EEG Diagnostic Index Using Linear Predictive Coding and PCA
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for diagnosing Parkinson's Disease are largely clinical and lack accuracy, with clinical diagnostic accuracy ranging from 73.8% to 80.6%, and are not significantly improved over the last 25 years, necessitating a more reliable and precise diagnostic system.
Innovation Solution
A system and method utilizing EEG data to calculate a Linear-predictive-coding Electroencephalogy Algorithm in PD (LEAPD) index, which provides a quantitative diagnostic index and level of confidence in diagnosis, employing Linear Predictive Coding and Principal Component Analysis to differentiate between Parkinson's Disease and healthy controls.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If clinical diagnostic methods are used for Parkinson's Disease, then the diagnostic process is simple and accessible, but the diagnostic accuracy is limited (73.8%-80.6%) and has not significantly improved over 25 years
Solution Approach 1:
The patent replaces manual clinical diagnostic methods with an automated computational system that processes EEG data through Linear Predictive Coding and classification algorithms. This substitution of mechanical/manual diagnostic processes with computational automation achieves superior diagnostic accuracy (90%+) while maintaining clinical accessibility
Solution Approach 2:
The patent transforms clinical diagnostic parameters by introducing quantitative EEG-based features (Linear Predictive Coding coefficients, spectral characteristics) that capture subtle neurological patterns invisible to clinical observation. This parameter transformation enables significantly improved diagnostic accuracy by measuring brain electrical activity patterns specific to Parkinson's Disease
2Reliability
If EEG-based automated diagnosis is implemented, then diagnostic accuracy is significantly improved, but the system complexity and computational requirements increase
Solution Approach 1:
The patent segments the diagnostic process into distinct modular stages: EEG signal acquisition, preprocessing (filtering, artifact removal), feature extraction (Linear Predictive Coding, spectral analysis), classification (supervised learning algorithms), and diagnostic output. This segmentation enables each module to be optimized independently, improving overall reliability while managing system complexity through modular architecture
Solution Approach 2:
The patent introduces intermediate computational representations (Linear Predictive Coding coefficients, spectral features, extracted EEG characteristics) that serve as mediators between raw EEG signals and final diagnostic decisions. These intermediaries transform complex neurological signals into standardized features that classification algorithms can process reliably, bridging the gap between signal acquisition and diagnostic interpretation
Data Source
AI summary
The disclosed apparatus, systems and methods relate to diagnosing Parkinson's disease from electroencephalography (EEG) data. Embodiments herein have practical applications, including diagnosing Parkinson's disease. The methods and systems of the various implementations herein generate a diagnostic index which reflects the probability of the patient having Parkinson's disease. It uses a novel feature extraction method based on Linear Predictive Coding (LPC) which is used to extract Parkinson's disease related features from EEG recordings of the patient and a novel classification method based on Principal Component Analysis (PCA) is used to calculate the diagnostic index from these features.