Instagram may give a better snapshot of your mental health than even your doctor, new research suggests.

The telltale signs lie in color scheme, filters and composition, according to Harvard University’s Andrew Reece and the University of Vermont’s Christopher Danforth: Depressed people tend to post “bluer, darker, and grayer” photos and are less likely to slap on a filter ― though those who did disproportionately used black-and-white “Inkwell.” (Healthy people, in contrast, favored warmer, lighter “Valencia.”)

Depressed study participants were also likelier than healthy ones to post pictures with faces, though their photos contained fewer faces on average. Posts with lots of comments were more likely to have been posted by depressed people, while the reverse held true for likes.

A computer tool analyzing 43,950 Instagram photos from 166 people who filled out mental health screening questionnaires correctly diagnosed depression in 70%, per the study published Tuesday in EPJ Data Science. Previous research has indicated that doctors accurately identify depression just 42% of the time.

“Health care providers may be able to improve quality of care and better identify individuals in need of treatment based on the simple, low-cost methods outlined in this report,” Reece and Danforth wrote. “Given that mental health services are unavailable or underfunded in many countries, this computational approach, requiring only patients’ digital consent to share their social media histories, may open avenues to care which are currently difficult or impossible to provide.”

Social media could be a diagnostic tool for depression, but it might also make us miserable: People who spent more than two hours a day on platforms like Facebook, Twitter, Instagram, Snapchat doubled their risk of feeling socially isolated, according to a recent University of Pittsburgh study, compared to those who spent a half hour or less. About 16.1 million U.S. adults ― or 6.7% ― experienced at least one “major depressive episode” in 2015, according to the National Institute of Mental Health.

iStock

Reece and Danforth acknowledged some limitations in their study ― “the non-specific use of the term ‘depression’ in the data collection process,” for one.

Another factor is participants having been crowdsourced from Amazon’s Mechanical Turk marketplace, which may not be representative of a broader population, John Torous, co-director of the digital psychiatry program at Harvard-affiliated Beth Israel Deaconess Medical Center, told Moneyish. And 43% of people who began the survey declined to share their Instagram data with the researchers, he noted.

“I certainly wouldn’t begin to label friends as depressed based on this,” said Torous, who was not involved in the study. “I wouldn’t reach out and say, ‘I think you’re depressed because your filters have changed on Instagram.’”

“Part of the point of the study … (is) that an individual person, a friend, a therapist, wouldn’t really be able to detect depression in a person’s Instagram because we’re not really able to compute the amount of metadata that this program can,” Vaile Wright, a psychologist at the American Psychological Association, told Moneyish.

But while the findings are unlikely to “make or break a diagnosis,” Torous said, they represent “an important puzzle piece” in mental health research. If you’ve noticed a dramatic change in a friend’s social media use, you could try reaching out in a way that’s not “accusatory or judgmental,” he said.

“There’s no harm in asking people if they want to talk, how things are going, if they want to check in,” Torous added. “I think as long as it’s in a nonjudgmental, open way, I think it can only help.”