Fast neuromorphic sound localization for binaural hearing aids

Area of Science:

• Neuroscience
• Electrical Engineering
• Signal Processing

Background:

• Hearing aid systems often struggle with accurately localizing brief or faint sounds.
• Synaptic circuit noise can degrade the performance of auditory processing systems.
• Neuromorphic engineering offers potential solutions for enhancing sensory perception.

Purpose of the Study:

• To develop and evaluate a neuromorphic sound localization circuit for hearing aid applications.
• To optimize neuron circuit parameters for noise suppression and improved detection.
• To demonstrate real-time sound localization capabilities for narrow-duration sound pulses.

Main Methods:

• Implementation of a neuromorphic circuit using leaky integrate-and-fire neuron models.
• Parameter optimization of individual neuron circuits to mitigate synaptic noise.
• Characterization of the circuit's detection range and resolution.

Main Results:

• The proposed circuit demonstrates a detection range of 500 microseconds and a resolution of 5 microseconds.
• Successful localization of sound pulses with durations as short as approximately 1 millisecond.
• Achieved real-time response for sound localization tasks.

Conclusions:

• The neuromorphic sound localization circuit significantly enhances auditory perception in hearing aid systems.
• The optimized neuron circuits effectively suppress noise, leading to improved localization accuracy.
• This technology enables real-time processing of transient sounds, offering a substantial advancement for hearing assistance devices.