Collaboration Stories

The purpose of this series is to tell the story of long-lasting, as well as emergent, stories of collaboration among clinicians and methodologists, content experts, and engineers, who have successfully worked on joint projects at the intersection of AI and health. Such collaborations are at the center of e-HAIL’s mission, and we offer these stories as a way of inspiring collaboration in other researchers.

“It’s About the Patient’s Need”: AI and Audiology Collaboration for Hearing Loss and Patient Accessibility

Article by Hossam Abouzahr, M.A., M.P.P., e-HAIL Grant Writer

Drs. Devin McCaslin and Dhruv “DJ” Jain didn’t know each other until e-HAIL’s program manager introduced them for a joint research project on pediatric hearing loss. Although they’ve known each other less than six months, talking to them together reveals the amicable relationship they’ve built by pursuing their joint research goals on hearing accessibility and hearing related conditions. Their first grant was recently awarded: creating an open-source database on dizziness and developing a machine learning algorithm to analyze symptom patterns and improve diagnosis. Jain and McCaslin have since developed their research and relationship, resulting in additional projects and grant applications.

With support from e-HAIL, they put together a team to explore opportunities to create a pediatric digital ear database that will include 3D ear scans and hearing tests from children with and without hearing loss. Good data is one of the biggest challenges for applying AI solutions to the medical field, and their database would set the groundwork to revolutionize the field of pediatric audiology by enabling research on predicting ear growth and changes in acoustics in children with hearing loss as well as optimal approaches for interventions to support those children. With support from e-HAIL’s grant writer, they submitted the grant application this January.

Jain and McCaslin are also eager to pursue a project related to balance testing. “Balance testing,” McCaslin explained, “is all done in the dark, and patients are asked to perform certain tasks like follow lights on a screen while we measure their eye movements.” However, the patients being tested often “have inner ear balance problems and also inner ear hearing problems,” and so they face particular challenges, such as when trying to hear the doctor’s instructions. McCaslin and Jain want to explore building an integrated, augmented transcription system to enable the doctor to communicate with the patient. Jain added, “For me, the joy is in deploying the solution in healthcare settings and greatly improving access for the patients who are deaf and hard of hearing, but relevant research questions include: what medium—augmented glasses, a tablet, or a phone—would work best? How should the medium be set up—mounted on the wall, between the patient and clinician?” 

McCaslin is the Director of Audiology at Michigan Medicine and completed his Ph.D. in hearing science at Ohio State University, where his mentor got him interested in the workings of the inner ear and its impact on dizziness. His interest in applied AI started when he was working at the Mayo Clinic. McCaslin suggested developing an algorithm that would review a patient’s record and figure out which appointments were needed and push it to a scheduler, saving time and adding consistency and transparency to a process originally done by hand. The system was so successful it was sold to a startup and became known as Dizzy Guide. He has continued this vein of research, and in 2023 he received a Research Scout award to develop an AI-powered triaging model for patients with dizziness and vertigo.

Jain’s Ph.D. is in human-computer interaction (HCI) with a focus on accessibility, although he actually sees technology as a “last resort” for solving accessibility challenges. Rather, he said, as head of the Soundability lab, “We first try to find a non-technical solution to every problem out there.” His lab adopts a broad definition of accessibility—it is not only for people with disabilities but includes tailored delivery of information for all users. He credits McCaslin with defining a key aspect of the Soundability lab’s direction: “AI for audio healthcare—that’s actually one of the research areas we put on the lab website now.” He continued, “The original goal for our lab was to design and evaluate technology targeting any area of accessibility. But now I think we are more focused in what we do, and we’ve renamed the lab to Soundability.” Thinking about HCI and accessibility requires understanding “the full pipeline,” from “understanding user needs to designing technology to meet those needs and then evaluating the success of the technology.” He adds: “This end-to-end user-centric research is the strength of our lab that leads to highly usable and deployable solutions.”

When asked to comment on AI’s prevalence, “There are a lot of opportunities you can use AI for, especially for sound, because sound is so varied, so diverse.” Under his direction, the lab researches how sound awareness technologies can model a user’s intent and deliver context-aware sound feedback, how generative AI can improve accessibility to audio-based media, and how to improve hearing-aids and earphones to extract desired sounds and suppress unwanted noise to create seamless hearing experience. 

When I asked about how they keep the relationship fulfilling for both of them, Jain was quick to state, “All the credit goes to Devin,” adding, “He’s always respectful of our research goals. He gives us space to innovate and explore our own research questions beyond the questions that are directly relevant to the project at hand.” On his end, McCaslin said, “DJ sees not just the AI and engineering, but also all aspects of sound and accessibility. He’s not just an engineer; he has a whole background in sound, design, and user research and understands the accessibility issues that we’re seeing in the clinic.” Reflecting back on how Jain and his lab first look for non-technical solutions, McCaslin added, “Jain and I work well together because he deeply understands the patients’ side. We both see technology as a means to get patients the fastest diagnosis for the least cost and with the best outcomes.”