Motor Event Detection Using Bipolar Re-referenced LFP Signals
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for detecting motor events in Parkinson's disease patients using Deep Brain Stimulation (DBS) lack efficient asynchronous detection techniques, which are essential for accurately identifying voluntary movements and improving treatment outcomes.
Innovation Solution
The system records and processes local field potential (LFP) signals from multiple probes in both brain hemispheres, applying bipolar re-referencing, nonlinear regression, and principal component analysis to determine optimal signal pairs, and uses template matching to detect motor events, enabling precise classification of voluntary movements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional motor event detection methods are used in DBS systems, then the system structure is relatively simple, but the detection accuracy and reliability of motor events are insufficient
Solution Approach 1:
The patent segments the motor event detection process into multiple independent processing stages: LFP signal acquisition from multiple electrodes, bipolar re-referencing to eliminate common-mode noise, nonlinear regression for feature extraction, and template matching for event classification. Each stage processes specific signal characteristics independently, improving detection accuracy while maintaining manageable system complexity through modular architecture.
Solution Approach 2:
The patent introduces an intermediary template matching mechanism that compares processed LFP signals against pre-established motor event templates. This intermediary step serves as a mediator between raw neural signals and final motor event detection, enabling accurate classification of voluntary movements by matching signal patterns without requiring complex real-time decision algorithms.
2Measurement precision
If multiple electrophysiological signals are recorded and processed using nonlinear regression and principal component analysis, then the detection precision is improved, but the computational complexity and processing time increase
Solution Approach 1:
The patent performs preliminary action by pre-processing LFP signals through bipolar re-referencing and establishing motor event templates before actual motor event detection. Nonlinear regression analysis is performed offline to characterize signal patterns, and principal component analysis is used to pre-identify relevant signal dimensions. This preliminary processing reduces the computational burden during real-time detection, minimizing processing time while maintaining high classification accuracy.
Solution Approach 2:
The patent applies parameter changes by transforming the high-dimensional LFP signal space into a reduced-dimensional feature space using principal component analysis. This parameter transformation retains the most informative signal characteristics while discarding redundant information, enabling accurate motor event classification with reduced computational complexity and faster processing speeds.
3Reliability
If bipolar re-referencing and nonlinear regression are applied to determine optimal signal pairs, then the reliability of motor event detection is improved, but the device complexity and processing requirements increase
Solution Approach 1:
The patent extracts the essential motor event detection function by isolating and processing only the most relevant LFP signal pairs determined through nonlinear regression. Instead of analyzing all possible electrode combinations, the system extracts and focuses on optimal signal pairs that show the strongest correlation with motor events, reducing algorithmic complexity while maintaining high detection reliability through targeted signal selection.
Data Source
AI summary
Systems and methods are disclosed to detect and/or classify electrophysiological signals as motor events. In some embodiments, a method may include: recording a first plurality of electrophysiological signals from a first plurality of probes inserted into the left hemisphere of the brain; recording a second plurality of electrophysiological signals from a second plurality of probes inserted into the right hemisphere of the brain; pre-processing the first plurality of electrophysiological signals and the second plurality of electrophysiological signals; bipolar re-referencing the first plurality of electrophysiological signals and the second plurality of electrophysiological signals; determining an optimal pair of electrophysiological signals from the bipolar re-referenced first plurality of electrophysiological signals and the bipolar re-referenced second plurality of electrophysiological signals; matching the optimal pair of electrophysiological signals with a template; and detecting motor events from the matching.