Speech coding strategies for Cochlear Implants
Research type
Research Study
Full title
Speech coding strategies for Cochlear Implants: Modification for real-time processing and validation
IRAS ID
330132
Contact name
Tobias Goehring
Contact email
Sponsor organisation
Cambridge University Hospitals NHS Foundation Trust & the University of Cambridge
Duration of Study in the UK
1 years, 9 months, 31 days
Research summary
Nearly 800,000 profoundly deaf or severely hard-of-hearing individuals rely on a cochlear implant (CI) to hear. Although CIs can often restore speech perception in quiet listening conditions, speech understanding can be considerably worse in conditions containing background noise, such as at work, school or in social settings. The reduced ability to communicate in these settings can lead to social isolation, negatively impacting the quality of life and mental health of the affected person and their peers. The interventions that can be feasibly applied within modern CI devices are limited since the CI is a wearable device and as such is power- and compute-constrained. Our project aims to alleviate communication difficulties and reduce power consumption with CIs, thereby improving the quality of life and ease of device use for CI recipients.
CIs transform an acoustic signal into a pattern of electric pulses used to stimulate the auditory nerve via a surgically implanted electrode array. During the transmission of acoustic information to the central auditory system via the CI, a major bottleneck occurs at the electrode-nerve interface. We leverage a model of perceptual masking, a phenomenon that occurs when a lower-magnitude signal is not perceived due to the proximity of a higher-magnitude signal, to remove pulses from the stimulation pattern that are imperceptible. Removing perceptually redundant pulses from the stimulation pattern improves the efficiency of information transmission while reducing the power required for stimulation with CIs. This project will improve this method to meet the computational and real-time constraints of modern CI processors and collect supporting evidence of its efficacy in ecologically-valid listening scenarios with CI recipients.REC name
South Central - Hampshire B Research Ethics Committee
REC reference
24/SC/0084
Date of REC Opinion
6 Mar 2024
REC opinion
Favourable Opinion