EEG Artifact Segmentation for Anesthesia Monitoring
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for evaluating anesthetic or intensive care EEGs struggle to accurately distinguish between biosignals and artifacts, particularly those caused by volatile anesthetics like sevoflurane, which can lead to misclassification due to interference from epilepsy-typical curves and other atypical signal components.
Innovation Solution
The method determines parameters from EEG curves in the time and/or frequency domain for multivariate classification, conducts artifact analysis, and uses deformation sensors to differentiate between artifacts and biosignals by comparing signal patterns with stored graphic features, thereby identifying epilepsy-typical interferences and reducing misclassification.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If mathematical-statistical methods are used to classify anesthetic EEG, then classification can be performed, but artifacts cause misclassification and reduce reliability
Solution Approach 1:
The patent segments the EEG signal analysis into multiple independent channels (EEG channels and EMG channels) that are processed separately. Each channel is analyzed for specific features, and the results are combined to make the final classification decision. This segmentation allows artifact detection in one channel to be distinguished from genuine brain activity in another channel, thereby improving classification reliability despite artifact presence.
Solution Approach 2:
The patent introduces an intermediary artifact detection and analysis step between raw EEG signal acquisition and final classification. This intermediary process identifies artifact-containing segments, analyzes their characteristics, and provides correction information to the classification algorithm. This mediator prevents artifacts from directly causing misclassification, thereby improving reliability.
2Measurement precision
If entropy of EEG and EMG signals is calculated to determine cerebral status, then anesthesia monitoring is enabled, but biosignals and artifacts cannot be distinguished
Solution Approach 1:
The patent applies different analysis methods to different signal components based on their local characteristics. EMG signals are analyzed for muscle artifact characteristics while EEG signals are analyzed for brain activity patterns. By treating different parts of the composite signal with specialized analysis techniques appropriate to their origin, the system can distinguish between genuine biosignals and artifacts even when both are present in the monitored signal.
Solution Approach 2:
The patent adds temporal and spectral dimensions to the analysis by examining signal characteristics across time and frequency domains. Artifact segments are identified by their distinctive temporal patterns and spectral features that differ from genuine brain activity. This multi-dimensional analysis provides additional discrimination power to differentiate biosignals from artifacts beyond simple entropy calculation.
3Reliability
If EEG curves are analyzed for interference signals, then potential biosignals can be identified, but device complexity increases
Solution Approach 1:
The patent performs preliminary artifact detection and characterization before final classification decision-making. By pre-identifying artifact segments and their characteristics, the system prepares correction information in advance that simplifies the subsequent classification process. This preliminary action prevents the need for complex real-time decision-making during classification, thereby managing device complexity while maintaining high detection accuracy.
Data Source
AI summary
The invention relates to a method and a device for evaluating an intensive EEG or an EEG during anaesthesia, according to which time domain and/or frequency range parameters are determined from the EEG graphs (112), the determined parameters being used in multivariate classification functions and the intensive EEG or EEG during anaesthesia being automatically divided into stages (116) as a result thereof. The EEG graphs are also analysed for interfering signal components from the quantity of biosignals of graphs characteristic of intensive EEGs or EEGs which are not performed during anaesthesia (114,118), and artifacts. If such interfering signal components are identified, the existence of biosignals of graphs characteristic of intensive EEGs or EEGs which are not performed during anaesthesia is verified by artifact analysis (120,122,126) in the absence of artifacts, and is not verified if artefacts are identified.