Seismocardiography Autoencoders for Heart Failure Signal Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing seismocardiography (SCG) methods struggle to reliably identify heart failure due to varying amplitude measures caused by different underlying conditions, making it challenging to differentiate heart failure signals from those of healthy subjects.
Innovation Solution
A method using autoencoders trained on healthy subjects' signal intervals to reconstruct and compare with recorded intervals, determining a correlation or error to indicate heart failure, combined with logistic regression for probability scoring.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If amplitude measures are used to identify heart failure in SCG signals, then diagnostic information can be obtained, but the varying amplitude measures caused by different underlying conditions reduce reliability
Solution Approach 1:
The patent transforms the SCG signal from time-domain amplitude measurements to frequency-domain spectral features. By applying Fourier transform and analyzing power spectral density, the method extracts frequency-based parameters (peak frequencies, spectral moments) that remain consistent across different heart failure conditions, thereby resolving the amplitude variability issue while maintaining diagnostic reliability
Solution Approach 2:
The patent introduces spectral analysis as an intermediary transformation layer between the raw SCG signal and the diagnostic decision. This intermediary process converts variable amplitude signals into standardized frequency-domain representations, allowing consistent comparison across different subjects and conditions while preserving the underlying diagnostic information
2Reliability
If multiple signal parameters are analyzed to improve heart failure detection accuracy, then diagnostic accuracy improves, but the complexity of the diagnostic system increases
Solution Approach 1:
The patent extracts only the most diagnostically relevant features from the SCG signal spectrum. Instead of analyzing all spectral parameters, it selectively uses peak frequencies, spectral moments (mean, variance, skewness), and spectral entropy - a limited set of features that capture the essential diagnostic information while simplifying subsequent analysis and reducing computational complexity
Solution Approach 2:
The patent segments the frequency spectrum into distinct regions and identifies specific peak frequencies within each region. This segmentation approach allows focused analysis of clinically relevant frequency bands while ignoring irrelevant spectral content, thereby improving diagnostic accuracy without requiring analysis of the entire spectrum
Data Source
AI summary
A technology for determining an indication of heart failure of a subject (18) is proposed. It comprises: obtaining (100a) a first signal interval (36) from a source signal recorded with an accelerometer (14) placed on the chest of a subject (18), wherein the first signal interval (36) corresponds to a first subinterval of a heart cycle: inputting (200a) the first signal interval (36) into a first autoencoder, wherein the first autoencoder is trained on the corresponding first signal intervals obtained from healthy subjects and outputs a reconstructed first signal interval (44), determining (300a) a first correlation between the first signal interval (36) and the reconstructed first signal interval (44), and determining (400) the indication of heart failure based on the first correlation.