Improving Cochlear Implant Success with Artificial Intelligence with Nancy Young, MD, FACS, FAAP

March 27, 2024

Since the early 1990s, under the direction of Nancy Young, MD, FACS, FAAP, Ann & Robert H. Lurie Children's of Chicago Cochlear Implant Program has provided cochlear implants to thousands of children with hearing loss, improving their hearing and quality of life. Now, new cochlear implant research at Lurie Children's led by Dr. Young is harnessing the power of artificial intelligence to help identify which children may benefit from more intensive therapy after their implant surgery. Dr. Young explains this research and how it could improve the lives of children beyond those with hearing loss.

“We're advancing science, but the research has direct clinical implications in improving the outcomes these kids get.”

Nancy Young, MD, FACS, FAAP
Head, Section of Otology/Neurotology and Medical Director, Audiology & Cochlear Implant Programs, Lurie Children's; The Lillian S. Wells Professorship in Pediatric Otolaryngology
Professor of Otolaryngology in the Division of Pediatric Otolaryngology, Northwestern University Feinberg School of Medicine; Professor & Fellow, Knowles Hearing Center, Department of Communication Sciences & Disorders, Northwestern University School of Communication; Faculty Affiliate Member, Institute for Innovations in Developmental Sciences, Northwestern University Feinberg School of Medicine

Show Notes

Dr. Young started the Lurie Children's Cochlear Implant Program in the early 1990s following the FDA approval of cochlear implants for children. The Program now includes 20 staff members including audiologists, therapists, social workers, and educators as well as community partners. [00:01:28]
Cochlear implants can enhance spoken language outcomes in children, preserve brain structure, and aid in necessary developmental progression. Potential outcomes following cochlear implant surgery depend on factors such as: age at implantation, whether or not hearing aids were used at an early age, and the degree of hearing loss. [00:05:14]
Dr. Young is committed to translational research, which often results in direct clinical implications. With key collaborator Patrick Wong, a neuroscientist with a background in speech pathology, Dr. Young received an R21 grant from NIH and NIDCD in 2018 as well as a much larger RO1 grant from NIDCD. [00:09:11]
Previous research on children with cochlear implants has shown variation in spoken language outcomes compared to children with normal hearing. However, new research is utilizing an understanding of children's brain structure and function, MRI technology, and machine learning to make individualized predictions, with new funding expanding this to both English and Spanish learners. [11:00]
This new research utilizes proven Parent Implemented Communication Treatment (PICT), which advocates the crucial role of a supportive family environment, where parents are coached on how to encourage language development in their children. [18:05]

Additional Reading

Nancy Young, MD [00:00:00] This could be creating a whole line of translational research that could improve the lives of children beyond those with hearing loss.

Erin Spain, MS [00:00:13] This is Precision: Perspectives on Children's Surgery from the Ann & Robert H. Lurie Children's Hospital of Chicago. I'm your host, Erin Spain. On this podcast, we introduce you to surgeons at one of the country's most renowned children's hospitals to find out how they're transforming pediatric medicine. Thousands of children with hearing loss have been coming to Lurie Children's since the early 1990s to receive cochlear implants that can provide useful hearing and improve their quality of life. Now new cochlear implant research at Lurie Children's is harnessing the power of artificial intelligence to help identify which children may benefit from more intensive therapy after their implant surgery. Here with details is Dr. Nancy Young, founder of Lurie Children's Cochlear Implant Program and medical director of Audiology. She, along with other Northwestern investigators, are leading this research. Welcome.

Nancy Young, MD [00:01:14] Thank you.

Erin Spain, MS [00:01:15] Tell me about the beginning of this program. You started the Lurie Children's Cochlear Implant Program in the early 1990s. Tell me about it and the children and families that your team sees every day.

Nancy Young, MD [00:01:27] I was recruited to start an auditory program including cochlear implantation, which was new. In fact, when I started, cochlear implants were not yet FDA approved for use in children. The first FDA approval occurred in 1991, and soon after that I performed the first implantation on a child that took place in the city of Chicago. Back then, the number of children who were cochlear implant candidates was limited because you couldn't have any residual hearing at all and be a candidate. But as it became clear that the children who got the implants, who had no hearing at all, were doing so much better than the children with hearing aids, who also had a lot of hearing loss, but not quite as much that the implanted kids were far surpassing them. It became clear over time that there were many more children who could be helped by the implant, who could learn to understand when people were speaking at normal conversational levels and themselves learn to talk. We've grown tremendously since then. When I started, it was me and one audiologist. Her name was Christy Riley Roney, and she's now part of the faculty in Northwestern's Communication Sciences and Disorders. But back in the day, it was just me and her at Lurie Children's. Now we have a team of about 20 people, including seven audiologists. We have therapists who provide children with listening and spoken language therapy. We work with many community partners. We have a social worker dedicated to working with our children to support parents and families. We also have an educator of the deaf and hard of hearing who works with families and with school systems and with early intervention therapists so that we're really providing comprehensive care and education to not only families, but the education to community professionals who are supporting and co-treating with us these patients.

Erin Spain, MS [00:03:33] Can you just tell me what it was like to be a pioneer in this field, to be one of the first doing this in Chicago? There has to be a lot of ups and downs that come with being the first to do this.

Nancy Young, MD [00:03:45] I always had tremendous faith that this technology would be transformative, and would change the lives of children and their families. I can't say that everyone else felt the same way in the beginning. I can remember many people thought this would never work for children who were born without hearing. I can remember both families and professionals expressing that point of view. There also back in those days was a lot of resistance from teachers. It was surprising, actually. And there also was a lack of therapists who knew how to help develop listening and spoken language. Now that is becoming easier for families to access individuals with those skill sets. But when I started, there were three people that I could find in the state who really had any background in this, and that's why we started our own program here at Lurie. And when I started, the state did not have an early intervention program. Now there is one, and that's been very helpful so that children under age three can get services often in their home. So we try to reach out and work with those people. Over the years, we've offered courses to help those types of therapists build their skills. So it's been very exciting to see things grow and change over the years.

Erin Spain, MS [00:05:02] And the technology has improved over the years. Can you tell me about that and explain the range of outcomes that you see now after surgery and what factors play a role in these outcomes?

Nancy Young, MD [00:05:14] There are many factors that can influence the outcome, and I think you also have to be cognizant of what are the possible outcomes you're after. I mean, the implant is for providing hearing, right? And it's focused on providing children with the sounds necessary to hear the components of speech as well as other sounds too. But that's an important part of it. So providing children with sound has many potential benefits. The first level, there's just connecting us to our environment, making us aware of sounds, objects, people in our environment that connects us to other people, it connects us to the environment. It's also an important safety issue. Is a car coming, right? At a higher level, because it takes more brain processing, it is for the brain to take the sounds that are important in speech and to learn to recognize words and what's being said. And then once you can do that, you're in a much better position for your brain to learn to talk, to mimic, to say the words that you're able to hear and to understand those words. Different children can receive different degrees of benefit. For children who are born with hearing loss, we want to get them implanted at a young age. Children who are implanted early in life tend to make progress more quickly and have a better spoken language outcome. So it's wonderful when we can implant the children before they're even a year of age. But of course we have to first make sure they're an appropriate candidate. So we want the kids to be identified, their degree of hearing loss, and figure it out. We want them fit with hearing aids. We want them to get therapy. Once that's done, we can then assess if they are really getting enough information from hearing aids? And often we can tell from the degree of loss, we can predict, oh, this one or that one. We expect them to be an implant candidate. But there's still a benefit to the process that I just described. And in fact, in a recent article that we've submitted for publication, we've shown having access to hearing aids, even for those that are very deaf, has a positive impact on preserving brain structure. That impact is useful for about a year of hearing aid use. But after that, if the hearing aids are not giving you enough, then that input preserving your brain structure wanes. So we think there's great benefit to fitting these children, no matter how deaf they are with hearing aids, as soon as possible and then working them up for the implant in a timely manner. And if they're not getting enough inputs, then we want to get them implanted sooner than later. Of course, there are going to be some children who are fit with hearing aids, who have a lesser degree of hearing loss but are still possibly implant candidates. Some of those children might initially make tremendous progress with hearing aids. So we're going to monitor them. And then when they start to fall off the curve, when it becomes clear that now they're not making the progress expected, that they're falling behind their hearing peers with their listening skills, that then is the time to move ahead with the implant, because, of course, as children age, what they need to hear and understand becomes harder for a baby. They might be hearing enough that they meet certain developmental milestones in terms of listening, but it might not be enough to take them to the next stage. So we have to keep a close eye on the kids who have some hearing, but maybe not enough to keep them moving forward in the direction that they need to go.

Erin Spain, MS [00:08:57] You mentioned submitting an article for publication on some research that you've done, and research is really critically important to advancing the field. Can you just tell me a little bit about your philosophy when it comes to research?

Nancy Young, MD [00:09:11] The kind of research that I'm very excited about is considered translational research. So we're advancing science, but the research has direct clinical implications in improving the outcomes these kids get. So I was very fortunate that about 13 or 14 years ago, I met Patrick Wong while he was faculty at Northwestern University in the Department of Communication Sciences and Disorders. He is a neuroscientist. He has a background in speech pathology as well, and he's very interested in the brain and language. And one of the papers he did early on that I found fascinating was he studied adults with normal hearing who were learning a second language. And he was able to show that by looking at these adults brains with MRIs, that there were different types of brains. And he could customize the type of second language training they received in order to improve their learning. And that's the type of thing he's interested in. So when we met, we decided we needed to work together and it really has been a very positive collaboration. We received an R21 grant from NIH and NIDCD in 2018, and that's a grant for very exploratory research and it's a smaller grant. With that we were able to expand our pilot data and use that to then obtain this more recent, much larger RO1 grant from an NIDCD.

Erin Spain, MS [00:10:45] That's right. And you're working with Dr. Wong, as you mentioned, using artificial intelligence in this research project that could offer more personalized therapy plans to your patients. Tell me about the study and the use of artificial intelligence and algorithms.

Nancy Young, MD [00:11:00] If you look at past studies that were done of children with cochlear implants, and even if you looked at children that had no known health issues who seem to be typically developing other than their hearing loss, and they were implanted at a young age and they had good therapy. So these are ideal candidates to develop a good understanding of spoken language and intelligible spoken language. If you look at their outcomes and you compare them, and this has been done previously, you compare them to children born with typical hearing, that the children with implants, they're more variable in their outcomes. Some of them do as well as the children born with typical hearing. But there's a greater range of how they perform, and many of them perform in the lower end of the normal range. And people have looked for many years to try to figure out how we know who's going to do well. And the two main predictors have been younger age is better, and the other one is more residual hearing, you know, a history of more residual hearing or more residual hearing at the time of implant, the brain having had access to more hearing. Those are the two things that were positive, but that did not account for all of the variance. And that type of prediction only works when you look at a group of children. It was not accurate on the individual level. If you had a new candidate in front of you, you couldn't use just age of implant and residual hearing to accurately predict the outcome. You couldn't customize a treatment program. You couldn't say your child's going to need more support, more help. We need to design a different type of habilitation plan for your child. Since language requires more than hearing, it requires the brain. What we do in our research is take into account brain structure and function. Right now, the mainstay of technology that we're using to look at the brain, and we call this neural prediction or brain prediction. We have been using MRI, done before the cochlear implant. There are other technologies that we're starting to explore as well, but that has been the technology for our funded research. And our data has shown that using the brain is more accurate than relying on these other factors that people had relied on for 30 years that don't allow individual prediction. Well, how do you do that prediction? Well, that's where artificial intelligence comes in. And machine learning is used to build predictive models. And the models are built, so far that we have our based on English learning children. With our new funding, we're going to use information about the brain gleaned from MRI scans before implantation to build models for both English learning and Spanish learning children. We think that the models will be very similar, but we want to see if that's true. We also have other centers that will be contributing that are around the U.S. and that's important because we want to be able to show that predictive models that we build at Lurie Children's, that the same type of models could be built at other centers and would be as accurate, that it's nothing unique to our center. Now, in terms of the treatment that you alluded to, that is a new treatment arm of the study. I would like to add that our overarching goal, the reason to do prediction is not for the sake of prediction. And one of the main messages I'd like to give about this research is that our goal is to use prediction to improve the outcome, not to deny care. Right? And that's one of the dangers of doing anything with prediction, is it can be misused. You want to use this to change the outcome in a positive way. We are very fortunate that at Northwestern University we have Megan Roberts, Ph.D. She runs a wonderful laboratory. She's all about early intervention. How early intervention for children can change the outcome, the language outcome. She's previously conducted a study, the first randomized clinical controlled study that showed a particular type of therapy, PICT, Parent Implemented Communication Therapy, that showed that for children with hearing loss and with cochlear implants, this is an effective therapy. So she is leading the treatment arm of our current study. Our hypothesis is that children who are predicted to struggle more to achieve a lower level of spoken language, that those children would benefit the most from this type of therapy. So we're going to test that hypothesis. We will have children who are predicted to perform in all different levels. She won't know what group they're in and they will get this intensive therapy in addition to the usual therapies they're already receiving. That's the next step in taking prediction to the next level. So we need to improve our prediction models. Yes, we have promising results, but we're only predicting six months out. Well, there's much more to language than what happens after six months. Language evolves over years and there are many complexities and parts to learning language. So we are going to study how well our predictive models can be used to predict language outcomes at various intervals down the road. And we're going to continue to build and improve these models by adding more and more patients because the models will have better training and then will be more predictive. I think the other thing that's very exciting about our research is there's no reason to think that this concept of using information that we're non-invasively being able to obtain about the brain to predict and improve the language for children with many disorders, not just hearing loss, but children with normal hearing who are at risk for difficulty in language development. So I think this could be creating a whole line of translational research that could improve the lives of children beyond those with hearing loss.

Erin Spain, MS [00:17:31] This really is personalized medicine because you're going to be able to tailor some of the therapies for patients. Is that right?

Nancy Young, MD [00:17:39] That's exactly right. We refer to it as a predict to prescribe approach.

Erin Spain, MS [00:17:45] I want to go back to something you said before about the type of intensive therapy that you're going to be working on in the treatment arm of this research. Is the Parent Implemented Communication Treatment, and the word "parent" is in there. Can you explain that to me and how important a supportive family environment is to facilitate this language development?

Nancy Young, MD [00:18:05] Your child can get therapy, but if that's once or twice a week for half hour or an hour, there's the rest of the child's life, which when they're infant and toddler that's with their family. So how the family interacts with the child can really make a difference. The therapy that they're receiving, it's actually the parent that's getting the therapy in the sense that the child will not be directly worked with; the parent will receive an education on how to more effectively work with their child. The therapy is really parent coaching and how to improve their skills at encouraging language in a way that's particularly important.

Erin Spain, MS [00:18:50] And you also mentioned that you would be looking in this new phase of the study of English and Spanish learning children. Why is it so important to now be expanding into Spanish learning children as well?

Nancy Young, MD [00:19:02] We'd like these models to someday be useful clinically and to be used across the country and across the world. We can't just assume that models that are built for English learning children are going to work for other languages. Now Spanish is a romance language. We think it will work equally well for Spanish, but we want to prove that, especially because in the United States, Spanish is a very common language. But there would be great benefit to doing the same thing for other languages, including, for instance, Chinese, which worldwide there are many people that speak Chinese. There are many languages that could be eventually evaluated.

Erin Spain, MS [00:19:45] What would you say to anyone who may be a little leery of using artificial intelligence and medical research? What's your experience been like?

Nancy Young, MD [00:19:54] Well, I think the devil's in the details. I think artificial intelligence is just a tool and it's not a single thing. I mean, we're using a specific type of artificial intelligence called machine learning, but it's really how you're using it. So, for instance, we're using it to predict for the purpose of improving the outcome. But someone else could say, well, we're going to predict and then we're going to deny care based on who's going to have a poor outcome. That's unethical, in my opinion. It would be wildly inaccurate when you're dealing with very young children, because how a child does has a lot to do with the intervention that they get. That's the whole reason for early intervention therapies, is because we can do so much to change the outcome because kids have so much neuroplasticity. But the other issue is, because we're interested in language outcomes, our research is focused on understanding speech and spoken language after an implant. Right now that's our focus. But I think it's important people realize that there are so many benefits to hearing beyond spoken language and even beyond understanding spoken language. So, Lurie Children's, our Cochlear Implant Team is one of the pioneers in providing cochlear implants to children with very complex medical problems, including problems associated with cognitive issues. I have always felt very strongly that if we can give a child back one of their senses, we should do that. It's about how they learn about the world. And we have children who have gotten the implant and some of them may never speak, but the parents will say, this is our way in, this is how we engage with our child. Our child rocks to music. Our child laughs when we talk to them. That is just so important. And you know, the way you engage with others in your life. And then I have children that would have been denied an implant because of these diagnoses. And some of them really surprise us that, in fact, the child we thought was wildly unlikely that they would ever understand speech, much less talk, in fact, does both. And I saw early on in my practice that there was a tremendous lack of predictability, especially when you started implanting these more complex children. And that's what really made it clear to me that this was much more about the brain than anything else. And that's why I'm so grateful that I met Dr. Wong at Northwestern. He's now in Hong Kong and has a very large research lab there. But we have been able to continue to collaborate and we actually have other collaborators around the world who are working with us.

Erin Spain, MS [00:22:57] What motivates you to keep pursuing these different ways to improve the lives of children beyond surgery?

Nancy Young, MD [00:23:03] It's extremely rewarding. I'm just very grateful that I was born at the right time. I mean, it is such a blessing to be able to be involved in something like this and to see it unfold, you know, to be part of learning about what is possible and to be able to give help to people when at the time no one thought it was possible. That is extremely motivating and extremely rewarding. Our implant team, very few people leave once they're on the implant team. It's extremely rewarding. No matter what role you play on the team, you really feel that you're making a difference.

Erin Spain, MS [00:23:40] Well, thank you so much, Dr. Nancy Young, for coming on the show and talking about not only the program where you're treating patients, but the cutting edge research taking place as well. I appreciate your time.

Nancy Young, MD [00:23:52] Thank you.

Erin Spain, MS [00:23:55] For more information, including how to make a referral or an appointment, visit LurieChildrens.org