Researchers have long sought to understand how people with depression share their feelings on Twitter, Instagram, Facebook, and Tumblr. Some hope to develop an algorithm that could spot the warning signals and provide a lifeline.
In 2016, Brown University released a study concluding that young people are at higher risk of depression if they experience cyber-bullying or meanness online. Earlier in 2016, another study suggested a correlation between Instagram users' filter choices and how miserable they feel about their lives.
Harvard University-trained data scientist Andrew Reece and University of Vermont professor Christopher Danforth sifted through 166 volunteers' 43,950 Instagram photos. The scientists found that the more depressed the volunteer, the more likely they were to post photos in darker tones with more grays and blues instead of bright and colorful pictures with friends. Depressed Instagrammers were also less likely to use filters—but if they did, they opted for the black and white filter, Inkwell.
Other researchers have analyzed language on social media. Megan Moreno, a researcher at Children's Hospital in Seattle, studies the ever-shifting words that self-harming communities use on Instagram. Of the 18 hashtags she has spotted—including seemingly innocuous words like #cat, #secretsociety123 and #blithe—only six generated a warning label on Instagram that redirected users to a site where they could seek help, according to her study.
"Big data approaches using text identification can be limited in that they typically need to have the right term and the right spelling, both of which are usually evolving over time as adolescents constantly refine and refresh their own slang," Moreno tells Broadly.
Read more: Why People Smoke Weed to Treat Depression
For instance, when Instagram started including warnings for words like #selfharm, users added two extra m's to the hashtag, her study found. Users may find comfort in sharing their pain, but Moreno's research suggests they can struggle to leave a self-harming community once they're known to other members. "Some teens can initially find self-harm communities supportive and welcoming, but then struggle with disengagement from that community when they are doing better," Moreno says. "Evidence shows exposure to images of cutting can also promote destructive behaviors by presenting self-harm as a viable way to deal with stress or negative emotions."
Some scientists want to harness the raw data of a phone's sensors instead of analyzing melancholic social media posts to look for signs of mental illness. Location-tracking appears to be the most promising development. Stephen Schueller, a behavioral psychologist at Northwestern University's Feinberg School of Medicine, told the Wall Street Journal that depressed people tend to travel more erratically around a city; doing so on future devices could trigger a pop-up asking if you need to get help. A Danish study recently analyzed the tone of voice as well as ambient noises recorded over microphone, finding that depressed people were more likely to have "slow, flat diction," according to Wired.
The jury is still out on the larger question of whether social media is beneficial or harmful to depressed people. A Depression and Anxiety study found that the most active social media users were twice as likely to be depressed as the general public. As Vocativ points out, depressed people may gravitate towards social media more often than happy people.
The way people present symptoms of depression tends to vary according to platform. Anecdotally, it seems like users on Twitter may juxtapose the broadcasting of morose feeling with something more lighthearted. "You are not going to listen to sad music and be depressed this week," one user posts, alongside a WikiHow illustration of a man pointing to himself in the mirror.
Twitter users, of course, often use the word "depressed" to describe a passing feeling or to add a little zest to a joke. Take Mira Gonzalez's recent tweet: "ariana grande exercising during her performance is very disrespectful to all the depressed, out-of-shape people watching at home #VMAs." Users seem post with more earnestness on Instagram. Hashtags like #cutting and #depressed are paired with black and white images. One user even repurposes a poem as a bio: "Roses are dead. Violets are dying. Outside I'm smiling. Inside I'm crying."
Many depressed users tag hashtags like #broken and #failure in a reminiscent of girls who post about their eating disorders with #ana and #thinspiration. Selfies are rare, as are colors, food, beaches, or any of the stereotypical trappings of the gramm-able lifestyle.
Writer Jamie Lauren Keiles described the appeal of Instagramming her depression. "On Instagram, I found a corner of the net where I was safe to shit out images of my terrible life in live time, without any imperative to express what I needed or interpret what it meant," she wrote in an essay. "It felt good and real to generate things that felt like plot points in a story as I laid in bed, waiting for the winds to change."
While scientists may be able to find a way for phones to provide effective intervention, it'll probably take years to develop an algorithm that takes into account all of the many, varied ways depressed people behave—with or without an audience.