Abstract
Inverse calculations were examined that sought the origin of a cardiac ectopic excitation sequence. Cardiac anatomy and its geometric relationships to sites on the body surface were adapted from human cross-sectional images to form a two-dimensional model, which included ventricular muscle and a primitive conduction system. The surrounding volume conductor was modelled in a simplified way as unbounded, homogeneous and isotropic. In a series of tests, one ectopic origin was designated the ‘true’ origin. The ECG for this true origin was compared to ECGs for 197 ectopic ‘trial’ origins, and differences between the wave forms for true versus trial origins were determined. Core issues were the magnitudes of changes in ECG wave forms as a function of the site of origin, whether these changes were sufficient to imply uniqueness, and what spatial resolution might be expected, in the presence of realistic noise levels. For a noise level of 10 μV RMS, the origin of excitation was localised to a single region of the muscle using one wave form from the body surface, with a resolution of 10 mm. The resolution was not improved significantly with a second electrode on the body surface, but was substantially improved with an endocardial electrode.