Computerized analysis and modelling of blood pressure: Dissertation

This thesis concerns the development of methods and techniques to analyse continuous, ambulatory recordings of blood pressure in order to study and ultimately diagnose the functioning mechanisms of the cardiovascular control system.A long term recording of the continuous blood pressure signal contains a huge amount of data which includes, among others, information about the responses of the cardiovascular control system to the various stimuli of everyday life.The study discusses the development of analysis environments and the development of methods to quantify the long term and short term features in blood pressure recordings.A feasibility analysis is presented in order to compare the continuous invasive method with an automatic cuff method. determination of the normotensive diurnal blood pressure profile and variability, utilizing the developed methods, is also discussed.Two computerized analysis environments were developed, one with a low cost microcomputer and the other with a powerful minicomputer.The developed methods enable an overall analysis with statistical indices.The analysis of long term features is based on signal decomposition in time domain with non linear median type processing.For the analysis of short term features a new method was introduced utilizing the cardiovascular closed loop system between heart rate and blood pressure, identified by using the multivariate autoregressive modelling technique.