Spatial approximation of the body surface potentials across standard pre-cordial leads for diagnosing myocardial electrical events

A mapping scheme for continuous spatial approximation of the body surface potentials across standard V1-V6 precordial leads is proposed. The study is based on standard 12 leads surface digital ECG (12 bit 500 Hz) from which six pre-cordial leads were selected for map construction. The approximation method employs artificial neural networks, termed Radial Basis Functions (RBF), that are trained by an original non-iterative training procedure. The aim of the study is to show that the proposed ECG imaging method is suitable for spatio-temporal localisation and quantification of ventricular repolarisation abnormalities. In particular its diagnostic value as a prognostic marker of VT is demonstrated. The proposed mapping scheme does not require any extra recording instrumentation yet suggests a new way of visualising information contained in the standard precordial leads.