Analyzing atrial electricity activity dynamical structure by recurrence complex network

The atrial electricity activity is a complex dynamical system and shares the recurrence property of a fundamental characteristic as many other dynamical systems. We transform multi-electrode epicardium signals into the recurrence complex network to quantify the structure properties of recurrence in the phase space. The analysis results suggest that the proportions of the motifs occurring in the networks will vary significantly with the change of the atrial dynamical structure. By using four-node motifs parameters, we can predict atrial fibrillation successfully with the total accuracy of 90.76%, which has a significant meaning for the postoperative evaluation.