New Network Model of the Musculoskeletal System Can Predict Compensatory Injuries

Anatomical diagrams that show the connections between sets of muscles and bones have existed for centuries, but the Bassett team’s network diagram is the first to map those connections over the entire body. Illustration by Brittney Bennett.

Network science examines how the actions of a system’s individual parts affect the behavior of the system as a whole. Some commonly studied networks include computer chip components and social media users, but University of Pennsylvania engineers are now applying network science to a much older system: the human body.

While detailed anatomical studies of the musculoskeletal system have existed since da Vinci, new research, led by Danielle Bassett, Eduardo D. Glandt Faculty Fellow and Associate Professor in the Department of Bioengineering, is the first to convert the entire body’s network of bones and muscles into a comprehensive mathematical model.

Bassett lab members Andrew C. Murphy, Sarah F. Muldoon, David Baker, Adam Lastowka, and Muzhi Yang, along with Brittney Bennett, a member of the Penn Network Visualization Program, contributed to the research. It was published in the Open Access journal PLOS Biology.

The simplified network treats bones as “balls” and muscles as “springs,” but can nevertheless provide sharper insight into how injury in one part of the body can lead to increased strain on another.

As the model becomes more realistic and tailored to individuals, it could help clinicians and physical therapists predict compensatory injuries and suggest ways of avoiding them.

“People who study biomechanics tend to focus on a single part of the body — the shoulder, the wrist, or the knee,” Bassett said. “Because that knowledge is so localized, they don’t have a way of connecting it to the rest of the body or to think about compensatory injuries that are far away.”

“We actually had to go back to Gray’s Anatomy-type texts,” Bassett continued. “It was a long, painstaking effort figuring out which bone connects to which muscle and collating that data into a full network.”

Even in its highly abstracted form, this network provides a picture of how forces are transmitted throughout the musculoskeletal system.

“We can say, ‘if this is the muscle you injured, here are the other muscles we should be most worried about,’” Bassett said.

The researchers also compared their network to the “motor homunculus,” a way of mapping brain regions to the parts of the body they control.

“We saw that the more impact that a muscle has on the rest of the body, the more real estate we use in our brain to control it,” Bassett said. “We think it’s a way for us to maintain robustness in those muscles — if a muscle can have a massive impact on the rest of the body, you don’t want any error in controlling it.”

Future work will refine the network, adding more realistic masses for individual bones and stretchiness for individual muscles. The modeling of tendons, and muscles that have more complicated flexing behaviors, will also be improved.