Suppression of Cardiogenic Oscillations in Esophageal Pressure Signals Using Ensemble Empirical Mode Decomposition

Abstract

Several clinical parameters associated with a patient’s respiratory health can be derived from esophageal pressure (Peso). However, cardiogenic oscillations (CGOs) are a major source of interference in Peso signals, which makes it difficult to accurately monitor respiratory mechanics. In this study, we present a CGO suppression scheme using Ensemble Empirical Mode Decomposition (EEMD). The proposed method was applied to clinically recorded Peso signals from four mechanically ventilated patients, and was used to decompose the tracings into their intrinsic mode functions (IMFs). Ignoring the IMFs associated with CGO during the reconstruction process resulted in a signal with significantly reduced amplitude fluctuations. The magnitude spectrum of the reconstructed signal further indicates that the higher frequency components of CGO have been removed. Preliminary results suggest that the method described in this study has the potential to be used in the clinical domain for denoising Peso signals.