Framework for Vocal Biomarkers in AI Disease Detection Established

Researchers from the Luxembourg Institute of Health (LIH) and the University of South Florida (USF) have led an international effort to establish the first consensus-based framework and definitions for vocal biomarkers. These voice-derived indicators are being studied for their potential to help detect and monitor diseases.

The study, published in Digital Biomarkers as part of the VOCAL initiative, brought together 24 experts from Europe and North America. It addresses the challenge of inconsistent terminology in voice-based disease detection, where terms like "voice biomarkers," "speech biomarkers," and "vocal biomarkers" have frequently been used interchangeably.

The new framework distinguishes between vocal measures and validated vocal biomarkers. It introduces a hierarchical model covering the domains involved in voice and speech production. The framework provides a scientifically grounded vocabulary intended to support collaboration among clinicians, speech specialists, engineers, data scientists, regulators, and industry stakeholders.

This work, driven by the LIH-coordinated eVoiceNet and USF-led Bridge2AI-Voice consortiums, is expected to accelerate the development of voice-based technologies for diagnosing and monitoring conditions such as Parkinson's disease, Alzheimer's disease, depression, heart failure, and type 2 diabetes. The initiative aims to establish international guidelines and standards for vocal biomarker research and implementation.