Real-world, Ambulatory Monitoring of Vocal Behavior to Improve the Assessment and Treatment of Voice Disorders
Daryush Mehta
March 31, 2021, Wednesday, 3:00 PM - 4:00 PM EDT
Abstract
An estimated 30% of the adult U.S. population suffer from a voice disorder at some point in their lives and often experience significant communication disabilities with far-reaching social, professional, and personal consequences. Most voice disorders are chronic or recurring conditions and are associated with inefficient patterns of vocal function, termed vocal hyperfunction. Thus, an ongoing clinical research goal is the prevention, diagnosis, and treatment of vocal hyperfunction through noninvasive, long-term monitoring of an individual's daily voice use. During this talk, I will present our team's ongoing work applying machine learning to the study of vocal hyperfunction in patients and matched healthy controls using wearable devices. Voice use and vocal function features are derived from neck-surface vibration recordings using vocal dose theory and novel impedance-based acoustic modeling. Big data approaches have proven critical to sifting through the real-world data streams. Results are encouraging as they bring us closer to improving the clinical treatment of vocal hyperfunction by deriving statistical models that can be integrated into real-time biofeedback paradigms, thereby facilitating learning of healthier vocal behaviors. I will also discuss our efforts to enhance clinical voice assessment through integrating signal features from wearables capturing electrodermal activity and environmental noise levels.
Bio
Daryush Mehta is Assistant Professor of Surgery, Massachusetts General Hospital, and Director, Voice Science and Technology Laboratory, Center for Laryngeal Surgery & Voice Rehabilitation, Department of Surgery, Massachusetts General Hospital.

I continue research efforts into the clinical analysis of normal and disordered voice production with particular emphasis on advanced statistical signal processing algorithms and ambulatory monitoring of daily voice use. My work bridges the areas of statistical signal processing and clinical voice assessment.

I investigate the details of the relationship between the motion of the vocal folds (the "voice box") and the acoustics of voice production. My expertise is in signal processing and acoustic voice analysis, and I bring these engineering tools to clinical voice research. We have developed a comprehensive laryngeal high-speed videoendoscopy system to image and quantify vocal fold vibratory characteristics and relate them to voice-related sensor measurements and mathematical models. Other major efforts develop a smartphone platform for tracking voice use using an accelerometer taped to the neck.

We hope that our results will aid voice surgeons and speech-language pathologists in better understanding the mechanisms of normal and disordered voice production.