Today, the biomedical research community generates a wealth of data; however, because the data are generated for analysis using other computational and statistical tools, most are not suitable for machine learning. The Bridge2AI program brings together AI and biomedical experts to collaboratively solve these problems. (Getty Images)
The National Institutes of Health, or NIH, today announced funding for three projects with Oregon Health & Science University leaders that will use data to improve health outcomes. The NIH Common Fund’s Bridge2AI program supports the widespread adoption of artificial intelligence, or AI, in biomedical research to address ambitious issues facing researchers today.
One of the program’s key goals is to generate new “flagship” data sets and best practices for machine-learning analysis. Machine learning is a type of AI that provides machines, such as computers, the ability to automatically learn and improve from their experiences. Today, the biomedical research community generates a wealth of data; however, because the data are generated for analysis using other computational and statistical tools, most are not suitable for machine learning. The Bridge2AI program brings together AI and biomedical experts to collaboratively solve these problems.
Shannon McWeeney, Ph.D. (OHSU)
Shannon McWeeney, Ph.D., chief data officer for the OHSU Knight Cancer Institute, will co-lead tool development to support one of the “grand challenge” projects in the Bridge2AI program, aimed at generating new biomedical and behavioral data sets that are ethically sourced, trustworthy, well-defined and accessible. The collaborative team will leverage $7.8 million in funding for the first year to develop software and standards to unify data attributes across multiple data sources and types.
“Moving the field of AI forward is essential to help detect and treat earlier diseases like cardiovascular disease, diabetes and cancer,” says McWeeney, a professor and head of the Division of Bioinformatics and Computational Biology in the OHSU School of Medicine. “The ability to understand and affect the course of complex, multi-system diseases has been limited by a lack of well-designed, high-quality, large, and inclusive multimodal datasets. We need transparency about how the data are generated with regard to any bias or uncertainties, and to ensure they are ethically sourced. We also need to lower the barrier for researchers to be able to use AI-based tools in their future research.”
McWeeney is co-leading these efforts with OHSU School of Medicine and Casey Eye Institute leaders Michelle Hribar, Ph.D., associate professor of medical informatics and clinical epidemiology (DMICE), and Hiroshi Ishikawa, M.D., professor of ophthalmology.
In addition to OHSU, other institutions collaborating on the Data Generation project include: University of Washington, California Medical Innovations Institute, Johns Hopkins University, University of California at San Diego, University of Pennsylvania, Stanford University, Native BioData Consortium, University of Alabama at Birmingham, University of Mississippi Medical Center, Henry Ford Health System and Microsoft.
This Bridge2AI Data Generation project, McWeeney says, is a natural extension of OHSU’s collaboration with Microsoft on the Cascadia Data Discovery Initiative, which is focused on enabling collaboration, data sharing and data-driven research.
David Dorr, M.D., M.S. (OHSU)
David Dorr, M.D., M.S., chief research information officer and professor of medical informatics and clinical epidemiology in the OHSU School of Medicine, will co-lead another of the grand challenge projects, “Skills and Workforce Development,” with a team from Washington University in St. Louis.
This module will be centered on bridging expertise across people in the biomedical and behavioral research domains to develop an AI/machine-learning research workforce. Dorr says this project is designed to enhance skill development and attract and develop a specialized workforce. In particular, he says, the project is focused on adding voice as a biomarker to AI models.
“Variations in sound production via our vocal cords, resonators and articulators, and the underlying cognitive processes that help us produce language can help diagnose a wide variety of illnesses, including laryngeal cancers, neurological disorders such as Parkinson’s, mood disorders, depression and schizophrenia, and speech and language differences in childhood, including neurodiverse conditions such as autism,” says Dorr. “We will aim to build an ethically sourced database of diverse human voices while protecting patient privacy.”
The “Voice as a Biomarker of Health” data generation project, co-led by Yael Bensoussan, M.D., from the University of South Florida and Olivier Elemento, Ph.D., from Weill Cornell Medicine, is projected to receive $3.8 million in the first year from the NIH Bridge2AI project. Using this data, Dorr says machine learning models will be trained to spot diseases from the human voice, which could empower clinicians with a low-cost diagnostic tool to be used alongside other clinical methods.
OHSU collaborators on this module include two DMICE leaders: professor and DMICE chair William Hersh, M.D., who will help co-create the curriculum and establish a competency framework; and, associate professor Steven Bedrick, Ph.D., who will focus his contributions in education around machine learning and AI, and will also bring computational expertise in speech and language disorders to the project.
Bridge Center: Tool optimization
Kyle Ellrott, Ph.D. (OHSU)
Kyle Ellrott, Ph.D., assistant professor of biomedical engineering in the OHSU School of Medicine, is part of the Bridge Center, which will coordinate the activities across all of the grand challenge projects in the Bridge2AI program. The tool optimization core will work to support open and coordinated tool development within and beyond Bridge2AI. Ellrott’s work will focus on ensuring tools developed for the grand challenges are reproducible and scalable for cloud infrastructure.
“The most innovative and cutting-edge AI methods are worthless if they can’t be replicated,” says Ellrott. “And with the scale of data that is becoming available, it is critical that tools can be efficiently deployed to cloud scale computational resources. Whether talking about heart disease, Alzheimer’s or cancer. these issues are constant. Our job will develop systems that enable all of the Bridge2AI data producers to work efficiently and reproducibly.”
OHSU will work with teams from UCLA and Sage Bionetworks to support this large ecosystem of AI data sources. This group is expected to receive more than $9.9 million over the next four years. This team will build off of the success of the Dialogue on Reverse Engineering Assessment and Methods (DREAM) challenges. These benchmarking efforts, established in 2006, have identified the best
methods for biomedical analysis by allowing anyone in the research community to participate in unbiased and open evaluations.
These projects are supported by the National Institutes of Health award numbers: OT2OD032644, OT2OD032720, and U54HG012517. More information on the Bridge2AI program: https://commonfund.nih.gov/bridge2ai/faqs