The images users share online speak a lot about them without saying a single word. These images can be a form of self-expression or a record of travel, and even more so, as a new study revealed that images that are shared through the Instagram service may carry evidence of mental health.
From the colors and faces in their images to the improvements they make before they are published, a study published days ago in the journal EPJ Data Science found that Instagram users with a history of depression picture the world differently from their peers.
"On average, people in our sample, who had previously suffered from depression, tended to publish images that were, on the contrary," said Andrew Rees, a postdoctoral researcher at Harvard University and co-author of the study with Christopher Danforth, a professor at the University of Vermont in the United States. The basis of each pixel separately, less visible and darker and more inclined to gray than healthy people. "
The researchers identified the participants as "depressed" or "healthy" based on whether they reported having received a clinical diagnosis of depression in the past. They then used automated learning tools to find patterns in images and create a predictive model of depression through publications.
Rees and Danforth found that depressed participants use Instagram filters, which allow users to change the brightness and color of the image digitally before publication, at a slower pace. When those users add any candidate, they tend to choose the "Inkwell" filter, which strips the image of its color and makes it black and white. While healthy users tend to prefer Valencia, which reduces the color of the image.
Participants with depression were more likely to post facial images. But when healthy participants published images containing faces, they tended, on average, to show more faces.
To be studied, participants had to meet several criteria, including having to be active on Instagram and willing to share their entire record of publications with researchers. They also had to disclose whether they had received a clinical diagnosis of depression or not.
Of the hundreds of responses they received, Reese and Danforth recruited 166 people, 71 of whom had a history of depression. It has collected nearly 44,000 images.
The researchers then used software to analyze the chromatography of each image, the saturation of colors and brightness, as well as the number of faces it contained. They also collected information on the number of publications per user and the number of comments and responses per post.
Using automated learning tools, the researchers found that the more comments a comment received, the more likely it would be from someone with depression. The opposite was true for the likes. They also found that depressed users tended to publish too much.
Although they cautioned that their findings may not apply to all Instagram users, Reese and Danforth argued that the findings suggest that a similar model of automated learning can someday be useful in conducting or promoting mental health testing.
0 Comments