The School of Medicine joins an initiative of the National Institutes of Health to expand the use of artificial intelligence in biomedical research – Washington University School of Medicine in St. Louis

Visit the news center

A project to develop a framework for using artificial intelligence to diagnose disease based on the voices of patients

Getty Images

Imagine that one day in the future, doctors could diagnose throat cancer, Alzheimer’s, depression, or other illnesses based on the patient’s voice. To help make this a reality, Washington University School of Medicine in St. Louis is joining the National Institutes of Health (NIH) Bridge2AI Program, an estimated $130 million initiative aimed at expanding the use of artificial intelligence (AI) in biomedical and behavioral research.

One of the first projects involves building a database of diverse human voices and harnessing artificial intelligence and machine learning tools to train computers to identify diseases based on the characteristics of the human voice. The effort – called Voice as a Biomarker of Health – will bring together researchers from 12 North American institutions, including the University of Washington, to build the database, which will be ethically sourced as well as protect patient privacy.

“There is evidence that well-designed computer models can predict who has dementia or cancer, for example, based on audio recordings, which will then complement additional diagnostic methods,” Philip Rowe Payne, Ph.D.Janet and Bernard Becker Professor, Chief Data Scientist and Director Informatics Institute. We will also lead new efforts in education and workforce development in artificial intelligence and its applications in biomedicine. As part of that, this project will help define an entirely new way to produce and share these kinds of complex data sets – in ethical ways that protect privacy – with a diverse group of scientists.”

Payne is leading the project at the University of Washington and collaborating with investigators across North America, including the University of South Florida, in Tampa, and Weill Cornell Medicine in New York, who are leading the project nationwide.

In addition to building this unique dataset, the University of Washington will co-lead the Center for Skills and Workforce Development for the national project. The core center, which it shares leadership with Oregon Health and Science University, will focus on training investigators — including scientists from academia, industry, government, and even citizen scientists — from across the country to be able to access and use audio data for research. According to Payne, any researcher seeking to learn how to use a data set would have an individualized learning plan with much of the learning presented in virtual form and then backed up with one-on-one mentoring.

“Often, citizen scientists, or the people we include in this category, are patients with diseases themselves or specialists in private practices who help patients with specific conditions, such as people who stutter and the speech pathologists who work with them,” Payne said. “We are developing outreach efforts to connect with people in the community to participate in this research and also to help us collect a rich and diverse data set of human voices. This is vital to building an ethical and representative data set that eliminates potential bias.”

Based on the current literature and ongoing research, the research team identified five disease categories in which voice changes have been associated with disease and there is an urgent need to improve early diagnosis. The data collected for this project will focus on the following disease categories:

  • Voice disorders (laryngeal cancers, vocal fold paralysis, benign laryngeal lesions).
  • Neurological and neurodegenerative disorders (Alzheimer’s, Parkinson’s, stroke, amyotrophic lateral sclerosis).
  • Mood and mental disorders (depression, schizophrenia, bipolar disorders).
  • Respiratory disorders (pneumonia, chronic pulmonary obstructionheart failure).
  • Voice and speech disorders in children (delayed speech and language, autism).

Although initial work with audio data has been promising, limitations in incorporating audio as a biomarker in clinical practice have been linked to small data sets, ethical concerns regarding data ownership, privacy, bias, and a lack of data diversity. To solve these issues, The Voice as a Health Biomarker Project is working to create a large, high-quality, multi-organization, and diverse audio database linked to biomarkers that are identity-protected and not identifiable from other data, such as demographics, medical imaging and genomics. Unified learning technology – a new framework for artificial intelligence that allows machine learning models to be trained on data without leaving the data from its source – will be deployed across multiple research centers by French-American AI biotech startup Oken to demonstrate that multi-center AI research can be conducted while maintaining on the privacy and security of sensitive voice data.

“Using voice to help diagnose disease becomes especially interesting when you consider the prevalence of virtual care and telemedicine during the pandemic,” Payne said. “Doctors are becoming more accustomed to seeing people from afar even though they cannot physically examine patients. But what if, during a virtual visit, there is an AI algorithm that can identify high blood pressure, for example, based on the patient’s voice. There is a dimension, but this may make telemedicine in the future more useful, with higher quality, better safety, and improved health outcomes, especially for people who live far from healthcare providers.”

Supported by AI experts, bioethicists, and sociologists, the project aims to transform basic understanding of diseases and introduce a new way of diagnosing and treating diseases in clinical settings. Because human voice recordings are cheap, easy to store, and readily available, diagnosing diseases through voice using artificial intelligence could be a transformative step in precision medicine and access to healthcare, Payne added.

This work is supported by the National Institutes of Health (NIH), grant number 1OT2OD032720-01.

About the University of Washington School of Medicine

Wash the medicine is a global leader in academic medicine, including biomedical research, patient care, and educational programs with 2,700 faculty members. The National Institutes of Health (NIH) research funding portfolio is the fourth largest among US medical schools, and has grown by 54% in the past five years. Combined with institutional investment, WashU Medicine commits over $1 billion annually to basic and clinical research, innovation, and training. . Its faculty practice is consistently among the top five in the country, with more than 1,790 faculty physicians working in more than 60 locations who are also the medical staff of the Barnes is a Jew And the St. Louis for children Hospitals BJC HealthCare. WashU Medicine has a rich history of MD/PhD training, recently committed $100 million in scholarships and curriculum renewals for medical students, and is home to top-tier training programs in every medical subspecialty as well as physical therapy, occupational therapy, audiology, and communication sciences.

Leave a Comment