Multidimensional Polynomial Fitting for Auditory Signal Landmark Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for determining neuronal action potential signals in cochlear implant systems are prone to high failure rates and inaccuracies due to signal artifacts, leading to unreliable amplitude measurements and fitting functions that require manual correction.
Innovation Solution
A multi-dimensional polynomial fitting approach is used to analyze tissue response signals, incorporating dimensions such as post-stimulus time, stimulus intensity, and inter-stimulation pulse intervals, to calculate the line of minimum principal curvature, thereby robustly determining physiological landmarks like extrema and neuronal action potentials.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional methods are used to determine neuronal action potential signals, then the measurement process is simple, but the measurement precision and reliability are low due to signal artifacts
Solution Approach 1:
The patent transitions from traditional single-dimensional time-based signal analysis to multi-dimensional analysis by incorporating stimulus intensity and inter-stimulation pulse intervals as additional dimensions. This dimensional expansion allows the system to distinguish physiological signals from artifacts by analyzing signal behavior across multiple parameters simultaneously, thereby improving measurement precision without requiring complex additional hardware
2Reliability
If traditional fitting methods are used for determining action potentials, then the process is quick, but the reliability of implant fitting parameters is poor due to high failure rates and inaccuracies
Solution Approach 1:
The patent replaces traditional manual correction methods and simple fitting algorithms with an automated multi-dimensional polynomial fitting system. This substitution eliminates the need for manual intervention while providing robust automated determination of physiological landmarks (extrema, action potentials) by fitting polynomials across multiple dimensions and identifying lines of minimum principal curvature, significantly improving reliability of implant fitting parameters
3Measurement precision
If manual correction is applied to fitting functions, then some inaccuracies can be corrected, but the process requires significant time and reduces productivity
Solution Approach 1:
The patent implements a self-correcting system where the multi-dimensional polynomial fitting automatically identifies and corrects inaccuracies in action potential determination. The system performs self-validation through mathematical consistency checks across multiple dimensions and automatically adjusts fitting parameters, eliminating the need for time-consuming manual correction while maintaining high measurement precision and improving overall productivity
Data Source
Figure 1
Figure 2~3
Figure 4
AI summary
Arrangements are described for determining a physiological characteristic of the auditory pathway (as whole or selected parts such as an inner ear). Electrical stimulation pulses are delivered to inner ear neural tissue and corresponding tissue response signals are developed by measuring over time response of the auditory pathway to each electrical stimulation pulse, with each tissue response signal forming a response curve including at least one physiological landmark such as a local maximum and a local minimum. A multi-dimensional polynomial is fit over the tissue response signals, and calculation starting points are defined based on prominent physiologic landmarks such as a local maximum and a local minimum for one selected tissue response signal. A line of minimum principal curvature of the multi-dimensional polynomial over the plurality of tissue response signals that intersect the calculation starting points is calculated to determine a physiological characteristic of the auditory pathway.