The design and implementation of high performance differential voltmeters for semi-parallel data acquisition are described. The general requirements and specific conditions encountered in electrical impedance tomography (EIT) for accurate measurements are analysed. The major parameters are common-mode rejection and bandwidth. A specific implementation of the voltmeters, with separate DC supplies and independent signal references, is described. This arrangement, in which each voltmeter follows the input signal, automatically cancels any common-mode voltage present at the input. The signal is fed to the remainder of the instrumentation through a transformer. The use of a reduced number of components contributes to the minimisation of the inter-channel variations. Furthermore, the geometrical distribution of the voltmeters around the object minimises the length of electrode wires, also reducing the input capacitance. The number of modular voltmeters and DC/DC converters is 32. The common-mode rejection of these voltmeters is greater than 72 dB in the frequency range 3.6-560 kHz. In conclusion, the proposed solution ensures a minimisation of common-mode errors and enables the use of a 250 kHz frequency.
In this paper, by solving the forward problem by the boundary element method, the relationship between the conductivity changes and the boundary potential differences is studied and the sensitivity coefficients are evaluated.
The main experimental problems (temperature, electrode, polarisation, non-uniformity of electric field, non-homogeneity of biological material) of impedance measurements of biological tissue are discussed. Minimisation of the inherent sources of experimental error is proposed.
In order to characterise small samples of tissue by complex impedance measurements between 10(-6) and 20 MHz, an experimental set-up is proposed. This paper describes the methodology used and presents some preliminary results.
The Southampton electrical impedance tomography (EIT) system used a Sheffield data acquisition unit and a PC based 'Harlequin' transputer card to reconstruct and display images of the distribution of internal conductivity within the thorax. The system produces real-time images relating to both cardiac and pulmonary function. As a first step towards diagnosis using these images neural nets have been applied to the identification of regions of interest in the EIT images for which some activity with time, such as ventricular ejection, is sought. This paper addresses the use of a back-projection network to identify characteristic regions within the images. The network facilitates the production of automated real-time activity plots by defining their effective extent in the images of specific organs. The application is novel within the medical imaging field as the aim is to use neural networks for real-time image analysis.
The objective of this study is to improve the quality of the hardware of the existing Lund impedance tomography system. Improvement in the current generator and different isolation proposals are presented.
A wideband high CMRR instrumentation amplifier is described. Based on a previously reported current-mode design, the improved amplifier features excellent CMRR performance of better than 60 dB up to 200 kHz with unity differential gain.
The design and implementation of high performance differential voltmeters for semi-parallel data acquisition are described. The general requirements and specific conditions encountered in electrical impedance tomography (EIT) for accurate measurements are analysed. The major parameters are common-mode rejection and bandwidth. A specific implementation of the voltmeters, with separate DC supplies and independent signal references, is described. This arrangement, in which each voltmeter follows the input signal, automatically cancels any common-mode voltage present at the input. The signal is fed to the remainder of the instrumentation through a transformer. The use of a reduced number of components contributes to the minimisation of the inter-channel variations. Furthermore, the geometrical distribution of the voltmeters around the object minimises the length of electrode wires, also reducing the input capacitance. The number of modular voltmeters and DC/DC converters is 32. The common-mode rejection of these voltmeters is greater than 72 dB in the frequency range 3.6-560 kHz. In conclusion, the proposed solution ensures a minimisation of common-mode errors and enables the use of a 250 kHz frequency.
Many methods have been proposed for reconstructing the electrical conductivity and permittivity inside a body from measurements made on the body's surface. We outline several of the different approaches and then describe in more detail some of the methods we have used to make images in two and three dimensions.
Two experimental studies are presented that attempt to assess the use of electrical impedance tomography (EIT) to map thermal changes deep in the body. The first was a joint study between our group in Sheffield and the Daniel den Hoed Cancer Center in Rotterdam, funded by the COMAC-BME Hyperthermia project. Phantom and in vivo experiments were conducted with a deep heating ring-capacitive hyperthermia device; the results highlighted some of the inherent problems (electromagnetic interference, nonuniform sensitivity) that remain to be solved for this application. The second experimental study involved heating the stomachs of volunteers by pumping known quantities of liquid (a salt/glucose solution, conductivity 5 mS) at controlled temperatures (25, 37 and 47 degrees C) in and out of the stomach via a nasogastric tube. Results indicated that the level of thermal change induced by these liquids could be reproducibly measured by impedance imaging. Both studies were a further step in assessing the capabilities of EIT for noninvasive monitoring of deep body hyperthermia. The results are encouraging and indicate the value of continued development of EIT for non-invasive thermometry.
Measuring the reactive component of the bio-impedance enables a full characterisation of the frequency response of a tissue. The amplitude of the reactive component is relatively small in the frequency range generally used in electrical impedance tomography (EIT). Its measurement is therefore more sensitive to errors. At higher frequencies, the amplitude of this component increases, which increases the signal-to-noise ratio. The stray capacitance, however, also increases and the front-end circuit must be designed carefully. The purpose of the present study is to show the feasibility of the collection of data at relatively high frequencies; 31.25 and 250 kHz were used. Both the real and reactive components were used to reconstruct images from capacitive targets. This study suggests that it may be possible to use multifrequency systems to determine the parameters of frequency loci and therefore tissue characterisation.
In electrical impedance tomography, two-dimensional (2D) finite element solutions are used in the imaging algorithms. It is assumed that a major part of the current flowing through the object is restricted to the measurement plane (i.e. the plane determined by the electrodes which are used for measuring voltage differences) and the current flowing elsewhere is negligible. However, there is usually a three-dimensional (3D) variation of the conductivity distribution and if there are regions of high contrast close to the measurement plane, the measured voltage values may be considerably affected. In this work a 3D finite element analysis is utilised to demonstrate the previously mentioned effects. Examples are given to show the measured voltage differences for conductivity distributions which are identical on the measurement plane but different elsewhere.
In order to circumvent the electrode position determination problem in static electrical impedance tomography, it is possible to insert the object to be imaged into a water-filled cylinder on which the electrodes are at fixed and known positions. It has previously been shown that if the boundary of the internally placed object and the conductivity of the salty water in the cylinder are known, then a significant improvement in the conductivity image of the object is obtained. An algorithm for finding the boundary of an internally placed object is developed based on the finite element method (FEM). The boundary is assumed to obey a parametric model and the parameters are estimated by inverting a matrix representing the sensitivity of the boundary voltage measurements to parameter variations. The algorithm assumes that the object's internal conductivity is uniform and known. Simulation studies show that if the internal conductivity is not uniform to the extent found in the arm cross-sections, up to 9% error in the boundary, as measured from a centrally placed reference point, may result. It is also shown that if previous knowledge about the boundary shape is used to model the boundary with fewer numbers of parameters, then the boundary may be found with less error.
One of the design considerations for electrical impedance tomography phantoms is that they must be easy to model accurately. This paper describes a phantom with this property. Experimental results from its evaluation and testing are given.
The 'sensitivity region' method of back projection is converted from polar drive to adjacent drive current injection strategy and individual sensitivity coefficients incorporated. For each projection sensitivity coefficients for each pixel were established and pixels were grouped together into the 'sensitivity regions'. These contain all pixels that have their largest sensitivity coefficients associated with a particular measurement pair. Back projection of measurement data was performed using these coefficients as the basis of the reconstruction and less sensitive measurement pairs were ignored. Formulae for the back projection are discussed and images from computer simulated data and raw data supplied by Sheffield as part of this concerted action are presented. Comparison with polar current injection strategy is discussed.
If a cylinder carries on its surface an electric charge of a density which is constant in the axial direction, but which varies sinusoidally with an integral number of wavelengths around the circumference, then inside the cylinder the electric field has its maximum value at points adjacent to the surface, and reduces continuously as the axis is approached. The rate of reduction increases with the number of wavelengths. A rotating cylinder of such charge may be simulated by a number (divisable by three) of equally spaced conducting bars mounted longitudinally on the surface of a stationary cylinder, the bars being connected in sequence to the lines of a three-phase voltage supply. The number of wavelengths of charge around the circumference can be varied by the connection together of a chosen number of adjacent bars effectively to form a single bar, and so fields of varying penetration depths can be produced from a fixed frequency supply. The variation of the resulting currents would be related to the corresponding variation in the complex conductivity of the cylinder material.
At the ESAT-MICAS research facility in Leuven, several EIT systems have been designed and realised. The latest hardware set-up makes use of a PC to control the data collection and to reconstruct the images. A voltage drive strategy and no common-mode feedback are some of its specific characteristics. The function-generator produces signals with a frequency between 10 and 100 kHz, so multifrequency images can also be produced. Static images have already been obtained and (semi-)real time imaging is possible with our latest mark IV system. This system has 16 bit analogue-to-digital convertors and is capable of taking 50 x 10(3) samples/s.
An investigation was conducted to determine and quantify the distortions in applied potential tomography (APT) images reconstructed from data originating from bodies of non-uniform reference conductivity distributions. The results show that the distortions in the images are dependent on the reference conductivity distribution and on whether the images are formed by back projection along the assumed equipotentials of a uniform reference conductivity distribution or along the equipotentials of the true conductivity distribution. We believe that this last finding is significant since our previously held expectation, similar to that of Yorkey and Webster (1987), that back projection along the true equipotentials of the reference conductivity distribution should result in an accurate reconstruction, is shown to be incorrect.
Respiration-related changes in the complex impedance were obtained on the thorax in three volunteers. The real part of the image clearly showed the lungs as regions of increased conductivity on expiration. The imaginary part of the image, reflecting changes in the ratio of permittivity to conductivity, showed a central negative region surrounded by a positive region extending to the periphery of the lungs. These features may be due to movement of the diaphragm and liver within the sensitive volume during respiration.
The main experimental problems (temperature, electrode, polarisation, non-uniformity of electric field, non-homogeneity of biological material) of impedance measurements of biological tissue are discussed. Minimisation of the inherent sources of experimental error is proposed.
