This programme is not yet complete.
Welcome to ISAAR 2021.
Most hearing impairment in adults arises from damage to the sensory cells and/or nerve fibers of
the inner ear. This talk will summarize recent research on animal models and human autopsy
material showing that, in both noise-induced and age-related hearing loss, the synaptic terminals
of cochlear nerve fibers degenerate first, leaving their peripheral targets, the inner hair cells,
partially disconnected from the brain. This primary neural degeneration has little effect on
hearing thresholds (the audiogram) but affects discrimination of complex sounds like speech.
Because the cell bodies and central projections of the cochlear neurons survive long after loss of
their synaptic connections, there is a therapeutic window for repair, as has been shown in animal
models using both local delivery, and virally mediated overexpression, of neurotrophins. The
endogenous capacity for repair is low in mice but high in guinea pigs, and the differences may
provide further clues for therapeutic approaches.
‘Hidden hearing loss’ has inspired a wealth of research, including what and how morphological changes in the auditory periphery might cause this phenomenon. For example, the stochastic undersampling model (Lopez-Poveda & Barrios, 2013) suggests that auditory deafferentation can potentially introduce internal noise in the subsequent auditory processing stages. In this model, auditory fibers (AF) are modelled as samplers, which sample the input sound at individual stochastic rates, and the loss of AFs is mimicked by reducing the number of samplers. However, the parameters used in this model do not capture the full complexity of physiological response characteristics, thus leaving unclear the quantity of information conveyed by the AFs. In our study, half-wave rectification, refractoriness, and three types of AFs are added to the original model to explicitly model AF (type) loss within a more realistic physiological setting. In addition, an artificial-neural-network-based stimulus reconstruction is used to decode the modelled AF responses back to an audio signal (Akbari et al., 2019, Morise et al., 2016). We conducted a pure tone in noise (PTiN) detection task and a modified version of HINT (Nilsson et al., 1994) via MTurk. The behavioral stimuli were degraded using our model, with 3 levels of AF loss (0, 90, 95%). Preliminary results indicate that the PTiN threshold increases significantly with a decrease in the number of fibers, at a rate that aligns well with predictions from Oxenham (2016). For the HINT, the results only showed a significant threshold shift between the 90% and 95% AF loss conditions. In conclusion, our model combines detailed physiological response properties with the stochastic undersampling model and thereby enables more realistic artificial introductions of lesions in the peripheral auditory pathway (e.g. selective frequency loss, or fiber type loss) and thus can benefit the study of auditory pathology for improving hearing restoration devices.
Hearing disorders are typically studied and treated from the perspective of wanting to make inaudible
sounds audible. Yet three of the most common and debilitating adult hearing complaints reflect just the
opposite problem: not what persons cannot hear, but what they cannot stop hearing. Older adults or
persons with a history of noise exposure often struggle to suppress the awareness of background noise
sources when listening to a target speaker, they are often assaulted by the irrepressible perception of
phantom sounds (tinnitus), and they can experience moderate intensity sounds as loud, distressing, or
even painful (hyperacusis). Although age, noise exposure, and hearing status are risk factors for these
perceptual disorders, their connection is indirect at best, prompting much speculation about the
intervening neural processes that may be more closely related. Work from our lab and others shows
that an underlying root cause for each of these disorders may be found in a dialog gone wrong between
cochlear primary afferent neurons and neurons in sound processing centers of the brain. Our work in
animal models has shown that cochlear neural degeneration (CND) triggers a compensatory plasticity at
higher stages of the central auditory pathway that often over-shoots the mark, rendering neurons
hyperactive, hypersensitive, hyper-synchronized, and internally ‘noisy’. Using in situ mRNA profiling,
optogenetics, single unit electrophysiology, and calcium imaging in behaving animals, I will show how
CND triggers excess central gain in the auditory cortex and how this central pathophysiology directly
underlies poor hearing in noise. I will also describe our ongoing efforts to develop physiological
biomarkers for these maladaptive central plasticity processes in human subjects and well as
interventions that improve multi-talker speech intelligibility in older adults with sensorineural hearing
loss by targeting noisy processing in the brain rather than focusing on the signal transmitted from the
Find more details by clicking on the ‘room’ tabs.
The goal of this project is to explore representations of pitch using physiological models for auditory-nerve (AN) and midbrain (inferior colliculus, IC) neurons. An established model for ‘central pitch’ (Goldstein, 1973, JASA) requires a robust neural response profile that corresponds to the spectrum of a harmonic tone complex. Representations based on auditory-nerve excitation patterns (rate vs. place profiles) change with sound level and are not robust in background noise. However, f0-related fluctuations in AN responses are robust both across levels and in noise. These peripheral fluctuations ultimately influence responses of IC neurons, for which a key property is amplitude-modulation tuning. Because IC neurons are sensitive to slow fluctuations of their inputs, the fluctuation profiles set up in the periphery map into rate-profiles across IC neurons (Carney, 2018, JARO). Thus, the population responses of IC neurons provide the input required by central pitch models. Here, the representation of periodicity pitch by model midbrain neurons will be tested for several stimuli. Estimates of pitch discrimination thresholds and pitch strength can be made based on model responses. Importantly, the fluctuation profiles in AN responses depend upon inner-ear nonlinearities that are affected by sensorineural hearing loss (SNHL). Specifically, the ‘flattening’ of fluctuations near harmonics in tone complexes, or near spectral peaks of complex sounds, is reduced when cochlear amplification and/or inner-hair-cell sensitivity is reduced. Thus, SNHL reduces the contrast in fluctuation amplitudes acrossAN frequency channels, and diminishes this mechanism for coding features of complex sounds. Effects of SNHL on pitch discrimination and pitch strength can thus be studied in this physiological-modeling framework. This effort builds on the pitch-discrimination modeling work in Bianchi et al. (2018, JARO), but takes advantage of the central pitch model to map physiological-model responses into decision variables for pitch-related tasks.
Any clinical trial for hearing treatments would benefit from improved stratification of the participants according to the pathophysiology underlying their hearing loss, but this need becomes more acute as molecular or small molecule approaches are developed. Three major categories of cochlear pathology in age-related hearing loss were proposed by Schuknecht & Gacek (1993): sensory (hair cell dysfunction), metabolic (stria vascularis dysfunction) and neural (auditory neuron defects). Non-invasive methods for distinguishing hair cell from synaptic/neural defects have been proposed, although there are concerns regarding sensitivity and specificity. However, methods for identifying a strial defect are not well-established beyond audiogram shapes (Dubno et al 2013), yet these are important to detect because they will require different treatment approaches, and there is little point in treating a hair cell or a synaptic defect if the primary pathology is a dysfunctional stria. In the mouse we have the advantage of a better understanding of the underlying pathology compared with humans. We are using a set of mouse mutants with known initial sites of lesion to search for diagnostic tools based on objective electrophysiological measures. The mutants have a primary strial dysfunction and reduced endocochlear potential (S1pr2stdf), a primary inner hair cell defect (Klhl18lowf), a primary outer hair cell defect (Slc26a5tm1(EGFP/cre/ERT2)Wtsi) and a primary synaptic abnormality with swelling of synaptic boutons under inner hair cells (Wbp2tm1a). So far, we have used features of ABRs and DPOAEs to distinguish between inner and outer hair cell defects but not strial dysfunction (Ingham et al. 2020). We continue investigating ABR waveform features, frequency tuning, forward masking responses, increasing click repetition rates, tone-in-noise responses and inter-trial coherence tests as non-invasive objectives measures that have potential to be translated into a human diagnostic test.
The acquisition of new skills, including aural communication, can be facilitated when a naïve
observer is exposed to a conspecific performing a well-defined behavior (i.e., social
learning). Although the neural bases for auditory social learning remain uncertain, one
plausible hypothesis is that social experience induces experience-dependent plasticity in
auditory cortex, as found for many forms of learning, thereby facilitating task acquisition. To
explore this idea, we developed a social learning paradigm in which naïve Observer gerbils
are exposed to a Demonstrator gerbil that is performing an amplitude modulation (AM) rate
discrimination task across an opaque divider. Observer gerbils subsequently acquire the AM
task more rapidly than controls (Paraouty et al., 2020). We first asked whether auditory
cortex activity is necessary for social learning by we transiently inactivating it during social
exposure to the Demonstrator. These Observers did not benefit from social exposure,
suggesting a necessary role for auditory cortex. To determine whether neural plasticity was
induced by social experience, we recorded from the Observer’s auditory cortex during social
exposure. Auditory cortex neurons displayed improved AM discrimination during the period
of social exposure. Furthermore, the magnitude of neural improvement correlated with an
animal’s subsequent rate of task acquisition. Together, these findings suggest that auditory
cortex plasticity plays a pivotal role in social learning.
It is commonly believed that the effects of exposure to noise cease once the exposure itself has ceased. If this is the case, exposure to noise relatively early in life, for example during military service, should not affect the subsequent progression of hearing loss. However, recent data from studies using animals suggest that noise exposure can accelerate the subsequent progression of hearing loss. In this paper I review data from published studies on the effects of noise exposure on the progression of hearing loss once noise exposure has ceased, particularly for the case of noise exposure during military service. I also present some new longitudinal data obtained from military personnel. The results are consistent with the idea that noise exposure during military service accelerates the progression of hearing loss at frequencies where the hearing loss is absent or mild at the end of military service, but has no effect on or slows the progression of hearing loss at frequencies where the hearing loss exceeds about 50 dB. Acceleration appears to occur over a wide frequency range, including 1 kHz. However, there is a need for further longitudinal studies using a larger number of subjects. Longitudinal studies are also needed to establish whether exposure to other types of sounds, for example at rock concerts, affects the subsequent progression of hearing loss.
Perception of a speaker’s voice is not only important for talker identification, but also for
assessing their emotional state and better understanding their speech, by adapting to their
speaking style, or segregating their speech from background interfering speakers. Our focus
on voice perception is further motivated by our interest in users of cochlear implants (CIs),
whose speech communication relies on electrically transmitted speech signals that are
inherently degraded in their spectro-temporal details. Previous research with post-lingually
deafened and implanted adult CI users had shown that this typical CI group seem to have
difficulties in tasks related to voice perception. Therefore, to be able to introduce any
improvement to this problem, we have to understand specific perceptual mechanisms
related to voice perception. For this purpose, we have been collecting data on development
of voice perception throughout childhood to adulthood. More specifically, in our PICKA
(Perception of Indexical Cues in Kids and Adults) project we have been investigating voice
cue perception, vocal gender categorization, voice emotion perception, and speech on
speech perception. In these experiments, we focus on vocal pitch (F0, related to glottal
pulse rate) and vocal tract length (VTL, related to formants) as these voice cues can largely
affect the perceived gender of a speaker and could be manipulated separately or together
from a single speaker using speech synthesis tools (e.g., STRAIGHT). Further, we have
been testing various groups as different learning models, such as early deafened children
and adults, implanted early or later in life, and who have to learn voice cues via their
implant directly. This systematic approach has revealed many interesting observations.
Namely, mechanisms related to voice cue use seem to develop slowly in childhood over
many years, and the effectiveness of use of voice cues seems to differ greatly across child
or adult users of CIs.
Find more details by clicking on the ‘room’ tabs above.
While many studies have reported a loss of sensitivity to interaural time differences (ITDs) carried in the fine-structure of low-frequency signals for listeners with hearing impairment (HI), relatively few data are available on the perception of ITDs carried in the envelope of high-frequency signals ITDs in this population. The few studies that exist found stronger effects of hearing loss at high frequencies than at low frequencies, but small subject numbers and several confounding effects prevented strong conclusions from being drawn. In the present study, we revisited this question while addressing some of the issues identified in previous studies. First, we focused on “rustle” stimuli that contain strong envelope fluctuations at high frequencies and thus have the potential to provide salient envelope ITDs. Second, we carefully equated sensation level across listeners and tested two different levels per listener, to better characterize effects of level. Third, we included young listeners in the HI group to tease apart effects of hearing loss and age. ITD discrimination thresholds were measured for 15 HI listeners and 10 listeners with normal hearing (NH). The stimuli were octave-band-wide rustle stimuli centered at 500 Hz or 4 kHz, which were presented at 20 dB or 40 dB sensation level. Broadband rustle stimuli and 500-Hz pure-tone stimuli were also tested. Overall, hearing loss had a detrimental effect on ITD discrimination. For the majority of HI listeners, their ITD deficit relative to the NH group was equivalent at low and high frequencies. For a handful of HI listeners, the deficit was strongly frequency-dependent. The results provide new data to inform binaural models that incorporate effects of hearing impairment.
The presence of congenital permanent childhood hearing loss (PCHL) reduces auditory access to
spectral and temporal cues in the speech signal, thereby influencing the development of
auditory processing and language abilities in children. Despite early detection and intervention,
weaknesses in listening in noise and language development in children with PCHL have been
documented. Even though concurrent relationships between these abilities have been
examined, there has been little research on how the ability to recognize speech in noise
develops from preschool to school age, and on the direction of the relationship between
speech recognition and language abilities. Increased knowledge about development of these
abilities and their potentially predictive relationship has important implications for theoretical
understanding of the mechanisms that underlie development after treatment of sensory
deprivation (in this case, providing cochlear implants to children with profound PCHL); and
clinical implications to guide management of children for improving outcomes. In this paper,
the influence of speech recognition at age 5 years on language ability at age 9 years will be
examined using cross-lagged correlation analyses of data from a group of children with
profound hearing loss who were followed as part of the Longitudinal Outcomes of Children with
Hearing Impairment (LOCHI) study.
Much interest surrounds the possible contribution of synaptopathy and/or neuropathy
(SNpathy) to hearing difficulties that occur despite normal audiometric thresholds. We made
extensive measurements in two groups of listeners with near-normal audiograms who were
expected to differ greatly in the likelihood of SNpathy: 19 Y(oung) adults aged 18-25 with
limited noise exposure and 23 M(iddle-aged) adults aged 44-61 with significant noise
exposure. Speech reception thresholds (SRTs) were measured binaurally in speechspectrum-
shaped noise for two tasks (recognition of complex sentences and consonant
identification in VCVs). To assess the use of temporal fine structure (TFS), target and
masker were presented either diotically (S0N0) or with the masker in phase at the two ears
and the target out of phase (SπN0). Stimuli were presented at both a high and a low level
(40 and 80 dB SPL) to assess claims that deficits due to SNpathy might be more prominent
at high levels. Although the relationships among various predictors and the outcome were
fairly complex, generally speaking: 1) The mean performance of the Y group was always
better than that of the M group, with differences ranging from 0.3 to 3.5 dB across the 8
conditions. 2) SRTs worsened across age in the M group for some conditions, meaning that
the younger M listeners could be performing similarly to the Y group. 3) Greater deficits
were found for SπN0 than for S0N0 conditions, implicating some deficits in the M group for
processing TFS. 4) The effect of level was small, especially when comparing the two groups.
5) Group differences in SRTs were not related to a small difference in audiometric
thresholds at frequencies ≤ 4 kHz. It therefore seems possible that SNpathy could be a
factor in these relatively small group differences, but that these effects differ little across
Understanding speech in noisy environments is important in everyday life and requires effortful listening. Older adults’ listening effort may be especially affected by noise, associated with a diminished ability to effectively track a target speaker among background speakers across longer periods of time. To further understand how different listening situations influence older adults’ ability to sustain effort to meaningful speech, the present study used 60-second audiobook segments. This experiment investigated the effect of 1) different signal-to-noise ratios (SNRs: quiet, 0dB, or -6dB) and 2) number of audiobook repetitions (three times each in the two noisy conditions) on listening effort, for 13 older adults with clinically normal hearing. We hypothesized that poorer SNRs would result in increased listening effort and that increased number of repetitions would result in decreased listening effort. Participants heard the audiobook segments while pupillary measures were recorded, with larger pupil dilations indicating greater listening effort. Generalized additive mixed model (GAMM) results revealed that these older listeners showed evidence of increased listening effort in the noisy conditions and decreased listening effort between the first and second presentation. The temporal precision of pupillometry measures also indexed time-specific changes in listening effort, with increased effort at the beginning and end of the audiobook segments. Furthermore, listening effort varied nonlinearly with listeners’ subjective intelligibility ratings of each SNR block. In comparison to the quiet condition, listening effort in the noisy conditions was associated with higher self-reported intelligibility ratings, suggesting that individuals only engaged in listening to speech that was at least moderately intelligible. Taken together, our findings illustrate how SNR and repetition influence listening effort when listening to sustained, continuous speech and how effort can deviate from intelligibility in some listening conditions. Thus, this study aims to provide an ecologically valid account of older adults’ sustained listening effort.
Compared to young adults, older adults often have increased difficulty comprehending
speech, especially in challenging acoustic environments. However, previous research has
surprisingly found that their cortical responses to speech demonstrate more robust tracking
of the acoustic speech envelope than those of younger adults, even though the opposite
result holds for subcortical responses to speech. Here we have analyzed
magnetoencephalography responses to continuous narrative speech in older and younger
listeners with clinically normal hearing, acquired in two separate experiments. Responses to
clean speech and to speech from two simultaneous talkers were used to distinguish between
bottom-up and task-related brain activity. We show multiple lines of evidence that older
adults show exaggerated cortical responses compared to younger, at several distinct cortical
processing stages. Exaggerated responses at early latencies are consistent with
excitation/inhibition imbalance seen in animal models, whereas exaggerated responses at
greater latencies (which are also strongly dependent on selective attention) are consistent
with the recruitment of additional neural resources in order to aid speech comprehension.
Exaggerated responses are only seen for cortical processing of slow speech features (≲ 10
Hz), however, and not at faster rates associated with pitch tracking (≳ 80 Hz). Additional
insight into the cortical processing of continuous speech is gained from the analysis of
sustained pupillometric measures and non-phase-locked alpha-band neural activity,
simultaneously obtained during the continuous speech listening.
Find more details by clicking on the ‘room’ tabs above.
The concept of a modulation filterbank has been shown to account well for psychophysical data from experiments assessing temporal envelope processing acuity in young normal-hearing (NH) listeners. Recent studies using functional imaging and physiological measurements observed a loss of modulation tuning in older listeners and acoustically-traumatized animals, suggesting that modulation frequency selectivity may be adversely affected by ageing or hearing impairment. However, behavioural evidence of reduced modulation frequency selectivity in older and/or hearing-impaired (HI) listeners has not yet been provided. The present study investigated modulation frequency selectivity in older NH and HI listeners, as compared to young NH listeners, using psychophysical paradigms. Data were collected in conditions of amplitude modulation (AM) detection, AM frequencydiscrimination, and modulation masking. All conditions used sinusoidal modulations applied on a sinusoidal carrier, with target modulation rates of 4, 16, 64, and 128 Hz. Masked modulation thresholds were obtained for fixed-bandwidth noise modulation maskers (bandwidth corresponding to 1⁄2 octave when on-frequency) centered at frequencies ranging from -5 to 2 octaves relative to the target modulation frequency. The results suggested a reduction in modulation frequency selectivity at all target modulation frequencies in older NH listeners as compared to the young NH group, particularly at the target modulation frequency of 4 Hz. Preliminary data indicate that modulation frequency selectivity is predominantly affected at low modulation rates in HI listeners. To quantify modulation frequency selectivity, the envelope power spectrum model of masking (EPSM) was used to derive modulation filters that account for the masking data. The differences in modulation filter shape and selectivity across listener groups are discussed and analyzed in connection to the AM detection and AM frequency discrimination data. A loss of modulation frequency selectivity, as observed in the present study, might have detrimental effects on higher-level tasks, such as speech intelligibility or stream segregation.
It is increasingly clear that age-related hearing impairment is a dysfunction of the entire
auditory system, from periphery to cortex. Poorer auditory peripheral function is associated
with a loss of neural inhibition along the auditory pathway, including auditory cortex, that
renders neurons hyperactive and hyperresponsive to sounds. This auditory-system
hyperactivity may impair speech intelligibility when background sound is present. Neural
hyperactivity has been extensively studied using electrophysiological recordings in nonhuman
mammals but is less explored in humans. In this presentation, we will describe work
on hyperactivity associated with aging and hearing loss in humans. Specifically, we will
describe how neural synchronization to low-frequency amplitude modulations in sounds
differs between younger and older adults, likely as a result of hyperexcitability, and provide
behavioral data that test predictions for speech-in-noise intelligibility derived from this
electrophysiological work. Neural synchronization in auditory cortex is enhanced in older
compared to younger adults. Further, the propensity of neural activity to synchronize with
different amplitude-modulation shapes in sounds changes with age: auditory cortex of older
adults is more sensitive to damped (sharp attack) compared to ramped (gradual attack)
envelope shapes, whereas younger adults show the opposite pattern. Our behavioral data,
in contrast, reveal better speech intelligibility when background noise is modulated with
damped compared to ramped envelope shapes in both age groups. We also present recent
work demonstrating that the way amplitude-modulated background maskers affect speech
intelligibility in older compared to younger adults critically depends on the naturalness of
speech (disconnected sentences vs. engaging stories). We will wrap up this presentation by
briefly talking about open questions and challenges related to the study of auditory-system
hyperactivity in humans.
Everyday communication depends on an interplay between coding the auditory and visual
signals reaching the ears and eyes and modulating the information contained in these
signals through cortical networks for attention, working memory, and language processing.
It is well established that aging affects both sensory coding and cognitive processes
However, teasing apart whether communication problems are due to sensory versus
cognitive issues can be difficult. As an example, many middle-aged listeners have problems
understanding speech in the presence of competing sound sources, even when they have no
clinically quantifiable sensory deficits. Moreover, it is surprisingly difficult to identify simple
psychophysical tasks that relate to difficulties understanding speech amidst competing
sounds in middle-aged listeners without identifiable hearing loss. Together, this pattern of
abilities seems to suggest that cognitive, rather than sensory, deficits are the root cause.
We argue, however, that subtle sensory deficits (such as cochlear synaptopathy) may be to
blame. Specifically, in naturalistic settings, sensory coding can impact how quickly a listener
can focus on one sound source, extract information from that source, store meaning in
memory, and switch attentional focus— all processes that are not exercised by simpler
tasks. Thus, performance on “cognitive” tasks depends directly on sensory fidelity in
complex, but not simple listening scenarios, a realization that has important implications for
diagnosis of communication difficulties.
Auditory selective attention is a crucial mechanism for understanding speech in everyday environments. Top-down selective attention allows expectations to enhance the neural representation of sounds collected by the auditory sensory system. As most cochlear implant (CI) users struggle to recognize speech in noise, it is imperative to understand if CI users exhibit auditory selective attention activities that involve modulation of neural responses to target speech and if such attentional ability predicts their speech-in-noise performance. Our experiment is designed to assess the strength of attentional modulation within the human auditory pathway. Both participants with normal hearing (NH) and with CI were given a pre-stimulus visual cue that directed their attention to either of two sequences in stationary background noise and asked to select a deviant syllable in the target stream. Due to the misaligned timings of the syllables in each stream, we are able to examine the event-related potential (ERP), a proxy for attention modulation, in response to each syllable provided in the stream. We hypothesized that the amplitude of ERPs would be greater when the syllable is attended if either group is capable of employing auditory selective attention and that the difference of ERP amplitude between attended and unattended trials predicts the performance in a speech-in-noise task. Our analysis showed that the amplitude of ERPs for the attended syllable was greater than that for the unattended syllable with the CI subjects, exhibiting that attention modulates CI users’ cortical responses to sounds. Moreover, the strength of attentional modulation showed a significant correlation with the same CI users’ speech-in-noise performance. The difference between ERP amplitudes for attended and unattended syllables existed but was weaker for NH subjects. These results show that the attentional modulation of cortical auditory evoked responses provides a valuable neural marker for predicting CI users’ success in real-world communications.
Age-related hearing loss in older adults is often perceived as being an unfortunate but
relatively inconsequential part of aging. However, the broader implications of hearing
loss for the health and functioning of older adults are now beginning to surface in
epidemiologic studies. This lecture will discuss recent epidemiologic research
demonstrating that hearing loss is independently associated with accelerated cognitive
decline, incident dementia, and brain aging. Mechanisms through which hearing loss
may be causally linked with cognitive decline and dementia will be discussed as well as
gaps in our current scientific knowledge. Current studies investigating the impact of
hearing rehabilitative interventions on reducing cognitive decline and the risk of
dementia in older adults will be explained and discussed.
1) To describe the mechanisms through which hearing loss may be related to risk of
cognitive decline and dementia in older adults
2) To discuss the epidemiological evidence demonstrating associations of hearing
loss and dementia
3) To explain gaps in our scientific knowledge of the relationship between hearing
loss and dementia
Find more details by clicking on the ‘room’ tabs above.
A listener’s ability to deal with challenging multi-talker situations hinges on her attention resources. While the neural implementation of target enhancement is comparably well understood, processes that enable distractor suppression are less clear. Typically, distractor suppression is quantified by the difference of the behavioural or neural response to distractors versus targets. However, such a difference can be driven by either target enhancement, distractor suppression, or a combination of the two. Here, we designed a continuous speech paradigm to differentiate target enhancement (enhanced tracking of target versus neutral speech) from active distractor suppression (suppressed tracking of distractor versus neutral speech). In an electroencephalography (EEG) study, participants (N = 19) had to detect short repeats in the to-be-attended speech stream and to ignore them in the two other speech streams, while listening also to the content of the to-be-attended audio stream. The ignored speech stream was task-relevant (to-be-attended) in the previous trial and was task-irrelevant in the present trial. The neutral speech stream was always task-irrelevant. We used phase-locking of the EEG signal to speech envelopes to investigate neural tracking via the temporal response function of the brain. Behavioural detection of repeats indicated the suitability of the paradigm to separate processes of attending and ignoring. Sensitivity of behavioral responses according to Signal Detection Theory revealed that the internal separation for attended versus neutral speech was larger than for attended versus ignored speech. Neurally, the attended stream showed a significantly enhanced tracking response compared to neutral and ignored speech. Unexpectedly, neural tracking did not reveal sizeable differences for neutral versus ignored speech. In sum, the present results show that the cognitive system processes to-be-ignored speech distractors different from neutral speech. However, this is not accompanied by active distractor suppression in the neural speech tracking response.
Mature brains are able to adapt to acquired sensory dysfunction. In case of hearing loss and
oral communication difficulties, postlingual deaf adults follow two main strategies. One relies
on a left-lateralized physiological, but relatively slow analytical pathway maintained by
efficient lipreading skills. The second strategy consists in engaging the right hemisphere in
an accelerated, non speech-dedicated network. This reorganization is efficient and rapid
thanks to direct interaction with Broca’s area bypassing regular phonological steps. It relies
on accelerated reading abilities. The first strategy will be beneficial in case of hearing
rehabilitation (cochlear implantation), while the second option relevant during the deprived
period will turn maladaptive. This active plasticity was evidenced in adults aged less than 65
years. When cochlear implantees are more than 65 years, they significantly perform less 2
years after surgery than a control sample aged 17-40 years. The reasons are still
hypothetical. Among them, the relation between severe hearing loss in aging brains and
dementia, and in particular Alzheimer’s disease, is questioned. Moreover, modification of
sensitivity to salient events presented at 40 Hz (i.e. belonging to the roughness range)
appears a promising potential biomarker of Alzheimer’s disease. Thus combining
audiological and neurophysiological screening in the presymptomatic phase of Alzheimer’s
disease may help preventing cognitive decline process by personalized interventions.
Our work focuses on patients with bilateral deafness who are eligible to receive bilateral
cochlear implants (BiCIs), and patients with single-sided deafness who receive a cochlear
implant (SSD-CI) in the deaf ear. In both the BiCI and SSD-CI populations there is a
potential benefit from the integration of inputs arriving from both ears. Benefits include
improved ability to localize sounds and to segregate speech from background noise,
compared with unilateral listening. However, patients typically perform worse than normal
hearing listeners. We use several approaches to understand mechanisms driving gaps in
performance. We assess their ability to process auditory cues that are most essential for
spatial hearing, and the role of age and auditory experience. We also use research
processors to test novel stimulation paradigms designed to restore binaural sensitivity and
speech understanding in noise. Our studies provide evidence for the role of auditory
plasticity in driving binaural hearing. In addition, patients report that bilateral hearing
reduces their cognitive load and fatigue, but few studies have addressed this issue.
Pupillometry studies and also functional near infrared spectroscopy might be used as
objective tools that can provide insight into the impact of integrating inputs from two ears,
whereby in some instances improved performance with two ears can be “costly” in the
listening effort domain.
Interfering speech has rapid spectrotemporal fluctuations that established noise reduction algorithms have difficulty suppressing without a concomitant loss, or distortion, of binaural cues for spatial hearing. An ongoing project at the University of Oldenburg (UOL) involves development and evaluation of prototype Short-Time Target Cancellation (STTC) assistive listening devices that can enhance speech intelligibility of a target talker, and attenuate interfering talkers, while still preserving binaural cues for spatial hearing. The STTC processing computes a ratio mask (i.e., a time-varying spectral gain) that can be applied to the binaural signals at the Left and Right ears, thereby attenuating the interfering talkers. The STTC processing is causal and memoryless, with low requirements in terms of memory size and computational power, and is designed to run online in real-time without training or any a priori knowledge about the number or locations of interfering sound sources; only an assumed “look” direction is needed. The STTC processing can be used either to filter the binaural signals at the Left and Right ears or as a postfilter for adaptive beamforming. Where adaptive beamforming processing computes a complex-valued filter-vector, the STTC processing computes a real-valued time-varying spectral gain; the two approaches are compatible and our evaluation results indicate that their combination has an additive effect. Although the STTC processing, and adaptive beamforming, can be implemented with standard in-ear or behind-the-ear (BTE) hearing aid earpieces, better performance can be achieved via a small microphone array integrated into the frame of a pair of eyeglasses. Evaluation results, using simulations in virtual acoustic environments, indicate that these prototype STTC assistive listening devices can effect enhancement of a target talker, and attenuation of interfering talkers, in both anechoic space and reverberation.
Clinical hearing aid fittings depend primarily on the pure-tone audiogram, and audiologists
can choose among well-validated prescriptions to create an appropriate frequency-gain
response for each listener. However, there are no definitive guidelines to guide other
aspects of the hearing aid response such as compression speed and strength of digital noise
reduction. This presentation describes a series of studies in which a “cue profile” test
based on synthetic speech sounds is used to assess a hearing-impaired listener’s use of
specific speech cues. The resulting profile quantifies how well individual listeners can utilize
higher-precision spectro-temporal information, or whether they rely on lower-precision
temporal (envelope) cues to consonant identification. We review consequences of different
cue profiles for perception of speech with different types of signal processing (fast- vs slowacting WDRC and strategies designed to preserve the speech envelope). Aided speech
recognition is influenced by the amount of hearing loss, the cue profile, and extent of
envelope modification in the signal. Listeners with more temporal-reliant cue profiles tend to
have poorer aided speech recognition. Those listeners receive greatest benefit when
modulation cues are preserved, compared to listeners with more spectral-reliant cue
profiles. These data suggest that better understanding of how different amplification
strategies interact with the listener’s auditory abilities may allow clinicians to target
strategies of greatest benefit to an individual.
Precise methods in audiological diagnostics and understanding the impact of threshold and
suprathreshold deficits on the audiological outcomes are crucial for the individualized and
successful treatment with hearing devices. Model-based approaches may support the
selection and fitting of hearing devices. This study aims at speech recognition and loudness
perception of hearing-impaired listeners in acoustic conditions with increasing complexity,
i.e., in well-controlled laboratory conditions like stationary masker in comparison to
acoustically more complex and ecologically valid scenes like cafeteria ambience.
Furthermore, in order to better understand the contribution of the individual loss in
sensitivity and suprathreshold deficits to speech recognition subjective data are simulated
with the framework of auditory discrimination experiments (FADE) considering different
components of hearing impairment. For aided measurements, two prescription rules, puretone
threshold-based NAL-NL2 and individual loudness perception-based trueLOUDNESS,
are compared in terms of speech recognition and loudness perception. The outcomes of
speech recognition measurements with hearing-impaired listeners show significant
correlations of unaided speech recognition thresholds across the “simple” laboratory
masking conditions. The performance in these conditions, however, show no significant
correlation with performance in realistic cafeteria scenes. The benefit from hearing devices,
defined as the difference in speech recognition threshold between the unaided and aided
condition, differs across maskers and shows no correlation between laboratory and cafeteria
maskers. While NAL-NL2 and trueLOUDNESS result in a comparable benefit in terms of
speech recognition, the loudness perception is restored better with the trueLOUDNESS
prescription rule. The accuracy of FADE simulations in unaided and aided conditions is
highest when both components of hearing impairment (sensitivity loss and suprathreshold
deficits) are accounted for. In summary, a model-based interpretation with a distinction
between threshold and suprathreshold distortion component might not only be useful for
diagnostic purposes but also helps to predict the benefit from a hearing device in
acoustically challenging conditions.
With age, our hearing ability starts to decline; communicating in noisy environments becomes challenging, and hearing faint sounds difficult. Part of this decline stems from outer-hair-cell damage, and another factor relates to synaptic damage at the auditory-nerve, i.e., cochlear synaptopathy (CS). Despite the suspected high prevalence of CS among people with self-reported hearing difficulties but normal audiograms, or those with impaired audiograms, conventional hearing-aid algorithms do not specifically compensate for the functional deficits associated with CS. Here, we present and evaluate a number of hearing restoration algorithms that maximally restore auditory-nerve coding in CS-affected peripheries. Using a biophysical model of the auditory periphery, we designed real-time signal-processing algorithms to three different CS types that operate on the time-domain signal. The algorithms preserve the stimulus envelope peaks but modify sound onsets and offsets to increase the resting periods between stimulation. We evaluated our developed algorithms in subjects with and without suspected age-related CS (N=30) to test whether they enhanced envelope-following-responses (EFRs), amplitude-modulation (AM) detection sensitivity and speech intelligibility. Volunteers with normal-hearing (NH) audiograms and ages between 18-25 (yNH) or 45-65 (oNH) y/o participated in our study and the difference between processed and unprocessed stimuli was assessed. Our data show that EFRs and perceptual AM sensitivity were enhanced in both yNH and oNH listeners when using our CS-compensation algorithms. Speech recognition in the Matrix test showed a small improvement that was not consistent across participants, with the yNH group and those with high AM detection sensitivity benefiting the most from the processed speech, suggesting that different approaches might be necessary when applying the algorithms to speech. This new type of sound processing may extend the application range of current hearing-aids and improve temporal envelope processing while leaving sound amplification unaffected.
In current clinical practice, hearing aids are typically fitted based on audiometric thresholds
only, even though research suggests that suprathreshold factors play a role for aided
outcome, too. In 2016, the Danish ‘Better hEAring Rehabilitation’ (BEAR) project was
initiated with the overall goal of improving hearing-aid rehabilitation. A focus area in that
project has been the development of a method for classifying hearing-impaired listeners into
four profiles capturing distinct differences in terms of audiometric hearing loss and
suprathreshold hearing abilities. Additional focus areas have been the development of an
auditory profile-based fitting strategy and the investigation of aided speech-in-noise
outcome in those profiles. In this contribution, we will provide an overview of these research
activities. Emphasis will be placed on insights gained with respect to the characterization of
individual hearing losses and the translation of the resultant findings into solutions that are
implementable in clinically available hearing devices.