Development of a cardiovascular monitor from hardware design to data analysis algorithms
Continuous ECG monitoring has proven to be useful in early detection of malignant arrhythmias and subsequent treatment of life threatening cardiac events. However, current medical instrumentation devices are not cost effective to be widely deployable to the general cardiac patients. Furthermore, algorithms needed for accurate diagnosis of malignant cardiac events are currently in a great demand. To this end, the objectives of this dissertation were to develop a low cost, personal digital assistant (PDA) based ECG monitor and novel data analysis algorithms that may lead to better quantification and differentiation of malignant cardiac dynamics from the normal ones. For the hardware development, a PDA-based ECG monitor, which can support data collection from two different signals, was developed. Moreover, software implementation for the data collection and subsequent data processing techniques that are applicable to both PDA and PC platforms were developed. For the first phase of the data analysis algorithm, a general-purpose algorithm for estimating high-resolution time-varying spectral contents as well as an instantaneous amplitude extraction was developed. The rationale for this development was based on the fact that cardiac events are highly nonstationary, and thus, may lead to better quantification and detection of dynamic that may differ from normal events. For the second phase of the data analysis algorithm, general-purpose algorithms that can extract linear and nonlinear dynamics underlying multi-variable cardiovascular signals were developed. While these methods are general-purpose algorithms that can handle many different cardiovascular related signals, verification of the algorithms was based on the ECG and blood pressure recordings obtained from human subjects. It is hoped that the new technology integrating both hardware and software development may lead to better understanding and more accurate diagnosis of malignant cardiac events via the means of a simple and noninvasive PDA-based cardiovascular monitor.