Research on predicting velocity-time integral of shunt How using time-frequency features of heart sound with correlation

Some cardiac diseases, such as ventricular or aortic septal defects, will cause shunt lesions accompanied by murmurs. The hemodynamic properties of shunt flow can he measured quantitatively through Doppler echocardiography. The velocity-time integral (VTI) of shunt flow is one of the important hemodynamic properties and is commonly used for the evaluation of cardiac function and treatment. On the other hand, heart sounds are used to predict cardiac diseases by mapping them to the time-frequency domain. This study aims to analyze correlations between different frequency hands of heart sounds and hemodynamic properties of shunt flow and predict VTI values using time-frequency features of heart sounds. The results showed that different frequency hands of heart sounds contributed significantly to the correlation and the support vector regression model has a certain potential in the prediction of VTI.