Implantable Microphone Signal Training for Consistent Hearing Quality

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Implantable microphones in medical devices, such as cochlear implants, exhibit unfavorable characteristics like high noise floor, non-flat frequency response, and sensitivity to internal vibrations, leading to poorer speech understanding and sound quality, especially when switching between external and implantable microphone modes.

Innovation Solution

Employing machine learning techniques, particularly deep neural networks, to transform implantable microphone signals to match external microphone signals by training a signal analysis module using simultaneous data from both types of microphones, thereby adjusting the implantable microphone output to resemble external microphone output.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of operation

If implantable microphones are used to provide invisible hearing, then the recipient can hear without external devices, but the sound quality and speech understanding deteriorate due to high noise floor and sensitivity to internal vibrations

Engineering Contradiction:

Improveinvisible hearing capabilityVSAvoidsound quality

Core Design Contradiction:

Ease of operationVSMeasurement precision

Solution Approach 1:

The system continuously compares signals from implantable and external microphones, using the external microphone signal as a reference to train the neural network. This feedback loop enables the implantable microphone signal to be adjusted to match the quality of external microphone signals, resolving the sound quality deterioration issue while maintaining invisible hearing capability

Inventive Principle:

Principle #23Feedback

Solution Approach 2:

The patent replaces traditional signal processing methods with machine learning-based neural networks. The neural network learns to transform implantable microphone signals to match external microphone signals, substituting conventional mechanical/acoustic signal processing with intelligent algorithms that can adaptively compensate for the implantable microphone's unfavorable characteristics

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

2Measurement precision

If machine learning techniques are used to transform implantable microphone signals, then sound quality consistency improves, but device complexity increases

Engineering Contradiction:

Improvesound quality consistencyVSAvoidsignal analysis module complexity

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The neural network is trained in advance using data from both implantable and external microphones before being deployed in the implantable device. This preliminary training phase allows the complex machine learning model to be prepared offline, reducing the computational complexity required during real-time operation in the implantable device while maintaining sound quality consistency

Inventive Principle:

Principle #10Preliminary action

Data Source

PatentUS20250381400A1Implantable sensor training

Publication Date: 2025.12.18 COCHLEAR LIMITED

AI summary

Presented herein are methods for training a signal analysis module. The methods include generating an external audio input based on sound signals captured by one or more external microphones: generating, with the signal analysis module, an implantable audio input based on sound signals captured by one or more implantable microphones; analyzing the implantable audio input relative to the external audio input; and adjusting operation of the signal analysis module based on the analyzing of the implantable audio input relative to the external audio input