Checking on your voice

Graph plotting our voice as we speak. By analysing the patterns it is possible to detect anomalies and underlying causes. Credit for image: Wevosys

It is quite normal to hear a friend speaking and noticing something is wrong by the tone in her voice. We get a hint on our friend feeling blue or being in the pink. Sometimes it is straightforward to detect a cold or a sore throat. No wonder in this.

Now researchers at several places including Mayo Clinic, SRI and Sonde Health in Boston are moving voice analyses one notch forward claiming to be able to detect a number of illness by looking at the vocal patterns picked up by a smartphone and analysing it through machine learning technologies.

In studies that have been going on for the last 5 years Charles Marmar, a psychiatrist specialised in assisting fighters returning from war zones, and Mayo Clinic researchers have been able to pick up 30 characteristics of voice patterns out of 40,000 they extracted from voice patterns and related them to a variety of conditions including PTSD (Post Traumatic Stress Disorders), TBI (Traumatic Brain Injuries) and depression.  In a way, these correlates to our abilities to detect the "mood" of our friend has she talks... The automatic analyses system proved 77% accurate in detecting PSTD in 2015 study on 39 men.

It doesn't stop here.  

At Mayo Clinic researchers are looking at vocal markers to detect heart diseases from remote, just by analysing the patient voice. They have enrolled 150 volunteers and analysed their vocal patterns. Using analytic software from Beyond Verbal they have identified 13 vocal features that can be associated to an increased risk of coronary disease.

The goal is to be able to detect the first signs of pathologies using your smartphone, as you talk everyday. A seamless guardian proactively monitoring our health.

Author - Roberto Saracco

© 2010-2019 EIT Digital IVZW. All rights reserved. Legal notice. Privacy Policy.

EIT Digital supported by the EIT