Adaptive estimation of time-varying late potentials technical and clinical aspects

Ventricular late potentials are low amplitude, high frequency signals present in the terminal portion of the QRS. The origin of late potentials are the myocardial zones with slow and inhomogeneous activation patterns that provide the substrate for reentrant excitation leading to ventricular tachycardia[1]. The signal averaging technique can only detect ventricular late potentials which are absolutely constant in duration, morphology, and timing relative to the QRS complex; therefore, this technique does not allow the detection of dynamic changes in ventricular late potentials which may occur either spontaneously or during various diagnostic and therapeutic interventions. The clinical advantage of detecting late potentials on a beat to beat basis are obvious as it facilitates the study of the relationship between late potentials and the occurrence of spontaneous or reentrant tachycardia. In this paper; a statistical model is developed characterizing the late potentials. Each QRS beat which contains late potentials is adaptively canceled and the output of the adaptive system would be an estimate of the late potentials.