The Bare Bones: Insights from Modeling our Musculoskeletal Matrix

The Bare Bones: Insights from Modeling our Musculoskeletal Matrix

The principles of network science have been used to study systems from computer chip components to social media users. And while we have been studying the human body since ancient times, application of these network principles has now built the first comprehensive mathematical model of our entire musculoskeletal system. I interviewed Danielle Bassett, Eduardo D. Glandt Faculty Fellow and associate professor in the University of Pennsylvania’s Department of Bioengineering, via email to find out more about her PLOS Biology study.

Your degrees are in physics. What led you to apply your skills to biological systems?

DB: Physicists have developed powerful tools to characterize complex systems. Although the tools are not traditionally applied to biological systems, in some cases, this cross-disciplinary application can open up new avenues for scientific inquiry. Biological systems – like the physical universe – are very complex, and we need formal mathematical approaches and theoretical insights to study them. My work aims to meet this need.

In this research, you used network science to analyze the human musculoskeletal system. What gave you this idea?

DB: I’ve always been fascinated by how we use mathematics to explain real-world phenomena; I can’t help but see networks wherever I go! Analyzing the human musculoskeletal system as a network was motivated by a bicycle accident I had while a graduate student in Cambridge, U.K. I tore the rotator cuff in my shoulder, and the primary injury led to secondary injury, with a very long rehabilitation regimen that took me off the King’s College racing team. Apart from the personal inconvenience, it illustrated to me that muscular injuries do not occur in isolation.

The human body is complex; what assumptions did you have to make to model it?

DB: We made many simplifying assumptions. A common joke about physicists is that they model cows as spheres, but these assumptions help us to obtain simple intuitions about large-scale phenomena. For this study, we modeled bones as point masses (rather like marbles), and we modeled muscles as springs. Our goal was to strip away any extraneous detail to isolate just the effects of the architecture of the network.

You also looked at how the brain controls the musculoskeletal network. What did you find?

DB: When we mapped brain regions to the groups of interconnected muscles that they control, we thought we could find brain control systems sending signals to multiple muscle groups, or multiple control systems sending signals to a single group. Instead, we were interested to find an ordered one-to-one mapping: each brain control system signals to a single muscle group, with important muscles requiring more brain “real estate” to control them. This suggests that the control system may have evolved and been optimized over evolutionary time scales.

What did you find most interesting about your results?

DB: We are particularly excited about how local network architecture predicts how a single muscle (or injury to it) can affect the whole musculoskeletal system. Our estimates of that impact are strongly correlated with observed recovery times after muscle injury, suggesting that the role of a muscle within the network is directly related to length of recovery.

What are compensatory injuries, and how might your model help us to understand them?

DB: Compensatory injuries are injuries that occur when we use other muscles to compensate for the loss of an initially injured muscle, as with my secondary injury that followed my rotator cuff tear. Our model seeks to predict where those compensatory injuries might occur, based on the local network that a single muscle sits within.

What else do you hope your study might lead to?

DB: Our next steps are to increase the sophistication of our model by incorporating more information about bone mass and muscle volume, and about how the musculoskeletal system differs (albeit slightly) in each of us. We are also hoping to work closely with orthopedists and physical therapists to develop models that they can use in clinical practice.

Research Article: Murphy AC, Muldoon SF, Baker D, Lastowka A, Bennett B, Yang M, et al. (2018) Structure, function, and control of the human musculoskeletal network. PLoS Biol 16(1): e2002811.

Image Credit: Adapted from image by Brittany Bennet


Beth works at PLOS as Journal Media Manager. She read Natural Sciences, specializing in Pathology, at the University of Cambridge before joining PLOS in 2013. She feels fortunate to be able to read and write about the exciting new research published by PLOS.

Leave a Reply

Your email address will not be published. Required fields are marked *