They examined more than 44,000 posts
Researchers from Harvard University and the University of Vermont have analysed close to 44,000 Instagram posts from 166 volunteers to identify what constitutes as an Instagram-coded call for help.
Andrew Reece and Chris Danforth recruited the volunteers through an Amazon service called MTurk. They have just published their findings in a report that also details their method.
"We applied machine learning tools to successfully identify markers of depression. Statistical features were computationally extracted from 43,950 participant Instagram photos, using color analysis, metadata components, and algorithmic face detection. Resulting models outperformed general practitioners’ average diagnostic success rate for depression. These results held even when the analysis was restricted to posts made before depressed individuals were first diagnosed. Photos posted by depressed individuals were more likely to be bluer, grayer, and darker. Human ratings of photo attributes (happy, sad, etc.) were weaker predictors of depression, and were uncorrelated with computationallygenerated features. These findings suggest new avenues for early screening and detection of mental illness."
They discovered a correlation between mood and colour by looking at hue, saturation and brightness of photographs.
“Inkwell”, which turns colour photos to black and white, was the most commonly used filter among depressed participants.
They were more likely to post on Instagram more frequently and upload pictures which included faces, albeit a smaller number of faces in each picture, indicating that they may “interact in smaller social settings”, the research found.
The fewer likes and comments on a post, the less engagement a participant enjoys and can also indicate depression.