Introducing: The Mobilize Center
Bringing big data to gait analysis
In 1891, when researchers first analyzed the mechanics of walking in three dimensions, the process was painstaking and cumbersome. It took six to eight hours just to prepare the subject for data collection. Today, such data are routinely collected for various therapeutic and research purposes using digital motion capture technology. At the same time, databases are filling with movement data collected by wearable fitness devices and smartphone apps. The availability of all these data raises the question: How can researchers best make use of it?
Enter the Mobilize Center, a new Big Data to Knowledge (BD2K) Center of Excellence funded by the National Institutes of Health (NIH). “Over the past decades, we’ve created human models and simulations that embed the mechanics and physiology of movement,” says Jennifer Hicks, PhD, Director of Data Science for the Center. “Combining these mechanistic models with advanced methods from statistics and machine learning opens up huge opportunities for understanding all these data,” she says.
For example, in collaboration with Michael Schwartz, PhD, and his colleagues from Gillette Children’s Specialty Healthcare, the Center will use statistical approaches to predict the outcome of surgery in children with cerebral palsy (CP)—a neurological disorder that affects motor control—using Gillette’s vast database of information about how children with CP move.
A few studies have used these data to understand why an individual walks with a certain gait. However, Hicks says, “We’ve only scratched the surface in terms of the types of questions we can answer.”
To dig deeper, the Center will need to figure out how to make more effective use of time-series data, such as the levels of muscle strength measured every few months for several years. In the past, researchers uncertain of how to make use of such data might simply select a peak or average value from the time period and discard the rest. The data science tools the Center is developing “will help us figure out ways to better condense these time-series signals or include more of them in our models,” Hicks says. That additional information, combined with a mechanistic understanding of how humans move, could identify new factors that better predict whether a child will benefit from a given surgery and thereby improve outcomes.
The Mobilize Center is also teaming up with industry partners, such as Azumio, to take advantage of the boom in physical activity monitoring. Azumio, a leader in health and fitness apps on mobile devices, has tons of data from some 70 million users. Indeed, says Bojan Bernard, PhD, Azumio’s CEO, “The distance walked in one day by users of our Argus app is equal to two trips to the moon and back.”
There is a big gap, though, between having all that data and having validated tools that actually encourage people to move more and in ways that reduce injuries. The Mobilize Center, with its experts in data science, biomechanics, and behavioral science, will address that gap. Using anonymized data from Azumio and other mobile applications and sensors, they will shed light on how various factors—environmental, social, biomechanics knowledge—impact behavior, particularly physical activity. Ultimately, that knowledge could be used to create more effective interventions.
“We’re very interested in collaborating with researchers to learn what’s hidden inside the data,” says Bernard. “In the long term, our goal is to use this information to improve human health. Working with the Mobilize Center is an important step towards this goal.”