Using Phoneme Perception to Modify CI Maps

Jane Madell
September 1, 2020

I have been teaching some online courses in Australia and Singapore this last week. I have been talking about the Medial Consonant Test, developed by Rich Tyler at the University of Iowa in the early days of cochlear implants. At that time, candidates for implantation were adults who had long term profound hearing loss.

The test was developed to help determine what phonemes an individual could hear and what they were missing. I am not sure if the test was originally intended to be used to help modify cochlear implant maps but that is how I started using it in the early 1990’s when I was at Beth Israel/New York Eye and Ear Medical Centers.

With my colleagues there, with other colleagues like Joan Hewitt since, we have used the information about errors, to modify CI programming.


Modifying CI Programming


As we know, obtaining threshold and comfort level is not letting us understand how much speech a person is understanding. Phoneme perception like that on the Medial Consonant Test is a good way to get information about perception. When we do this test, we record what the error confusion was, not just whether the it was right or wrong.

For example If the stimulus was /d/ and the person said /t/ I assume that the person is not hearing the voicing cue at around 300 Hz. If the stimulus is /s/ and the person says /sh/ I assume they are missing high frequency information at around 5000 Hz.

By recording the phoneme confusion, and by evaluating all the formant bands for each of the consonants, we can determine what is missing and how to change the map.  Results should be immediate.

This is just an extension of doing speech perception testing and using it to modify technology.

Everyone who tries this loves it. Please try it.

Leave a Reply