Identifying and removing low quality seismocardiogram cycles

点击翻译按钮获取中文摘要

Cardiovascular diseases (CVDs) affect the lives of 1.6 million Canadians, and are the second leading cause ofdeath in Canada. CVDs cost the Canadian economy more than $20 Billion annually. Heart Force Medical Inc.(HFM) is a Vancouver-based company developing software and hardware technologies for the noninvasivecardiac monitoring to prevent and diagnose CVDs. In particular, HFM's proprietary technology measures heartvibration, which is referred as the seismocardiogram (SCG). SCGs can be captured by placing an accelerometeron the sternum, and provide information about mechanical aspects of the heart. Due to the inherent sensitivityof the SCG signal to motion artifacts caused by respiration or body movements, sophisticated algorithms haveto be developed to remove or minimize the effect of these artifacts. Another source of inadequate reliability ofthe SCG signal is the inappropriate placement of the accelerometer sensor on the sternum. In these situations,the SCG signals are of low quality (LQ) and the extraction of cardiac information is unreliable. Therefore, thepurpose of this proposal is to develop an algorithm that (1) can accommodate motion artifacts, and (2) canautomatically identify and remove LQ SCG cycles prior to analysis. Such an algorithm has to learn themorphology of recognizable and acceptable SCG cardiac cycles and discard those cycles when important datacannot be extracted. This step is crucial in the overall electromechanical assessment of cardiac function. Thedeveloped algorithm will be deployed in relevant HFM devices to increase their accuracy, reliability, andusability to evaluate cardiac performance.