Study finds link between poor mental health and browsing negative online content

People with poorer mental health are more prone to browsing negative content online, which further exacerbates their symptoms, finds a study led by UCL researchers.

The relationship between mental health and web-browsing is causal and bi-directional, according to the Wellcome-funded study published in Nature Human Behaviour.

The researchers have developed a plug-in tool that adds 'content labels' to webpages-;similar to nutrition labels on food-;designed to help users make healthier and more informed decisions about the content they consume. These labels emphasize the emotional impact of webpage content, along with its practicality and informativeness.

Co-lead author Professor Tali Sharot (UCL Psychology & Language Sciences, Max Planck UCL Centre for Computational Psychiatry and Ageing Research, and Massachusetts Institute of Technology) said: "Our results show that browsing negatively valenced content not only mirrors a person's mood but can also actively worsen it. This creates a feedback loop that can perpetuate mental health challenges over time."

Over 1,000 study participants answered questions about their mental health and shared their web browsing history with the researchers. Using natural language processing methods, the researchers analysed the emotional tone of the webpages participants visited. They found that participants with worse moods and mental health symptoms were inclined to browse more negative content online, and after browsing, those who browsed more negative content felt worse.

In an additional study, the researchers manipulated the websites people visited, exposing some participants to negative content and others to neutral content. They found that those exposed to negative websites reported worse moods afterward, demonstrating a causal effect of browsing negative content on mood. When these participants were then asked to browse the internet freely, those who had previously viewed negative websites-;and consequently experienced a worse mood-;chose to view more negative content. This finding highlights that the relationship is bi-directional: negative content affects mood, and a worsened mood drives the consumption of more negative content.

Co-lead author, PhD student Christopher Kelly (UCL Psychology & Language Sciences, Max Planck UCL Centre for Computational Psychiatry and Ageing Research, and Massachusetts Institute of Technology), said: "The results contribute to the ongoing debate regarding the relationship between mental health and online behaviour.

"Most research addressing this relationship has focused on the quantity of use, such as screen time or frequency of social media use, which has led to mixed conclusions. Here, instead, we focus on the type of content browsed and find that its emotional tone is causally and bidirectionally related to mental health and mood."

To check whether an intervention could be used to change web-browsing choices and improve mood, the researchers conducted a further study. They added content labels to the results of a Google search, which informed participants whether each search result would likely improve their mood, make it worse, or have no impact. Participants were then more likely to choose the positively-labelled sites deemed likely to improve their mood-;and when asked about their mood after, those who had looked at the positive websites were indeed in better moods than other participants.

In response, the researchers have developed a free browser plug-in that adds labels to Google search results, providing three different ratings of how practical a website's content is, how informative it is, and how it impacts mood.

We are accustomed to seeing content labels on our groceries, providing nutritional information such as sugar, calories, protein, and vitamins to help us make informed decisions about what we eat. A similar approach could be applied to the content we consume online, empowering people to make healthier choices online."

Professor Tali Sharot, UCL Psychology & Language Sciences, Max Planck UCL Centre for Computational Psychiatry and Ageing Research, and Massachusetts Institute of Technology