The photo on the right has a higher hue (bluer), lower saturation (grayer), and lower luminance (darker) than the photo on the left. Instagram photos posted by people with depression had HSV values shifted towards those in the right photograph, compared to photos posted by healthy individuals.
When you feel sad, the people around you probably know it. Moody playlists, hunched shoulders, long sighs – there are many ways we signal to the rest of the world when we’re having a bad day. It’s not all that hard to imagine that your Instagram posts can look happier when you’re feeling happy and saddest when you’re feeling sad.
What if you’ve been feeling depressed, but didn’t quite know it yet: Would your depression still show up in some way in the photos you’ve shared online? This possibility got us thinking: How could we combine what psychologists know about depression, with what data scientists know about analytics, to develop a quantifiable approach to assessing mental health on Instagram?
The results of our work suggest that early warning signs of emerging mental health problems such as depression can be seen in Instagram posts, even before any clinical diagnosis is made.
We asked people to share their stories of Instagram posts with us, along with details about their mental health history. By design, approximately half of our study participants reported having been clinically diagnosed with depression within the past three years. Overall, we collected 43,950 photos posted on Instagram for analysis.
People with depression in our sample tended to post photos that were, on average, bluer, darker, and grayer than those posted by healthy individuals.
Using findings from clinical psychology research, we identified several visual and behavioral markers associated with depression that appeared to be good candidates for measurement. For example, individuals with depression show different preferences for color, shading, and brightness of images than healthy individuals. Pixel analysis of photos in our dataset revealed that people with depression in our sample tended to post photos that were, on average, bluer, darker, and grayer than those posted by healthy individuals.
Depression is also characterized by low or avoidant social engagement. Social engagement involves other people, so we speculated that a rough measure of sociability might be the average number of people that appear in the photos you post. We wrote a face detection algorithm to count the number of faces that appeared in each published photograph. People with depression were found to post significantly fewer faces per photograph than healthy people.
The way depressed and healthy people chose to present their photos on Instagram also differed. Instagram offers a set of ready-made filters that adjust the look of a photo. Among healthy users, we observed that the most popular filter was Valencia, which gives photos a warmer, brighter look. Among depressed users, however, the most popular filter was Inkwell, which turned it black and white. In other words, people suffering from depression were more likely to prefer a filter that literally strips all the color out of the images they wanted to share.
We were able to reliably observe these differences, even by only looking at posts from depressed users that were posted before receiving a clinical diagnosis of depression. These and other recent discoveries (Here, HereAND Here) indicate that social media data can be a valuable resource for the development of efficient, low-cost, and accurate predictive mental health screening methods.
We strongly believe that there is an important ethical discussion that must occur in tandem with these technological developments, regarding data privacy and the implications of applying sophisticated analytical tools in an online medium that does not forget. Even so, the possibility that social media analytics could offer a means of getting help faster to people in need is important and should be further explored.
Read the full article here.
