Neural Encoding of Speech and Music: Implications for Hearing Speech in Noise

“Neural Encoding of Speech and Music: Implications for Hearing Speech in Noise”

The publisher regrets an error with the pictures used for [Figures 1] [2] [3] [4] [5] [6] [7] [8] in the above article in Seminars in Hearing, Volume 32, Number 2, 2011. pgs. 131–137. The correct Figures are indicated on the following page.

Figure 1 The grand average brain stem response waveform obtained from 38 typically developing children (middle) is visually similar to the [da] stimulus waveform (top). The spectrum of the brain stem response contains energy at the fundamental frequency (F0) and its integer multiples (bottom). Modified from Anderson et al, Hear Res 2010.⁷⁰

Figure 2 (Left) Years of musical practice are related to a measure of auditory attention (subtest of the Institute of Hearing Research (IHR) Multicentre Battery for Auditory Processing (IMAP); Medical Research Council Institute of Hearing Research, Nottingham, UK). (Right) Musicians (black) have faster reaction times on the auditory attention subtest than nonmusicians (gray) (*p < 0.05). Modified from Strait et al, Hear Res 2010.³³

Figure 3 (Top) Compared with nonmusicians (black), musicians (grey) have higher amplitudes for the harmonics of the upper tones of a musical chord (*p < 0.05, **p < 0.01). Notably, this difference was not seen for the fundamental frequency. (Bottom) Musicians again have higher amplitudes in response to the complex portion of the baby's cry compared with nonmusicians, but this enhancement was not seen in response to the periodic region of the response (**p < 0.01). Modified from Lee et al, J Neurosci 2009³¹; Strait et al, Eur J Neurosci 2009.²⁶

Figure 4 The ability to take advantage of stimulus regularities is a factor in speech-in-noise (SIN) perception. (Top) Brain stem responses of typically developing children are enhanced in the repetitive, regularly occurring presentation of the [da] syllable (grey) compared with the variable presentation (black). (Bottom) The degree of enhancement in the repetitive condition is related to SIN performance (Hearing in Noise Test). Modified from Chandrasekaran et al, Neuron 2009.⁴⁴

Figure 5 Time domain grand average responses to three speech syllables ([ga], [da], and [ba]) of 20 typically developing children (bottom panels) reflect differences in stimulus timing and cochlear tonotopicity that are not seen in the stimulus waveforms (top panels). The 52- to 57-millisecond region is magnified to highlight these timing differences. The scatter plot on the right demonstrates the relationship between speech-in-noise performance (Hearing in Noise Test [HINT]) and differentiation scores (the degree to which the response latencies of the three syllables correspond to the expected pattern). Modified from Hornickel et al, Proc Natl Acad Sci U S A 2009.⁵¹

Figure 6 The magnitude of the F0 is a factor in good speech-in-noise (SIN) perception. (Top) The top SIN perceivers in a group of 38 children (based on Hearing-in-Noise Test [HINT]) and a group of 17 young adults (based on the Quick Speech-in-Noise [QuickSIN] test) had higher F0 magnitudes than the bottom groups (*p < 0.05) during the formant transition in brain stem responses to the speech syllable [da] presented in quiet (children) and in background babble (young adults). (Bottom) In both children and young adults, F0 magnitude is related to scores on clinical measures of SIN perception (children, HINT; adults, QuickSIN). Modified from Anderson et al, Hear Res 2010⁷⁰ and Song et al, J Cog Neurosci 2010.⁶⁹

Figure 7 Brain stem processing of speech (including enhancement of stimulus regularities, magnitude of pitch cues, and the degree of noise induced timing shifts) predicted 56% of the variance in speech-in noise perception (based on Hearing-in-Noise Test, front and Left scores) using a structural equation model. Modified from Hornickel et al, Behav Brain Res 2011.⁷³

Figure 8 In response to the syllable [da], peak timing is less delayed by noise (six-talker babble) in musicians (black) compared with nonmusicians (grey), and the overall morphology (assessed by the degree to which the response correlates with the stimulus) is less degraded by noise in musicians than in nonmusicians. (A) Grand average responses of 15 young adult nonmusicians. The circled peaks correspond to the onset and transition peaks of the response that are more delayed in noise in nonmusicians compared with musicians. (B) Noise both delays and degrades the response. (C) In quiet, the onset peaks are essentially equivalent between groups, but noise causes a greater latency delay in nonmusicians (**p < 0.01). (D) The stimulus-to-response correlation r values are essentially equivalent in noise between groups, but noise causes greater degradation, as indicated by a decrease in r value in nonmusicians compared with musicians (**p < 0.01). When speech is presented in noise, Hearing-in-Noise Test scores are related to subcortical onset peak latencies (E) as well as stimulus-to-response correlation r-values (F). Modified from Parbery-Clark et al, J Neurosci 2009.⁷⁸