Cognitive Computing Research for Children with Cochlear Implants
Posted on: 17 June 2016
Trinity College Dublin, The National Centre for Cochlear Implants at Beaumont Hospital and IBM have launched a research project using cognitive computing to help predict speech perception in children with cochlear implants. It will improve device tuning and speed, but most critically it will improve overall patient outcomes in children with severe to profound hearing loss.
To obtain a high level of speech perception, the software within the cochlear implant must be customised and adjusted for each individual child. This is a complex and time consuming rehabilitation programme managed by audiological scientists over numerous sessions in order to obtain optimal access to speech and environmental sounds for the patient.
The project aims to use predictive modelling to help detect the subtle signs of vital changes in a patient's sound and speech perception to enable earlier proactive intervention.
IBM data scientists and multidisciplinary teams at the Trinity Centre for Bioengineering at Trinity College Dublin and The National Centre for Cochlear Implants at Beaumont Hospital will use cognitive computing to gain new insights into speech perception in cochlear implants recipients.
In addition, by analysing the diverse data from healthcare management systems, researchers can use machine based learning to aid the personalised approach to patient care and help to optimise the clinical decision-making processes while reducing costs.
“Combining our neural engineering research with cognitive computing will help us to gain new insights from our electrophysiology data of speech perception in cochlear implant patients and be a powerful complement in the development with our clinical colleagues of more personalised rehabilitation plans,” says Professor Richard Reilly, Professor of Neural Engineering at Trinity College Dublin.
The cognitive algorithms that the IBM data scientists will develop are based on machine learning that has drawn heavily on knowledge of neuroscience, statistics and applied mathematics. Professor Reilly added “the focus of our research at Trinity College Dublin is on clinical neural engineering based on signal processing of neuroimaging and physiological data for specific clinical problems. This collaboration with IBM helps to develop new understanding of how the brain processes audio information in cochlear implant recipients.”
Vice President of IBM Ireland Lab, Bill Kearney says: “The data scientists will use IBM Deep Learning platform that can automatically analyse large datasets to help discover patterns that people cannot discern. Cognitive computing applications are designed to learn from human experts. The data discoveries will enable the clinical researchers to help identify new relationships between the way specific individuals respond and allow for more personalised tailoring of treatments and lifelong management for the patients.”
The research project is supported by an IBM Faculty Award, a competitive worldwide programme intended to foster collaboration between researchers at leading universities worldwide and those in IBM research, development, and services organisations.