ECG-Based Diastolic Function Assessment Using Time-Frequency Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current heart testing methods, such as electrocardiograms (ECGs), fail to detect subtle abnormalities in diastolic function, leading to misdiagnoses, while echocardiograms, the gold standard, are expensive and typically reserved for confirmed heart issues.
Innovation Solution
A system and method that uses ECG measurements, combined with machine-learned computational models, to estimate echocardiogram parameters indicative of diastolic function, utilizing time-frequency transforms and patient demographics to provide quantitative and qualitative indicators of diastolic function.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If echocardiography is used to measure diastolic function parameters, then measurement precision is improved, but device complexity and cost increase
Solution Approach 1:
The patent replaces the mechanical/electronic imaging system of echocardiography with an electrocardiogram-based computational model. The system uses electrical signal processing and machine learning algorithms to estimate diastolic function parameters from ECG data, eliminating the need for complex ultrasound hardware while achieving comparable measurement precision.
Solution Approach 2:
The patent creates a computational copy or digital twin of the echocardiography measurement process. By training machine learning models on data from patients with known diastolic function (obtained through echocardiography), the system learns to predict these parameters from ECG signals alone, effectively copying the diagnostic capability without requiring the original complex equipment.
2Measurement precision
If echocardiography is used to assess diastolic function, then measurement precision is improved, but ease of operation deteriorates
Solution Approach 1:
The patent extracts the essential diagnostic information from the complex echocardiography process and encapsulates it within a simplified ECG-based computational model. By separating the complex image acquisition and analysis steps from the final diagnostic output, the system provides precise measurements through a much simpler operational interface that requires only standard ECG recording.
3Ease of operation
If conventional ECG analysis methods are used, then ease of operation is maintained, but measurement precision deteriorates
Solution Approach 1:
The patent fundamentally changes the parameters used for ECG analysis by introducing time-frequency transform domain features and machine learning models rather than traditional time-domain measurements. This transformation enables the extraction of subtle patterns and relationships in the ECG signal that are invisible to conventional analysis methods, significantly improving measurement precision while maintaining operational simplicity.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Diastolic function may be assessed by operating one or more machine-learned computational models on electrocardiograms or electrocardiogram-derived features to compute quantitative diastolic indicators, including estimates of echocardiography parameters conventionally measured by echocardiography. Parameters such as those derived from time-frequency transforms of the electrocardiograms are used as input to the model(s) and/or computed within the model(s).