App Detects Voice to Indicate If You're Suffering From Depression

Mental illness affects more people than ever before, with one out of every six adults dealing with symptoms of mental illness at any given time.

From insomnia to libido loss, tearfulness to anxiety, depression can manifest itself in a variety of ways. One company, however, has devised an unusual method of determining whether or not someone is depressed: listening to their voice.

According to research, people suffering from depression produce vowels with a narrower frequency range, resulting in a flatter-sounding voice. Paraverbal features are measurable traits in someone's voice that can also be detected in other mental illnesses, such as posttraumatic stress disorder.

Kintsugi

Kintsugi, a California-based app startup, is attempting to tap into these telltale signs of depression by developing a machine-learning model that, after listening to a speaker's voice, can assess their likelihood of having depression using the PHQ-9 and GAD-7 scales, which rate a patient's depression and anxiety severity on a 0 to 21 scale, respectively.

Users use the Kintsugi app to talk about their emotions regularly, answering pre-set prompts or challenges, and the machine learning AI scores them on these scales with each recording.

Kintsugi users can track their PHQ-9 and GAD-7 depression and anxiety scores in tandem with their journal entries over time, according to the company's website. This allows people to see progress without exerting much more effort than simply discussing current issues.

The National Science Foundation awarded Kintsugi multiple Small Business Innovation Research grants to develop this novel AI software that can detect signs of clinical depression and anxiety from short speech clips. The more people who use the app, the better the machine-learning model gets at detecting mental-health-related traits in voices.

Kintsugi co-founder Rima Seiilova-Olson explained that their neural network model was trained using tens of thousands of depressed voices. Later, he likened their app to a group of psychiatrists, but far more sensitive.

The Potential

The data from the voice clips, according to Kintsugi, will help patients get diagnoses more quickly, allowing them to be treated more quickly if they need it.

According to a study published in April in the journal Frontiers in Psychology, Cogito and Ellipsis Health have both developed AI systems that analyze the mental health markers of a person's voice. The Ellipsis Health app, on the other hand, shows the feasibility of using voice recordings to screen for depression and anxiety among various age groups.

Prentice Tom, Kintsugi's Chief Medical Officer, stated that real-time data improve a clinician's ability to provide better care and can be easily integrated into current clinical workflows. He went on to say that Kintsugi's voice biomarker tool is a crucial component in moving to a more efficient, quality-driven, value-based health care system.

Kintsugi Voice is a business API that allows applications to communicate with one another. The software, which can be integrated into clinical call centers and remote patient monitoring apps, allows the person on the other end to assess the speaker's emotional state.

The Kintsugi website advertises that by uniquely focusing on how people are speaking versus what they say. This reduces language bias inherent across socioeconomic classes and protects patient privacy by only analyzing features of the voice signal, their website states.