Title: Evaluating the efficacy of Generative Model Based Resource Efficient Monitoring for Electrocardiogram and Photoplethysmogram sensing Specific Aims: Nearly 1 in 150 adults in America are suffering from some form of congenital heart disease. A recent survey by the center for disease control (CDC) has concluded that nearly $2 billion hospital expenditures are related to congenital heart disease. Long term physiological monitoring of the heart is essential in diagnosis, treatment and rehabilitation of congenital heart disease patients as well as determining the optimal time for surgery. The most common physiological signals that are monitored in ICU or at home settings for individuals with heart defects are blood oxygen levels (Photoplethysmogram, PPG) and electrocardiogram (ECG) signals. In a hospital telemetry setting although the focus is on remote monitoring of patients, the initial link from the patient to the monitoring unit is wired. The state of the art medically approved wireless technique to monitor the heart is the Holter monitor which can at most sustain 48 hours of continuous monitoring due to storage and energy limitations. Commercially available heart sensors have been reported to last nearly 84 hours without needing battery replacements . These lifetimes are clearly not enough since long term monitoring for congenital heart diseases are typically prescribed for months if not years . In our previous work, we have proposed generative model based resource efficient monitoring (GeMREM), a novel sensing algorithm that can result in a 40 fold increase in lifetime of ECG and a 300 fold increase in the lifetime of a PPG sensor. In principle, the GeMREM technique considers the periodicity of PPG or ECG signals and develops a generative model. The generative model if supplied with the correct parameters can generate synthetic physiological signals that are equivalent to the original signal in diagnostic content. In an embodiment of a sensor, if the sensed physiological signal matches the model then the sensor does not store or transmit signals to the monitoring station or base station. If the signal does not match the model then sample by sample signal is stored or transmitted. Theoretical study with MIT BIH data shows that this technique can result in huge increase in lifetime and reduction in storage requirements while maintaining required accuracy. However, to propose a medical grade GeMREM enabled sensor we need to ensure a) effectiveness: a trained physician should not be able to distinguish between GeMREM synthesized ECG signal and the signal obtained from a Holter monitor, and b) efficacy: the lifetime increase and storage reductions proposed in theory are exhibited in practice. In an ongoing clinical study at the ICU of St. Lukes hospital under supervision of Dr. Richard Houser we have deployed GeMREM enabled Shimmer ECG sensors on 25 patients each monitored for 20 hours. The main aim of the study is to establish the effectiveness of the technique. The proposed research is aimed towards establishing the efficacy in terms of lifetime increase, and storage decrease for both ECG and PPG sensors. The specific aims of this research are as follows: a) Deploy GeMREM enabled ECG sensors for efficacy study The effectiveness study considered equivalence of the synthetic signal generated by GeMREM enabled Shimmer sensors and original signal collected by Holter monitors in terms of diagnostic properties established by a heart specialist Dr. Richard Houser. In this research, we will keep the same diagnostic quality and evaluate the lifetime and storage requirements of the sensor. We will deploy the sensors on patients for 24 hours and monitor the initial battery level before the study and the final battery level after the study of the sensors. A battery model will then be used to estimate the lifetime. The data will be stored in a smartphone database and the size of the data base will be used to determine the storage requirements. b) Develop GeMREM enabled PPG sensors In this aim we will implement the GeMREM algorithm for PPG sensing in the open source Pulse sensor by Arduino. The Arduino sensor will directly communicate with the smartphone via Bluetooth. c) Deploy PPG sensors for effectiveness study We will perform a similar clinical study as the ongoing study at St Luke, to determine the effectiveness for the PPG sensing technique. We will compare the GeMREM generated synthetic signals with medical grade Smithsoem pulseoximeters. Public Health Relevance: If the effectiveness and efficacy of this technique is proven through the proposed clinical study then this technique can revolutionize long term remote monitoring of chronic diseases by increasing the lifetime of sensors from days to months. Further, such a sensing technique will bring down the energy consumption of the sensors to a level that they can be powered using energy scavenged from body heat. Thus, battery less sensing will come closer to reality than ever before. This will help reduce sensing cost as well as sensor size
|Effective start/end date||9/1/14 → 5/31/19|
- HHS: National Institutes of Health (NIH): $400,053.00
Explore the research topics touched on by this project. These labels are generated based on the underlying awards/grants. Together they form a unique fingerprint.