The team then analyzed the volunteers' Instagram photos using a statistical computer model, looking for several visual markers associated with depression. A computer program trained to diagnose depression from the images was correct 70 per cent of the time, a study led by Harvard University and Dr Christopher Danforth, of Vermont University, found. "This study is not yet a diagnostic test, not by a long shot", Danforth says, "but it is a proof of concept of a new way to help people".
"Imagine an app you can install on your phone that pings your doctor for a check-up when your behaviour changes for the worse, potentially before you even realise there is a problem". These and other recent findings (here, here, and here) indicate that social media data may be a valuable resource for developing efficient, low-cost, and accurate predictive mental health screening methods. Ideas from well-established psychology studies on color, brightness and shading preferences of people were used for analyzing the images.
Prof Danforth and USA colleague Andrew Reece from Harvard University wrote in a blog post accompanying the study: "Pixel analysis of the photos in our dataset revealed that depressed individuals in our sample tended to post photos that were, on average, bluer, darker, and greyer than those posted by healthy individuals". People who aren't depressed, on the other hand, preferred the Valencia filter, which brightens the image. And when they did use filters, they were especially likely to chose Inkwell, which turns photos black and white, compared to healthy people.
When the researchers analyzed the almost 44,000 images, they found that posts from users who had a diagnosis of depression were likely to be bluer, grayer and darker than posts from users without the condition.
From studying the posting patterns of people suffering with mental health issues, they found that sharing lots of pictures on social media was an indicator of users that suffer with depression.
The study, published Tuesday in the journal EPJ Data Science, analyzed almost 44,000 Instagram photos from 166 volunteers, who also shared their mental health history.
They were able to up to a point, but not as effectively as the computer software.
The scientists claim that artificial intelligence performed better, with Danforth saying: "Obviously you know your friends better than a computer". "This algorithm can sometimes detect depression before a clinical diagnosis is made".
"It shows some promise to the idea that you might be able to build a tool like this to get individuals help sooner", Danforth told HuffPost, adding that the program could have some utility for doctors when it comes to diagnosing patients who may only come in once every few years for a checkup.