Measuring Post Traumatic Stress Disorder in Twitter

By Glen A. Coppersmith, Craig Harman, Mark Dredze. International Conference on Web and Social Media, 2014

351 citations recorded by the source corpus.

Traditional mental health studies rely on information primarily collected through personal contact with a health care professional. Recent work has shown the utility of social media data for studying depression, but there have been limited evaluations of other mental health conditions. We consider post traumatic stress disorder (PTSD), a serious condition that affects millions worldwide, with especially high rates in military veterans. We also present a novel method to obtain a PTSD classifier for social media using simple searches of available Twitter data, a significant reduction in training data cost compared to previous work. We demonstrate its utility by examining differences in language use between PTSD and random individuals, building classifiers to separate these two groups and by detecting elevated rates of PTSD at and around U.S. military bases using our classifiers.

Subfields: Artificial Intelligence, Clinical Psychology, Social Psychology

View Paper: https://www.semanticscholar.org/paper/ea24d85e059d7d1dc201bd0380c76caf1f78f1e4

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Clinical Psychology

In Social Psychology

Citations in the corpus

Quantifying Mental Health Signals in Twitter (2014, background): “Such topics have already been the focus of several studies (Coppersmith et al., 2014; De Choudhury et al., 2014; De Choud-hury et al., 2013d; De Choudhury et al., 2013b; De Choudhury et al., 2013c; Ayers et al., 2013).”
A Social Media Study on the Effects of Psychiatric Medication Use (2019, background): “…naturalistic patterns of mood, behavior, cognition, social milieu, and even mental and psychological states, both in real-time and longitudinally (Chancellor et al. 2016, Coppersmith et al. 2014, De Choudhury et al. 2013, Dos Reis and Culotta 2015, Saha et al. 2019a, Yoo and De Choudhury 2019).”
The role of personality, age, and gender in tweeting about mental illness (2015, methodology): “As Coppersmith et al. (2014b) acknowledge, this method of collection is susceptible to multiple biases, but represents a simple way to build a large dataset of users and their textual information.”
Discovering Shifts to Suicidal Ideation from Mental Health Content in Social Media (2016, neutral): “Linguistic attributes of shared content and social interactional patterns have been utilized to understand and infer risk to major depressive disorder [24, 49, 32, 16, 60, 70], postpartum depression [21, 22], addiction [47, 44], and other mental health concerns [35, 18, 17, 46].”
Detecting Changes in Suicide Content Manifested in Social Media Following Celebrity Suicides (2015, background): “These include utilizing social media to understand conditions and symptoms related to diseases [37, 38], influenza propagation [49] and prediction [39], cyberbullying and teenage distress [14], substance abuse [29, 30], postpartum depression [10], mental health [12, 36, 22, 8, 9, 50], and insomnia [”
Dynamic Wordclouds and Vennclouds for Exploratory Data Analysis (2014, methodology): “Further, some linguistic psychometric analysis depends on the relative distribution of pronouns and other seemingly contentless words (Coppersmith et al., 2014a; Chung and Pen-nebaker, 2007).”
Data ex Machina: Introduction to Big Data (2017, methodology): “Twitter data have been used to study people who suffer from PTSD, suicidal ideation, and depression (Coppersmith et al. 2014, De Choudhury et al. 2013, 2016).”
Leveraging Large Language Models for Mental Health Prediction via Online Text Data (2023, background): “From the computational technology perspective, early research started with basic methods [19, 24, 53].”
Deep Learning for Depression Detection of Twitter Users (2018, background): “We do the same at the sentence level by predicting the surrounding sentences (Hill et al., 2016), as well as their possible sense (i.e., depressed, PTSD, or neither).”
Forecasting the onset and course of mental illness with Twitter data (2016, methodology): “In addition to physical ailments, predictive screening methods have successfully identified markers in social media data for a number of mental health issues, including addiction (6), depression (712), PostTraumatic Stress Disorder (PTSD) (13,14), and suicidal ideation (15).”
Explainable Sentiment Analysis with Applications in Medicine (2018, methodology): “In psychology, Sentiment Analysis applications have been proposed to: assess concern levels in suicide-related tweets, using both human coders and an automatic machine classifier [71]; to detect depression symptom in tweets [72],and also for the early detection of a whole range of depressionrelated ”
Depression and Self-Harm Risk Assessment in Online Forums (2017, background): “This makes social media a signiﬁcant resource for studying language re-lated to depression, suicide, and self-harm, as well as understanding the authors’ reasons for making such posts, and identifying individuals at risk of harm (Coppersmith et al., 2014a).”
Natural Language Processing of Social Media as Screening for Suicide Risk (2018, background): “Summarized from health record data presented in Padrez et al. 16 Coppersmith et al 3”
Enhancing Crowd Wisdom Using Explainable Diversity Inferred from Social Media (2018, neutral): “In [27], chances of Post-traumatic stress disorder is high for military personnel compared to civilians was evaluated using binomial-test on the Twitter text.”