How effective can a programmed depression algorithm be at finding the early stages of depression based on the content we upload and share? A new case study shows that our posts and likes on Instagram can speak wonders about how close we are to the edge.
The new recently published case study pitted a programmed algorithm against human perception and compared the results when it came to accuracy in detecting signs of depression of varying intensities.
Harvard University’s Andrew Reece and University of Vermont’s Christopher Danforth, along with their respective teams developed a depression algorithm that can observe and analyze all activity and content posted by an Instagram, such as posts, likes, and comments. For this particular case study, the algorithm focused mainly on the actual photos posted by users.
It is worth mentioning that before the case study officially started, the Instagram users selected for the depression algorithm had given their consensus and freely shared details about their medical history in regards to depression.
Instagram and the Depression Algorithm
The computer algorithm “looks” at an image. It takes into account the base colors present in images, the brightness, the contrast, the saturation, what filters are used, how many people are present in each image, and facial expressions, as well as a multitude of other factors. The depression algorithm compiles the data, makes its logical deduction, and then estimates the threshold possibility for that account’s owner to be suffering from depression or chronic social anxiety.
In order to equivalate the results, Reece and Danforth showed the same images to real humans and asked them to make the same deduction and estimation. The viewers did not have any prior psychological training. They did not personally know the subjects.
When it came to perception, awareness, and insight, the algorithm was the winner. Nevertheless, the Instagram depression meter has a limited range and set of abilities and cannot manage to keep up with licensed medical professionals.
Instagram and other broadly used social media services could use a similar depression algorithm to detect the issue in its developing stages. When it comes to depression, not everybody manages to find the strength to get professional help. The people who do seek out help do so after months of being at the whim of their inner monsters.