Scientists have created an algorithm that can determine whether an Instagram user is showing signs of depression based on their posts to the app, according to a study published Monday by EPJ Data Science.
Researchers used almost 44,000 pictures from 166 people. Of the sample, 71 participants had a history of depression. The computer algorithm successfully identified markers of depression 70 percent of the time, according to the study.
It was able to spot markers of depression based on Instagram posts even before participants were clinically diagnosed.
The photos were examined based on their colors, the number of faces and the number of likes a post received. Researchers concluded that participants who posted photos with blue, gray or dark light tended to be depressed. Depressed Instagram users were also more likely to post photos with faces, but fewer faces per photo than their less-depressed counterparts. Depressed users also tended to receive fewer likes and were more likely to post photos without a filter.
However, the co-authors of the study, Andrew Reece and Christopher Danforth, caution that their study was limited by its relatively small sample size. Roughly 43 percent of their initial participants refused to share their Instagram data out of privacy concerns. Reece and Danforth did not immediately return NBC's request for comment.
The findings cannot be generalized to every Instagram user, but could serve as a "blueprint for effective mental health screening in an increasingly digitalized society."
Reece and Danforth concluded that their algorithm helped prove that mental illness and social media use have a scientifically calculable correlation.