Study: 'Vocal biomarkers' can predict likelihood of heart disease

Voice analysis specialist Beyond Verbal and Mayo Clinic unveiled results from a new study demonstrating a strong correlation between some voice characteristics and the presence of coronary artery disease (CAD).

Beyond Verbal, which bills itself as an emotions analytics company, focuses on extracting as much information as possible from the tone of voice of a speaker, said CEO Yuval Mor. Voice analysis may be used to understand emotions and well-being, he said. We can tell with the human ear, for example, if a person is feeling well or not. The new study data confirms that we can also understand a medical condition by using voice feature analysis.

The double-blind study involved 120 patients who had been referred for elective coronary angiography and 120 control patients. Beyond Verbal provided Mayo Clinic with acoustic features, and the latter zeroed in on some characteristics that could “strongly and independently” be associated with CAD, Mor said. In particular, one feature was linked to a 19-fold higher likelihood of CAD, the companies said in a statement.

BD&L Summit

Deal-Making Insights for the Life Sciences Industry

Bringing together key deal-makers and serving as an open forum for cross-functional business development and legal teams to share valuable insights and actionable strategies on successfully managing alliances, licensing agreements, and M&A deals.

Beyond Verbal has previously discovered links between voice characteristics and neurological conditions, such as Parkinson’s disease, but this study is the company’s first foray into deploying the tech in a non-neurological condition. It is also the world’s first to indicate a link between voice features and CAD, according to the statement. It could provide the foundation of a novel, noninvasive and simple test to estimate a patient’s risk of CAD.

Beyond Verbal will continue to work with Mayo Clinic, eventually testing the voice-analysis tech in different languages to confirm that voice-characteristic analysis transcends language, Mor said. Additionally, this research shows that voice analysis can be useful in conditions that are not traditionally associated with the voice. To drive research into identifying vocal biomarkers in other conditions, Beyond Verbal recently launched its mHealth Research Platform.

Other players looking into vocal biomarkers include PureTech’s Sonde Health, which licensed its audio-analysis tech to an MIT lab to help develop its diagnostic tool. Like Beyond Verbal’s tech, Sonde’s platform analyzes nonlinguistic characteristics of voice samples. The company hopes to apply it to mental disorders, such as depression, as well as physical conditions, including cardiovascular conditions.