Extraction of respiratory activity from ECG and PPG signals using vector autoregressive model

Respiratory signal is usually recorded with techniques like spirometry, pneumography or whole body plethysmography. These techniques require the use of cumbersome devices that may interfere with natural breathing, unmanageable in certain applications such as ambulatory monitoring, stress testing, and sleep studies. Infact, the joint study of cardiac and pulmonary systems is of great interest in most of these applications. Particularly the methods for extraction of respiratory information from physiological signals are attractive to pursue. In this present work we are addressing a method for extraction of respiratory activity from commonly available physiological signals such as ECG and Photoplethysmogram (PPG) using vector auto regressive (VAR) modelling technique. To test the efficacy of the proposed technique, the method is applied on a set of fifteen data records with different breathing rates and respiration amplitudes of physiobank archive for extraction of respiratory activity from the ECG and PPG signals. Extracted respiratory signal using the proposed bivariate VAR model is compared with the original respiratory signal present in the record and is considered as reference signal for comparison. Correlation analysis done in both frequency and time domains has shown a high degree of acceptance for the extracted respiratory signal with respect to the original reference respiratory signal. Higher values of accuracy rate clearly indicated significance of the extracted respiratory signal from the ECG and BP signals, when compared with the original recorded signal and could become a better alternative to the classical methods for recording respiratory signals.