How Instagram Filters Can Predict Whether Your Kid Is Depressed
It's worth a thousand words, and that may make a difference.
Scientists have created an algorithm that detects symptoms of depression based on a user’s Instagram posts. The new tool, described in a study published this week in the journal EPJ Data Science, can identify depressed people based on their preferences toward grayer, bluer, and darker hues with 70 percent accuracy. “This could help you get to a doctor sooner,” coauthor of the study Chris Danforth of the University of Vermont, told The Independent. Or, imagine that you can go to the doctor and push a button to let an algorithm read your social media history as part of the exam.”
Past research has connected mood to color—specifically, studies have linked lack of color and darker or grayer colors to depression. Despite the fact that depression has also been linked to less social activity, many depressed people still post on social media. As a result, researchers have been working for some time to utilize social media to predict and, ideally prevent, depression and other mental health issues by examining what healthcare professionals don’t always see. “Doctors don’t have visibility into our lives the way our mobile phone does,” Danforth told Mashable.
“It knows a lot more about us than we know about ourselves.”
For the current study, Danforth and his research partner Andrew Reece of Harvard University tested their algorithm on 43,950 Instagram posts from 166 individuals, 71 of whom had been previously diagnosed with depression. Results showed that depressed participants were less likely to appear in photos with friends and less likely to use filters. When they did use filters, they were more likely to opt for Inkwell, a feature that turns images to black and white. The algorithm itself correctly identified which participants had depression 70 percent of the time and ruled out depression with 81 percent accuracy. Conversely, doctors tend to diagnose depression accurately only 42 percent of the time.
“Although we had a relatively small sample size, we were able to reliably observe differences in features of social media posts between depressed and non-depressed individuals,” Reece said in a statement. “We also demonstrate that the markers of depression can be observed in posts made prior to the person receiving a clinical diagnosis of depression.”
If computers can detect things humans ultimately miss, it might mean earlier and more effective treatments and intervention options for everyone. Maybe your kid getting into social media won’t be so bad for them, after all.