Vigorous research activity in mHealth has resulted in an ever-growing list of physiological and behavioral markers. However, translation of these biomarkers into real time intervention lagged behind the observational research studies that led to their development due to computation, storage, and communication bottlenecks faced by wearables and smartphone platforms. Further, the next generation of wearables is emerging with the ability to sample data from multiple sensors at rates several orders of magnitude higher than current-generation devices, exacerbating the computational and communication bottleneck. They can image structure, motion, and function, to provide visibility into physiology previously possible only in clinics.
Traditionally, such imaging sensors use post-processing algorithms for feature identification, co-registration, alignment and enhancement. However, high frequency-high volume imaging data from wearables cannot be transported to cloud computing for post-processing. Finally, researchers have shown that the high-dimensionality sensor data needed to compute biomarkers present immense privacy risks. Advances in machine learning are leading to an ever-growing list of surprising inferences about user identity and activities that can be made from seemingly innocuous sensors, particularly when data are captured over long durations. Simplistic methods such as stripping personally identifiable information and addition of noise that focus on anonymizing the data have been ineffective for mHealth, both from privacy and utility perspectives, particularly with the availability of vast amounts of side information (e.g. metadata), computational power, and advanced algorithms.
To address these growing challenges, we propose a hierarchical computing framework that reduces the data into minimal modular abstractions called Micromarkers computed at the edge devices. Micromarkers can be used directly as features in new biomarker inferences or can be adapted to support legacy algorithms. TR&D3 is developing hardware, software, and computational techniques to implement privacy-aware, efficient, and embedded intelligence support into wearables. They will enable continuous, high-throughput, low latency biomarker captures across wearable, mobile, and cloud platforms to support large scale and long-term research studies, and eventual real-life rollout.
TR&D3 is conducting the following innovative research to address the technological challenges described above:
- Develop modular and reusable micromarker abstractions to enable resource-efficient concurrent computation of a growing collection of biomarkers.
- Create signal processing architectures combining Compressive Sensing and Machine Learning algorithms to support biomarker computations on resource constrained high data rate sensor arrays.
- Enable optimization of privacy-utility tradeoffs in biomarker computations via cross-layer mechanism design.
TR&D3 is producing the following technological resources for the community:
- mDOT Center applications and software development kits (SDKs) on popular wearables, personal devices, and smartphones with embedded micromarker-based implementation of biomarkers for stress, fatigue, speaking, smoking, craving, eating, brushing, and new biomarkers from collaborative projects
- Reference design and prototypes of mDOT Center radio-frequency (RF) Patch sensors, modular hardware modules and embedded software cores to power wearable sensor arrays
- A toolbox for exploring privacy implications of sensor and biomarker choices and enabling run-time control over privacy-utility trade-off in biomarker implementations.
Impact on Science & Society
TR&D3 will make it possible for temporally-precise mHealth interventions for maintaining health and managing the growing burden of chronic diseases, to be realized on participants’ personal devices with minimal impact on battery life so as not to interfere with routine usage of these devices. TR&D3 will also enable much greater personalization of such interventions by expanding access to many new and emerging biomarkers, resulting in a higher likelihood of continued engagement and improved efficacy. By enabling the computation of novel biomarkers from the next generation of high data rate sensors that provide unprecedented visibility into physiology, previously possible only in clinical settings, TR&D3 will enable remote care for patients via mHealth interventions who have traditionally required close involvement of clinicians, marking a major transformation in the care delivery, especially to the neediest. By providing fine-grained and low-burden privacy control to patients over their data being shared with third parties, TR&D3 will ensure that the health impact of temporally-precise mHealth interventions is not adversely impacted by privacy concerns.
Dr. Emre Ertin is a Research Associate Professor with the Department of Electrical and Computer Engineering at The Ohio State University. He received the B.S. degree in Electrical Engineering and Physics from Bogazici University in Turkey in 1992, the M.Sc. degree in Telecommunication and Signal Processing from Imperial College, U.K. in 1993, and the Ph.D. degree in Electrical Engineering from Ohio State in 1999. From 1999 to 2002 he was with the Core Technology Group at Battelle Memorial Institute. His current research interests are biomedical sensor design and statistical signal processing with application to sensor networks and mobile health. Visit Google Scholar page
Dr. Mani Srivastava is a Professor of Electrical Engineering and Computer Science at the University of California, Los Angeles. His research is broadly in the area of networked human-cyber-physical systems, and spans problems across the entire spectrum of applications, architectures, algorithms, and technologies. His current interests include issues of sensing, privacy, security, data quality, and variability in the context of applications in mHealth and sustainability. He is a deputy director of NSF Expeditions on Variability and is the lead investigator on an NSF Cyber Physical Systems Frontier Project called RoseLine. His works have been cited extensively (over 30,000 times) and have won several best paper awards. He has served as editor-in-chief of IEEE Transaction on Mobile Computing and the ACM Mobile Computing and Communication Review. He is a Fellow of IEEE. Visit Google Scholar page
Dr. Ida Sim is a is Professor of Medicine, Co-Director of Biomedical Informatics at the University of California, San Francisco’s Clinical and Translational Sciences Institute. She is also co-founder of Open mHealth, a non-profit organization that is breaking down barriers to mobile health app and data integration through an open software architecture. Dr. Sim received her M.D. and her Ph.D. in Medical Informatics from Stanford University, and is an international leader in informatics for health care and clinical research. Her research work is focused on knowledge-based technologies for evidence-based practice, especially in the ontological representation of clinical trials. In policy work, Dr. Sim was the founding Project Coordinator of the World Health Organization's International Clinical Trials Registry Platform, which sets global standards on clinical trial registration and reporting. She is a Fellow of the American College of Medical Informatics, and a member of the American Society for Clinical Investigation. She is also a practicing primary care physician. Visit Google Scholar page