Predicting speech in noise performance using evoked responses

Lead Research Organisation: University of Southampton
Department Name: Faculty of Engineering & the Environment


Our team is currently researching the response of the brain to speech and we were recently funded by the UK Engineering and Physical Science Research Council (EPSRC) for the project 'Personalized fitting and evaluation of hearing aids with electroencephalography (EEG) responses'. Evoked responses in brain signals, measured from electrodes placed on the head, are routinely used to test people who cannot cooperate with standard hearing tests, such infants or people with learning difficulties, but are also used in more advanced diagnostic studies. Our new work is focusing on the responses to speech rather than the clicks or short tone stimuli that are currently routinely used. Speech responses are more challenging to measure, but also relevant in every-day listening environments. The goal is for evoked responses to guide optimised hearing aid fittings for patients who may otherwise struggle to understand speech in noise.

It is well know that a standard hearing test using pure tones (the audiogram) is not a good predictor of speech in noise performance. Some people with normal audiograms can struggle to hear in noise and in some cases may be suffering from 'Auditory Processing Disorder'. Conversely patients with hearing loss may show similar ability to deal with speech in noise, even if they have considerable variations in their audiograms. Clearly more sophisticated tests and a better understanding of brain activity during listening are needed to predict speech in noise performance.

The aim of the current project is to explore whether and how speech evoked response measurement can improved test approaches. We will make subjective measurements of speech perception in different levels of background noise and compare results to measurements of evoked responses to sound using surface EEG. The research work will involve the development, testing and optimization of signal processing methods and the analysis of signals collected from healthy volunteers and patients with a range of hearing impairments.


10 25 50

Studentship Projects

Project Reference Relationship Related To Start End Student Name
EP/N509747/1 01/10/2016 30/09/2021
1938440 Studentship EP/N509747/1 25/09/2017 31/07/2021 Samuel Perry
Description A compact and inexpensive virtual acoustics system for clinical and research evaluation of spatial hearing in cochlear implant and hearing aid users
Amount kr 660,000 (DKK)
Organisation Oticon Foundation 
Sector Charity/Non Profit
Country Denmark
Start 03/2020 
End 03/2021
Description Electro-haptic hearing: Using tactile stimulation to improve speech-in-noise performance and music perception in cochlear implant users
Amount kr 3,219,313 (DKK)
Organisation Oticon Foundation 
Sector Charity/Non Profit
Country Denmark
Start 07/2019 
End 06/2022