Group Research Projects
Moveit! : Incentivizing a Healthy Lifestyle through Mobile Just-In-Time Adaptive Interventions
Group members: Joshua D'Arcy, Karnika Singh, Vishal Dubey, Harish Yerra, Christina Le, Sabrina Qi
MoveIt! aims to provide a personalized "Just-in-time Adaptive Interventions" (JITAI) model to nudge users toward healthier dietary and physical activity behaviors. MoveIt! aims to revolutionize cardiometabolic preventative care through novel methods for mobile prediabetes prevention, including recommender models based in Natural Language Processing.
Dynamic Time Warping: Biomedical Signal Processing
Group members: Yihang Jiang, Yuankai Qi, Will Wang
DTW group aims to develop metrics to evaluate the performances of different DTW algorithms to matching time-series signals. Yihang also proposed a novel algorithm called eventDTW based on the peak information of the signals, and we demonstrated its better performance in aligning signals with different sampling.
Individual Research Projects
Learning Digital Biomarkers of Prediabetes from Longitudinal Wearable Sensors
Brinnae's thesis work focuses on extracting digital biomarkers from longitudinal wearable sensor data for chronic disease diagnosis and management. She is currently working on mining digital biomarkers for glycemic variability in prediabetic populations using a combination of machine learning and time series methods.
Digital Biomarkers for Coronary Artery Disease Prescreening
Peter's thesis work applies a data-driven approach to uncover previously unknown CAD risk factors from both implantable and consumer wearable devices for remote CAD prescreening.
Digital Biomarkersof Circadian Disruption and their Utility in Cardiovascular Risk Assessment
Will's thesis project aims to detect and quantify cardiovascular responses to circadian disruption and compare digital biomarkers against more conventional circadian metrics for evaluating the cardiovascular effects of circadian disruptions. Such digital biomarkers are meant to provide longitudinal, accessible and individualized metrics for evaluating the cardiovascular risk resulting from circadian disruptions (e.g. shift workers).
Karnika's research is focused on understanding the physiologic effects of weight-loss interventions and factors involved in adherence to interventions and outcomes resulting from participation in them. Through partnerships with the Stanford StrongD study and the Duke Hearts and Parks initiatives, Karnika will compare cardiometabolic, physical activity, and sleep digital biomarkers derived from weight-loss interventions in both adults and children.
Yuankai's research combines signal processing and feature engineering (e.g. Discrete Wavelet Transforms) and machine learning approaches (e.g. CNNs) for multi-class arrhythmia detection from consumer wearable devices photoplethysmography (PPG) sensors. Kai also worked with summer student Yihang Jiang to develop a novel method for multimodal signal alignment by modifying traditional Dynamic Time Warping methods.
Blood Pressure Variability
Connor's project is focused on exploring how variability in clinical blood pressure measurements affects hypertension diagnosis and treatment. Connor has developed an in silico cohort using a Bayesian approach to simulate potential misdiagnoses arising from measurement inaccuracies.
Heart Rate Variability & Meditation
Mengjie's master thesis focuses on exploring whether short and repeated meditation sessions are associated with changes in cardiovascular digital biomarkers. These brief meditation sessions may help to restore autonomic balance and could be incorporated into daily living through minor lifestyle changes.
Prometheus Flu Study
Emilia's project is focused on developing digital biomarkers of early infection, including digital biomarkers of resilience to contracting the infection. Her particular area of interest is on H1N1 influenza infection through collaboration with the Duke Biochronicity project. Such digital biomarkers could transform the role of consumer wearable devices into tools for continuous, real-time infection diagnostics.