The role of envelope periodicity in the perception of masked speech with simulated and real cochlear implants

Abstract

In normal hearing, complex tones with pitch-related periodic envelope modulations are far less effective maskers of speech than aperiodic noise. Here, it is shown that this is diminished in noise-vocoder simulations of cochlear implants (CIs) and further reduced with real CIs. Nevertheless, both listener groups still benefitted significantly from masker periodicity, despite the lack of salient spectral pitch cues. The main reason for the smaller effect observed in CI users is thought to be an even stronger channel interaction than in the CI simulations, which smears out the random envelope modulations that are characteristic for aperiodic sounds. In contrast, neither interferers that were amplitude-modulated at a rate of 10 Hz nor maskers with envelopes specifically designed to reveal the target speech enabled a masking release in CI users. Hence, even at the high signal-to-noise ratios at which they were tested, CI users can still exploit pitch cues transmitted by the temporal envelope of a non-speech masker, whereas slow amplitude modulations of the masker envelope are no longer helpful.

A direct comparison of voice pitch processing in acoustic and electric hearing

“…For the CI ears, there was even a slight trend towards greater activity in the left hemisphere. With current CI systems, the access to spectral information is severely limited, whereas temporal envelope modulations are conveyed fairly accurately ( Chatterjee and Peng, 2008 , Macherey and Carlyon, 2014 , Steinmetzger and Rosen, 2018 ). Models of auditory processing propose that the temporal resolution of the left auditory cortex is higher compared to the right one, which is in turn characterised by a greater spectral resolution ( Poeppel, 2003 , Zatorre et al, 2002 ).…”

“…Although a few PET ( Giraud et al, 2001a , Giraud et al, 2001b , Giraud et al, 2000 , Mortensen et al, 2006 ) and fNIRS studies ( Chen et al, 2017a , Chen et al, 2017b , Mushtaq et al, 2020 , Olds et al, 2016 ) also used vowels or non-speech stimuli, systematic comparisons of basic auditory perception processes in CI users and normal-hearing (NH) controls are lacking. This is particularly striking with regard to pitch perception, which is very limited when listening through a CI and a crucial factor underlying the difficulties CI users encounter when listening to speech and music, especially in background noise ( Green et al, 2004 , Oxenham, 2008 , Steinmetzger and Rosen, 2018 ). The only pitch cues available to CI users are the periodic fluctuations of the temporal envelope of the acoustic input and hence their ability to identify the voice pitch contours that constitute a major part of prosody is strongly reduced ( Chatterjee and Peng, 2008 , Everhardt et al, 2020 ).…”

Predicting the effects of periodicity on the intelligibility of masked speech: An evaluation of different modelling approaches and their limitations

“…8. These representations were generated by computing the envelope power, as implemented in the front end of the mr-sEPSM, for each combination of auditory and modulation filters, time-averaged over the entire set of stimulus materials in each condition (see also Steinmetzger and Rosen, 2018). Figure 9, in contrast, shows modulation spectra of the eight different combinations of target speech and masker at each of the seven SNRs used for model prediction.…”

Model-based hearing-enhancement strategies for cochlear synaptopathy pathologies

“…Since temporal-envelope coding is degraded by CS, a prime candidate for acoustic intervention is thus to modify the speech envelope which is essential for robust speech intelligibility (SI) [19][20][21][22][23][24][25][26][27]. Acoustic modifications of the speech-envelope shape and periodicity can improve SI [17,[28][29][30][31][32], especially in connection with hearing through cochlear implants [33][34][35][36][37][38]. Here, we go beyond these conventional approaches to develop a model-based type of auditory envelope processing that operates directly on the signal waveform to counteract the functional consequences associated with CS.…”

“…Since temporal-envelope coding is degraded by CS, a prime candidate for acoustic intervention is thus to modify the speech envelope which is essential for robust speech intelligibility (SI) [23][24][25][26][27][28][29][30][31]. Acoustic modifications of the speech-envelope shape and periodicity can improve SI [18,[32][33][34][35][36][37], especially in connection with hearing through cochlear implants [38][39][40][41][42][43]. Several studies have shown that envelope enhancement can be beneficial for speech recognition in individuals with temporal processing deficits [22,36,[44][45][46][47][48].…”