Evolution of ventricular fibrillation revealed by first return plots

We examine computational simulations of the propagation of activation and break-up of a scroll wave in a 100×100×50 array under Fitz-Hugh Nagumo dynamics. Calculating the temporal evolution of a spatial average in each of the x-, y-, and z-directions produces a 'pseudo-ECG'. We apply a combination of high-dimensional embedding, singular value decomposition and projection, to transform the data and extract a derived time series. The 'inter-beat' intervals of this derived time series are used to compute first return plots, the temporal evolution of which shows characteristic time-dependent dynamics. This technique provides a pictorial representation of the scroll wave break-up. We show that the same characteristic pattern may be observed in episodes of VF recorded from human subjects. Furthermore, this characteristic evolution may be observed from each of three scalar 'pseudo-ECG' time series derived from the three orthogonal directions in the same computational simulation. This technique offers a new approach to analyze and track the evolution of scroll wave break-up in ventricular fibrillation.