Nonlinear dynamic approaches to identify atrial fibrillation progression based on topological methods. (August 2019)

• Record Type: Journal Article
• Authors: Safarbali, Bahareh; Hashemi Golpayegani, Seyed Mohammad Reza
• Is Part Of: Biomedical signal processing and control. Volume 53(2019)
• Publication Date: 2019-08
• Subjects: Atrial fibrillation -- Topological data analysis -- Fractal dimension -- Nonlinear signal processing -- Dynamical system theory; Signal processing -- Periodicals; Biomedical engineering -- Periodicals; Signal Processing, Computer-Assisted -- Periodicals; Image Processing, Computer-Assisted -- Periodicals; Biomedical Engineering -- Periodicals; 610.28
• DOI: 10.1016/j.bspc.2019.101563
• Languages: English
• ISSNs: 1746-8094

Abstract

Highlights: Two topological approaches (fractal dimension and persistent homology) based on HRV phase space geometry was suggested to replicate the changes in AF progression. A new feature based on the initial and end Čech radius of each barcode in β 1, is defined as "Time of Life" (TOL) to classify AF stages qualitatively. A neural network was implemented to prove the effectiveness of both TOL and fractal dimension as classification features. The classification performance was accomplished as 93% accuracy of datasets. TOL could be used as a tool to predict AF rhythm in patients, who have asymptomatic paroxysmal AF episodes. Abstract: In recent years, atrial fibrillation (AF) development from paroxysmal to persistent or permanent forms has become an important issue in cardiovascular disorders. Information about AF pattern of presentation (paroxysmal, persistent, or permanent) was useful in the management of algorithms in each category. This management is aimed at reducing symptoms and stopping severe problems associated with AF. AF classification has been based on time duration and episodes until now. In particular, complexity changes in Heart Rate Variation (HRV) may contain clinically relevant signals of imminent systemic dysregulation. A number of nonlinear methods based on phase space and topological properties can give more insight into HRV abnormalities such as fibrillation. Aiming to provide a nonlinear tool to qualitatively classify AF stages, we proposed two geometrical indices (fractal dimension and persistent homology) based on HRV phase space, which can successfully replicate the changes in AF progression. The study population includes 38 lone AF patients and 20 normal subjects, which are collected from the Physio-Bank database. "Time of Life (TOL)" is proposed as a new feature based on the initial and final Čech radius in the persistent homology diagram. A neural network was implemented to prove the effectiveness of both TOL and fractal dimension as classification features. The accuracy of classification performance was 93%. The proposed indices provide a signal representation framework useful to understand the dynamic changes in AF cardiac patterns and to classify normal and pathological rhythms.