Researchers at the Luddy School of Informatics, Computing, and Engineering and the Department of Psychological and Brain Sciences have found that the language used by individuals on social media platforms may be predictive of a diagnosis of depression for the user.
In a study published in Nature Human Behavior, the researchers found that social media users who had received a clinical diagnosis of depression and posted about the diagnosis on Twitter also exhibited higher levels of cognitive distortions. Cognitive distortions, a crucial tenet that is targeted in cognitive behavioral therapy, are patterns of thought that are maladaptive, overly negative, and unrealistic, and that can have a negative effect on someone’s mood and motivation.
The study, “Individuals with depression express more distorted thinking on social media,” identified structural language patterns that are indicative of depression-related distorted thinking.
“I was wondering about how social media may have an effect on how people communicate and how they feel,” said Johan Bollen, a professor of informatics and computing and one of the lead researchers in the study. “This coincided with a conversation I had with my colleagues at the Psychology and Brain Sciences department, where I learned that people suffering from depression, anxiety, and other internalizing disorders often exhibit cognitive distortions. For example, someone may say something like: ‘I won’t go to the party because I will have a bad time. All my friends think I am a total loser.” The distorted nature of thought, not the fact that it is about a party, will make the individual feel bad and may prevent them from socializing and having a good time.”
Researchers studied two cohorts of Twitter users—individuals with depression and a randomly selected group—and compared the prevalence of expressions of cognitive distortions in the language of the users.
“Our results were very surprising,” Bollen said. “We found that the group of individuals with depression on Twitter expressed much higher levels of some types of cognitive distortions than a random sample. The effect was very significant, indicating that the language of individuals that suffer from depression is actually quite different in nature. This effect could not be explained by the topics discussed (e.g., depression or its symptoms), the emotional content of the language, nor the use of personal pronouns. It seemed to be specific to how the language was structured, and what it reveals about the individual's way of thinking.”
The study is a first step toward a better understanding of how online language interacts with depression and how this interaction plays out on social media and shapes its effects on societal well-being. It opens the potential for improved online diagnostics and therapeutics by automatically analyzing the structure of people’s language, regardless of the topic discussed and the emotional content of the language, identifying precisely whether an online message contains cognitive distortions, and whether this is a risk factor for depression, anxiety, and a range of other internalizing disorders for that individual or group.
“We will look at whether we can generalize this approach to other internalizing disorders, e.g., anxiety, seasonal affective disorders, and whether we may improve (and potentially automate) therapeutics and diagnostics,” Bollen said.
The effort was a collaboration between Bollen, postdoctoral researcher Marijn ten Thij, and Ph.D. student Krishna Bathina from the Luddy School, and Assistant Professor Lorenzo Lorenzo-Luaces and Associate Research Scientist Lauren Rutter from IU’s Department of Psychological and Brain Sciences.
“Social media is a powerful force in our society, and developing a way to use technology to help people based on what they say opens huge opportunities to make a positive impact,” said Kay Connelly, associate dean for research at the Luddy School. “This kind of collaborative research is at the heart of everything we do at our school, and I’m excited for the next steps in this work.”