Dr. Francis Collins, director of National Institutes of Health, announced a new national initiative, the National Center of Excellence for Mobile Sensor Data-to-Knowledge (MD2K), with the main focus on developing computational tools to facilitate the collection and analysis of large-scale health data generated by mobile and wearable sensors. Deepak Ganesan sensor systems expert, and Benjamin Marlina machine learning expert, will co-lead the center at the University of Massachusetts Amherst (UMass Amherst).

The national MD2K team includes computer scientists, engineers, statisticians and biomedical researchers from 11 universities and one nonprofit organization. The main problems this group needs to resolve relates to the complexities of mobile sensor data to accelerate biomedical discovery and optimizing care delivery.

For the MD2K Center at the University of Massachusetts Amherst, the main goal will be to work on inferring measures of patient health, as well as markers of behavioral, physical, social, and environmental risk factors from mobile sensor data. Here, researchers will focus on smoking and congestive heart failure as health problems with high mortality risk.

The national MD2K team includes computer scientists, engineers, statisticians and biomedical researchers from 11 universities and one nonprofit organization.“The promise of mobile health sensing is that we can use body-worn sensors to detect various behavioral and environmental cues that will help predict adverse health events in real-time,” Ganesan says. “For example, it has long been known that smoking relapse is related to stress, alcohol consumption and other cues. With wearable sensors, we can aim to detect these cues in real-time and offer interventions to patients before relapse occurs.”

“The challenge,” Marlin adds, “is making the data analytics highly robust and scalable while taking into account energy use and communications costs as well as the security and privacy of the data. The MD2K Center will explore solutions to all of these problems with the immediate goal of developing accurate and reliable computational tools that biomedical researchers will be able to easily incorporate into health and behavior studies.”