Atrial Fibrillation Detection Using Time-Varying Coherence Functions
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for detecting atrial fibrillation, particularly paroxysmal rhythms, face challenges due to their short-lasting and asymptomatic nature, making accurate detection difficult, especially in Holter monitoring applications where motion and noise artifacts confound P-wave detection, leading to the need for improved detection technologies.
Innovation Solution
The use of time-varying coherence functions (TVCF) estimated by multiplying two time-varying transfer functions derived from adjacent R-R interval data segments, with the determination of whether the TVCF is less than a predetermined quantity to detect atrial fibrillation, combined with Shannon entropy for increased accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If P-wave detection methods are used for AF detection, then detection capability is provided, but motion and noise artifacts confound the accuracy of detection
Solution Approach 1:
The patent extracts and eliminates the harmful components (motion and noise artifacts) from the ECG signal by using artifact rejection algorithms that identify and remove contaminated segments, allowing P-wave detection to proceed on clean intervals only
Solution Approach 2:
The patent changes the detection parameter from direct P-wave amplitude detection to P-wave fiducial point localization using template matching and dynamic thresholding, which makes the detection more robust to artifacts by comparing signal characteristics against known P-wave patterns
2Measurement precision
If RR interval variability methods are used for AF detection, then sensitivity and specificity are improved, but large amounts of histogram data and threshold values must be stored
Solution Approach 1:
The patent extracts only the essential statistical features (mean, standard deviation, skewness, kurtosis) from the RR interval distribution, discarding the need to store complete histograms while retaining the discriminatory power for AF detection
Solution Approach 2:
The patent transforms the detection approach from comparing complete density histograms to comparing simplified statistical moments (first four moments of the distribution), which dramatically reduces storage requirements while maintaining detection accuracy
3Measurement precision
If traditional AF detection algorithms are used, then detection capability is provided, but they cannot accurately detect paroxysmal rhythms due to short-lasting and intermittent nature
Solution Approach 1:
The patent applies preliminary artifact rejection and data quality assessment before the main AF detection algorithm, preparing clean data segments that enable accurate detection even in short paroxysmal episodes
Solution Approach 2:
The patent uses dynamic thresholding and adaptive template matching that adjust to local signal characteristics, enabling accurate P-wave detection in varying conditions throughout paroxysmal episodes of different durations
Data Source
AI summary
Methods and systems for automatic detection of Atrial Fibrillation (AF) are disclosed. The methods and systems use time-varying coherence functions (TVCF) to detect AF. The TVCF is estimated by the multiplication of two time-varying transfer functions (TVTFs).