EEG-Based Hearing Aid Isolates Attended Sound Source

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Hearing assistance systems face challenges in identifying and isolating relevant sound sources in noisy environments, as existing technologies struggle to accurately determine which sound sources a user is interested in, particularly due to the complexity of processing multiple sound sources and the limitations of current brain-wave signal processing methods.

Innovation Solution

A hearing assistance system utilizing a sparse model and dynamic finite impulse response (FIR) filters to analyze EEG signals and sound inputs, selectively determining the most relevant EEG electrodes and time intervals, and employing an alternating direction method of multipliers (ADMM) to identify the sound source the user is attending to, thereby reducing computational resources and energy consumption.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If traditional brain-wave signal processing methods are used to identify sound sources, then the system can process multiple sound sources, but the computational resources and energy consumption increase significantly

Engineering Contradiction:
Improvesound source identification accuracyVSAvoidenergy consumption
Core Design Contradiction:
Measurement precisionVSUse of energy by moving object

Solution Approach 1:

The patent extracts only the most relevant EEG channels and time intervals from the complete signal sets using sparse modeling techniques. By identifying and isolating the specific channels and time periods that contain the most informative brain responses to sound sources, the system processes only a fraction of the original data, significantly reducing computational load and energy consumption while maintaining identification accuracy.

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The patent applies local quality by assigning different weights and processing priorities to different EEG channels and time intervals based on their informational value. Rather than uniformly processing all channels, the system identifies channels with higher correlation to sound source activity and allocates more computational resources to those specific locations in the signal space, optimizing the trade-off between accuracy and energy use.

Inventive Principle:
Principle #3Local quality

2Measurement precision

If all EEG channels and time intervals are processed to ensure comprehensive sound source analysis, then the measurement accuracy improves, but the device complexity and computational requirements increase

Engineering Contradiction:
Improvesound source identification accuracyVSAvoidprocessing unit complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The processing unit extracts and focuses only on the most informative subsets of EEG channels and time intervals through sparse modeling. This extraction approach reduces the dimensionality of the problem from processing all channels to processing only those with significant correlation to sound source activity, thereby simplifying the device complexity while preserving measurement precision.

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The system performs partial action by processing only a selected subset of EEG data rather than the complete dataset. By applying sparsity constraints, the system identifies and processes only the necessary portion of channels and time intervals required for accurate sound source identification, avoiding the excessive computation that would result from analyzing all available data.

Inventive Principle:
Principle #16Partial or excessive action

3Use of energy by moving object

If a sparse model is used to select relevant EEG channels and time intervals, then computational resources and energy are reduced, but the processing time for model selection increases

Engineering Contradiction:
Improvecomputational energyVSAvoidmodel selection time
Core Design Contradiction:
Use of energy by moving objectVSLoss of time

Solution Approach 1:

The patent applies preliminary action by pre-computing and storing the sparse model parameters, channel selection weights, and time interval priorities during an initial calibration phase. Once the sparse model structure is established, the system can rapidly apply these pre-determined selections during actual sound source identification without repeating the computationally intensive model selection process, thus reducing real-time processing time while maintaining energy efficiency.

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentEP3267698B1A hearing assistance system comprising an EEG-recording and analysis system
Publication Date: 2024.10.09 OTICON
  • EP3267698B1 patent drawingFigure 1A~1B
  • EP3267698B1 patent drawingFigure 1C~1D
  • EP3267698B1 patent drawingFigure 1E~1F

AI summary

A hearing assistance system comprises an input unit for providing electric input sound signals ui, each representing sound signals U; from a multitude nu of sound sources Si, an electroencephalography (EEG) system for recording activity of the auditory system of the user's brain and providing a multitude ny of EEG signals yj, and a source selection processing unit receiving said electric input sound signals ui and said EEG signals yj, and in dependence thereof configured to provide a source selection signal Ŝx indicative of the sound source Sx that the user currently pays attention to using a selective algorithm that determines a sparse model to select the most relevant EEG electrodes and time intervals based on minimizing a cost function measuring the correlation between the individual sound sources and the EEG signals, and to determine the source selection signal Ŝx based on the cost functions obtained for said multitude of sound sources.