COSEn Algorithm for Atrial Fibrillation Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for detecting atrial fibrillation from heart rate time series alone are limited by their inability to distinguish between sinus rhythm with frequent ectopy and atrial fibrillation, and require large datasets and invasive procedures, leading to inaccurate diagnosis and inappropriate treatments.
Innovation Solution
The development of a novel algorithm using coefficient of sample entropy (COSEn) that analyzes the order and dynamics of heart beat intervals, allowing for accurate detection of atrial fibrillation in short records without the need for invasive measures, and incorporating entropy measures to improve diagnostic performance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If heart rate time series analysis is used to detect atrial fibrillation, then non-invasive monitoring is achieved, but diagnostic accuracy deteriorates due to inability to distinguish sinus rhythm with ectopy from atrial fibrillation
Solution Approach 1:
The patent transforms the heart rate time series into a new parameter space using recurrence plots and entropy calculations. Instead of analyzing raw heart rate values directly, the system creates recurrence quantification parameters (RQPs) that capture the temporal dynamics and patterns of RR intervals, enabling accurate differentiation between sinus rhythm with ectopy and atrial fibrillation while maintaining non-invasive monitoring
Solution Approach 2:
The patent introduces recurrence plots as an intermediary representation between the raw heart rate time series and the diagnostic decision. These recurrence plots serve as a mediator that transforms complex temporal patterns into quantifiable entropy measures, which then serve as reliable indicators for rhythm classification without requiring invasive EKG signals
2Device complexity
If traditional entropy measures are used to analyze heart rate variability, then computational simplicity is maintained, but diagnostic performance deteriorates due to inability to capture order and dynamics of heartbeat intervals
Solution Approach 1:
The patent segments the heart rate time series into recurrent patterns using recurrence plots, dividing the continuous signal into discrete temporal structures that can be individually analyzed. This segmentation allows the system to capture local dynamics and order properties that traditional global entropy measures miss, while maintaining computational efficiency through systematic processing of these segments
Solution Approach 2:
The patent transitions from analyzing heart rate data in one dimension (time series values) to two dimensions (recurrence plots with axes representing time and RR interval values). This dimensional transformation reveals hidden patterns and structures in the data, enabling more accurate detection of atrial fibrillation while the subsequent entropy calculation on these 2D patterns maintains computational simplicity
3Reliability
If large datasets are required for accurate rhythm detection, then diagnostic reliability is improved, but monitoring efficiency deteriorates due to increased data requirements and processing time
Solution Approach 1:
The patent performs preliminary transformation of the heart rate time series into recurrence plots and calculates entropy measures on these transformed representations. This preliminary action extracts the essential diagnostic information from the raw data in advance, allowing for reliable rhythm detection using smaller, more manageable datasets and reducing the overall data processing requirements
Solution Approach 2:
The patent extracts key diagnostic features from the heart rate time series by calculating recurrence quantification parameters and entropy measures. This extraction process isolates the most informative aspects of the cardiac rhythm dynamics, enabling reliable detection of atrial fibrillation using condensed feature sets rather than requiring analysis of large volumes of raw data
Data Source
AI summary
A method for analysis of cardiac rhythms, based on calculations of entropy and moments of interbeat intervals. An optimal determination of segments of data is provided that demonstrate statistical homogeneity, specifically with regard to moments and entropy. The invention also involves calculating moments and entropy on each segment with the goal of diagnosis of cardiac rhythm. More specifically, an absolute entropy measurement is calculated and provided as a continuous variable, providing dynamical information of fundamental importance in diagnosis and analysis. Through the present invention, standard histograms, thresholds, and categories can be avoided.