Evaluation of auto-regressive modeling procedures for the detection of abnormal intra-QRS potentials using a boundary element electrocardiogram model

Auto-regressive modeling procedures (AR) have been used to quantify low magnitude components of abnormal activation within the QRS complex, called abnormal intra-QRS potentials (AIQPs). Research is still needed to identify the range in which the AR model can detect AIQP sources in different locations in the heart. A ventricular source model and forward model of electrocardiography called ECGSIM was used to test the ability of the AR model in detecting delays at various locations in the ventricles. A high resolution ECGSIM heart model was created to provide greater temporal and spatial control of the AIQP sources. The overall optimal model orders in this study were at an intermediate range seen previously in the literature. The ability of the AR model to detect the AIQP sources was dependent on the resolution of the heart model and the size and location of the AIQP sources.