AI Can Now Detect Depression in a Child's Speech

The AI's results matched those received from a structured clinical interview and parent questionnaire.

Picture

Article from Mangmoo Blog

The AI's results matched those received from a structured clinical interview and parent questionnaire.

A new artificial intelligence (AI) can now detect signs of anxiety and depression in the speech of children. The machine learning algorithm has the potential to provide an effective way of diagnosing conditions that are difficult to spot.

Difficult to diagnose

It is estimated that around one in five children suffer from anxiety and depression. However, children under the age of eight can't properly articulate their emotions making it difficult for adults to diagnose.

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

"The majority of kids under eight are undiagnosed."

When it comes to mental health disorders, particularly internalizing disorders, early diagnosis is key. If caught and treated while their brains are still developing then children have positive outcomes. If, however, left untreated till later in life, they are at greater risk of substance abuse and suicide. 

The Trier-Social Stress Task

To test their diagnosing AI, the researchers put 71 children between the ages of three and eight through the Trier-Social Stress Task. This is a protocol designed to induce psychological stress.

The kids were told to improvise a three-minute story and advised that they would be judged based on how interesting it was. During the test, they were given only neutral or negative feedback. 

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

The researchers then used a machine learning algorithm to analyze audio recordings of each kid's story. The AI's results matched those received from a structured clinical interview and parent questionnaire.

"The algorithm was able to identify children with a diagnosis of an internalizing disorder with 80% accuracy, and in most cases that compared really well to the accuracy of the parent checklist," said University of Vermont biomedical engineer and study senior author Ryan McGinnis. And it did it in just a few seconds.

Researchers now say the next step will be to develop their new algorithm into a universal screening tool for clinical use. It could be turned into a smartphone app that could both record and analyze results immediately, warning parents of any potential problems.

 

Article from Mangmoo Blog