Comparison of five regularization methods for the inverse electrocardiography problem

Download
2013
Güçlü, Alperen
Understanding heart’s electrical activity is very important because coronary problems -such as heart attacks, arrhythmia and stroke- are the leading cause of death in the world. Forward and inverse problems of electrocardiography (ECG) are methods that provide detailed information about the electrical activity of the heart. Forward problem of electrocardiography is the estimation of body surface potentials from equivalent cardiac sources. Inverse problem of electrocardiography can be described as estimation of the electrical sources in the heart using the potential measurements obtained from the body surface. Due to spatial smoothing and attenuation that occur within the thorax, inverse ECG problem is ill-posed and the transfer matrix is ill-conditioned. Thus, regularization is needed to find a stable and accurate solution. In this thesis, epicardial potentials used as equivalent cardiac sources to represent electrical activity of the heart and performances of five different regularization methods are compared. These regularization methods are Tikhonov regularization, truncated singular value decomposition, least squares QR factorization, truncated total least squares, and Lanczos truncated total least squares. Results are assessed qualitatively using correlation coefficient (CC) and relative difference measurement star (RDMS) measures. In addition, real and reconstructed surface potential distributions are compared qualitatively. Body surface potential measurements are simulated with different levels of measurement noise. Geometric errors are also included by changing the size and the location of the heart in the mathematical torso model. According to our test results, the performances of the regularization methods in solving the inverse ECG problem depend on the form and amount of the noise.

Suggestions

Model-based evaluation of the control strategies of a hand rehabilitation robot based on motor learning principles
Yağmur, Onur Can; Turgut, Ali Emre; Arıkan, Kutluk Bilge; Department of Mechanical Engineering (2022-1-28)
Stroke is an important health problem that occurs after blockage or bleeding in the vessels feeding the brain. It is one of the leading causes of death in the world. Patients with surviving hemiplegia often have a loss or decrease in voluntary movement of the right or left side of the body. The disease reduces the quality of life of patients with its spasticity and limits their independence; This situation poses an important problem for patients, their relatives, and the whole society. This thesis aims to d...
Investigation of on skin surface response due to acoustic radiation from stenosed blood vessels
Salman, Huseyin E.; Yazıcıoğlu, Yiğit (2015-11-02)
Arterial stenosis is a form of cardiovascular disease which leads to highest rate of fatalities worldwide. When stenosis is present in coronary arteries, it leads to heart attack and often sudden death. Arterial disease is usually not confined to critical organs such as the heart and the brain but observed throughout the peripheral cardiovascular system. When an artery has a diameter reduction, acoustic radiation originating from this constriction propagates through soft tissues and reaches to skin surface....
Comparison of ML and MAP parameter estimation techniques for the solution of inverse electrocardiography problem
Erenler, Taha; Serinağaoğlu Doğrusöz, Yeşim; Department of Electrical and Electronics Engineering (2018)
This study aims to determine the cardiac electrical activity from body surface potential measurements. This problem is called the inverse problem of electrocardiography. Reconstruction of the cardiac electrical activity from the body surface potential measurements is not an easy task, since this problem has an ill-posed nature due to attenuation and spatial smoothing inside the medium between the source and the measurement sites, meaning that even small errors in the mathematical model or noise in the measu...
Investigation of the association between genetic and activity polymorphisms of paraoxonase 1 and ischemic stroke risk
Demirdöğen Can, Birsen; Adalı, Orhan; Department of Biochemistry (2007)
Stroke is the third leading cause of death. Atherosclerosis in the carotid arteries is a risk factor for ischemic stroke. Oxidized low density lipoprotein (LDL) plays a central role in the progression of atherosclerosis. Human paraoxonase 1 (PON1), a high-density lipoprotein (HDL) associated serum esterase/lactonase, protects HDL and LDL from oxidative modifications. Thus, PON1 is protective against the development of atherosclerosis. PON1 gene has two functional coding region (192Q/R and 55L/M) and one pro...
Virtual Heart Models: Multi-Physics Approaches to Computational Cardiology (VHEART)
Göktepe, Serdar(2015-8-31)
Heart disease is the number one cause of death in industrialized nations. Despite the broad class of treatment techniques such as medication, surgery and tissue-engineered therapies, heart disease remains to be one of the most frequent, disabling, and life-threatening diseases. In Europe it accounts for almost half of overall annual mortality rate. In the European Union (EU) alone, cardiovascular disease causes over 2 million deaths per year. The cost of cardiovascular disease to the EU economy is €192 bill...
Citation Formats
A. Güçlü, “Comparison of five regularization methods for the inverse electrocardiography problem,” M.S. - Master of Science, Middle East Technical University, 2013.