Spatiotemporal QRST cancellation techniques for analysis of atrial fibrillation: methods and performance

A new approach to QRST cancellation in ECGs with atrial fibrillation is presented. The method compensates for morphological changes caused by variations in the electrical axis and uses linear transformations to combine multiple lead information. Based on an ECG database of 15 subjects with simulated atrial fibrillation, the authors' results show that the spatiotemporal technique produces QRST cancellation which is superior to that obtained by straightforward average beat subtraction (ABS). The average energy in QRST-related errors, as compared to ABS, was reduced to 58 percent. The performance results obtained for recorded ECGs containing atrial fibrillation agreed very well with those of the simulated AF signals.