Performance variability on perceptual discrimination tasks in profoundly deaf adults with cochlear implants

Abstract

Objectives: The purpose of this study was to evaluate performance on two challenging listening tasks, talker and regional accent discrimination, and to assess variables that could have affected the outcomes.

Study design: A prospective study using 35 adults with one cochlear implant (CI) or a CI and a contralateral hearing aid (bimodal hearing) was conducted. Adults completed talker and regional accent discrimination tasks.

Methods: Two-alternative forced-choice tasks were used to assess talker and accent discrimination in a group of adults who ranged in age from 30 years old to 81 years old.

Results: A large amount of performance variability was observed across listeners for both discrimination tasks. Three listeners successfully discriminated between talkers for both listening tasks, 14 participants successfully completed one discrimination task and 18 participants were not able to discriminate between talkers for either listening task. Some adults who used bimodal hearing benefitted from the addition of acoustic cues provided through a HA but for others the HA did not help with discrimination abilities. Acoustic speech feature analysis of the test signals indicated that both the talker speaking rate and the fundamental frequency (F0) helped with talker discrimination. For accent discrimination, findings suggested that access to more salient spectral cues was important for better discrimination performance.

Conclusions: The ability to perform challenging discrimination tasks successfully likely involves a number of complex interactions between auditory and non-auditory pre- and post-implant factors. To understand why some adults with CIs perform similarly to adults with normal hearing and others experience difficulty discriminating between talkers, further research will be required with larger populations of adults who use unilateral CIs, bilateral CIs and bimodal hearing.

Keywords: Bimodal hearing; Cochlear implants; Speech discrimination.

Figure 1

D-prime results for the talker discrimination task are shown in this figure. The horizontal edges of each box represent the 25^th and 75^th percentiles, the solid line within the box represents the median and the dotted line represents the mean. The whiskers represent the 10^th and 90^th percentiles and the solid circles show the suspected outliers. The outcomes for all of the participants are shown in the far right box, the results for the participants who used one cochlear implant are provided in the center, and finally, the outcomes from the participants who used bimodal hearing are provided in the left box plot.

Figure 2

Outcomes from the talker discrimination task are displayed in this figure. The overall percent correct is provided on the far left, and each of the remaining plots display the percent correct for the listening condition when both sentences were spoken by the same talker, by talkers of the same gender, and by talkers of different genders. The horizontal edges of each box represent the 25^th and 75^th percentiles, the solid line within the box represents the median and the dotted line represents the mean. The whiskers represent the 10^th and 90^th percentiles and the solid circles show the suspected outliers.

Figure 3

The d-prime results for the accent discrimination task are presented. The results for the adults who used bimodal hearing are provided on the left, the center box shows the results for the adults who used one CI and the results for all study participants are provided in the right box plot. The horizontal edges of each box represent the 25^th and 75^th percentiles, the solid line within the box represents the median and the dotted line represents the mean. The whiskers represent the 10^th and 90^th percentiles and the solid circles show the suspected outliers.

Figure 4

Percent correct outcomes from the accent discrimination task are presented in this figure. The horizontal edges of each box represent the 25^th and 75^th percentiles, the solid line within the box represents the median and the dotted line represents the mean. The whiskers represent the 10^th and 90^th percentiles and the solid circles show the suspected outliers. Within the figure, the overall performance is shown along with the performance when northern and southern talkers were paired, when both talkers were northern, and when both talkers were southern.