Method for classifying cardiac arrhythmias using photoplethysmography.

Abstract

Advances in mobile computing and miniature devices have contributed to the accelerated development of wearable technologies for clinical applications. The new trend of wearable technologies has fostered a growth of interest for sensors that can be easily integrated into wearable devices. In particular, photoplethysmography (PPG) is especially suitable for wearable sensing, as it is low-cost, noninvasive, and does not require wet electrodes like the electrocardiogram. Photoplethysmograph signals contain rich information about the blood pulsating variation which is strongly related to the electrical activities of the heart. Therefore, in this paper we hypothesize that the ambulatory PPG monitoring could be employed for arrhythmia detection and classification. This paper presents a method for classifying ventricular premature contraction (VPC) and ventricular tachycardia (VT) from normal sinus rhythm (NSR) and supraventricular premature contraction (SVPC) recorded in patients going through ablation therapy for arrhythmia. Although occasional VPCs are benign, the increase in the frequency of VPC events may lead to VT, which in turn,could evolve into ventricular fibrillation and sudden cardiac death. Therefore the accurate measurement of VPC frequency and early detection of VT events becomes essential for patients with cardiac disease.