Non-invasive Imaging of Electrical Activity of the Heart in terms of 3 Dimensional Transmembrane Potential Distributions

(Kalbin Elektriksel Aktivitesinin 3 Boyutlu Transmembran Potansiyel Dağılımları Cinsinden Girişimsiz Olarak Görüntülenmesi)

This work was supported by TUBITAK – 1001, Grant Number 111E258

Project Personnel

Assoc. Prof. Dr. Yeşim Serinağaoğlu Doğrusöz (PI)

Graduate students supported by the project grant:

Alireza Mazloumi Gavgani

Mir Mehdi Seyedebrahimi

Gizem Bedir

Fourough Gharbalchi No

Other graduate students with contributions to the project:

Uğur Cünedioğlu

Summary of the Project

Inverse electrocardiography is the estimation of cardiac electrical sources from body surface potential (BSP) measurements. Inverse solutions can guide the physicians for diagnosis and treatment planning of lethal heart diseases. However, inverse problem is ill-posed and even small perturbations in the measurements yield unbounded errors in the solutions. To overcome this difficulty, many regularization approaches have been proposed in literature. However, these methods have been applied and tested under varying conditions in different studies; there is no consensus among researchers on the method with the best performance.

Lately, solutions in terms of transmembrane potentials (TMP) have become popular, since they provide information about the electrical activity of the three dimensional myocardium. There are few studies in this area and it is still an open question how different methods will perform under different arrythmia conditions. The main goal in this project is to solve the inverse problem in terms of TMPs, using different approaches but under the same (and diverse) cardiac conditions. First, we obtained TMP distributions for various cardiac electrical activity assumptions using Aliev-Panfilov model. Then we solved the forward ECG problem to obtain the corresponding BSPs, which were later used in the inverse problem solutions. Among the five inverse approaches, Bayesian MAP estimation had the best performance under all conditions. TTLS, LTTLS and LSQR were also successful in finding the initial stimulation points and recovering the wavefront.

We made contributions in two more areas in this project. The first one is our study of fiber orientation effects on TMP distributions. We found that even using fiber orientations from a different heart is much better than using the isotropic assumption. The second one is the analytical verification of the FEM based forward problem; with an appropriate mesh size, we showed that the numerical solution converges to the analytical solution.

Figure 1. Transmebrane potential (TMP), epicardial and body surface potential distributions in a heart with normal tissue physiological parameters. Initial stimulation is given to mimim WPW syndrome.

Figure 2. Transmebrane potential (TMP) distributions for a heart with ischemic region at 5 different time instances. (A) View from the outer surface, (B) x-, y- and z-axes cross-sectional view, (C) Multiple x-axis cross-sectional view.

Figure 3. True transmebrane potential (TMP) distributions and inverse problem solutions for a heart with ischemic region, and ectopic stimulation, at 6 different time instances. From top to bottom: True (original) maps, Tikhonov regularization solution, truncated total least squares (TTLS) solution, Lanczos bidiagonalization based TTLS solution, Lanczos diagonalization based least squares QR solution, and Bayesian MAP estimation solution.

Publications, Talks and Theses from the Project Work

Mir Mehdi Seyedebrahimi, “Simulation of transmembrane potential propagation in three dimensional ventricular tissue using Aliev Panfilov model,” Graduate Program of Biomedical Engineering, METU 2015 (co-supervised by Dr. Ferhat Eyyüpkoca)

Gizem Bedir, “Forward problem of electrocardiography in terms of 3D transmembrane potentials using COMSOL,” Graduate Program of Biomedical Engineering, METU 2015 (co-supervised by Dr. Barbaros Çetin)

Gizem Bedir, Barbaros Cetin, Yesim Serinagaoglu Dogrusoz, “Forward Problem of Electrocardiography in terms of 3D Transmembrane Potentials using COMSOL Multi-physics,” in review.

Mir Mehdi Seyedebrahimi, Yesim Serinagaoglu Dogrusoz, “Effects of Anisotropy and Errors in Fiber Orientations on 3D Transmembrane Potential Distributions: A Simulation Study,” in review.

Yeşim Serinagaoglu Dogrusoz, “Estimation of Ectopic Foci in the Heart using Epicardial and Transmembrane Based Solutions of Inverse Electrocardiography,” ECG Imaging Workshop, (invited talk) 25-28 March, 2015, Bad Herrenalb, Karlsruhe, Germany.

Yesim Serinagaoglu Dogrusoz, “Statistical Approaches in Electrical Imaging of the Heart,” IEEE International Symposium on Biomedical Imaging (ISBI), April 29-May 2, 2014, Beijing, China, (invited talk) 2014.

Mir Mehdi Seyedebrahimi, Yesim Serinağaoğlu Doğrusöz, Uğur Cünedioğlu, “Aliev-Panfilov Modeli Kullanılarak Normal ve Kısmi İskemik Dokularda Transmembran Potansiyellerinin Yayılımının Simülasyonu,” BIYOMUT, Ekim 2014, İstanbul, 2014.

Gizem Bedir, Yeşim Serinağaoğlu Doğrusöz, “COMSOL Çoklu Fizik Ortamı Kullanılarak Elektrokardiyografide İleri Problem Çözümü,” TIPTEKNO, Antalya, Turkey, 2013.

Mehdi Seyedebrahimi, Uğur Cünedioğlu ve Yeşim Serinağaoğlu Doğrusöz, “Transmembran Potansiyel Dağılımlarının 3 Boyutlu Normal ve Kısmi İskemik Ventriküler Geometride Simülasyonu,” TIPTEKNO, Antalya, Turkey, 2013.

Other Publications Related to Project Work

Fourough Gharbalchi No, “Body surface lead reduction algorithm and its use in inverse problem of electrocardiography,” Graduate Program of Biomedical Engineering, METU, 2015 (co-supervised by Dr. Gerhard Wilhelm Weber)

A. Mazloumi Gavgani, Y. Serinagaoglu Dogrusoz, “Noise Reduction using Anisotropic Diffusion Filter in Inverse Electrocardiology,” 33rd Annual International Conference of the IEEE EMBS, San Diego, California, USA, August 30 - September 3, 2012