Elevation and depression in the ST segment of the electrocardiogram is commonly used as part of a diagnosis of myocardial ischaemia, although there is not yet a clear correlation between these observations and partial thickness ischaemia. In this work, we use a half-ellipsoid bidomain model of subendocardial ischaemia in a ventricle to study the effect of changes in model parameters on ST segment epicardial potential distributions (EPDs). We use generalised polynomial chaos techniques to produce mean EPDs, where the six bidomain conductivity values are varied, as well as blood conductivity and fibre rotation, for a number of different representations of the ischaemic region.We find that, as the thickness of the ischaemic region (i.e. the ischaemic depth) increases, the character of the mean EPD first changes from a single minimum approximately above the ischaemic region, to a maximum over the ischaemic region, with the minimum moving to a border of the ischaemic region. Next a second minimum develops, in addition to the previous maximum and minimum. In contrast, the strength of the maximum and the minima is only affected in a minor way by changes in the width of the ischaemic border and the position of the ischaemic region, provided it is not near the apex or base of the ventricle. When the size of the ischaemic region is increased, the magnitudes of both the maximum and the minima increase, but their character does not change.In summary, the qualitative progression of the mean EPD with increasing ischaemic depth, from single minimum through to a maximum surrounded by two minima, is the same, regardless of the size and position of the ischaemic region.
IntroductionSimulation studies are an important tool for studying the effects of myocardial ischaemia on the electrocardiogram, and, in particular, the effects of subendocardial ischaemia, since it is less well-understood than transmural ischaemia.If these studies are to be clinically useful, it is important that modellers are able to quantify, in some fashion, the effects of the various modelling assumptions and parameter choices they make in their models.Some previous work in this area has used simplified model geometries for the left ventricle (e.g. slabs, cylinders, half-ellipsoids [1]), while other work uses more realistic geometries [2,3]. Another study [4] has looked at the effect of the shape of the ischaemic region on the epicardial potential distributions (EPDs) that are produced.The majority of recent work represents the cardiac tissue as consisting of interpenetrating bi-domains, intracellular (i) and extracellular (e). In these the current is assumed to flow in three mutually orthogonal directions, longitudinal (l), transverse (t) and normal (n), that is along, across and between the sheets of cardiac fibres, respectively. This leads to six bidomain conductivity parameters (g pq , p = i, e, q = l, t, n) in the model, in addition to other parameters, such as the blood conductivity g b and the angle (ROT) that represents the overall rotation of ...