EEG Spike Detection With Neural Networks and Artifact Filtering

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing techniques for analyzing electroencephalogram (EEG) data lack effective methods for training neural networks to accurately identify epileptiform abnormalities and present the data to physicians or technicians in a user-friendly manner, making it difficult to distinguish artifacts from underlying signals.

Innovation Solution

A system and method for training neural networks to generate probability values for epileptiform abnormalities by processing EEG signals, incorporating artifact removal algorithms and user interfaces that allow for selective filtering and visualization of EEG data, enabling clearer interpretation.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If artifact filters and neural network algorithms are applied to remove artifacts from EEG signals, then the clarity and accuracy of underlying brain signals is improved, but the complexity of the system increases

Engineering Contradiction:

Improveaccuracy of seizure detectionVSAvoidsystem complexity

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The EEG signal processing is divided into discrete epochs (time segments), allowing artifact removal to be applied systematically to each segment. This segmentation enables the complex processing to be managed in manageable units while maintaining overall signal integrity and detection accuracy.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

A neural network algorithm acts as an intermediary between the raw EEG signal and the final seizure detection output. The neural network processes the signal through learned patterns, effectively removing artifacts while preserving genuine neurological events, thus improving accuracy without requiring manual intervention.

Inventive Principle:

Principle #24Intermediary (Mediator)

2Reliability

If multiple processing steps are applied to clean EEG data, then the reliability of seizure detection is improved, but the time required for processing increases

Engineering Contradiction:

Improvereliability of seizure detectionVSAvoidprocessing time

Core Design Contradiction:

ReliabilityVSLoss of time

Solution Approach 1:

The neural network is trained in advance on large datasets of labeled EEG signals containing both normal and seizure patterns. This preliminary training allows the system to perform rapid artifact removal and seizure detection during actual use, maintaining high reliability without excessive processing time during critical detection phases.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The processing system operates continuously on incoming EEG data streams, applying artifact removal and detection algorithms in real-time rather than batch processing. This continuous operation maintains reliability through constant monitoring while minimizing time loss by avoiding repeated start-stop processing cycles.

Inventive Principle:

Principle #20Continuity of useful action

Data Source

PatentUS12616421B2System and method for generating a probability value for an epileptiform abnormality by neural network to identify signal spikes

Publication Date: 2026.05.05 PERSYST DEVELOPMENT CORP

AI summary

A method and system for generating a probability value for an event. The system includes a source for generating a plurality of digital input signals, a processor connected to the source to receive the plurality of digital input signals from the source, and a display connected to the processor for displaying a final output. Preferably, the method further includes validating the probability value.