Harmonic Regression for MRI EEG Artifact Removal
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for removing artifacts from electrophysiologic recordings in MRI scanners, such as ballistocardiogram (BCG) artifacts, are inadequate due to reliance on corrupted reference signals and subjective separation criteria, limiting the quality of brain function monitoring and clinical applications.
Innovation Solution
A harmonic regression technique using physically motivated parametric models of artifacts and underlying physiological signals, which performs an iterative optimization process to estimate and remove artifacts without requiring reference signals or subjective criteria, effectively decoupling BCG artifacts from true EEG signals.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If reference signal-based methods (ECG or motion sensors) are used to remove BCG artifacts, then artifact removal can be performed, but the reference signals are corrupted in magnetic fields greater than 1.5 Tesla or during long EEG recordings, making peak detection and adaptive filtering difficult
Solution Approach 1:
The patent introduces an intermediary model-based approach that does not rely on corrupted reference signals. Instead of using ECG or motion sensor data as intermediaries, the invention creates a mathematical model of BCG artifacts that can be directly fitted to the EEG data, bypassing the need for external reference signals that become corrupted in high magnetic fields.
Solution Approach 2:
The patent replaces the mechanical/reference signal-based approach with a computational model-based approach. Instead of relying on physical reference signals (ECG electrodes, motion sensors) that are susceptible to corruption, the invention uses mathematical modeling and iterative optimization to estimate and remove artifacts, substituting physical measurement with computational analysis.
2Measurement precision
If reference signal-free methods (independent component analysis or wavelet basis decompositions) are used, then no corrupted reference signals are needed, but many basis elements contain substantial overlap between signals and artifacts, requiring subjective and case-specific criteria for separation
Solution Approach 1:
The patent changes the parameters used for signal separation by modeling BCG artifacts with specific physiological parameters (heart rate, pulse rate, respiratory rate) rather than using generic basis functions. This parametric approach allows for objective separation based on known physiological ranges, eliminating the need for subjective criteria while maintaining accuracy.
Solution Approach 2:
The patent introduces dynamics into the artifact removal process by allowing the model parameters (amplitudes, frequencies, phases) to vary over time through iterative optimization. This dynamic adaptation enables the model to track changing artifact characteristics without requiring manual intervention or subjective judgment, reducing post-processing complexity.
3Reliability
If BCG artifacts are removed using traditional methods, then some artifact reduction is achieved, but BCG artifacts have significantly larger amplitudes (150-200 microVolts) than underlying EEG activity (10-100 microVolts), obscuring EEG activity up to 20 Hz
Solution Approach 1:
The patent performs preliminary action by fitting the BCG artifact model to the EEG data before final analysis. The iterative optimization process预先 estimates artifact parameters and removes BCG components, allowing the true EEG signal to be recovered with high fidelity across the full frequency spectrum including the 0-20 Hz range.
Data Source
AI summary
Systems and a method for reducing periodic artifacts in an electrophysiologic signal are provided. The method includes receiving a time-series electrophysiologic signal acquired from a subject and providing a regression model that defines an interference signal caused by periodic artifacts using a harmonic representation. The method also includes applying the regression model using the time-series electrophysiologic signal to define a cost function and performing an iterative optimization process to estimate regression parameters that minimize the cost function. The method further includes determining, using the regression parameters, the interference signal, and generating a corrected time-series electrophysiologic signal by reducing the interference signal.