As its first year wraps up, the Center of Excellence for Mobile Sensor Data-to-Knowledge (MD2K) has made significant progress towards its goal of using mobile sensor data to realize Precision Medicine.
The MD2K Center is tasked with developing the means to gather, analyze, visualize and interpret health-related mobile sensor data. This capability is critical to discovering new insights on the role of behavioral and environmental context in the onset and progression of disease. The ultimate goal is to develop timely and personalized mobile health interventions for early detection and prevention of adverse health events, which will help realize the vision of Precision Medicine articulated by President Obama in January.
The MD2K team has made remarkable progress in converting mobile sensor data into markers of adverse health events and risk factors. This includes automatically detecting when a newly abstinent smoker lapses for the first time in a smoking cessation attempt. This computational model, called puffMarker, uses hand gestures from smartwatches and the breathing signature from respiration sensor. This work was led by Nazir Saleheen, a PhD student working with MD2K director Dr. Santosh Kumar, and was recently published at ACM UbiComp.
A second work demonstrated the feasibility of detecting stress from heart beats and breathing patterns. This computational model, called cStress, was led by Dr. Karen Hovsepian, a former postdoc of Kumar (now a faculty at Troy University) and published at ACM UbiComp. The cStress model has recently been shown to be informative and actionable in improving interaction among couples.
These two papers were among 59 that the MD2K team has published in its first year. Most of these articles have appeared in high-profile venues that include ACM UbiComp, ACM CHI, ACM Sigcomm, ACM MobiCom, IEEE CVPR, ACM SIGMOD, and ACM KDD in data science research, and Circulation, Nature Reviews Cardiology, Journal of Consulting and Clinical Psychology, Nicotine & Tobacco Research, Drug & Alcohol Dependence, Psychophysiology, and Translational Behavioral Medicine, and in health research. In addition, MD2K team has given 93 talks at 63 meetings in the past year.
 During the MD2K Annual Meeting, students presented posters based on MD2K research from the past year. |
A list of papers published by MD2K researchers can be found here. Also, posters explaining both puffMarker and cStress can be found here.
“Development of the puffMarker model has fulfilled a longstanding need in smoking cessation research to precisely pinpoint the timing of first lapse. This work has opened the doors to discover potent markers in the data collected by mobile sensors that can be used to deliver just-in-time mHealth intervention to improve smoking cessation rates.” Kumar said.
Also, studies are underway in MD2K’s key research areas, Congestive Heart Failure management (at The Ohio State University) and smoking cessation (at Northwestern Medical School and Rice University). All the models and markers developed by MD2K (including the ones described above) will be included in the open-source software (for mobile phone and cloud) soon to be released by MD2K, that can be used widely by researchers, industry professionals or individuals to collect, analyze, visualize and interpret mobile sensor data for health.
"It's a pleasure and an honor for the University of Memphis to be hosting a Center as innovative, productive, and impactful as the MD2K has been in its first year,” said U of M President Dr. M. David Rudd. “The contributions of such a talented group of scientists has been nothing short of remarkable and, if the first 12 months are any indication, the next few years will result in some impressive advances in precision medicine."
The MD2K team comprises 20-plus faculty, 20-plus students, four staff and three software engineers from 12 institutions and the non-profit Open mHealth. The center is one of 11 national Centers of Excellence funded through the National Institutes of Health (NIH) Big Data to Knowledge (BD2K) initiative, designed to support advances in research, policy and training needed for the effective use of Big Data in biomedical research.