Predicting speech improvement of individual paediatric cochlear implant recipients using high‐accuracy neural measures
Refereed conference paper presented and published in conference proceedings


Despite the use of cochlear implants (CI) in childhood deafness for over 30 years, high variability in postsurgical speech improvement continues to exist with 30% of recipients not meeting expected benefits. Neural markers have been increasingly shown to predict individual treatment outcomes more accurately than behavioural measures alone and could inform paediatric speech habilitation strategies. However, CI’s are incompatible to most brain imaging tools. The recently developed fNIRS provided a solution.

By comparing postsurgical functional brain imaging of CI recipients younger than 3.5 years at implant to typically developing peers, affected and unaffected brain regions were revealed. Templates of found regions were then created and used to predict speech outcomes of CI recipients by machine learning algorithms built from previously observed speech improvements.

It is expected that the previously auditory deprived auditory cortices show the highest prediction accuracy, sensitivity, and specificity. No results are available yet at this time of writing but are expected by Aug‐Sep.

This study’s findings are the first of its kind to demonstrate a low‐cost and clinically available speech improvement predictor with high accuracy. It is expected this study will provide a stepping stone towards an objective tool for choosing the best personalised speech habilitation strategy per individual CI recipient within the clinic.
著者Burghardt R. E., Lee K. Y. S., Wong P. C. M., Tong M. C. F.
會議名稱Hong Kong Speech and Hearing Symposium 2018
會議地點Hong Kong
會議論文集題名Hong Kong Speech and Hearing Symposium 2018
關鍵詞cochlear implant

上次更新時間 2019-21-06 於 15:10