Episode 413: How AI Will Impact MS Patient Care with Dr. Brad Willingham

Summary

---

RealTalk MS — Episode 413: How AI Will Impact MS Patient Care with Dr. Brad Willingham

Date: July 28, 2025
Host: Jon Strum | Guest: Dr. Brad Willingham

---

Episode Overview

This episode of RealTalk MS explores the transformative role of artificial intelligence (AI) in multiple sclerosis (MS) care and research. Host Jon Strum welcomes Dr. Brad Willingham, Director of MS Research at Shepherd Center, to discuss how AI is moving from a conceptual idea to practical implementation in MS research, clinical practice, and rehabilitation. They discuss the ways AI is improving everything from data interpretation and personalized care to how patients and clinicians communicate and make decisions.

---

Key Discussion Points & Insights

The Three Domains of Medicine and AI’s Role

[15:34]

• Science of Medicine: AI enhances scientific discovery by processing vast amounts of data—especially imaging—to uncover patterns undetectable by humans.
  • “Our scientific knowledge grows directly as a function is our ability to see and observe things. And so one of the things AI is really good at is ingesting large amounts of information and recognizing patterns...” — Dr. Willingham [15:52]
• Practice of Medicine: AI supports medical professionals by helping them keep up with new research and providing clinical decision support.
  • “AI tools such as open evidence are making that a reality...they have access in lay language...and get guidance.” — Dr. Willingham [17:37]
  • Dr. Willingham prefers the term “augmented intelligence,” emphasizing AI as a support for clinicians, not a replacement.
• Delivery of Care: AI bridges gaps between clinicians and patients by making data and health information accessible and understandable to both.
  • “Arguably, one of the greatest challenges...is to make sense of all the data we have available to us.” — Dr. Willingham [18:38]

Patient Empowerment & Shared Decision-Making

[20:03–22:39]

• AI's capacity to break down and explain complex medical language means patients can better understand their condition and options.
  • “It will make that communication...that's usually full of all sorts of buzzwords and jargon...a thing of the past.” — Jon Strum [20:03]
• Patients become “co-creators” in their care, engaging in feedback loops via self-monitoring technologies.
  • “Patients are not merely an endpoint user but an active participant and a co creator in their health data, in their monitoring and decision making.” — Dr. Willingham [20:35]
• AI can tailor information output: clinicians get clinical insight, patients get actionable summaries geared to their goals and understanding.

Demonstrating AI’s Transformative Power

[22:39–24:33]

• Jon recommends trying Google’s NotebookLM with a difficult academic article to experience how AI translates dense research into understandable summaries.
  • “Ask...‘I'm someone living with MS, I'm a non-scientist. Can you explain this to me?’...it is transformative.” — Jon Strum [22:39]
• Dr. Willingham likens the impact of large language models (LLMs) to the rise of the iPhone—ubiquitous implementation is “just a couple years” away.

Digital Twin Technology & Personalized Care

[24:44–27:35]

• Digital Twins: Creating a comprehensive digital model of a patient, integrating diverse health data, to simulate responses to various treatments.
  • “We can rapidly simulate thousands of different treatment strategies for a single patient...we're already way ahead in terms of clinical practice.” — Dr. Willingham [25:37]
• Potential for clinical research: using digital twins as control groups in “in silico” trials to address variability in human studies.
• This technology accelerates precision rehabilitation—applying the principles of precision medicine (long present in genomics) to rehabilitation and functional goals.

AI in Telerehabilitation and Remote Care

[28:24–31:05]

• Remote Strategies: AI can synthesize data from multiple at-home devices and apps and deliver actionable summaries or next steps to patients, minimizing barriers like transportation, access, or geography.
  • “We're looking at...can AI not only summarize data but can it serve as a remote care coach...to provide encouraging updates or provide suggestions about your next steps.” — Dr. Willingham [29:14]
• Clinical decision support can be enhanced with AI “coaches.” There’s a need for caution regarding safety and accuracy, leading to rigorous testing and customization.
• Real-time feedback using AI-driven computer vision: e.g., video-based platforms analyzing movement and giving corrective cues.
  • “The AI agent...will give you cues...‘squat a little lower, raise that left arm a little higher.’” — Dr. Willingham [31:05]

Current MS Rehabilitation AI Research at Shepherd Center

[31:46–32:51]

• Focused on large language models for data summarization and digital twins for tailoring rehabilitation.
• Dr. Willingham emphasizes the importance of engaging both patients and clinicians as technologies develop, ensuring new solutions meet real-world needs.
  • “As we're innovating...I think it's important that we also think about how we make that data meaningful and engage folks living with MS early and often in your research.” — Dr. Willingham [32:37]

---

Notable Quotes & Memorable Moments

• “One of the things AI is really good at is ingesting large amounts of information and recognizing patterns and making observations in there.” — Dr. Brad Willingham [15:56]
• “I like to think more of it as augmented intelligence rather than artificial intelligence. You still need that clinician there, but it definitely can facilitate.” — Dr. Brad Willingham [17:40]
• “AI has immense potential to help extract meaning from all the data we have in today's healthcare environment, but also make the information accessible to both clinicians and patients.” — Dr. Brad Willingham [18:15]
• “Patients are not merely an endpoint user but an active participant and a co creator in their health data...actively engaged in those feedback loops.” — Dr. Brad Willingham [20:35]
• “When you do [try out LLMs], it is transformative. You will see everything in your world a little differently going forward.” — Jon Strum [24:33]
• “A digital twin is essentially a digital model of either a person or a system...We can use, we can train that system on historical data...simulate different treatment strategies with that digital twin.” — Dr. Brad Willingham [25:07]
• “This is the realization of precision rehabilitation...getting a deep understanding of the individual's needs, preferences and goals, and using that to create targeted treatment strategies in a data driven manner.” — Dr. Brad Willingham [27:41]
• “Are you doing it right? Am I improving? What are the next steps? ... AI is going to be able to...help guide a patient and provide information about, hey, this is how you're doing towards your goals.” — Dr. Brad Willingham [28:37]
• “Squat a little lower, raise that left arm a little higher. There are ways we can also get real time feedback through AI applications such as computer vision.” — Dr. Brad Willingham [31:05]

---

Important Timestamps

• Intro & Setting the Stage: [00:18–01:33]
• AI Overview & Domains of Medicine: [15:34–19:26]
• Patient Communication & Empowerment: [20:03–22:39]
• Experiencing AI in Practice (NotebookLM Example): [22:39–24:33]
• Digital Twin Technology Explained: [24:44–27:35]
• Precision Rehabilitation with AI: [27:35–28:24]
• AI in Telerehabilitation: [28:24–31:41]
• Current and Future AI Research Directions: [31:46–32:51]
• Closing Thoughts: [32:53–33:10]

---

Summary Takeaway

This episode delivers an insightful roadmap for how AI is poised to revolutionize MS care—from turbocharging research, aiding clinicians, personalizing rehabilitation, to empowering patients to be true partners in their own health journeys. Dr. Willingham and Jon Strum break down both the practical innovations arriving today and the future possibilities—from digital twins and “AI coaches” to real-time feedback and data clarity—making it clear that the AI revolution in MS care is not just hype, but an emerging reality.

---

Transcript

[A]

I'm john strum and this is real talk, mississippi.

It's July 29th and we have a lot to talk about. Over the past couple of years, we've often talked about ways that artificial intelligence could impact Ms. Research and Ms. Care. Now we're seeing the beginning of that transition when things are starting to move from the conceptual phase to the actual phase, with much more to come. And joining me in a freewheeling conversation about some of the ways that AI will impact, well, so many areas of healthcare research and patient care is Dr. Brad Willingham. As the Director of Ms. Research at Shepherd center in Atlanta, Dr. Willingham focuses on the development of innovative technologies and strategies to advance neurorehabitation rehabilitation, and that role puts him at the forefront among those pioneers who are crafting new paradigms in research and patient care, paradigms that will leverage artificial intelligence to create better outcomes for people living with Ms. And I'll venture to add, just about every other chronic illness. But before we get to my conversation with Dr. Willingham, there are a few other things that you should know about.

Phenibrutinib is an investigational oral disease modifying therapy being developed by pharmaceutical maker Genentech. Like Tolebrutinib, which is awaiting FDA and EMA approval this fall, phenibrutinib is part of a new category of DMTs called BTK inhibitors. BTK inhibitors have been shown to be an effective treatment for some forms of cancer and they represent a new category of disease modifying therapies with some interesting characteristics. Now the BTK stands for Brutons tyrosine kinase, which is a protein that plays an important role in B cell activation and inflammation. A BTK inhibitor blocks the activity of this protein and by inhibiting btk, phenobrutinib is expected to reduce inflammation and prevent further damage in Ms. The phase 2 clinical trial for phenobrutinib showed that it significantly reduced the occurrence of new brain lesions with active inflammation in people with relapsing multiple sclerosis. Now, results of an open label extension study have shown that after a year of treatment, 90% of the extension study participants had no new lesions, no relapses and and no sustained disability worsening. Currently, there are two Phase 3 clinical trials testing phenibrutinib against Obagio to determine whether phenobrutinib is better at reducing relapse rates after two years. There's also a clinical trial underway comparing phenobrutinib with ocrevus as a treatment for primary progressive Ms. We'll keep you posted as pharmaceutical manufacturer Genentech provides updates on these clinical trials. And if you'd like to learn more about this open label extension study, you'll find that link in today's show. Notes.

About half the people diagnosed with Ms. Will at some point be diagnosed with clinical depression, and I've occasionally heard someone say that yes, they felt really sad after being diagnosed with Ms. Or after experiencing a relapse. But just to add some clarity here, sadness and depression are two very different things. Sadness is a normal human emotion that's experienced in response to difficult, unpleasant or unhappy situations or circumstances. And it's perfectly normal and reasonable to feel sad after experiencing an Ms. Relapse or a worsening of a symptom. But depression is something else. Depression can be characterized by persistent feelings of sadness that lead to a loss of interest in other areas of your daily life. I've characterized depression as that 600 pound gorilla that sits in the middle of your living room and keeps you from leaving the house. Depression keeps you from going to the gym, meeting your friends, going out on a date, or pursuing any of the kinds of things that would otherwise normally interest you. Left untreated, depression can keep you from getting out of bed in the morning. Depression is a serious condition that requires intervention like talk therapy or prescription medication or exercise. Newly published results of a meta analysis of 12 separate studies suggests that exercise that follows established guidelines for physical activity in multiple sclerosis leads to clinically meaningful gains in depression, making exercise a viable non pharmacological treatment for people with Ms. Now, just as a quick refresher, a meta analysis is an examination of data from a number of independent studies of the same subject in order to try to identify overall trends. Now, in this case, researchers at the University of Kansas Medical center and the University of Michigan combined data from published randomized clinical trials. Of the over 2000 studies that were reviewed, 12 were analyzed. These clinical trials involved 458 adults between the ages of 28 and 48, all of whom were living with Ms. 233 of these adults received exercise interventions and 225 were in control groups, meaning they did not receive any exercise intervention. The research team set out to determine whether exercise interventions that followed the physical activity guidelines for Ms. Produced better outcomes than exercise interventions that didn't follow those guidelines. The guidelines themselves recommend engaging in at least 150 minutes of exercise or lifestyle physical activity, which is defined as physical activity that's done during occupational, household or leisure activities each week. Or people with Ms. May engage in 10 to 30 minutes of moderate aerobic activity twice a week, along with strength training from major muscle groups. All of the 12 studies analyzed followed the physical activity guidelines for MS, but the programs themselves they were widely varied. Three studies used yoga and three others used aquatic exercise. Some studies combined aerobic and resistance training, other studies combined aerobic exercise and Pilates, while others combined high intensity interval training with balance training. Still other studies combined video games and physical activity, cycling or home based programs. Program duration ranged anywhere from five to 109 weeks and and study participants exercised somewhere between two to five times a week in sessions that ran from 15 minutes to an hour and 45 minutes. Eleven of the 12 studies that used the Beck Depression Inventory, which is a common scale for measuring depression, showed that exercise that met the physical activity guidelines for Ms. Significantly reduced depression. So managing Ms. Related depression doesn't necessarily mean getting a prescription for one more medication. It can mean simply adopting the physical activity guidelines for Ms. And if you'd like to review those guidelines and learn more about the benefits of exercise and physical activity for everyone with ms, regardless of ability level, I'll recommend that you visit the National Ms. Society's website where you'll even find helpful videos to coach you in your workouts. You'll find that link in today's Show Notes. And if you'd like to review the details of this study, you'll also find that link in today's Show Notes.

By now you've heard the advice warning that early intervention with disease modifying therapy leads to better outcomes. And since we're living in an age when opinion often masquerades as fact, you may be curious about this. Is it simply someone's well meaning advice like eat your vegetables and you'll grow up big and strong? Well, no, it's not. There's no shortage of data evidence that demonstrates that the sooner someone diagnosed with Ms. Can begin treatment with a disease modifying therapy and then stay on that treatment, it slows disease progression, preserving a higher quality of life. Over the past few years, we've even shared study results with you that demonstrate that starting with a high efficacy, disease modifying therapy produces even better outcomes. Now, newly published research results demonstrate what happens when you don't have the advantage of early intervention or you stop taking your dmt. And the authors of this particular study point out that not starting disease modifying therapy following an Ms. Diagnosis isn't always a choice. There can be other real world factors at play. Economics, affordability and the healthcare system itself can all be barriers to accessing care. This retrospective single center Observational study focused on 88 people with Ms. Being treated in a major medical facility in Sao Paulo, Brazil. Of these 88 patients, 69 started treatment with Tysabri more than four years after the onset of Ms. And this group of patients was three times more likely to attain an EDSS score of 6 or higher compared to patients who were able to start treatment sooner. As a quick refresher, EDSS stands for Expanded Disability Status Scale, and it's a scale that ranges from 0 to 10 that neurologists use to measure a patient's level of disability. A score of 6 indicates that someone requires a cane or crutch to walk about 109 yards without resting. It indicates that overall daily activities will be more challenging without accommodation or assistance. So in this study, 32% of the late treatment group attained an EDSS score of 6 or higher, compared to 11% of the early treatment group. The study also looked at adherence to a treatment plan, and the researchers found that 25% of the patients stopped taking Tysabri due to the perceived risk of progressive multifocal leukoencephalopathy, or pml. Now, PML is a rare but serious brain infection caused by the JC virus. And it's been shown that people who are carrying the JC virus have previously received immunosuppressant medications and had a longer duration of Tysabri therapy, have a higher risk of developing PML while which can be fatal. That's why before starting Tysabri, patients are routinely screened to make sure they aren't carrying the JC virus. But back to our study. So 25% of the patients stopped taking Tysabri due to the perceived risk of pml. But not a single case of PML was ever reported and these patients had a three times higher risk of experiencing an Ms. Relapse compared to the patients who stayed on their therapy. The authors of this study fairly make the case that the efficacy of a disease modifying therapy that's demonstrated in a clinical trial may not reflect its real world effectiveness. If a patient has no access to the drug for years while their Ms. Potentially worsens, or they have to stop therapy because their access to a drug changes, well, the outcome changes as well. And the evidence clearly shows that outcome changes for the worst. Healthcare access issues aren't limited to Brazil. In fact, they occur to some degree everywhere, including the United States. But if you have the option to start treatment early and your access to your disease modifying therapy isn't arbitrarily changed, the evidence indicates that you're hurting yourself, risking unnecessary relapses, facing higher levels of disability, and those outcomes only make living with Ms. Worse than it has to be. So if you're lucky enough to be able to choose, please choose to put yourself first. Start treatment with a DMT as early as possible and stay on that treatment. You'll be giving yourself a better outcome. If you'd like to review the details of this study, you'll find a link in today's show. Notes Is there something better than early intervention? There is, and I like to think of it as proactive intervention. Getting someone on a disease modifying therapy before they've developed a single Ms. Symptom. Artificial intelligence could make that a reality. And in a moment, my guest, Dr. Brad Willingham will join us to discuss some of the ways AI is going to impact Ms. Research and and Ms. Patient care.

If you're a regular listener, you've probably heard me say on multiple occasions that artificial intelligence is going to be a game changer in Ms. Research and Ms. Clinical practice. Last month, while I was attending the International Progressive Ms. Alliance Digital Tools workshop, I was fortunate to run into Dr. Brad Willingham, the director of Ms. Research at shepherd center, where he focuses on the development of innovative technologies and strategies to advance neurorehabilitation. And after talking with Dr. Willingham for just a few minutes, I realized that AI is also going to be a game changer in Ms. Rehabilitation as well. Dr. Willingham is joining me today to discuss how artificial intelligence may change things for researchers, clinicians, and especially for patients. Welcome back to the podcast, Dr. Willingham.

[B]

Thank you so much, Sean. Thank you for having me. It's definitely an exciting topic and it can be a little nebulous and intimidating. So hopefully we can bring some clarity around the issue of AI and clinical research in Ms. Care.

[A]

I've heard it said that there are three domains for us to consider. The science of medicine, the practice of medicine, and the actual delivery of care. So I'm hoping you can give us a view from 40,000ft of some of the ways that AI might impact attacked each of these domains.

[B]

Yeah, absolutely. I think that was Nigam Shah at Stanford who first said that. And it was well said. Right, because when you talk about AI, it truly is, you know, general purpose technology and it has applications to so many different fields and nuances of each field. And yeah, so we talk about, you know, the science of medicine, we talk about the evidence, you know, the scientific knowledge which clinical practice is based on. And so how can AI advance research and it can help us do that by seeing Things that we couldn't necessarily see before. You know, our scientific knowledge grows directly as a function is our ability to see and observe things. And so one of the things AI is really good at is ingesting large amounts of information and recognizing patterns and making observations in there. And so we have a lot of sophisticated instrumentation in the laboratory today. Imaging being a great example of that. We have very high resolution imaging. AI is allowing us to now harness that data from these advanced scientific instruments to recognize patterns and extract novel insights from that. We've seen tremendous strides in the imaging field of our ability to extract information from these high resolution images that we are just limited in humans to see then with the medical practice side. So actually practicing medicine, we see different areas. You know, imaging is definitely one of those, well, that bleeds into the practice of medicine, but also clinical decision support. And so there's a lot of great research that's been done and available, readily available online. But it's really difficult for clinicians to harness that at once. Right. You know, constantly new papers are being published every day. How do they keep up to date with, you know, the state of science and the contemporary medical knowledge? And AI tools such as open evidence are making that a reality. Right. So just at, you know, at clinicians fingertips, they have access in lay language where they can communicate with the latest and greatest research and pose questions and get guidance. I like to think more of it as augmented intelligence rather than artificial intelligence. You still need that clinician there, but it definitely can facilitate.

That point of medical practice. And in terms of healthcare delivery, this is where I think you and I, when we saw each other in Philly, we were talking about AI as a dissemination tool making health information more accessible to both clinicians and patients. And I think that's an area. And some of our research here at shepherd is starting to demonstrate that where AI has immense potential to help extract meaning from all the data we have in today's healthcare environment, but also make it the information accessible to both clinicians and patients. Arguably, one of the greatest challenges we have in modern medicine and contemporary biomedical research is to make sense of all the data we have available to us. We saw at the Digital Tools workshop. Every day there's an amazing new digital health technology coming online, but with that comes continuous streams of data. How do we extract clinical meaning from that? But then how do we communicate that in a way that's meaningful to the clinician, or in a way a patient understands and is meaningful to them? And I think that's an area where AI is really going to make an impact in the near future.

[A]

That sounds really exciting. As we've been saying, it's going to be such a game changer. And I think it's worth pointing out not only for researchers, not only for clinicians, but for patients as well, because of AI's innate ability to take complex subject matter and simply break it down and do the explaining. As I'm saying that, I realize there will be less of a need for my podcast because we will have AI agents who can do that. But at least for now, I think I'm okay.

But you know, it will make that communication that sometimes isn't always clear, that's usually full of all sorts of buzzwords and jargon that are not part of a patient's everyday vocabulary. It kind of makes that a thing of a past. And so it allows them to better understand their status. And somebody who understands their status is in much better position to participate in that shared decision making with their doctor, aren't they?

[B]

And you're kind of alluding to patient in the loop models of care, where patients are not merely an endpoint user, but an active participant and a co creator in their health data, in their data monitoring decision making, and they're actively engaged in those feedback loops. And so digital health technologies are a great opportunity for that, to allow patients to engage in self monitoring and be more informed about changes in their health. But it's difficult to comprehend the amount of data coming in. And it's often in numerical streams. So you have these spreadsheets and spreadsheets of data generated on a daily basis or even an hourly basis. How do you get from that, from these dense numerical data sets to meaning, to something that's understandable, that's clear and that's actionable.

[A]

And.

[B]

And we have some early research showing how we can feed LLMs. It's a specific type of LLM that's in a secure healthcare environment. But we can feed these large language models, large numerical data streams from wearable sensors and mobile apps and medical records. And the LLM is able to quickly and efficiently ingest all of that structured numerical data and then it responds with a clear natural language summary of the highlights. But what the highlights are depends on who the user are, as you point out. That's the other beautiful thing about AI is it's highly plastic and it can be tailored to the needs of each individual. And so the output that a AI agent LLM may give a clinician may be different than the output it gives the patient from the same data set. Because a clinician is looking for clinical decision support, whereas the patient is looking for what empowers them, what helps them achieve their goals, what motivates them, et cetera. And so it really is a great solution to bridge that gap.

[A]

You know, if anybody listening to this conversation is interested in experiencing something like this, here's what I would suggest. And it will be a moment you remember if you do this. So go to Google and download their free app called NotebookLM. Then go ahead and find the most complex, most difficult to understand academic paper you can find online. And there's no shortage of them. Find one, import it into NotebookLM and ask the notebook, tell the notebook, I'm someone who's living with ms, I'm a non scientist. Can you explain this to me? And in a matter of seconds I think they will experience what you have just described.

[B]

Absolutely. And I love that you're encouraging people to go out and try this for themselves because it really is transformative and experiencing it is the only way to realize the potential. You know, I see that the invention and the public release of LLMs similar to that of the iPhone, where it happens, but you don't really realize the extent of the impact of the technology at the time. It'll take a couple years, but it will ultimately be ubiquitous. There's so many different use cases for it. I love that example too because it allows folks in your prompt. When you say tell me about this article, you can say your interest area, as I am a educator, I am a scientist. You can also say your goals, if I am interested in these goals for myself. And it not only will tell you about the scientific information in the article, but it will contextualize it within a framework that's meaningful to you. I love that you're encouraging your listeners to do that.

[A]

Oh, absolutely. And as you point out that moment you do the first time, it is transformative. You will see everything in your world a little differently going forward.

Beyond just talking about how things become easier to understand for people who are living with Ms. I'm wondering if we can talk about another aspect that AI will certainly could impact someone's Ms. Treatment plan. Can you explain what a digital twin is?

[B]

Absolutely. So this is another fantastic topic. A digital twin is essentially a digital model of either a person or a system. In healthcare, when we talk about a digital twin, we talk about feeding a comprehensive data set of a patient. This is things about their physiology, their physical function, their cognition, their socioeconomic demographics, feeding it in to an AI system to create a digital representation of that patient. But then the powerful thing of digital twin technology that we can do is then we can use, we can train that system on historical data so it knows how various patient models, digital twin models, respond to different therapies based on these different parameters. And so what we can do is then we can simulate different treatment strategies with that digital twin to see how the patient may respond to those various treatment strategies. So you could see right off the bat, this is a very powerful tool where if we can rapidly simulate thousands of different treatment strategies for a single patient to get an idea of which handful of treatments may be most optimal, we're already way ahead in terms of, you know, clinical practice and, you know, what works best for each patient. Also the area. Digital twin technology also holds great promise for scientific research to expedite scientific discovery. You know, one of the challenges in scientific research is the heterogeneity we have in our participant populations. You know, we, we practice evidence based medicine. It's great. This group approved more on average, but you do have heterogeneity in there. Right. And so what if we can conduct in silico trials where the digital twin is the control group? Right. And so it's just an example of how that might expedite scientific research as well. So that's another area of active research here at shepherd center where we're exploring, we are using comprehensive evaluation strategies to get a deep understanding of a patient's health, function and physiology. We're creating digital models and then we're trying to determine how and for whom optimal treatment strategies work.

[A]

It really begins to deliver on the promise of personalized medicine, doesn't it?

[B]

Absolutely. This is the realization of precision rehabilitation. Precision medicine has been around for a while and it's been rapidly advanced by the field of genomics. But now taking that model of precision medicine and applying it to precision rehab, so it requires a different suite of measurements. Right. So in rehab, we're looking more at evaluating physiology and function and personal goals in terms of activities of daily living. But the model is the same, right? Getting a nuanced understanding, a deep understanding of the individual's needs, preferences and goals, and using that to create targeted treatment strategies in a data driven manner.

[A]

While we're talking about rehab, I'm wondering if you can talk about how AI might be used at home when it comes to telerehabilitation.

[B]

So just backing up telerhabilitation, remote care strategies have immense potential to overcome many barriers that folks living with Ms. Face on a daily basis when attempting to access care. Right. And we've gone through those lists before. It can be anything from symptoms or transportation, finances, geography, access, expertise, et cetera. Remote care strategies overcome many of those barriers and promote access to care. But one of the challenges is remote care strategies require the use of several different technologies. You have a digital platform you're getting your therapy from, you have a mobile app that's tracking your symptoms. You have wearable sensors, so you have all of this dense, complex data, but then it's also in different systems. Right? And so one of the things we've been exploring here is how we can use AI to harmonize that data, bring in data from different sources and synchronize it, and then use the LLMs we're talking about to provide a clear summary of progress and next steps. And so in terms of remote care, one of the challenges is the lack of clinical supervision, the lack of clinical interaction. Right? You're doing your therapy, you're sticking to it, but are you doing it right? Am I improving? What are the next steps? And this is where AI is going to be able to use the data from all these wonderful remote technologies, integrate it and help guide a patient and provide information about, hey, this is how you're doing towards your goals. And, and this may be some next steps. And so we have research looking on summarizing the data, but we're also looking at this idea of clinical decision support with providing suggestions. Right. And so actually what we're looking at is can AI not only summarize data, but can it serve as a remote care coach to coach you during your remote care strategies, where, whether it's telexercise or telerehab, to provide encouraging updates or provide suggestions about your next steps. You have to be more cautious when you're providing suggestions. And so the experiments, the research we're conducting around that work is a little bit more rigorous in terms of safety, risk, and again, controlling what the AI is saying through very customized large language models.

[A]

I'm imagining an AI bot telling me, I need to lift my arm a little higher.

[B]

Yeah. So computer vision is a form of AI, right? And so we have platforms like Kentai that, you know, you just have a iPhone or laptop and you know, you conduct movements in front of it. And the AI agent will just using the video stream, it will give you cues, and so it'll analyze your biometrics based on the video and it'll say, okay, squat a little lower, raise that left arm a little higher. There are ways we can also get real time feedback through AI applications such as computer vision.

[A]

I'm wondering if there's anything else that you have in the works that you want to share with us.

[B]

We talked about.

Large language models for summarizing information. We talked about digital twin technology for precision rehab. These are two of the major areas of rehabilitation research we're working on right now in the Ms. Research program. The only other thing I would mention is that, you know, as the volume and complexity of, you know, this health information grows, it becomes increasingly challenging for our patients and our clinician clinicians to navigate it. So as we're innovating, you know, we have a lot of wonderful engineers out there innovating new technologies every day. I think it's important that we also think about how we make that data meaningful and engage folks living with Ms. Early and often in your research, engage the clinicians early and often in your research to see, okay, this is what's meaningful, but this is how we communicate it. This is how we represent it. And it's going to be important that we have these multidisciplinary teams involved.

That's the only other thing I might bring up.

[A]

Well, Dr. Brad Willingham, I want to thank you for all you do to improve the quality of life for people living with Ms. Thank you for giving us a glimpse into how AI is going to impact Ms. Rehabilitation. And thanks so much for talking with me today.

[B]

Thank you for having me, John. It's always a pleasure.

[A]

That's going to wrap up this episode of Real Talk. Ms. Real Talk, Ms. is powered by the National Ms. Society, and you can share this episode of the podcast by letting your friends or family members know that all they have to do is is point their web [e-mail] 413. You'll find that link in today's show notes so you can easily copy and paste it right into an email or a text. And if you have a minute, I hope you'll visit the Apple App Store or the Google Play Store and Download the free RealTalk Miss app for your iOS or Android smartphone or tablet. It's the best way for us to stay connected. The app will automatically download the latest episode of Real Talk Ms. You'll be able to access any of our past episodes. You'll be able to save your favorite episodes, and it's a great way for me to share bonus content with you. The app is free, so I hope you'll download it today. I'm John Strum. Thanks for listening. Stay safe and make healthy choices.

Sam.