Mechanisms of vowel recognition for ineraid patients fit with continuous interleaved sampling processors

Vowel recognition was assessed for eight, cochlear implant patients who use the Ineraid's six-electrode array. Recognition was tested in three conditions: with the Ineraid after years of experience; with a CIS processor at fitting of the processor; and with the CIS processor after 1 month's experience. At the time of fitting of the CIS processor, vowel recognition was not superior to that with the Ineraid. Recognition improved significantly over the period of a month. At 1 month, performance was significantly better with the CIS processor than with the Ineraid. This outcome is interpreted to mean that ramapping of the vowel space is necessary following fitting with the CIS processor and some of the remapping occurs over a time period of days or weeks, rather than hours. Vowel errors at one month could be accounted for by two mechanisms. One is that patients attended to low-frequency channels at the expense of high-frequency channels, or could not use information in high- frequency channels. The second is that, for diphthongs, patients could not detect frequency change over the course of the utterance.