Atrial fibrillation classification method for patients with different pharmacological or surgical therapies

This method aimed to assess electrophysiologists to choose the most suitable therapy for patients suffering from atrial fibrillation (AF), depending on whether a paroxysmal or persistent episode is presented. Since the surface ECG masks the differentiation between subtypes of AF, an early detection of paroxysmal episodes allows a clinically preventive treatment to stop recurrence and the natural progression towards persistent AF. Features of the General Fourier-family time-frequency transform were used as inputs of a Linear Discriminant Analysis classifier. Accuracy, sensitivity and specificity were measured to evaluate performance. AF episodes are mostly correctly classified, having into account that, from a clinical point of view, it is more important to detect almost every paroxysmal episode than viceversa, in order to stop the progression of these patients towards persistent AF.