New wavelet transform based EGM morphology discriminationalgorithms for implantable ICDs are presented. The algorithms aresimilar to correlation waveform analysis (CWA) and area of difference(AD) but computations are performed in the wavelet domain. The use ofthe wavelet transform allows more efficient signal processing due to itsinformation compression and filtration properties. Also, waveletcoefficients corresponding to different time scales are weighted toreflect the relative importance of corresponding time scales for EGMmorphology discrimination and to facilitate the computation. Thealgorithms have been implemented in ICD firmware and are capable ofprocessing morphology at up to 250 beats per minute. Holter dataanalysis showed that morphology measurements were stable over 48 hoursof electrogram recording
展开▼
