Separation of ventilation and perfusion related signals within EIT-data streams

Abstract

Electrical impedance tomography is a widely established technique for ventilation monitoring in the case of clinical research. To assess the capability of a patient's gas exchange, it is necessary to determine the regional supply of ventilation V as well as the local perfusion Q. Therefore, many approaches have been investigated to extract both signals from EIT-data, which would allow a V/Q-mapping. The challenge of this issue depends on the very weak perfusion related signal compared to the signal of ventilation. First approaches to visualize the perfusion were performed inducing apnea, ECG-gating or were based on Fourier series. Rather new techniques use principle component analysis (PCA) defining eigenvectors as a signal in time and thus try to determine "typical" ventilatory as well perfusion related signal streams. In this work, the algorithms of separation will be analyzed and compared (using apnea needs no separation and will not be discussed in this work). © 2010 IOP Publishing Ltd.