s41598-024-59294-y\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSchreiner M, Fischer T, Riedl R. Impact of content type and communication channel on emotional engagement in social media marketing. J Interact Mark. 2019;47:38\u0026ndash;55. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s10660-019-09353-8\u003c/span\u003e\u003cspan address=\"10.1007/s10660-019-09353-8\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://link.springer.com/article/\u003c/span\u003e\u003cspan address=\"https://link.springer.com/article/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSteinberg L. (2014). Age of opportunity: Lessons from the new science of adolescenc Houghton Mifflin Harcourt. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.researchgate.net/publication/277621882_Laurence_Steinberg_Age_of_O\u003c/span\u003e\u003cspan address=\"https://www.researchgate.net/publication/277621882_Laurence_Steinberg_Age_of_O\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e pportunity_Lessons_from_the_New_Science_of_Adolescence.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSteinert S, Dennis MJ. (2022). Emotions and Digital Well-Being: on Social Media\u0026rsquo;s Emotional Affordances. Philosophy \u0026amp; Technology, 35(2), Article 36. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s13347-022-00530-6\u003c/span\u003e\u003cspan address=\"10.1007/s13347-022-00530-6\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e, \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://link.springer.com/article/10.1007/s13347-022-00530-6\u003c/span\u003e\u003cspan address=\"https://link.springer.com/article/10.1007/s13347-022-00530-6\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSteinert S, Dennis M. Emotional affordances: How digital technologies shape emotion and emotion regulation. Emot Rev. 2022;14(3):181\u0026ndash;93. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://pubmed.ncbi.nlm.nih.gov/35450167/\u003c/span\u003e\u003cspan address=\"https://pubmed.ncbi.nlm.nih.gov/35450167/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTang H, Tang W, Zhu D, Wang S, Wang K, Wang J. EMFSA: Emoji-based multifeature fusion sentiment analysis. PLoS ONE. 2023;18(9):e0310715. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1371/journal.pone.0310715\u003c/span\u003e\u003cspan address=\"10.1371/journal.pone.0310715\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWahl-Jorgensen K. Emotion and journalism: Theory, practice, and culture. Polity; 2018.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYang G, King SG, Lin H, Goldstein RZ. Emotional expression on social media support forums for substance cessation: Observational study of text-based Reddit posts. J Med Internet Res. 2023;25:e45267. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.2196/45267\u003c/span\u003e\u003cspan address=\"10.2196/45267\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhang Z, Liew K, Kuijer R, She WJ, Yada S, Wakamiya S, Aramaki E. Differing content and language based on poster-patient relationships on the Chinese social media platform Weibo: Text classification, sentiment analysis, and topic modeling of posts on breast cancer. JMIR Cancer. 2024;10:e51332. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.2196/51332\u003c/span\u003e\u003cspan address=\"10.2196/51332\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"discover-mental-health","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"dimh","sideBox":"Learn more about [Discover Mental Health](https://www.springer.com/44192)","snPcode":"","submissionUrl":"","title":"Discover Mental Health","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Discover Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"emotional representation, emotional distress, social media, digital representation","lastPublishedDoi":"10.21203/rs.3.rs-9047352/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-9047352/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eThe present systematic review aims to synthesise the existing evidence on tracking the digital representation of emotions through social media platforms, thereby providing scope for identifying users' emotional vulnerability and emotional psychopathology. Emotions are central to human experience, and their expression has undergone a major shift with the rise of digital media platforms. Social media currently serve as a primary medium of communication and emotional exchange. The PRISMA, 2020 guidelines were followed in this systematic review. The review was conducted on four databases, including Scopus, Web of Science, PubMed, and APA Psycnet. The final list consists of 28 articles, specifically focusing on emotional representation through social media, published in English and selected based on the inclusion and exclusion criteria. The findings show an acceleration in interdisciplinary research on emotional representation and digital media, with 90% of studies published after 2022. The most widely researched social media platform from the reviewed research articles was Twitter, with predominantly textual analysis using linguistic and semantic markers. A strong association was found from the reviewed articles on the recognition of the emotional patterns and the early detection of mood disorders and suicidal ideation. The review indicates the function of social media as an emotional expression outlet and a diagnostic mirror of the users' affective and psychological processes.\u003c/p\u003e","manuscriptTitle":"Digital Representation of Emotions through Social Media: A Systematic Review on Tracking Emotions","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-03-25 19:33:44","doi":"10.21203/rs.3.rs-9047352/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"reviewerAgreed","content":"59218131213226275269116651602981162250","date":"2026-04-16T15:58:26+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2026-03-20T10:10:04+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-03-16T12:19:17+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2026-03-16T12:18:41+00:00","index":"","fulltext":""},{"type":"submitted","content":"Discover Mental Health","date":"2026-03-06T07:08:17+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"discover-mental-health","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"dimh","sideBox":"Learn more about [Discover Mental Health](https://www.springer.com/44192)","snPcode":"","submissionUrl":"","title":"Discover Mental Health","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Discover Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"2fea78a6-095a-4d42-b106-4bf5d36c5cba","owner":[],"postedDate":"March 25th, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[],"tags":[],"updatedAt":"2026-03-25T19:33:44+00:00","versionOfRecord":[],"versionCreatedAt":"2026-03-25 19:33:44","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-9047352","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-9047352","identity":"rs-9047352","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}