Detection of Cardiac Alternans via Statistical Analysis of Eigenvalues of a Dynamical System

• Activity Number: 636
• Type: Contributed
• Date/Time: Thursday, August 5, 2010 : 8:30 AM to 10:20 AM
• Sponsor: Section on Statistical Computing
• Author(s): Adam Petrie*+
• Companies: University of Tennessee
• Keywords: alternans ; dynamical system ; eigenvalue ; maximum likelihood ; genetic algorithm

Abstract

Predicting cardiac alternans is a crucial step toward detection and prevention of ventricular fibrillation, a heart rhythm disorder that kills hundreds of thousands of people in the US each year. We view the heart in terms of some complicated and intractable dynamical system whose true state vector is unknown. Alternans occurs when an eigenvalue of this system becomes -1. We treat the observed time sequence of action potential durations in paced cardiac tissue as a projection of the true state vector, and present a maximum likelihood method to estimate the dominant eigenvalues of the dynamical system using this pseudo-state space. We present numerical examples based on a toy model and a cardiac model which verify that our approach can accurately find dominant eigenvalues and can predict when alternans may occur.