Pseudo-EGM Generation From ECG Using Machine Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Intracardiac electrograms (EGMs) provide precise electrical conduction pathway views but are invasive, causing patient discomfort and requiring recovery, while surface electrocardiograms (ECGs) lack the necessary precision.
Innovation Solution
An apparatus and method generate pseudo-EGM data from ECG data using a processor and memory to train an EGM machine-learning model, synchronizing and correlating ECG and EGM data to predict EGMs non-invasively.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If intracardiac electrograms (EGMs) are used to obtain precise electrical conduction pathway views, then measurement precision is improved, but patient comfort and ease of operation deteriorate due to invasive procedures
Solution Approach 1:
The patent creates a copy of EGM data by training a machine learning model to translate ECG data into pseudo-EGM data. The model learns the mapping between ECG and EGM representations and then generates synthetic EGM-like signals from new ECG inputs, providing a non-invasive alternative that replicates the diagnostic value of invasive EGMs
Solution Approach 2:
The patent introduces ECG data as an intermediary that bridges the gap between non-invasive monitoring and invasive EGM measurements. By using ECG as the input modality and training a model to translate it into EGM representations, the system mediates between the comfort of surface electrodes and the diagnostic precision of intracardiac sensors
2Measurement precision
If intracardiac electrograms (EGMs) are used to obtain precise electrical conduction pathway views, then measurement precision is improved, but patient discomfort and harmful factors increase due to invasive catheter insertion
Solution Approach 1:
The system generates a synthetic copy of EGM data through machine learning translation from ECG inputs, eliminating the need for physical catheter insertion while preserving the diagnostic information contained in EGM signals
Solution Approach 2:
The patent converts the limitation of ECG (lower measurement precision) into a benefit by using it as training data for a machine learning model. The model learns to compensate for ECG's lower fidelity by predicting EGM-like signals, thereby turning the non-invasive ECG's weakness into a strength that enables safe, repeatable monitoring
3Measurement precision
If intracardiac electrograms (EGMs) are used to obtain precise electrical conduction pathway views, then measurement precision is improved, but recovery time and loss of time increase due to invasive procedures
Solution Approach 1:
By generating pseudo-EGM data through machine learning translation from ECG inputs, the system provides immediate diagnostic information without requiring recovery time from invasive procedures, eliminating the time loss associated with catheter insertion and patient recuperation
Data Source
AI summary
Apparatus and method for generating pseudo-EGM data from ECG data are disclosed. The apparatus includes at least a processor and a memory communicatively connected to the at least a processor, wherein the memory contains instructions configuring the at least a processor to generate EGM model training data, wherein generating the EGM model training data includes receiving the EGM model training data, wherein the EGM model training data includes exemplary ECG data correlated to exemplary EGM data and time synchronizing the exemplary ECG data and the exemplary EGM data, train an EGM machine-learning model using the EGM model training data, receive subject data, wherein the subject data includes subject ECG data and generate subject EGM data as a function of the subject ECG data using the trained EGM machine-learning model.