Global Editions

Sound mind: Detecting depression through your voice

by TR Pakistan

Artificial Intelligence (AI) algorithms can now more accurately detect a depressed mood using the sound of your voice, according to new research by the University of Alberta computing scientists. The research was conducted by PhD student Mashrura Tasnim and Professor Eleni Stroulia in the Department of Computing Science.

The paper, “Detecting Depression from Voice,” was presented at the Canadian Conference on Artificial Intelligence in 2019.

The study builds on past research that suggests that the timbre of our voice contains information about our mood. Using standard benchmark data sets, Tasnim and Stroulia developed a methodology that combines several machine-learning algorithms to recognize depression more accurately using acoustic cues.

The ultimate goal is to develop meaningful applications from this technology, Stroulia explained in a press release.

Read more: A moody gut often accompanies depression — new study helps explain why

“A realistic scenario is to have people use an app that will collect voice samples as they speak naturally. The app, running on the user’s phone, will recognize and track indicators of mood, such as depression, over time. Much like you have a step counter on your phone, you could have a depression indicator based on your voice as you use the phone.”

According to a recent news report in, between 15 to 35 people commit suicide in Pakistan every day. That’s roughly one person every hour. In 2012, the World Health Organisation estimated that the rate of suicide in the country was 7.5 per 100,000 people. This meant that around 13,000 people killed themselves that year.

In 2016, the estimate was 2.9 per 100,000 i.e. over 5,500 ended their lives. Experts say the number of people dying is likely somewhere between the two figures, but the exact number remains unknown.

Such a tool could prove useful to support work with care providers or to help individuals reflect on their own moods over time. “This work, developing more accurate detection in standard benchmark data sets, is the first step,” added Stroulia.