Session P7C.8

Variability of ECG Features on the Torso Plane: A New Approach to Myocardial Infarction (MI) Detection

H SadAbadi*, M Ghasemi, A Jalali, M Atarod, A Ghaffari

K.N. Toosi University of Technology
Tehran, Iran

Hypothesis/Objective: An increased diagnostic accuracy of the exercise ECG can be achieved by improving the signal processing or developing the ECG leads. Development of ECG leads has been studied by changing the location of the electrodes or recording the body surface potential maps to provide a more representative picture of cardiac activity and hence improve the overall process of patient diagnosis. Body surface potential mapping (BSPM) is a technique used to record the ECG from a large number of recording sites on both anterior and posterior surfaces of the torso. Such an approach has the potential to provide the basis for improved cardiac diagnosis. In this paper, we conduct a method based on BSPM for accurate detection of MI size and location.
Method: In this paper, we address a new method base on behavior of some ECG’s features, which are Q, R, and T amplitudes and ST dispersion. We call these Q, R, T, and ST curves. At the first step, by plotting the variability of Q, R, and T amplitudes and ST dispersion for horizontal and vertical nodes or simply horizontal and vertical lines of torso plane, these curves are obtained. All the nodes of torso plane which mostly refer to the infarcted area of the heart are called abnormal nodes else are called normal nodes. Similarly, all the lines that cross only normal nodes are called normal lines, and else are called abnormal lines. The behavior of the mentioned curves for normal lines is smooth. However, it is disturbed for abnormal ones. Finally, we use two neuro-fuzzy networks for clustering normal and abnormal lines, one for horizontal and the other for vertical ones. The overlap of horizontal and vertical abnormal lines determines an area which refers to the infarcted myocardial.
Results and Conclusion: We evaluated the method on CinC/PhysioNet Challenge 2007 data and looking forward to submitting our first entry to the Challenge soon. The simplicity of the method is its major advantage.

(Abstract Control Number: 303)