Peter Rex Johnston

Selecting the corner in the L-curve approach to Tikhonov regularization

The performance of two methods for selecting the corner in the L-curve approach to Tikhonov regularization is evaluated via computer simulation. These methods are selecting the corner as the point of maximum curvature in the L-curve, and selecting it as the point where the product of abcissa and ordinate is a minimum. It is shown that both these methods resulted in significantly better regularization parameters than that obtained with an often-used empirical Composite REsidual and Smoothing Operator approach, particularly in conditions where correlated geometry noise exceeds Gaussian measurement noise. It is also shown that the regularization parameter that results with the minimum-product method is identical to that selected with another empirical zero-crossing approach proposed earlier.

A new method for regularization parameter determination in the inverse problem of electrocardiography

Computing the potentials on the hearts epicardial surface from the body surface potentials constitutes one form of the inverse problem of electrocardiography. An often-used approach to overcoming the ill-posed nature of the inverse problem and stabilizing the solution is via zero-order Tikhonov regularization, where the squared norms of both the surface potential residual and the solution are minimized, with a relative weight determined by a so-called regularization parameter. This paper looks at the composite residual and smoothing operator (CRESO) and L-curve methods currently used to determine a suitable value for this regularization parameter, t, and proposes a third method that works just as well and is much simpler to compute. This new zero-crossing method selects t such that the squared norm of the surface potential residual is equal to t times the squared norm of the solution. Its performance was compared with those of the other two methods, using three simulation protocols of increasing complexity. The first of these protocols involved a concentric spheres model for the heart and torso and three current dipoles placed inside the inner sphere as the source distribution. The second replaced the spheres with realistic epicardial and torso geometries, while keeping the three-dipole source configuration. The final simulation kept the realistic epicardial and torso geometries, but used epicardial potential distributions corresponding to both normal and ectopic activation of the heart as the source model. Inverse solutions were computed in the presence of both geometry noise, involving assumed erroneous shifts in the heart position, and of Gaussian measurement noise added to the torso surface potentials. It was verified that in an idealistic situation, in which correlated geometry noise dominated the uncorrelated Gaussian measurement noise, only the CRESO approach arrived at a value for t. Both L-curve and zero-crossing approaches did not work. Once measurement noise dominated geometry noise, all three approaches resulted in comparable t values. It was also shown, however, that often under low measurement noise conditions none of the three resulted in an optimum solution.

Application of sigmoidal transformations to weakly singular and near‐singular boundary element integrals

Accurate numerical determination of line integrals is fundamental to reliable implementation of the boundary element method. For a source point distant from a particular element, standard Gaussian quadrature is adequate, as well as being the technique of choice. However, when the integrals are weakly singular or nearly singular (source point near the element) this technique is no longer adequate. Here a co-ordinate transformation technique, based on sigmoidal transformations, is introduced to evaluate weakly singular and near-singular integrals. A sigmoidal transformation has the effect of clustering the integration points towards the endpoints of the interval of integration. The degree of clustering is governed by the order of the transformation. Comparison of this new method with existing co-ordinate transformation techniques shows that more accurate evaluation of these integrals can be obtained. Based on observations of several integrals considered, guidelines are suggested for the order of the sigmoidal transformations. Copyright

The effect of conductivity values on ST segment shift in subendocardial ischaemia

The aim of this study was to investigate the effect of different conductivity values on epicardial surface potential distributions on a slab of cardiac tissue. The study was motivated by the large variation in published bidomain conductivity parameters available in the literature. Simulations presented are based on a previously published bidomain model and solution technique which includes fiber rotation. Three sets of conductivity parameters are considered and an alternative set of nondimensional parameters relating the tissue conductivities to blood conductivity is introduced. These nondimensional parameters are then used to study the relative effect of blood conductivity on the epicardial potential distributions. Each set of conductivity parameters gives rise to a distinct set of epicardial potential distributions, both in terms of morphology and magnitude. Unfortunately, the differences between the potential distributions cannot be explained by simple combinations of the conductivity values or the resulting dimensionless parameters.

The importance of anisotropy in modeling ST segment shift in subendocardial ischaemia

In this paper, a simple mathematical model of a slab of cardiac tissue is presented in an attempt to better understand the relationship between subendocardial ischaemia and the resulting epicardial potential distributions. The cardiac tissue is represented by the bidomain model where tissue anisotropy and fiber rotation have been incorporated with a view to predicting the epicardial surface potential distribution. The source of electric potential in this steady-state problem is the difference between plateau potentials in normal and ischaemic tissue, where it is assumed that ischaemic tissue has a lower plateau potential. Simulations with tissue anisotropy and no fiber rotation are also considered. Simulations are performed for various thicknesses of the transition region between normal and ischaemic tissue and for various sizes of the ischaemic region. The simulated epicardial potential distributions, based on an anisotropic model of the cardiac tissue, show that there are large, potential gradients above the border of the ischaemic region and that there are dips in the potential distribution above the region of ischaemia. It could be concluded from the simulations that it would be possible to predict the region of subendocardial ischaemia from the epicardial potential distribution, a conclusion contrary to observed experimental data. Possible reasons for this discrepancy are discussed. In the interests of mathematical simplicity, isotropic models of the cardiac tissue are also considered, but results from these simulations predict epicardial potential distributions vastly different from experimental observations. A major conclusion from this work is that tissue anisotropy and fiber rotation must be included to obtain meaningful and realistic epicardial potential distributions.

MATHEMATICAL MODELLING OF FLOW THROUGH AN IRREGULAR ARTERIAL STENOSIS

A mathematical model of flow through an irregular arterial stenosis is developed. The model is two-dimensional and axi-symmetric with the stenosis outline obtained from a three-dimensional casting of a mildly stenosed artery. Agreement between modelled and experimental pressure drops (obtained from an axi-symmetric machined stenosis with the same profile) is excellent. Results are also obtained for a smooth stenosis model, similar to that used for most mathematical modelling studies. This model overestimates the pressure drop across the stenosis, as well as the wall shear stress and separation Reynolds number. Also, the smooth model predicts one instead of three recirculation zones present in the irregular model. The original stenosis is modified to increase the severity from 48 and 87% areal occlusion, while maintaining the same general shape. This has the effect of increasing the pressure drop by an order of magnitude and decreasing the number of recirculation zones to one, with a lower separation Reynolds number.

Semi–sigmoidal transformations for evaluating weakly singular boundary element integrals

Accurate numerical integration of line integrals is of fundamental importance to reliable implementation of the boundary element method. Usually, the regular integrals arising from a boundary element method implementation are evaluated using standard Gaussian quadrature. However, the singular integrals which arise are often evaluated in another way, sometimes using a different integration method with different nodes and weights. Here, a co-ordinate transformation technique is introduced for evaluating weakly singular integrals which, after some initial manipulation of the integral, uses the same integration nodes and weights as those of the regular integrals. The transformation technique is based on newly defined semi-sigmoidal transformations, which cluster integration nodes only near the singular point. The semi-sigmoidal transformations are defined in terms of existing sigmoidal transformations and have the benefit of evaluating integrals more accurately than full sigmoidal transformations as the clustering is restricted to one end point of the interval. Comparison of this new method with existing coordinate transformation techniques shows that more accurate evaluation of weakly singular integrals can be obtained. Based on observation of several integrals considered, guidelines are suggested for the type of semi-sigmoidal transformation to use and the degree to which nodes should be clustered at the singular points. Copyright

A new method for the numerical evaluation of nearly singular integrals on triangular elements in the 3D boundary element method

A new method (the sinh-sigmoidal method) is proposed for the numerical evaluation of both nearly weakly and nearly strongly singular integrals on triangular boundary elements. These integrals arise in the 3D boundary element method when the source point is very close to the element of integration. The new polar coordinate-based method introduces a sinh transformation in the radial direction and a sigmoidal transformation in the angular direction, before the application of Gaussian quadrature. It also uses approximately twice as many quadrature points in the angular direction as in the radial direction, in response to a finding that the evaluation of these types of integrals is particularly sensitive to the placement of the quadrature points in the angular direction. Comparisons with various other methods demonstrate its accuracy and competitiveness. A major advantage of the new method is its ease of implementation and applicability to a wide class of integrals.

A cylindrical model for studying subendocardial ischaemia in the left ventricle

In this paper a mathematical model of a left ventricle with a cylindrical geometry is presented with the aim of gaining a better understanding of the relationship between subendocardial ischaemia and ST depression. The model is formulated as an infinite cylinder and takes into account the full bidomain nature of cardiac tissue, as well as fibre rotation. A detailed solution method (based on Fourier series, Fourier transforms and a one dimensional finite difference scheme) for the governing equations for electric potential in the tissue and the blood is also presented. The model presented is used to study the effect increasing subendocardial ischaemia has on the epicardial potential distribution as well as the effects of changing the bidomain conductivity values. The epicardial potential distributions obtained with this cylindrical geometry are compared with results obtained using a previously published slab model. Results of the simulations presented show that the morphologies of the epicardial potential distributions are similar between the two geometries, with the main difference being that the cylindrical model predicts slightly higher potentials.

Error estimation of quadrature rules for evaluating singular integrals in boundary element problems

The efficient numerical evaluation of integrals arising in the boundary element method is of considerable practical importance. The superiority of the use of sigmoidal and semi-sigmoidal transformations together with Gauss-Legendre quadrature in this context has already been well-demonstrated numerically by one of the authors. In this paper, the authors obtain asymptotic estimates of the truncation errors for these algorithms. These estimates allow an informed choice of both the transformation and the quadrature error in the evaluation of boundary element integrals with algebraic or algebraic/logarithmic singularities at an end-point of the interval of integration. Copyright