One study shows that the photos we post on social media can be used as markers of depression.

Do you like to publish black and white photos or add filters to highlight colors? Be that as it may, the photos you post to your Instagram account can provide clues about your mental health, according to a study by researchers at the University of Vermont (USA).

Using automatic learning, the artificial intelligence used in the study was able to detect who had depression and who do not based on their Instagram photos and randomly selected volunteers. 

“This points to a new method for the early detection of depression,” said Christopher Danforth, the leader of the study.

In the study, researchers examined Instagram’s photographs of more than 160 volunteers, recruited from the Amazon Mechanical Turk, an online crowdsourcing platform. Participants provided experts with information about their past depression diagnosis and responded to a questionnaire designed to assess a person’s current depression level.

About half of the people in the study had been diagnosed with depression in the past three years.

When the researchers analyzed the nearly 44,000 images, they found that messages from users who had a probable diagnosis of depression were bluer,

The use of a photographic filter was less common among individuals diagnosed with depression than among those who did not. But when individuals with a diagnosis of depression used filters, many opted for black and white filters such as “Inkwell”. In fact, some of the photographic characteristics that the researchers identified “coincide with common perceptions regarding the effects of depression on behavior,” the authors noted. For example, earlier studies had already suggested an association of depression with the preference for darker, more blue and monochromatic colors.

People inclined to depression were also more likely to post photos without people or with fewer people per photo compared to other users.

Using Instagram photos and the mental health history collected in the first part of the study, the researchers tested the algorithm for detecting depression while simultaneously with a different group of volunteers to see which of them, whether human or intelligence Artificial, was closer to the actual result of the mental health status of the participants.

The results revealed the machine learning algorithm did a better job than the human partners, as AI was able to correctly identify individuals with possible depression 70% of the time. 

“Obviously, Despite the findings, the study has limitations. For example, experts claimed to have used a broad definition of depression, so examining specific types of depression could lead to different results than those obtained. In addition, much more research is needed before this type of technology can be used to diagnose mental health conditions. ”

This study is not yet a diagnostic test, but it is a proof of concept of a new way of helping people, ” concludes Danforth. Despite the findings, the study has limitations. For example, experts claimed to have used a broad definition of depression, so examining specific types of depression could lead to different results than those obtained. In addition, much more research is needed before this type of technology can be used to diagnose mental health conditions. ” This study is not yet a diagnostic test, but it is a proof of concept of a new way of helping people, ” concludes Danforth. In addition, much more research is needed before this type of technology can be used to diagnose mental health conditions. ” This study is not yet a diagnostic test, but it is a proof of concept of a new way of helping people, ” concludes Danforth. In addition, much more research is needed before this type of technology can be used to diagnose mental health conditions. ” This study is not yet a diagnostic test, but it is a proof of concept of a new way of helping people, ” concludes Danforth.

 

Reference: Instagram photos reveal predictive markers of depression. 2017 EPJ Data Science.Doi.org/10.1140/epjds/s13688-017-0110-z

LEAVE A REPLY

Please enter your comment!
Please enter your name here