Detection of Mental Stress through Digital Communication Patterns: A User-Centric Survey Analysis

Abstract

Mental health awareness has become increasingly critical in the digital age, where text-based communication dominates interpersonal interaction. This study investigates the feasibility of detecting mental stress through the analysis of digital chat messages by examining user behaviour patterns, linguistic indicators, and perceptions regarding automated stress detection systems. A comprehensive survey was conducted with 54 participants from diverse demographics to understand how stress manifests in digital communication and user acceptance of artificial intelligence-based detection mechanisms. The findings revealed that 63.0% of participants experienced mental stress while engaging in chat conversations, with 68.5% acknowledging that their message characteristics changed during stressful periods. The key stress indicators identified were short or abrupt replies (59.3%), increased negative word usage (46.3%), and reduced emoji frequency (72.2%). The study demonstrated that 81.5% of participants experienced emotional overwhelm during digital communication, indicating the significant potential for early stress detection. Although 51.9% of respondents expressed comfort with AI-based analysis, universal privacy concerns (100%) highlighted the critical need for privacy-preserving detection systems. This research contributes empirical evidence to support the development of ethical, user-controlled mental health monitoring tools that leverage natural language processing and machine learning techniques. These findings provide a foundation for designing non-intrusive stress detection systems that respect user privacy while offering valuable insights into mental health.

Introduction

The study explores how digital communication platforms like WhatsApp and Instagram can be used to detect mental stress through chat messages. With rising mental health issues, especially among young people, traditional detection methods are limited, making non-intrusive approaches like analyzing messaging behavior highly valuable.

The research identifies a gap in existing studies, which mainly focus on public social media rather than private messaging, and lack insights into user acceptance and privacy concerns. To address this, the study aims to analyze behavioral and linguistic stress indicators, user perceptions of AI-based detection, and ethical issues.

Using a survey of 54 participants, the study found that most users frequently use chat applications, and 63% experience stress during communication. Around 68.5% reported noticeable changes in their messages when stressed. Common stress indicators include short replies, negative words, repeated messages, reduced emoji use, and typing errors. Emotional states like anger, sadness, and frustration are also commonly expressed.

The findings show that stress detection requires a multi-factor approach, as patterns like message length vary among individuals. While about 52% of users are comfortable with AI analyzing their messages, strong privacy concerns exist, including fears of data misuse, lack of consent, and algorithmic bias.

The study concludes that AI-based stress detection systems can be effective if they incorporate multiple indicators, ensure privacy protection, maintain transparency, and allow user control. However, limitations such as small sample size and reliance on self-reported data highlight the need for further research, including real-time analysis, larger datasets, and cross-cultural studies.

Conclusion

A. Summary of Key Findings This paper offers extensive information on the mental stress detection through the digital communication patterns. This study illustrates that: 1) Digital communication is vulnerable to mental stress, with only 63.0% of participants being not effected. 2) The stress is seen to have patterns in the content of messages and 68.5% of them admit that they make changes in messages when they are stressed. 3) There are several linguistic and behavioural clues that can be used in the process of detection, such as the length of the message, emotion, the use of emojis, and the choice of words. 4) The issue of privacy and ethics is universal, and so privacy-saving system structures are required. 5) Acceptance by users is also ambivalent with 51.9 percent being comfortable and 22.2 percent out of place, which means that there is necessity of transparent systems that are user-controlled.

References

[1] Pennebaker, JW, Mehl, MR, &Niederhoffer, KG (2003). Psychological aspects of natural language use: Our words, our selves. Annual Review of Psychology 54:547-577. https://doi.org/10.1146/annurev.psych.54.101601.145041 [2] De Choudhury M, Gamon M, Counts S, Horvitz E (2013) Predicting depression via social media. Proceedings of the International AAAI Conference on Web and Social Media 7(1):128-137. https://doi.org/10.1609/icwsm.v7i1.14432 [3] Coppersmith G, Dredze M, Harman C (2014) Quantifying mental health signals in Twitter. Proceedings of the Workshop on Computational Linguistics and Clinical Psychology: From Linguistic Signal to Clinical Reality, pp 51-60. https://doi.org/10.3115/v1/W14-3207 [4] Guntuku SC, Yaden DB, Kern ML, Ungar LH, Eichstaedt JC (2017) Detecting depression and mental illness on social media: an integrative review. Current Opinion in Behavioral Sciences 18:43-49. https://doi.org/10.1016/j.cobeha.2017.07.005 [5] Chancellor S, De Choudhury M (2020) Methods in predictive techniques for mental health status on social media: a critical review. NPJ Digital Medicine 3(1):1-11. https://doi.org/10.1038/s41746-020-0233-8 [6] Hovy D, Spruit SL (2016) The social impact of natural language processing. Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics, pp 591-598. https://doi.org/10.18653/v1/P16-2096 [7] Calvo RA, D\'Mello S (2010) Affect detection: An interdisciplinary review of models, methods, and their applications. IEEE Transactions on Affective Computing 1(1):18-37. https://doi.org/10.1109/T-AFFC.2010.1 [8] Tausczik YR, Pennebaker JW (2010). The psychological meaning of words: LIWC and computerized text analysis methods. Journal of Language and Social Psychology 29(1):24-54. https://doi.org/10.1177/0261927X09351676 [9] Martínez-Miranda J, Aldea A (2005) Emotions in human and artificial intelligence. Computers in Human Behavior 21(2):323-341. https://doi.org/10.1016/j.chb.2004.02.010 [10] Chancellor S, Baumer EPS, De Choudhury M (2019) Who is the \"human\" in human-centered machine learning: The case of predicting mental health from social media. Proceedings of the ACM on Human-Computer Interaction 3(CSCW):1-32. https://doi.org/10.1145/3359249 [11] Harika Juluri, Deepender Singh, T. R. Yograj Singh, and G. Sai Kiran Yadav, “Text-Based Stress Detection and Classification Using Machine Learning,” International Journal of Creative Research Thoughts (IJCRT), vol. 12, no. 7, 2024 [12] “Stress Detection System Using Machine Learning,”International Journal for Research in Applied Science & Engineering Technology (IJRASET), 2024. [13] “Stress Detection Using AI and Machine Learning,”International Journal of Engineering Research & Technology (IJERT), vol. 13, no. 8, Aug. 2024. [14] “Stress Detection Using Machine Learning and Deep Learning,” International Journal for Research in Applied Science & Engineering Technology (IJRASET), 2024. [15] Ríssola EA, Losada DE, Crestani F (2020)“A survey of computational methods for online mental health monitoring.” ACM Computing Surveys, 53(6):1–35. [16] “Automated Stress Detection Using Machine Learning,” International Journal of Engineering Research & Technology (IJERT), 2022. [17] De Choudhury M, Counts S, Horvitz E (2013)“Social media as a measurement tool of depression in populations.”Proceedings of the ACM Web Science Conference, pp. 47–56. [18] Preotiuc-Pietro D, Eichstaedt J, Park G, Sap M, Smith L, Tobolsky V, Schwartz HA, Ungar L (2015) “The role of personality, age, and gender in tweeting about mental illness.” Proceedings of the Workshop on Computational Linguistics and Clinical Psychology, pp. 21–30. [19] Gaur M, Kursuncu U, Alambo A, Sheth A, Daniulaityte R (2018) “Characterizing mental health from social media using deep learning.” Proceedings of the IEEE International Conference on Big Data, pp. 5075–5077. [20] Saha K, Torous J, Ernala SK, Rizuto C, Stafford A, De Choudhury M (2019) “A computational study of mental health awareness campaigns on social media.” Translational Behavioral Medicine, 9(6):1197–1207. [21] Losada DE, Crestani F, Parapar J (2017) “eRisk 2017: CLEF lab on early risk prediction on the internet.” Experimental IR Meets Multilinguality, Multimodality, and Interaction, pp. 346–360. [22] Losada DE, Crestani F, Parapar J (2017) “eRisk 2017: CLEF lab on early risk prediction on the internet.” Experimental IR Meets Multilinguality, Multimodality, and Interaction, pp. 346–360. [23] Orabi AH, Buddhitha P, Orabi MH, Inkpen D (2018) “Deep learning for depression detection of Twitter users.” Proceedings of the Fifth Workshop on Computational Linguistics and Clinical Psychology, pp. 88–97. [24] Shen G, Jia J, Nie L, Feng F, Zhang C, Hu T (2017) “Depression detection via harvesting social media: A multimodal dictionary learning solution.” Proceedings of IJCAI, pp. 3838–3844. [25] Resnik P, Armstrong W, Claudino L, Nguyen T, Nguyen VA, Boyd-Graber J (2015) “Beyond LDA: Exploring supervised topic modeling for depression-related language in Twitter.” Proceedings of the 2nd Workshop on Computational Linguistics and Clinical Psychology, pp. 99–107. [26] Lin H, Jia J, Guo Q, Xue Y, Huang J, Cai L (2014) “User-level psychological stress detection from social media using deep neural network.” Proceedings of the 22nd ACM International Conference on Multimedia, pp. 507–516. [27] Sriram B, Fuhry D, Demir E, Ferhatosmanoglu H, Demirbas M (2010) “Short text classification in Twitter to improve information filtering.” Proceedings of the 33rd International ACM SIGIR Conference, pp. 841–842. [28] 28.Mohammad SM (2012)“From once upon a time to happily ever after: Tracking emotions in novels and fairy tales.”Proceedings of the 5th ACL-HLT Workshop on Language Technology for Cultural Heritage, pp. 105–114. [29] Baziotis C, Pelekis N, Doulkeridis C (2017) “DataStories at SemEval-2017 Task 4: Deep LSTM with attention for message-level and topic-based sentiment analysis.” Proceedings of the 11th International Workshop on Semantic Evaluation (SemEval), pp. 747–754. [30] Zhang L, Wang S, Liu B (2018) “Deep learning for sentiment analysis: A survey.”Wiley Interdisciplinary Reviews: Data Mining and Knowledge Discovery, 8(4):e1253.

Copyright

Copyright © 2026 Prof. Rupali Khade, Anshika JIndal , Pravin Bachute, Kartik Kale, Aditya Bhavsar, Varsha Bhagwat. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.