An app that detects depression by “looking at your face”

He diagnosis of depression It is not simple. Firstly, because when a person feels the mental apathy with which he is usually associated, he may refuse to seek psychological or medical assistance. And also because it is a rather vague diagnosis. Someone may have depressive traits but not depression per se, and the line separating one problem from another is not always clearly defined. What is now clear is that there are vital times when people need psychological help in order to succeed. Detecting this in time is important, especially for what has already been mentioned. Someone who is already feeling very bad may not want to seek help. Therefore, the mobile application Mood Capturedeveloped by scientists from Dartmouth Universitycould become a very useful tool in the future.

This is an application that combines artificial intelligence With facial processing technologies Receive a diagnosis of depression before the first obvious symptoms appear. This is achieved by randomly capturing images of the user using the front camera of the mobile phone. These images are analyzed for characteristics of a person or environmentwhich may be associated with this mental health problem.

So far, the first tests carried out using the app have been able to accurately diagnose depression in people. 75% of participants. Its creators want to increase the scale to 90%, so more research is needed. However, they believe that after 5 years may be ready to use.

Why is a cell phone so useful for this type of diagnosis?

According to a study conducted by IKEA Spain in 2023, one in three people look at their mobile phone more than 100 times a day. In another study of this type, but conducted in the United States, it was estimated that adults in this country do this on average 344 times a day.

Leaving aside the consequences that this may have on a psychological level, what is clear is that this is an object that is completely part of our day to day. Perhaps if we placed a camera on an iron and observed facial expressions, it would take us years to collect four or five images. But with him mobile phone no problem.

Andrew Campbell, one of the developers of the depression diagnostic app, estimates that he unlocked his phone himself. more than 800 times a week. So he and his colleagues agreed that it could be an ideal tool for analyzing the evolution of users’ facial expressions.

Training on an application for diagnosing depression

This application is mainly based on artificial intelligence. Therefore, as always happens with these algorithms, it was initially necessary train her.

This was done through a group of volunteers who had to answer a specific question while looking at their cell phone: “Have you felt apathetic, depressed or hopeless?” The question is part of a questionnaire known as PQH-8which is used specifically to diagnose depression.

While they were answering, the mobile phone took a random series of photos with front-camera. In this way, he collected information about his facial expression and environment, which he associated with reactions about his mood.

The algorithm was able to match factors such as gaze, eye movement, muscle rigidity or ambient lighting with user’s mood.

Very promising 75%

After the algorithm was trained, it was tested using 175 people who were informed that their mobile phone would take photographs of them with the front camera. However, they were not told exactly when. So they collected 125,000 images in 90 dayswith the help of which artificial intelligence launched a possible diagnosis of depression.

These same participants then answered the PQH-8 question, which was used in the training to check if the application results were correct. The diagnosis of depression was correct in 75% of cases. This is very promising, although the authors of the study They want to reach 90%.

Of course, this application will just be a wake-up call. A diagnosis of depression or depressive traits should be made by a professional. Yes, they hope it might be useful for encourage users to seek out this professional before symptoms become severe. This would be an interesting option and nothing intrusive, since the images never leave the mobile phone. They are processed directly in it.

Notably, the study was originally published without peer review. However, having appeared on Association for Computing Machinery 2024 CHI Conference has undergone the necessary changes. For this reason, it will finally be published in May next year, at the same time as the congress.

Of course, everything is going smoothly, but it’s too early to ring the bells. For now, we have to trust our judgment and the judgment of those around us. If you feel like something is wrong with you mood or you see that people around you are worried about her, do not hesitate to contact a psychologist. Your mental health is just as important as your physical health, don’t forget that.