When we hear the term “voice recognition”, most of us think of virtual assistants such as Google, Siri, and Alexa. While voice command can be useful in certain circumstances, it often remains just a fun feature to play around with for several days.

But what would you say if voice recognition software became capable of detecting disease based on changes in your speech patterns?

Disease detection through speech analysis could revolutionize diagnostics in young adults with an active lifestyle, who are always busy. They have neither the time, nor the inclination to see a doctor. Friend of mine once said:

“I’m a grown-up 32-year old man, so when there’s something wrong with my health, I will, just like any of my peers, go straight to intensive care.”

If you wish to avoid critical illness, it’s important to detect early signs of disease on a casual basis”.

What does the research say?

A recent study at Mayo Clinic in the US found a strong correlation between voice signal characteristics and cardiac ischemia, which is increasingly often diagnosed in people aged under 40, who live in cities and deal with a lot of stress. This technology opens up unique opportunities to remotely monitor people’s health risks.

Here’s another example. Depression is the most common mental disorder. The World Health Organization reports that globally more than 300 million people suffer from depression! It affects nearly 1 in 12 US adults. There are people suffering, but they are not getting treatment—80% of patients with depression don’t seek professional help unless their condition becomes dangerous. According to the same data, depression is highly associated with suicide. Given the above, consider the fact that general practitioners only identify this mental disorder in half of all cases.

How can smartphones help to treat depression?

There are apps that help assess your risk for depression using questionnaires. Several years ago researchers found a way to do that based on voice and speech analysis. For example, as feelings of depression get worse, vocal features become more gravelly, hoarse, and less fluent.

Cogito, a Boston-based startup, created an app that analyzes the human voice to help detect psychological states. In the future, it will be able to send data to the doctor if their patient’s voice patterns change significantly.

Doctor Smart has partnered with South Korea’s LangNet to research healthcare applications of voice recognition. Users of the app will be able to use their voice to ask questions and describe their problems. AI will then analyze the audio recording and provide a summary of the health issue, with keywords, to the doctor.

Besides being difficult to diagnose, depression presents another challenge, which has to do with medication.

Doctors write prescriptions, but unfortunately patients don’t always take their pills, especially those who refuse treatment. Speech recognition algorithms are capable, for instance, of detecting changes signaling that the patient has stopped taking antidepressants. 

As House M.D. says, everybody lies. It’s a good thing if the doctor is aware.