Silent Paroxysmal Atrial Fibrillation Detection by Neural Networks Based on ECG Records

Atrial Fibrillation (AF) is one of the most common arrhythmias in the world and it is a life-threatening disorder which increases the risk of stroke with time. PAF is a special type of AF which is usually seen as temporarily AF that could last less than a week and terminated by itself. The disorder increases with age and is associated with cardiac morbidity. In this study, silent AF (SAF) detection is done by using different neural network schemes. At first step, features selection is done by utilizing Genetic algorithm. This step results in obtaining 8 HRV features. In the next step, obtained feature space is given to neural networks. The proposed approach provides good classification performance in detecting PAF events.