Skip to main navigationSkip to main content
The University of Southampton
Web Science Institute

Better Hearing Healthier Aging; Use of Machine Learning for Early Detection of Cochlear Implant Underperformance


Hearing and interpretation of sounds with a cochlear implant relies on the integrity of the implant, the integration of the implant array (electrodes) with the tissue/neurons in the cochlea and appropriate stimulation of the auditory pathways in the brainstem and brain. Measures exist for each of these “compartments” and are collected at clinical appointments, or by the implant itself across its lifetime. People using cochlear implants show variable outcomes, ranging from exceptionally good hearing to those who have not achieved the anticipated level of benefit from their implant. The reasons for this variable outcome are not understood.

Aim: to develop a machine learning algorithm that uses multiple longitudinal data sets from the different “compartments” for the earliest possible identification of a drop in hearing performance, indicative of future cochlear implant failure.

Intended outcome: to develop this to enable real-time clinical use of this early warning algorithm to develop and test interventions that may prolong the effective use of the implant, and in parallel to increase the opportunity for more people to be implanted due to increased effectiveness and lower cost/individual.

Motivation: Hearing loss is a major contributor to disability affected life-years and costs @ £25bn/annum in the UK alone. It is identified as one of the strongest risk factors for the development of dementia. UKRI have set a grand challenge to stimulate innovation in routes to healthier lives, a reflection of the need for a step-change in approach to conditions of later life. Cochlear implants (CI), neuro-prosthetics, restore hearing function in otherwise severely deaf people and are associated with a range of benefits.  Once implanted, the device is supported for the lifetime of the user by the health service. Despite their proven effectiveness, the numbers of individuals being implanted is a small fraction of those who may benefit, this is partly due to the healthcare costs associated with undergoing surgery for, and lifetime care costs after, implantation. In parallel there is a need to ensure that the CI is effective for as long as possible across the remaining lifetime of the user, to reduce the need for, and burden of, re-implantation.


Principal Investigator: Dr Tracey Newman

Co Investigator: Professor Carl Verschuur, Dr Stephen Crouch, Mr HM Yuen, Dr Mary Grasmeder

Privacy Settings