SenSE: Multimodal noninvasive wearable sensors and machine learning for predicting critical glycemic events

点击翻译按钮获取中文摘要

Proposal Title:SenSE: Multimodal noninvasive wearable sensors and machine learning for predicting critical glycemic eventsNSF-2037383: Coté, Gutierrez, Erraguntla, McKay and DeSalvoNon-technical Abstract:The main objective of this research is to develop a wearable sensor system that allows people with diabetes to monitor their blood glucose (sugar) levels without having to draw blood. The approach relies on measuring subtle changes in physiological signals that occur when blood glucose deviates from its normal range. Predicting these deviations is critical for diabetes management, as hypoglycemia (low blood glucose) can have severe health consequences in the short term, and hyperglycemia (high blood glucose) for the long term. The signals are measured using six different sensors, all on a wearable arm band, and are combined with machine-learning algorithms to predict sugar levels and help keep patients in the normal blood sugar range. This wearable system could eventually be used to predict and monitor other adverse health outcomes such as heart problems (e.g., congestive heart failure or hypertension). This project also provides research training to diverse students at the graduate and undergraduate level. The research team will work with the university programs to broaden participation in science and engineering. In addition, high school students and teachers from underserved and underrepresented areas will be involved to work on university-based laboratory research to excite them about science and engineering. Technical Abstract:This project seeks to develop a multimodal wearable sensing platform and machine learning algorithms that can be used to predict glycemic events non-invasively. Managing diabetes requires balancing the long-term risks of diabetes complication related to hyperglycemia (blindness, kidney failure, amputation, stroke, and heart disease) and the acute risks of hypoglycemia (seizure, coma, or death). At present, however, this can only be achieved with continuous glucose monitors (CGMs), which are expensive, invasive, and typically only prescribed to 5% of the patients with diabetes (type 1). Thus, the specific aims of this project are to: 1) Develop a multimodal system of non-invasive sensors including optimal placement and integration of six sensors; 2) Develop predictive models of glycemic events including statistical and deep-learning models capable of predicting future hypoglycemic events and characterize time in hyperglycemic range; and 3) Characterization and validation with human subjects, including two studies to be conducted in the clinic and at home.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.