Researchers are exploring how voice biometrics could offer clues into clinical conditions like Post Traumatic Stress Disorder (PTSD) and heart disease, reports MIT Technology Review‘s Emily Mullin.
A team at New York University’s Langone Medical Center, led by psychiatry head Charles Marmar, is working on machine learning algorithm software that can detect vocal patterns related to PTSD, depression, and traumatic brain injuries. A preliminary test using 39 volunteers showed that the technology could identify diagnosed PTSD cases with an accuracy of 77 percent; and the technology has been advancing since those results were published in 2015, so there’s a good chance that there are even stronger results to come.
Meanwhile, researchers at the Mayo Clinic has teamed up with Israel-based Beyond Verbal to develop a machine learning voice biometrics system that could detect heart issues, based on the reasoning that associated chest pain could affect the voice in subtle ways. An initial study with 150 patients found 13 unique vocal patterns associated with heart disease, with one researcher asserting that “specific segments of the voice can be predictive of the amount or degree of the blockages found by the angiography,” or x-ray scanning.
The studies complement other medical research exploring the as-yet-undiscovered clinical applications of biometric technology, which can detect medical cues that human perception might miss. That’s good news for doctors and patients alike.
Source: MIT Technology Review
January 24, 2017 – by Alex Perala