Cochlear implantation with hearing preservation yields significant benefit for speech recognition in complex listening environments

Abstract

Objective: The aim of this study was to assess the benefit of having preserved acoustic hearing in the implanted ear for speech recognition in complex listening environments.

Design: The present study included a within-subjects, repeated-measures design including 21 English-speaking and 17 Polish-speaking cochlear implant (CI) recipients with preserved acoustic hearing in the implanted ear. The patients were implanted with electrodes that varied in insertion depth from 10 to 31 mm. Mean preoperative low-frequency thresholds (average of 125, 250, and 500 Hz) in the implanted ear were 39.3 and 23.4 dB HL for the English- and Polish-speaking participants, respectively. In one condition, speech perception was assessed in an eight-loudspeaker environment in which the speech signals were presented from one loudspeaker and restaurant noise was presented from all loudspeakers. In another condition, the signals were presented in a simulation of a reverberant environment with a reverberation time of 0.6 sec. The response measures included speech reception thresholds (SRTs) and percent correct sentence understanding for two test conditions: CI plus low-frequency hearing in the contralateral ear (bimodal condition) and CI plus low-frequency hearing in both ears (best-aided condition). A subset of six English-speaking listeners were also assessed on measures of interaural time difference thresholds for a 250-Hz signal.

Results: Small, but significant, improvements in performance (1.7-2.1 dB and 6-10 percentage points) were found for the best-aided condition versus the bimodal condition. Postoperative thresholds in the implanted ear were correlated with the degree of electric and acoustic stimulation (EAS) benefit for speech recognition in diffuse noise. There was no reliable relationship among measures of audiometric threshold in the implanted ear nor elevation in threshold after surgery and improvement in speech understanding in reverberation. There was a significant correlation between interaural time difference threshold at 250 Hz and EAS-related benefit for the adaptive speech reception threshold.

Conclusions: The findings of this study suggest that (1) preserved low-frequency hearing improves speech understanding for CI recipients, (2) testing in complex listening environments, in which binaural timing cues differ for signal and noise, may best demonstrate the value of having two ears with low-frequency acoustic hearing, and (3) preservation of binaural timing cues, although poorer than observed for individuals with normal hearing, is possible after unilateral cochlear implantation with hearing preservation and is associated with EAS benefit. The results of this study demonstrate significant communicative benefit for hearing preservation in the implanted ear and provide support for the expansion of CI criteria to include individuals with low-frequency thresholds in even the normal to near-normal range.

Figures

FIGURE A1

Individual and mean audiometric thresholds, in dB HL, as a function of signal frequency, in Hz, for the implanted and the non-implanted ears obtained on the date of testing. Error bars represent +/− standard error measurement.

FIGURE A2

EAS benefit as a function of ITD threshold, in microseconds.

FIGURE 1

R-SPACE™ 8-loudspeaker system.

FIGURE 2

Mean pre- and post-implant audiometric thresholds for the implanted ears of the Polish- and English-speaking participants are shown as filled and unfilled circles, respectively. Mean thresholds for the non-implanted ears of the Polish (filled squares) and English (filled stars) participants are also displayed. Error bars represent +/− 1 standard deviation.

FIGURE 3

Individual and mean speech reception thresholds (SRT) in dB SNR are shown for the bimodal (gray bars) and best aided EAS (black bars) listening conditions. Unfilled circles represent SRT data for the binaural aided condition. Error bars represent +/− 1 standard error.

FIGURE 4

Individual and mean speech recognition scores in percent correct are shown for fixed level SNR of +6 dB. The bimodal and best aided EAS listening conditions are represented by gray and black bars, respectively. Error bars represent +/− 1 standard error.

FIGURE 5

Individual and mean speech recognition scores in percent correct are shown for fixed level SNR of +2 dB. The bimodal and best aided EAS listening conditions are represented by gray and black bars, respectively. Error bars represent +/− 1 standard error.

FIGURE 6

Normalized EAS benefit for speech recognition at +6 and +2 dB SNR as a function of low-frequency pure tone average (LF PTA) in dB HL in the implanted ear postoperatively, in the non-CI ear, as well as the degree of LF PTA elevation. Polish and English subject data are shown by filled and unfilled circles, respectively.

FIGURE 7

Individual and mean reverberant speech recognition, in percent correct, is shown for reverberation time of 0.6 seconds. The bimodal and best aided EAS listening conditions are represented by gray and black bars, respectively. Unfilled circles represent data obtained in the binaural aided condition. Horizontal dashed lines represent mean performance for the listeners with normal hearing. Error bars represent +/− 1 standard error.

FIGURE 8

Normalized EAS benefit for reverberant speech recognition as a function of low-frequency pure tone average (LF PTA) in dB HL in the implanted ear postoperatively, in the non-CI ear, as well as the degree of LF PTA elevation. Polish and English participant data are shown by filled and unfilled circles, respectively.