Image reconstruction and spectral expansion analysis in electrical impedance tomography

Electrical impedance tomography (EIT) is a recently developed technique for producing cross-sectional images of electrically conducting objects. Its applications in medicine stem from the fact that physiological events cause changes in the electrical resistivity of biological tissues. The development of an iterative image reconstruction technique is described, its properties are examined and it is shown that it performs better than an algorithm based on backprojection between equipotential lines. It is also shown that the iterative algorithm can be described by the more general technique of spectral expansion which expands the EIT image domain into a set of basis images. The versatility and flexibility of this method is demonstrated and it is shown that it provides an appropriate framework for comparing existing reconstruction algorithms and also furnishes a means of designing new ones. A study of some criteria which may be used in constructing an appropriate algorithm has been presented. Spectral expansion provides unique insight into the characteristics of the imaging system, for example the resolution versus noise trade-off for different parts of the image can be easily assessed. A preliminary study of a sample population of normal volunteers is reported and the methods for extracting quantitative information and images of cardiac ventricular ejection and pulmonary perfusion are described. The results presented show how appropriate processing can avoid the long data collection times previously required to separate the weak cardiac related and the larger ventilation signals. The study has shown that the technique can be useful in a clinical environment such as continuous monitoring of intensive care patients.