gms | German Medical Science

Understanding the reasons underlying variation in speech intelligibility of cochlear implant (CI) listeners is crucial for providing individualized treatment to these listeners. Speech intelligibility of CI users is influenced by several factors such as etiology, demographics, cognition, insertion angle and implant type. The number and the status of the surviving spiral ganglion neurons, i.e. cochlear neural health, may also be a particularly important factor [1]. Despite its apparent importance for speech perception of CI users [2], it is still a matter of ongoing debate how to reliably estimate cochlear health during a routine clinical visit. The goal of this study was to investigate the potential of cochlear health measures for predicting speech intelligibility alone and together with demographic data as well as electrode impedances. Speech intelligibility was assessed in 24 ears (13 CI users) by measuring speech reception thresholds (SRT) for the German matrix sentence test in stationary speech-shaped noise. ECAP amplitude growth functions (AGF) were measured on each electrode using stimuli with an initial cathodic phase and forward masking artefact reduction. The cochlear health measure of choice was the change in the slope of the AGF when the interphase gap (IPG) of the biphasic pulse was increased from 2.1 µs to 30 µs, which was defined as the IPG effect on slope (IPGEslope). To consider the relative importance of each frequency for speech intelligibility, the measured IPGEslope of each electrode was weighted according to the band importance function introduced by ANSI (1997). The averaged across-site mean value of these weighted IPGEslope values were reported as the cochlear health measure for each ear. This information, together with information about electrode impedances and demographic data were used to fit a multiple linear regression model. The results showed a mild but significant correlation (R2=0.33*) between the weighted IPGEslope and SRT. Employment of weighted IPGEslope together with CI experience and electrode impedances in a multiple linear regression model explained a larger range (R2=0.63**) of the variation in speech intelligibility. A significant correlation (R2= 0.48***) was observed between age and IPGEslope. Therefore, inclusion of age as independent variable did not provide substantial additional information to the multiple linear regression model. It can be concluded that IPGEslope does have the potential to be used as a clinical measure of cochlear neural health. Major parts of the interindividual variance in speech intelligibility of CI users may be explained by cochlear health, CI experience and electrode impedance. Inclusion of additional individual information such as insertion angle may help explaining variations in speech intelligibility to a higher extent which would help provide individual treatment and rehabilitation plans to CI users.