Symmetric Projection Attractor Reconstruction Analysis as a Method to Assess Seismocardiogram Quality in a Healthy Population

Abstract

Signal quality significantly affects the processing, analysis, and interpretation of biomedical signals. There are many procedures for assessing signal quality that use averaged numerical values, thresholding, analysis in the time or frequency domain, or nonlinear approaches. An interesting approach to the assessment of signal quality is using symmetric projection attractor reconstruction (SPAR) analysis, which transforms an entire signal into a two-dimensional plot that reflects the waveform morphology. In this study, we present an application of SPAR to evaluate the quality of seismocardiograms (SCG signals) from the CEBS database, a publicly available seismocardiogram signal database. Visual inspection of symmetric projection attractors suggests that high-quality (clean) seismocardiogram projections resemble six-pointed asterisks (∗), and any deviation from this shape suggests the influence of noise and artifacts.Clinical relevance - SPAR analysis enables quick identification of noise and artifacts that can affect the reliability of the diagnosis of cardiovascular diseases based on SCG signals.