Exploration, processing and visualization of physiological signals from the ICU

Abstract

This report studies physiological signals measured from patients in the Intensive Care Unit (ICU). The signals explored include heart rate, arterial blood pressure, pulmonary artery pressure, and central venous pressure measurements. Following an introduction to these signals, several methods are proposed for visualizing the data using time and frequency domain techniques. By way of a patient case study we motivate a novel method for data clustering based on the singular value decomposition and present some potential applications based on this method for use within the ICU setting.