Under the Hood

Depression Prevention: AI Detects ‘Emotional Suffering’ Of Children

Parents may soon better understand their children’s emotions through the use of a new artificial intelligence tool. The technology can read speech patterns of young children and predict early signs of anxiety and depression.

A new study published in the Journal of Biomedical and Health Informatics suggests a machine learning algorithm could help speed up the diagnosis and treatment for kids with signs of depression. To date, it is difficult for health care providers to spot mental issues in young people.

Psychologists say children under the age of eight commonly can't reliably articulate their emotional suffering. Such challenge to understand their mental state then contributes to the growing number of children with undiagnosed anxiety or depression. 

"We need quick, objective tests to catch kids when they are suffering," Ellen McGinnis, lead author of the study and a clinical psychologist at the University of Vermont Medical Center's Vermont Center for Children, Youth and Families, said in a statement. "The majority of kids under eight are undiagnosed."

Researchers said that early diagnosis of mental conditions is important since children respond well to treatment because their brains are still developing. But when left untreated, they are at risk of substance abuse and suicide later in life.

AI: New Key To Fight Depression

For their study, the researchers used the Trier-Social Stress Test that triggers feelings of stress and anxiety in 71 children aged three to eight. The participants were asked improvise a story, which will be later judged based on how interesting it was. 

The researchers gave only neutral or negative feedback during the speech of each child. 

"The task is designed to be stressful, and to put them in the mindset that someone was judging them," McGinnis said.

The team then used a machine learning algorithm to analyze the audio recordings of each child's story and compared the results to the child's initial diagnosis. They found that the tool effectively diagnosed children with 80 percent accuracy.

The algorithm also only took seconds to process data and provide a diagnosis.

The researchers plan to develop the speech analysis algorithm into a universal screening tool and offer it for clinical use. The algorithm may be used on a smartphone app to record and analyze results immediately.