Atrial Fibrillation Classification via Spectral Dispersion Metrics
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for detecting atrial fibrillation are inadequate as they rely heavily on patient perception and lack consistency, making it difficult to determine treatment effectiveness and patient risk for stroke, especially since up to one-third of patients are unaware they have AF.
Innovation Solution
An apparatus and method that utilize a real-time dynamically adjustable signal transformation of electrocardiogram signals to form frequency, time, or phase domain representations, extracting features like Spectral Dispersion Metrics to accurately classify heart rhythm states, including atrial fibrillation, through an input stage, analysis stage, and classification stage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If current detection methods relying on patient perception are used, then the detection process is simple, but the measurement precision and reliability are insufficient
Solution Approach 1:
The patent segments the ECG signal analysis into distinct frequency bands (0-16Hz, 16-32Hz, 32-48Hz, 48-64Hz) and calculates power values for each band separately. This segmentation allows precise identification of AF by comparing power distribution across bands, achieving high measurement precision while maintaining manageable system complexity through modular processing steps.
Solution Approach 2:
The patent transforms the time-domain ECG signal into the frequency domain using Fast Fourier Transform (FFT), converting temporal waveforms into spectral power distributions. This dimensional transformation enables detection of AF through power ratio comparisons across frequency bands, providing a new perspective that significantly improves detection accuracy beyond simple time-domain analysis.
2Reliability
If simple detection methods are used, then the device complexity is low, but the consistency and reliability of detection results are poor
Solution Approach 1:
The patent applies preliminary signal conditioning including filtering and rectification before power calculation. The ECG signal undergoes bandpass filtering to remove artifacts, followed by full-wave rectification to ensure all power values are positive. These preliminary actions standardize the input signal, ensuring consistent and reliable detection results across different patients and recording conditions.
Solution Approach 2:
The system calculates the ratio of power in the 0-16Hz band to the total power across all bands, creating a normalized metric that provides feedback on the overall power distribution pattern. This feedback mechanism allows the system to reliably identify AF by comparing the calculated ratio against established thresholds, ensuring consistent detection regardless of absolute signal amplitude variations.
Data Source
AI summary
An atrial fibrillation classification system collects celectrocardiogram signals and converts them to a frequency, time, or phase domain representation for analysis. An evaluation stage extracts energy density profile over a range of frequencies, time intervals, or phases, which is then summed and normalized to form dispersion metrics. The system then analyzes the dispersion metrics, in their respective domains, to determine whether a patient is experiencing an arrhythmia and then to classify the type of arrhythmia being experienced.