Frequency domain characterization of artifact and tachyarrhythmias in the surface electrocardiogram

The power spectra of an excessive high-frequency artifact (that obscures the QRS complex), narrow complex or supraventricular tachycardia (SVT), and wide complex tachyarrhthmias, including ventricular tachycardia (VT) and ventricular fibrillation (VF) are examined quantitatively for distinguishing features of these signals. Of 12 features examined, four differ significantly among these groups. It is found that SVT has more peaks spaced at longer intervals than monomorphic VT (mVT). Polymorphic VT (pVT) and VF have higher primary frequencies than mVT. Severe and excessive ripple artifacts have more peaks than SVT, mVT, and pVT, and are more irregular than SVT and mVT. Although the populations differ significantly, there is considerable overlap by individual samples. These features may be useful in the development of a noise-characterizing preprocessor that would allow QRS detection and delineation algorithms to vary their parameters automatically to optimize sensitivity and specificity for the incoming electrocardiogram signal