Spectral Analysis of 12 Lead ECG Recordings of Ventricular Fibrillation in Humans

Abstract

Background: The ECG remains a key tool to assess cardiac electrical activity. Analysis of malignant rhythm disturbances such as VF commonly uses single lead data. Important information from the other leads may be lost. This study evaluated the information contained in a 12 lead ECG during induced VF to assess for distinct information particular to each lead. Methods: Standard 12 lead ECGs were collected from 11 patients undergoing implantation of a cardioverter defibrillator system. VF was induced during implantation. Power spectra were analysed for each ECG lead as well as measures of rate (dominant frequency) and regularity (spectral bandwidth). Using partial spectral analysis the proportion of information that was distinct in each lead was assessed. Results: VF frequency differed on average by 0.19 Hz within an individual across the 12 leads. Bandwidth measures showed an average variation of 0.98 Hz. Average residual power from partialisation with lead II was 10.6%, indicating that 89.4% of the spectral activity was common across all leads. The average maximum residual power was 24.7%. Conclusion: Measures of frequency and regularity derived from the VF power spectrum were consistent across the 12 lead ECG. An average of almost 90% of the power spectrum was common across the leads. Multiple ECG leads are not required for characterisation of VF. Measures derived from the power spectrum are not dependent on a particular lead. Single lead data such as that recorded from external defibrillators or telemetry monitoring systems can therefore be used without loss of important mechanistic information.