The study of the impact of murmurs on heart sounds by using multiple signal classification pseudo-spectrum

The aim of this study was to present a novel framework for the analysis of the impact of murmurs on heart sounds recorded in real clinical environment. Heart sound records were rigorously selected from the PASCAL dataset using an automated signal quality assessment algorithm. Recordings from 159 patients were analyzed for spectral differences in normal, systolic and diastolic murmurs using a Multiple signal classification algorithm pseudo-spectrum. The spectral features evaluated for the first heart sound (S1) and the second heart sound (S2) were: energy, frequency and frequency density. Results show increased energy of fundamental heart sounds in systolic and diastolic murmurs similarly, whilst frequency is decreased inversely. Furthermore, the frequency density of the first and second heart sounds decreases in murmurs and it is shown to be the lowest in systolic murmur cases.