Neural Network Hearing Aid Individualization for Synaptopathy

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Current auditory signal processing methods fail to accurately compensate for different types of hearing impairment, particularly synaptopathy, and do not effectively individualize processing algorithms for each listener, leading to inadequate speech intelligibility restoration.

Innovation Solution

The development of an artificial neural network-based method that generates individualized auditory signal processing models by accounting for the integrity of auditory nerve fibers and synapses, incorporating metrics like OAEs and AEPs, and using NN-based models to simulate the auditory periphery and compensate for hearing impairments.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If standard hearing-aid algorithms (NAL-NL or DSL) are used to compensate for frequency-specific outer-hair-cell damage, then frequency-specific gain compensation is improved, but compensation for synaptopathy and individualization for different hearing impairment types deteriorates

Engineering Contradiction:

Improvefrequency-specific gain compensationVSAvoidcompensation for different hearing impairment types

Core Design Contradiction:

Measurement precisionVSAdaptability or versatility

Solution Approach 1:

The patent applies local quality by creating specialized processing branches for different hearing impairment types (outer-hair-cell loss, synaptopathy, normal hearing) within the hearing aid system. Each branch has tailored signal processing algorithms optimized for its specific impairment type, allowing frequency-specific compensation for OHC damage while simultaneously providing synaptopathy-appropriate processing without compromising either function.

Inventive Principle:

Principle #3Local quality

Solution Approach 2:

The hearing aid system segments the auditory processing into distinct pathways based on diagnosed impairment type. The system divides the single processing chain into multiple parallel processing streams (normal-hearing pathway, OHC-loss pathway, synaptopathy pathway), each handling specific impairment characteristics independently, thereby resolving the contradiction between specialized frequency compensation and broad adaptability.

Inventive Principle:

Principle #1Segmentation

2Device complexity

If a single processing algorithm is used for all listeners, then device complexity is reduced, but individualization for different hearing impairments deteriorates

Engineering Contradiction:

Improveprocessing algorithm structureVSAvoidindividualization for different listeners

Core Design Contradiction:

Device complexityVSAdaptability or versatility

Solution Approach 1:

The system dynamically adapts its processing complexity based on the diagnosed impairment type. Rather than maintaining fixed high complexity for all users, the system activates only the necessary processing branches for each individual's specific hearing condition. This dynamic configuration allows the device to scale complexity appropriately, reducing it for simple cases while providing full individualization for complex impairments.

Inventive Principle:

Principle #15Dynamics

Solution Approach 2:

The hearing aid employs a universal processing framework that can handle multiple impairment types through a single integrated system architecture. The multi-functional processing unit can operate in different modes (normal hearing, OHC loss, synaptopathy) depending on the diagnosed condition, eliminating the need for entirely separate devices or algorithms for each impairment type while maintaining full individualization capability.

Inventive Principle:

Principle #6Universality (Multi-functionality)

3Measurement precision

If psychoacoustic experiments are conducted to measure speech perception capabilities, then individualization accuracy is improved, but time consumption and cost increase

Engineering Contradiction:

Improvespeech perception capability measurementVSAvoidtime for individualization process

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

The system performs preliminary automated diagnosis and impairment characterization during the initial fitting process using objective measures (audiograms, OAEs, ABR) rather than requiring extensive subsequent psychoacoustic testing. By establishing the impairment type and parameters in advance through efficient automated assessments, the system reduces the time required for individualization while maintaining accurate measurement of speech perception capabilities through the neural network models.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The patent replaces time-consuming manual psychoacoustic experimentation with automated neural network-based assessment systems. The neural networks automatically analyze auditory responses and predict speech perception capabilities without requiring lengthy behavioral tests, thereby substituting the mechanical/experimental process with an automated computational approach that maintains precision while dramatically reducing time and resource requirements.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Data Source

PatentEP4128224B1A closed-loop method to individualize neural-network-based audio signal processing

Publication Date: 2025.06.04 UNIV GENT

Figure 1
Figure 2
Figure 3

AI summary

The present invention is in the field of auditory devices. In particular, the present invention provides a method for converting an auditory stimulus to a processed auditory output. The present invention also relates to uses of the method, auditory devices configured to perform the method, and computer programs configured to perform the method for converting an auditory stimulus to a processed auditory output.