Prof. Washington receives $2.18M NIH New Innovator Award

Original story from https://www.hawaii.edu/news/2023/10/03/artificial-intelligence-detect-conditions/:

Wouldn’t it be helpful to have a tool that can help identify potential cases of mental health conditions and developmental issues? Imagine the amount of people that could be helped earlier with powerful instruments and aids as part of that diagnosis.

Diagnosing mental health conditions and developmental delays such as autism and ADHD in individuals through video may one day be possible using artificial intelligence (AI) and is the focus of a new University of Hawaiʻi at Mānoa project. This public impact research by Assistant Professor Peter Washington from the Information and Computer Sciences Department is funded by a $2.18-million New Innovator Award from the National Institutes of Health (NIH).

The New Innovator Award is part of NIH’s High-Risk, High-Reward Research program. Washington’s research grant is one of 85 announced by the program in 2023 totaling approximately $187 million to support highly innovative scientists who propose visionary and broadly impactful behavioral and biomedical research projects.

This grant will provide five years of funding to Washington’s research lab, the Hawaiʻi Digital Health Lab. Washington’s lab develops new AI techniques together with smartphone apps and websites to help advance diagnostic and treatment options for a variety of healthcare conditions.

AI-diagnosed mental health conditions, developmental disorders

While AI is increasingly used to diagnose a variety of health conditions, current AI cannot understand complex social human behavior from videos, and therefore is unable to diagnose mental health conditions and developmental disorders. Washington’s project will create AI models that work together with a group of people to establish computational diagnoses of subjective, complex and social conditions in psychiatry using videos recorded while playing video games. The subjects who are participating in the human-AI collaboration process receive extensive privacy and health ethics training.

Using the data collected, the project aims to create efficient, remote and accessible psychiatric evaluations from the comfort of a home computer.

“Diagnostics for childhood and adolescent psychiatric conditions are inaccessible to many populations due to long waiting lists, lack of health insurance and the absence of proximity to the nearest available and affordable clinician,” Washington said. “We hope that the project will result in convenient and affordable psychiatric healthcare for families living in places such as rural Hawaiʻi.”

In particular, Washington’s project will focus on diagnosing adolescent developmental delays, such as autism and ADHD as a starting point towards building a method that can work for a broader range of psychiatric conditions. Providing diagnoses remotely through a computer in a mostly automated way can help families receive quicker care. According to Washington, the use of video games that are useful for AI analysis can make the traditionally burdensome diagnostic process fun and convenient for children and adolescents.

“In addition to its healthcare impact, this project is significant from a computer science perspective because it involves using AI working together with a remote crowd of humans to measure human behavior precisely enough to predict a psychiatric diagnosis,” Washington said. “Achieving precise measurements about complex social human behavior from structured videos using a combination of AI and crowdsourcing can have far-reaching applications beyond diagnosing psychiatric conditions, including multimedia and social media analytics as well as educational technology.”