Quantitative Evaluation of Temporal Episode Patterns in Paroxysmal Atrial Fibrillation
(2018) 45th Computing in Cardiology Conference, CinC 2018 45.- Abstract
Flow velocity in left atrial appendage decreases when paroxysmal atrial fibrillation (PAF) progresses to longer episodes, suggesting that the temporal PAF episode pattern may be related to risk of thrombus formation. This study investigates the feasibility of discriminating episode patterns based on two descriptors: the aggregation characterizes the temporal distribution of PAF episodes, whereas the Gini coefficient characterizes differences in episode duration. The descriptors were studied on three PhysioNet databases with annotated PAF episodes, resulting in a total of 102 recordings. Three types of patterns were defined: congregation of several episodes in a single and multiple clusters, and episodes dispersed over the entire... (More)
Flow velocity in left atrial appendage decreases when paroxysmal atrial fibrillation (PAF) progresses to longer episodes, suggesting that the temporal PAF episode pattern may be related to risk of thrombus formation. This study investigates the feasibility of discriminating episode patterns based on two descriptors: the aggregation characterizes the temporal distribution of PAF episodes, whereas the Gini coefficient characterizes differences in episode duration. The descriptors were studied on three PhysioNet databases with annotated PAF episodes, resulting in a total of 102 recordings. Three types of patterns were defined: congregation of several episodes in a single and multiple clusters, and episodes dispersed over the entire monitoring period. The results show that the aggregation descriptor achieves large values for patterns with a single and multiple clusters (0.76± 0.07 and 0.60± 0.08, respectively). In contrast, much lower values are obtained for dispersed episode patterns (0.10± 0.05). The Gini coefficient is better suited for discriminating among the patterns with high PAF burden and, therefore, represents a descriptor which is complementary to aggregation. Both descriptors may have relevance when studying the relationship between episode pattern and the risk of thrombus formation.
(Less)
- author
- Simaityte, Monika ; Petrenas, Andrius ; Marozas, Vaidotas ; Henriksson, Mikael LU ; Bacevicius, Justinas ; Aidietis, Audrius and Sornmo, Leif LU
- organization
- publishing date
- 2018-09-01
- type
- Chapter in Book/Report/Conference proceeding
- publication status
- published
- subject
- host publication
- Computing in Cardiology Conference, CinC 2018
- volume
- 45
- article number
- 8743941
- publisher
- IEEE Computer Society
- conference name
- 45th Computing in Cardiology Conference, CinC 2018
- conference location
- Maastricht, Netherlands
- conference dates
- 2018-09-23 - 2018-09-26
- external identifiers
-
- scopus:85068781339
- ISBN
- 9781728109589
- DOI
- 10.22489/CinC.2018.059
- language
- English
- LU publication?
- yes
- id
- c62b6e0d-5cfe-47e5-8ef7-6b0906bc44a9
- date added to LUP
- 2019-07-25 13:02:58
- date last changed
- 2022-04-02 20:42:10
@inproceedings{c62b6e0d-5cfe-47e5-8ef7-6b0906bc44a9,
abstract = {{<p>Flow velocity in left atrial appendage decreases when paroxysmal atrial fibrillation (PAF) progresses to longer episodes, suggesting that the temporal PAF episode pattern may be related to risk of thrombus formation. This study investigates the feasibility of discriminating episode patterns based on two descriptors: the aggregation characterizes the temporal distribution of PAF episodes, whereas the Gini coefficient characterizes differences in episode duration. The descriptors were studied on three PhysioNet databases with annotated PAF episodes, resulting in a total of 102 recordings. Three types of patterns were defined: congregation of several episodes in a single and multiple clusters, and episodes dispersed over the entire monitoring period. The results show that the aggregation descriptor achieves large values for patterns with a single and multiple clusters (0.76± 0.07 and 0.60± 0.08, respectively). In contrast, much lower values are obtained for dispersed episode patterns (0.10± 0.05). The Gini coefficient is better suited for discriminating among the patterns with high PAF burden and, therefore, represents a descriptor which is complementary to aggregation. Both descriptors may have relevance when studying the relationship between episode pattern and the risk of thrombus formation.</p>}},
author = {{Simaityte, Monika and Petrenas, Andrius and Marozas, Vaidotas and Henriksson, Mikael and Bacevicius, Justinas and Aidietis, Audrius and Sornmo, Leif}},
booktitle = {{Computing in Cardiology Conference, CinC 2018}},
isbn = {{9781728109589}},
language = {{eng}},
month = {{09}},
publisher = {{IEEE Computer Society}},
title = {{Quantitative Evaluation of Temporal Episode Patterns in Paroxysmal Atrial Fibrillation}},
url = {{http://dx.doi.org/10.22489/CinC.2018.059}},
doi = {{10.22489/CinC.2018.059}},
volume = {{45}},
year = {{2018}},
}