In the near future, when people on all seven continents can use AI to heal or prevent musculoskeletal pain, we could very well look to a large room at the back of the Education Complex at 麻豆映画传媒 and say this is one of the places where everything began to change. It鈥檚 an interesting space. In fact, if you miss the sign over a threshold in the lobby (Institute of Exercise Physiology and Rehabilitation Science) you might wonder if you鈥檝e walked into a lab, a rehab clinic, or a fitness center. There are treadmills and mats, resistance bands and treatment tables, and voices of researchers in mid-instruction.
In an area set apart with privacy dividers, Assistant Professor Colby Mangum rolls a chair up to a portable ultrasound machine and a laptop. Aside from the signage, these are the first giveaways that you鈥檝e entered an important intersection of research, technology, physical health, and profound possibilities.
鈥淲e want the lab to be as 鈥榬eal world鈥� as possible, so we can address real problems with real patients,鈥� Mangum says of what she calls the REhabilitation, Athletic assessment and DYnamic imaging (READY) Lab. 鈥淚t鈥檚 the only way to make a genuine impact on the future of musculoskeletal health.鈥�
鈥淲e want the lab to be as 鈥榬eal world鈥� as possible, so we can address real problems with real patients.鈥� 鈥� Colby Mangum, assistant professor
Before diving into her research, it鈥檚 helpful to know something about the researcher. Mangum鈥檚 career aligns with the cues in this active lab. She started as an athletic trainer while also becoming an expert in the use of ultrasound imaging as a researcher so she could blend all of her interests into finding better answers for people with chronic pain and for those who want to prevent it. In 2018, that personal mission led her to 麻豆映画传媒 where she was asked to grow the athletic training research agenda at the College of Health Professions and Sciences. Today, thanks to internal seed funding, she鈥檚 ramping up her research focus on improving diagnostics and treatment courses for the most widespread pain in the world: Low back pain, which affects more than 600 million people globally.
Mangum has seen enough cases of lower back pain to spark her curiosity about the source of the pain and how muscles around the lower back and spine are affected.
鈥淚t oftentimes seems to connect to the core,鈥� she says. 鈥淚鈥檓 talking about well-conditioned athletes, younger people and older people. It goes back to an imbalance in those three layers of abdominal muscles: external obliques, internal obliques, and transverse abdominis.鈥�
MRI and CT are the most often used imaging modalities to assess musculoskeletal conditions, but MRI is expensive and CT exposes patients to radiation. In addition, neither is good at imaging tissue movement, which can yield important information about the physical properties of tissues. Ultrasound, on the other hand, is safe, low cost, portable, and can capture tissue dynamics during rehab sessions.
鈥淯ltrasound has the potential to unlock answers, especially for low back pain for a large population scale,鈥� Magnum says, 鈥渂ut first we need to overcome some issues we face.鈥�
Those issues primarily come down to the training it takes for clinicians and researchers to acquire good quality ultrasound images and the time it takes for them to manually measure ultrasound readouts. There鈥檚 also high subjectivity in those measurements. Mangum has believed for years that some ultrasound automation would be the only way to reach hundreds of millions of people and begin making a transformational impact. She also knew she couldn鈥檛 do something that big alone in this lab.
鈥淚t became clear that if I could collaborate with an AI expert,鈥� Mangum says, 鈥渋t would be the gamechanger.鈥�
The Power of Two
鈥淭he first time I spoke with Laura, I sensed her enthusiasm for this project,鈥� Mangum says of Laura Brattain, the College of Medicine associate professor with whom she鈥檒l collaborate over the course of the next year. 鈥淲e recognize the value of our complementary expertise 鈥� mine with ultrasound and patient care in a clinical setting and hers in applying AI to healthcare.鈥�
Mangum calls Brattain 鈥渁 powerhouse鈥� because of her training as a biomedical engineer and fascination with medicine. While at MIT, Brattain developed portable emergency care technologies and, not so coincidentally, AI algorithms for analyzing ultrasound images for disease diagnostics and procedure guidance. She joined the in 2024 as an associate professor in the College of Medicine with an affiliated appointment in the College of Engineering and Computer Science 鈥� a position conducive for translational AI research.
鈥淭he AI initiatives at 麻豆映画传媒 provide me with a bigger platform to innovate with clinicians and turn research into something impactful,鈥� Brattain says. 鈥淎s soon as I arrived, I found myself in this great ecosystem, collaborating with hospitals and with researchers who have clinical experience, like Colby.鈥�
The combined skillsets of Mangum and Brattain and the far-reaching potential of their low back pain project will start with a seed grant of $12,000 from 麻豆映画传媒 and a matching grant from the College of Medicine. The funding will allow them to acquire ultrasound data from both individuals with and without a history of lower back pain, use AI and human raters to analyze the images, then establish if there is reliability and agreement between the manual and automated analysis. AI is certainly faster, but is it just as accurate and consistent? Could it one day eliminate the training needed by ultrasound operators as well as improve the data and ultimately, the outcomes?
鈥淭he AI initiatives at 麻豆映画传媒 provide me with a bigger platform to innovate with clinicians and turn research into something impactful.鈥� 鈥� Laura Brattain, associate professor
The researchers will work with students (a health sciences major and a biomedical AI doctoral student) to generate solid preliminary results before seeking larger external financial support. However, the numbers behind these initial dollar signs are not the story.
鈥淚 already have the necessary space and equipment for the work,鈥� Mangum says. 鈥淭he strength of the seed funding is in how it merges our expertise so we can stretch the boundaries of what people are capable of doing with ultrasound.鈥�
Mangum demonstrates the way a typical ultrasound has been used to address low back pain. A patient lies on a table while the technician uses a transducer to take pictures of the three layers of abdominal muscles. While the study will take similar static images, the 麻豆映画传媒 researchers will also capture additional, dynamic images with the patient鈥檚 body in motion 鈥� an approach that until now has not been fully explored by others.
鈥淲e know the problem is typically a muscle out of balance somewhere in there,鈥� she says, pointing to the entire picture on her screen. To pinpoint the 鈥渟omewhere,鈥� the technician has to find the correct orientation of the ultrasound probe and take time to make the manual measurements, which can potentially lead to inconsistencies and delays in rehab.
鈥淎s soon as I started doing ultrasound imaging in graduate school, I thought, 鈥榯here has to be a better way to get these measurements,鈥欌�� Mangum says. 鈥淎I could certainly be the way. By implementing it into our existing tools, it seems there is a strong potential for improving speed, accuracy and outcomes.鈥�
It鈥檚 easy to understand why these possibilities excite a clinician like Mangum. But what about Brattain, the biomedical engineer?
鈥淚鈥檓 currently the chair of AI Community of Practice at the American Institute of Ultrasound for Medicine. Just like Colby, I鈥檓 a strong proponent of AI-driven ultrasound applications. I鈥檓 an avid pickleball player and work out daily. Before taking on this project, I did a survey on the most common physical issues for active people 鈥� low back pain kept coming up. If my skills with AI can help standardize musculoskeletal ultrasound, that means we have potential to improve the current workflow for low back pain rehabilitation and we can eventually scale this up to personalized at-home care and expand to other musculoskeletal injuries. The potential return on investment is high.鈥�
Better Lives for People Far and Near
Mangum and Brattain鈥檚 skill sets are as complementary as their perspectives. For example, take their concept of scale, which Brattain mentioned.
鈥淓nd users might access this kind of ultrasound technology on something like a smartwatch,鈥� she says. 鈥淪o, it would be available anywhere, even in rural communities and in extreme environments like the battlefield and in space.鈥�
From her lab space near the center of campus, Mangum says, 鈥淲e could walk just beyond this door and find people on campus with low back pain.鈥� This is why her lab will become a low back pain research clinic in the coming months. The 40 participants for the project will be a mix of people at 麻豆映画传媒 and from the greater community. Mangum and her student researcher will perform ultrasounds and send the images to Brattain and her team, who will conduct the AI analysis.
鈥淎nd then we鈥檙e off to the races,鈥� Mangum says. 鈥淏y leveraging each other鈥檚 strengths, we can improve rehab, apply what we learn to other musculoskeletal needs, and help more people. That鈥檚 where we can go with this because of the power of collaboration.鈥�
