佛蒙特大学的克里斯·丹佛斯和哈佛大学的安德鲁·里斯在EPJ Data Science期刊上发表了他们的发现。他们称自己开发的工具会扫描Instagram发布的消息，受试者被检测出抑郁的成功率高达70%。相反，医生只有42%的概率能准确判断情况。
It's often said that eyes are a window into the soul. Instagram, it turns out, can be that window, too, as researchers say they've developed an algorithm that uses posts on the social media network to identify depression more effectively than doctors can.
Chris Danforth of the University of Vermont and Andrew Reece of Harvard University published their findings in the EPJ Data Science journal. The tool they built to scan Instagram posts, they said, could accurately identify depression in 70% of their study's participants. Doctors, on the contrary, are successful only 42% of the time.
To arrive at these results, Danforth and Reece's team looked at roughly 44,000 Instagram photos posted by 166 study participants — 71 of whom were diagnosed with depression in the past. Analyzing factors such as hue, the use of filters and the presence of people, researchers were able to determine what they call "depression markers." Depressed people, for example, were more likely to post photos with darker, grey colors. Of those tones, the black and white Inkwell filter was more likely to be chosen, though depressed people on the whole were less likely to choose filters to begin with. People who aren't depressed, on the other hand, preferred the Valencia filter, which brightens the image.
Surprisingly, depressed people were more likely to post photos with faces in them. (Because the authors didn't determine who the photos were of, they couldn't say if that correlation meant the depressed people took more selfies.)