Time-Delay Neural Network for ECG Reconstruction from Single EGM

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing methods for reconstructing surface electrocardiograms (ECG) from endocardial electrograms (EGM) using neural networks are not robust, particularly failing to account for changes in the QT interval and new morphologies, requiring multiple EGM derivations, and involving complex processing steps that can lead to false reconstructions.

Innovation Solution

A time-delay neural network is used to process EGM signals and their delayed versions to reconstruct ECG signals, allowing for the generation of acceptable ECG signals even with new unlearned beats and capable of producing all twelve ECG derivations from a single EGM derivation, with the network learning from simultaneous EGM and ECG signals over multiple cardiac cycles.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If traditional neural networks are used to reconstruct ECG from EGM, then the reconstruction can be performed, but the method is not robust and fails to account for changes in QT interval and new morphologies

Engineering Contradiction:

Improverobustness of ECG reconstructionVSAvoidability to handle new beat morphologies and QT interval changes

Core Design Contradiction:

ReliabilityVSAdaptability or versatility

Solution Approach 1:

The patent applies dynamics by making the neural network adaptive through online learning capabilities. The network can dynamically adjust its weights and parameters in real-time to accommodate changes in QT interval and new beat morphologies, transforming a static reconstruction model into a dynamic one that evolves with the patient's cardiac conditions.

Inventive Principle:

Principle #15Dynamics

Solution Approach 2:

The invention changes the parameters of the neural network during operation through online learning. By continuously updating the network weights based on incoming EGM data and corresponding ECG labels, the system adapts to parameter changes in cardiac morphology and QT interval without requiring retraining from scratch.

Inventive Principle:

Principle #35Parameter changes

2Reliability

If multiple EGM derivations are used to improve reconstruction accuracy, then the reliability improves, but the device complexity and number of electrodes required increases

Engineering Contradiction:

Improveaccuracy of ECG reconstructionVSAvoidnumber of electrodes and derivations required

Core Design Contradiction:

ReliabilityVSDevice complexity

Solution Approach 1:

The patent applies universality by designing a single EGM derivation to perform the work of multiple derivations. Through sophisticated neural network processing and online learning, one EGM channel can reconstruct all twelve ECG derivations, making the system universal in its capability despite reduced hardware requirements.

Inventive Principle:

Principle #6Universality (Multi-functionality)

Solution Approach 2:

The invention uses copying by generating synthetic ECG waveforms from a single EGM derivation. The neural network creates copies of the electrical activity patterns that would be observed at the body surface, effectively replicating the information that would normally require multiple physical electrodes to capture.

Inventive Principle:

Principle #26Copying

3Measurement precision

If complex processing steps including intermediate detection are used, then the reconstruction detail improves, but the risk of false reconstructions and processing errors increases

Engineering Contradiction:

Improvedetail of ECG reconstructionVSAvoidaccuracy of reconstruction without false positives

Core Design Contradiction:

Measurement precisionVSReliability

Solution Approach 1:

The patent applies continuity by performing online learning and reconstruction in a continuous, uninterrupted manner. The neural network continuously processes EGM data and updates its parameters without stopping for intermediate detection steps, maintaining a continuous stream of accurate ECG reconstructions without the discontinuities that cause false positives.

Inventive Principle:

Principle #20Continuity of useful action

Solution Approach 2:

The invention uses the neural network as an intermediary that directly maps EGM to ECG without requiring intermediate detection of QRS complexes or other cardiac events. This direct mapping approach eliminates the need for intermediate processing steps that could introduce errors, while still achieving high measurement precision.

Inventive Principle:

Principle #24Intermediary (Mediator)

Data Source

PatentUS9681815B2Reconstruction of a surface electrocardiogram from an endocardial electrogram using non-linear filtering

Publication Date: 2017.06.20 SORIN CRM

AI summary

The present invention relates to an active medical device that uses non-linear filtering for reconstructing a surface electrocardiogram from an endocardial electrogram. At least one endocardial EGM electrogram signal is collected from of samples collected from at least one endocardial or epicardial derivation (71′, 72′, 73′), and at least one of a reconstructed surface electrocardiogram (ECG) signal through the processing of collected EGM samples by a transfer function (TF) of a neural network (60′). The neural network (60′) is a time-delay-type network that simultaneously processes said at least one endocardial EGM electrogram signal, formed by a first sequence of collected samples, and at least one delayed version of this EGM signal, formed by a second sequence of collected samples distinct from the first sequence collected samples. The neural network (60′) provides said reconstructed ECG signal from the EGM signal and its delayed version.