Atrial heart fibrillation modeling: a tool for adequate choice of signal processing functions

Framingham Study showed that Atrial Fibrillation (AF) exposes the patient to an increased stroke risk and to low cardiac output. It is well known that the mechanism of Atrial Fibrillation is reentry. In order to facilitate AF's studies, we propose a deterministic model that can be used for cardiac signal characterization. The model is based on a linear combination of several "Atrial Flutter-like signals. We obtained AF-close aspects. To demonstrate its usefulness we performed a power spectrum transform (i.e. scalar Fast Fourier Transform) of this model. Thanks to the model we show that this transform cannot be used as is, and may induce errors in results interpretation. We also show based on the same example how to overcome the limitation with help of the model. We expect that this model will help in mathematical transforms relevance evaluation. This may lead to new findings in AF classification, and to improvement in patients' treatment, follow-up and survival.