Denoising RF defibrillator waveforms for intracardiac atrial substrate impedance characterisation using digital filtering techniques

Intra-cardiac impedance (ICI) is one of the major determinants of success during internal cardioversion of atrial fibrillation(AF) and is consequently of significant importance in the development of more effective cardioversion therapies. In this study, the voltage and current waveform delivered during internal cardioversion of patients with persistent AF treated using a novel low-tilt rectilinear waveform (generated by a radiofrequency (RF) defibrillator) and following a step up energy protocol (50V-300V in 50V steps) were recorded. However, within the clinical environment these waveforms are typically compromised with electrical noise. The objective of this work was to investigate if it was possible to develop an algorithm, using digital signal processing (DSP)in the MATLAB environment, to denoise the AF defibrillation voltage and current waveforms by applying spectral analysis and digital filtering techniques. Preliminary results indicated that spectral analysis of the processed rectilinear waveform signals allowed estimation of the most appropriate cutoff frequency as 1125Hz-2250Hz. Initial results also indicated that application of a Hanning windowing was more appropriate than Blackman windowing in the spectral preservation of the rectilinear waveform signals. Furthermore, for this particular type of waveform, noise reduction using a 7th order Butterworth filter design was found to be more efficacious than a 200 order FIR digital filter. In conclusion, within this explorative study, the processing algorithm developed was effective for denoising voltage and current signals recorded during cardioversion and therefore provides a useful tool for characterisation of intra-cardiac impedance during cardioversion.