An algorithm has been developed by researchers to help diagnose depression and other mental illnesses in users of Instagram.
The online photo-sharing, video-sharing and social networking service has been utilised in a study by Andrew Reece of Harvard University and Chris Danforth of the University of Vermont to diagnose mental illness according to the colours a person uses in their photos.
Instagram has approximately 500 million users as of June 2016 and according to the study people who post darker coloured or greyer images on the site are more likely to suffer from depression, compared to those who have vibrant-coloured pictures.
Additionally, Reece and Danforth created a machine-learning algorithm that finds links between image properties and depression.
An image on Instagram that has vibrant colours can be modified with a feature called Inkwell that turns colour into black and white – the study spotted that more depressed people use Inkwell than non-depressed people.
The researchers asked 170 employees of Amazon's Mechanical Turk service who had Instagram accounts to fill out a questionnaire. Included was a standard clinical depression survey and the participants were also asked to share their Instagram images.
According to Tech Times, about 100 photographs were chosen from each participant and people were asked to rate them on a scale of 0 to 5 based on how interesting, sad or happy the photos looked. Each image was also categorised based on hue, saturation and the number of faces.
The machine-learning algorithm discovered that decreased brightness and saturation and increased hue foretold depression. The study also found that depressed and non-depressed people used photo filters differently. For instance, the researchers found that depressed people were less likely to use filters.
The algorithm had a 70 per cent chance of finding a person with depression and the researchers said: “More generally, these findings support the notion that major changes in individual psychology are transmitted in social media use, and can be identified via computational methods.”
In their paper, Reece and Danforth reported: “Statistical features were computationally extracted from 43,950 participant Instagram photos, using colour 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, greyer, and darker. Human ratings of photo attributes (happy, sad, etc.) were weaker predictors of depression and were uncorrelated with computationally generated features.”
Reece and Danforth believe that their new technique will help us better recognise and understand mental illness in people. It could also detect early-stage depression, which would aid in an effective diagnosis.