Noise-Induced Changes of the Auditory Brainstem Response to Speech—a Measure of Neural Desynchronisation?

Abstract

It is commonly assumed that difficulty in listening to speech in noise is at least partly due to deficits in neural temporal processing. Given that noise reduces the temporal fidelity of the auditory brainstem response (ABR) to speech, it has been suggested that the speech ABR may serve as an index of such neural deficits. However, the temporal fidelity of ABRs, to both speech and non-speech sounds, is also known to be influenced by the cochlear origin of the response, as responses from higher-frequency cochlear regions are faster and more synchronous than responses from lower-frequency regions. Thus, if noise caused a reweighting of response contributions from higherto lower-frequency cochlear regions, the temporal fidelity of the aggregate response should be reduced even in the absence of any changes in neural processing. This 'place mechanism' has been demonstrated for non-speech ABRs. The aim of this study was to test whether it also applies to speech ABRs. We used the so-called 'derived-band' method to isolate response contributions from frequency-limited cochlear regions. Broadband and derived-band speech ABRs were measured both in quiet and in noise. Whilst the noise caused significant changes to the temporal properties of the broadband response, its effects on the derived-band responses were mostly restricted to the response amplitudes. Importantly, the amplitudes of the higher-frequency derived-band responses were much more strongly affected than those of the lower-frequency responses, suggesting that the noise indeed caused a reweighting effect. Our results indicate that, as for non-speech ABRs, the cochlear place mechanism can represent a potentially substantial confound to speech-ABR-in-noise measurements.

Differences between children and adults in the neural encoding of voice fundamental frequency in the presence of noise and reverberation

“…Results from the present study are not directly comparable with the majority of previous EFR studies evaluating the effects of noise and reverberation due to the use of broadband vowel stimuli. In the two known studies that used a frequency‐specific approach, EFRs at f 0 generated by F1, but not higher formants, were found to be robust or unaffected by noise (Boer et al, 2020; Laroche et al, 2013). Evidence from guinea pigs suggest that the robustness of f 0 encoding in the F1 region is supported by phase‐locking to individual resolved harmonics (i.e., fine structure).…”

“…Given the good correspondence between our experimental data and simulations with identical stimuli at least for F1 EFRs (Figure 6), we speculate that the discrepancies between our experimental data and previous studies arise from multiple methodological differences, at least in part. Influential methodological factors may include varied excitation patterns for filtered formants presented independently (Laroche et al, 2013), noise type (white noise in Laroche et al, 2013, vs. equal‐excitation noise per auditory filter in Boer et al, 2020, vs. speech shaped noise in the present study) and SNR (−5 dB in Laroche et al, 2013, vs. 20 and 10 dB in Boer et al, 2020 vs. 5 dB in the present study).…”

“…A constraint with EFRs elicited by vowels that are naturally broadband is the difficulty ascertaining the cause of noise‐induced deterioration on EFR amplitude. The attenuation may not only reflect neural desynchronization but may also reflect a shift in the dominant cochlear place of EFR initiation (Boer et al, 2020). While EFRs in quiet conditions are mostly dominated by contributions from mid‐to‐high frequency harmonics that are unresolved in the cochlea (i.e., multiple harmonics pass through an auditory filter; Easwar et al, 2018; Nuttall et al, 2018), the presence of noise could shift the dominance to low frequency resolved harmonics (i.e., only one harmonic passes through an auditory filter) because EFRs from the resolved harmonics may be less affected or attenuated by noise (Boer et al, 2020; Laroche et al, 2013).…”