Seminario 19/17: M. Victoria Caballero (U. Murcia) – Symbolic Recurrence Rate to detect atrial fibrillation

Información

Atrial fibrillation (AF) is an abnormal rhythm of the heart that is felt as an irregular heartbeat or pulse. This disease is the most common sustained cardiac arrhythmia and increases the risk of heart failure and other heart complications. Paroxysmal AF (PAF) is defined as an intermittent AF that terminates spontaneously o with intervention in less than seven days.

In this work we present a new tool, namely symbolic correlation integral, for AF detection in RR intervals time series. This method is based on symbolic analysis. This kind of analysis may reveal physiological properties of AF and contain important information that can be used to discriminate PAF and no-PAF.

To obtain the symbolic time series we consider the time series of RR intervals embedded in an m-dimensional space. Next, we define the symbolization map which transforms the embedded time series in a sequence of symbols. The set of symbols is the ensemble of all the permutations of length m that represent ordinal patterns, and the symbolization map assigns every element of embedded m-dimensional time series, namely m-history, to a permutation that sort out from the smallest to the greatest the entries of this vector.

Two m-histories are symbolic recurrence states to a symbol when by the symbolization map have the same image. Then we define he Symbolic Recurrence Rate (SRR) to a symbol as the   probability of symbolic recurrence states to this symbol.

We have calculated SRR by using moving windows procedure to RR interval time series, with the aim of identifying the most important changes in the dynamics of RR interval time series from a subject.