Frequency Following Response and Speech Recognition Benefit for Combining a Cochlear Implant and Contralateral Hearing Aid

Abstract

Multiple studies have shown significant speech recognition benefit when acoustic hearing is combined with a cochlear implant (CI) for a bimodal hearing configuration. However, this benefit varies greatly between individuals. There are few clinical measures correlated with bimodal benefit and those correlations are driven by extreme values prohibiting data-driven, clinical counseling. This study evaluated the relationship between neural representation of fundamental frequency (F0) and temporal fine structure via the frequency following response (FFR) in the nonimplanted ear as well as spectral and temporal resolution of the nonimplanted ear and bimodal benefit for speech recognition in quiet and noise. Participants included 14 unilateral CI users who wore a hearing aid (HA) in the nonimplanted ear. Testing included speech recognition in quiet and in noise with the HA-alone, CI-alone, and in the bimodal condition (i.e., CI + HA), measures of spectral and temporal resolution in the nonimplanted ear, and FFR recording for a 170-ms/da/stimulus in the nonimplanted ear. Even after controlling for four-frequency pure-tone average, there was a significant correlation (r = .83) between FFR F0 amplitude in the nonimplanted ear and bimodal benefit. Other measures of auditory function of the nonimplanted ear were not significantly correlated with bimodal benefit. The FFR holds potential as an objective tool that may allow data-driven counseling regarding expected benefit from the nonimplanted ear. It is possible that this information may eventually be used for clinical decision-making, particularly in difficult-to-test populations such as young children, regarding effectiveness of bimodal hearing versus bilateral CI candidacy.

Keywords: bimodal benefit; bimodal hearing; cochlear implants; electrophysiology; hearing aids.

Figure 1.

Audiometric thresholds (dB HL) in the nonimplanted ear from 125 to 8000 Hz for each participant.

Figure 2.

Speech recognition scores (rau) for CNC (a), AzBio sentences in +5 dB SNR (S₀N₀) (b), and AzBio sentences in +5 dB SNR (S₀N_45–315) (c) listening with the CI-only (x-axis) and in the bimodal condition (y-axis). See the main text for further details. CNC = consonant–nucleus–consonant; CI = cochlear implant; HA = hearing aid.

Figure 3.

Average acoustic benefit (a) measured in rau and normalized acoustic benefit (b) measured in percent for CNC words in quiet, AzBio sentences at +5 dB SNR (S₀N₀ and S₀N_45–315). Error bars represent ± 1 standard deviation. CNC = consonant–nucleus–consonant.

Figure 4.

Grand average envelope FFR waveform (a), grand average fine structure FFR waveform (b), FFR envelope spectrum (c), and FFR fine structure spectrum (d). Note that the envelope and fine structure spectra were calculated over the steady state (60–180 ms) range of the response waveform. Shaded regions represent SEM. Arrows point to amplitude at F0 (100 Hz; c) and F1 (700 Hz; d).

Figure 5.

Acoustic benefit (left column) measured in rau and normalized acoustic benefit (right column) measured in percent for CNC words (a and b), AzBio sentences in +5 dB SNR (S₀N₀) (c and d), and AzBio sentences in +5 dB SNR (S₀N_45–315) (e and f) as a function of envelope spectrum amplitude of the frequency following response at the fundamental frequency (F0, 100 Hz). CNC = consonant–nucleus–consonant.