Spatial Distribution of Seismocardiographic Signal Clustering

Abstract

Seismocardiographic (SCG) signals are chest wall vibrations that correlate with cardiac activity and often measured using accelerometers on the chest surface [1]. SCG may be generated by valve movements, heart muscle contraction and blood flow momentum changes. Respiration is a source of variability and studying the SCG signal “clustering” with respiration may help define physiological mechanisms related to this SCG variation [2]. Grouping similar SCG events into clusters may also help reduce SCG variability and possibly increase its diagnostic utility [1], [3]. Previous studies often measured SCG signals at one location that varied among studies (e.g., xiphoid process, 4th ICS, etc.) [2]–[4]. However, SCG clustering may depend on SCG measurement location. The objective of the current study is to investigate the dependence of SCG clustering on the measurement location. This distribution may be of diagnostic value and can help compare results from different studies. It may also help define locations where clusters are best separated, which may help optimize choices of SCG measurement location.