Dataset Information

Forecasting the onset and course of mental illness with Twitter data.

ABSTRACT: We developed computational models to predict the emergence of depression and Post-Traumatic Stress Disorder in Twitter users. Twitter data and details of depression history were collected from 204 individuals (105 depressed, 99 healthy). We extracted predictive features measuring affect, linguistic style, and context from participant tweets (N?=?279,951) and built models using these features with supervised learning algorithms. Resulting models successfully discriminated between depressed and healthy content, and compared favorably to general practitioners' average success rates in diagnosing depression, albeit in a separate population. Results held even when the analysis was restricted to content posted before first depression diagnosis. State-space temporal analysis suggests that onset of depression may be detectable from Twitter data several months prior to diagnosis. Predictive results were replicated with a separate sample of individuals diagnosed with PTSD (Nusers?=?174, Ntweets?=?243,775). A state-space time series model revealed indicators of PTSD almost immediately post-trauma, often many months prior to clinical diagnosis. These methods suggest a data-driven, predictive approach for early screening and detection of mental illness.

SUBMITTER: Reece AG

PROVIDER: S-EPMC5636873 | biostudies-other | 2017 Oct

REPOSITORIES: biostudies-other

ACCESS DATA

Publications

Forecasting the onset and course of mental illness with Twitter data.

Reece Andrew G AG Reagan Andrew J AJ Lix Katharina L M KLM Dodds Peter Sheridan PS Danforth Christopher M CM Langer Ellen J EJ

Scientific reports 20171011 1

We developed computational models to predict the emergence of depression and Post-Traumatic Stress Disorder in Twitter users. Twitter data and details of depression history were collected from 204 individuals (105 depressed, 99 healthy). We extracted predictive features measuring affect, linguistic style, and context from participant tweets (N = 279,951) and built models using these features with supervised learning algorithms. Resulting models successfully discriminated between depressed and he ...[more]

PMID: 29021528

Similar Datasets

Project description:Importance:People with mental illness are more likely to have contact with the criminal justice system, but research to date has focused on risk of offense perpetration, while less is known about risk of being subjected to crime and violence. Objectives:To establish the incidence of being subjected to all types of criminal offenses, and by violent crimes separately, after onset of mental illness across the full diagnostic spectrum compared with those in the population without mental illness. Design, Setting, and Participants:This investigation was a longitudinal national cohort study using register data in Denmark. Participants were a cohort of more than 2 million persons born between 1965 and 1998 and followed up from 2001 or from their 15th birthday until December 31, 2013. Analysis was undertaken from November 2016 until February 2018. Exposures:Cohort members were followed up for onset of mental illness, recorded as first contact with outpatient or inpatient mental health services. Diagnoses across the full spectrum of psychiatric diagnoses were considered separately for men and women. Main Outcomes and Measures:Incidence rate ratios (IRRs) were estimated for first subjection to crime event (any crime and violent crime) reported to police after onset of mental illness. The IRRs were adjusted for cohort member's own criminal offending, in addition to several sociodemographic factors. Results:In a total cohort of 2 058 063 (48.7% male; 51.3% female), the adjusted IRRs for being subjected to crime associated with any mental disorder were 1.49 (95% CI, 1.46-1.51) for men and 1.64 (95% CI, 1.61-1.66) for women. The IRRs were higher for being subjected to violent crime at 1.76 (95% CI, 1.72-1.80) for men and 2.72 (95% CI, 2.65-2.79) for women. The strongest associations were for persons diagnosed as having substance use disorders and personality disorders, but significant risk elevations were found across almost all diagnostic groups examined. Conclusions and Relevance:Onset of mental illness is associated with increased risk of exposure to crime, and violent crime in particular. Elevated risk is not confined to specific diagnostic groups. Women with mental illness are especially vulnerable to being subjected to crime. Individual's own offending accounts for some but not all of the increased vulnerability to being subjected to crime.

Project description:BackgroundSymptoms of serious mental illness are multidimensional and often interact in complex ways. Generative models offer value in elucidating the underlying relationships that characterise these networks of symptoms.AimsIn this paper we use generative models to find unique interactions of schizophrenia symptoms as experienced on a moment-by-moment basis.MethodSelf-reported mood, anxiety and psychosis symptoms, self-reported measurements of sleep quality and social function, cognitive assessment, and smartphone touch screen data from two assessments modelled after the Trail Making A and B tests were collected with a digital phenotyping app for 47 patients in active treatment for schizophrenia over a 90-day period. Patients were retrospectively divided up into various non-exclusive subgroups based on measurements of depression, anxiety, sleep duration, cognition and psychosis symptoms taken in the clinic. Associated transition probabilities for the patient cohort and for the clinical subgroups were calculated using state transitions between adjacent 3-day timesteps of pairwise survey domains.ResultsThe three highest probabilities for associated transitions across all patients were anxiety-inducing mood (0.357, P < 0.001), psychosis-inducing mood (0.276, P < 0.001), and anxiety-inducing poor sleep (0.268, P < 0.001). These transition probabilities were compared against a validation set of 17 patients from a pilot study, and no significant differences were found. Unique symptom networks were found for clinical subgroups.ConclusionsUsing a generative model using digital phenotyping data, we show that certain symptoms of schizophrenia may play a role in elevating other schizophrenia symptoms in future timesteps. Symptom networks show that it is feasible to create clinically interpretable models that reflect the unique symptom interactions of psychosis-spectrum illness. These results offer a framework for researchers capturing temporal dynamics, for clinicians seeking to move towards preventative care, and for patients to better understand their lived experience.

Project description:BackgroundThe COVID-19 pandemic led to unprecedented mitigation efforts that disrupted the daily lives of millions. Beyond the general health repercussions of the pandemic itself, these measures also present a challenge to the world's mental health and health care systems. Considering that traditional survey methods are time-consuming and expensive, we need timely and proactive data sources to respond to the rapidly evolving effects of health policy on our population's mental health. Many people in the United States now use social media platforms such as Twitter to express the most minute details of their daily lives and social relations. This behavior is expected to increase during the COVID-19 pandemic, rendering social media data a rich field to understand personal well-being.ObjectiveThis study aims to answer three research questions: (1) What themes emerge from a corpus of US tweets about COVID-19? (2) To what extent did social media use increase during the onset of the COVID-19 pandemic? and (3) Does sentiment change in response to the COVID-19 pandemic?MethodsWe analyzed 86,581,237 public domain English language US tweets collected from an open-access public repository in three steps. First, we characterized the evolution of hashtags over time using latent Dirichlet allocation (LDA) topic modeling. Second, we increased the granularity of this analysis by downloading Twitter timelines of a large cohort of individuals (n=354,738) in 20 major US cities to assess changes in social media use. Finally, using this timeline data, we examined collective shifts in public mood in relation to evolving pandemic news cycles by analyzing the average daily sentiment of all timeline tweets with the Valence Aware Dictionary and Sentiment Reasoner (VADER) tool.ResultsLDA topics generated in the early months of the data set corresponded to major COVID-19-specific events. However, as state and municipal governments began issuing stay-at-home orders, latent themes shifted toward US-related lifestyle changes rather than global pandemic-related events. Social media volume also increased significantly, peaking during stay-at-home mandates. Finally, VADER sentiment analysis scores of user timelines were initially high and stable but decreased significantly, and continuously, by late March.ConclusionsOur findings underscore the negative effects of the pandemic on overall population sentiment. Increased use rates suggest that, for some, social media may be a coping mechanism to combat feelings of isolation related to long-term social distancing. However, in light of the documented negative effect of heavy social media use on mental health, social media may further exacerbate negative feelings in the long-term for many individuals. Thus, considering the overburdened US mental health care structure, these findings have important implications for ongoing mitigation efforts.

Dataset Information

Forecasting the onset and course of mental illness with Twitter data.

Publications

Forecasting the onset and course of mental illness with Twitter data.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets