Speech Recognition and Listening Effort in Cochlear Implant Recipients and Normal-Hearing Listeners

Abstract

The outcome of cochlear implantation is typically assessed by speech recognition tests in quiet and in noise. Many cochlear implant recipients reveal satisfactory speech recognition especially in quiet situations. However, since cochlear implants provide only limited spectro-temporal cues the effort associated with understanding speech might be increased. In this respect, measures of listening effort could give important extra information regarding the outcome of cochlear implantation. In order to shed light on this topic and to gain knowledge for clinical applications we compared speech recognition and listening effort in cochlear implants (CI) recipients and age-matched normal-hearing listeners while considering potential influential factors, such as cognitive abilities. Importantly, we estimated speech recognition functions for both listener groups and compared listening effort at similar performance level. Therefore, a subjective listening effort test (adaptive scaling, "ACALES") as well as an objective test (dual-task paradigm) were applied and compared. Regarding speech recognition CI users needed about 4 dB better signal-to-noise ratio to reach the same performance level of 50% as NH listeners and even 5 dB better SNR to reach 80% speech recognition revealing shallower psychometric functions in the CI listeners. However, when targeting a fixed speech intelligibility of 50 and 80%, respectively, CI users and normal hearing listeners did not differ significantly in terms of listening effort. This applied for both the subjective and the objective estimation. Outcome for subjective and objective listening effort was not correlated with each other nor with age or cognitive abilities of the listeners. This study did not give evidence that CI users and NH listeners differ in terms of listening effort - at least when the same performance level is considered. In contrast, both listener groups showed large inter-individual differences in effort determined with the subjective scaling and the objective dual-task. Potential clinical implications of how to assess listening effort as an outcome measure for hearing rehabilitation are discussed.

Keywords: cognition; dual-task; effort scaling; listening effort; speech recognition; working memory.

FIGURE 1

Estimated speech recognition functions for the CI recipients (green) and the NH listeners (violet). Thin lines show individual functions, bold lines show the group mean. SNR = signal-to-noise ratio.

FIGURE 2

Listening effort (LE) assessed by ACALES as a function of the signal-to-noise ratio (SNR) in CI recipients (green) and NH listeners (violet). Bold lines represent the mean. ESCU = Effort Scaling Category Units. The dashed horizontal line shows the value of 7 ESCU (“moderate effort”).

FIGURE 3

Listening effort (LE) in CI recipients (green) and NH listeners (violet) at SNRs associated with 50 and 80% speech recognition. ESCU = Effort Scaling Categorical Units.

FIGURE 4

Outcome of the dual-task paradigm. (A,B) Primary task (speech recognition) in CI and NH listeners, (C,D) secondary task (reaction time) in CI and NH listeners, (E) proportional dual-tasks costs for the secondary task. CI, cochlear implant recipients; NH, normal-hearing listeners. Squares = single-task, crosses = dual-task.