AUTOREGRESSIVE MODELING OF EPICARDIAL ELECTROGRAMS DURING VENTRICULAR-FIBRILLATION

During ventricular fibrillation (VF), electrograms from bipolar epicar dial electrodes generally appear to have little organization or struct ure. We sought to identify any well defined organization or structure in these signals by determining if they could be modeled as an autoreg ressive stochastic process with a white noise excitation during the sh ort time period (6.5-8 s) typically used by automatic implantable defi brillators. The autoregressive model is then used to synthesize VF sig nals using a white noise excitation with the same probability distribu tion function as the estimated excitation determined from the autoregr essive model for that particular true VF episode. Both the original an d ten synthesized VF signals for each patient are then compared using root mean square (rms) amplitude, the number of zero crossings per sec ond, the amplitude distribution of the signals, the rate, and percent variation of rate. The results of examining the synthesized VF wavefor ms indicate that the rms amplitudes are similar to the true VF wavefor ms. While the synthesized VF signals had higher rate, more regular RR intervals, more zero crossings per second, and spent less time at base line than the VF signal from which they were generated, these differen ces are generally not significant (p greater-than-or-equal-to 0.05). T he use of such synthesized VF signals may allow more thorough testing of VF detection algorithms than is possible with the present limited l ibraries of human VF recordings.