2019 Seminars
February 2019
| HRC Seminar with Sebastien Santurette | February 1st |
Sebastien Santurette, Oticon
Title: N/A
Abstract: N/A
| HRC Seminar with Daniel Fogerty | February 15th |
Daniel Fogerty
Title: Toward an understanding of speech glimpsing
Abstract: In everyday listening environments speech is often heard in the presence of background sound sources. These environmental sounds often interact with the target speech through masking or interference, resulting in listeners experiencing greater difficulty understanding the target message. Addressing this problem of understanding speech in adverse listening conditions entails the ability to resolve partial information. One model for defining the availability of partial information involves defining “glimpses” of target speech at momentarily favorable signal-to-noise ratios. The contribution of these speech glimpses to intelligibility involves understanding properties of the preserved target speech, interactions with the competing background, and abilities of the listener. Findings suggest that access to slow amplitude modulation properties (under 8 Hz) is critical for speech recognition. Furthermore, interference of these speech modulations by background noise is detrimental to recognition. Acoustic analyses of speech glimpses that index the proportion and distribution of information can capture average performance for speech recognition in speech-modulated noise. However, significant individual variability is often observed, particularly among older adults. Initial work assessing multiple perceptual and cognitive abilities is underway to explain this individual variability in glimpsing abilities. Overall, these findings support temporal sampling models of speech perception and the crucial role of speech amplitude modulation necessary for maximal speech recognition in adverse listening conditions.
| HRC Seminar with Malcolm Slaney | February 20th |
Malcolm Slaney
Title: N/A
Abstract: N/A
| HRC Seminar with Rachel Theodore | February 15th |
Rachel Theodore
Title: Listener sensitivity to structured phonetic variation
Abstract: Memory for spoken language is not a veridical representation of experience. Instead, memory reflects integration across our interlocutors’ messages, resulting in robust memory for meaning with relatively poor memory for the specific form of the message. This is striking considering that in the process of mapping from speech to meaning, listeners show exquisite sensitivity to the acoustic-phonetic structure of speech. In this talk, I will review selected findings from work in our laboratory that examines listeners’ ability to dynamically adapt to structured phonetic variation, focusing on variation associated with talkers’ idiolects. These studies examine mechanisms that allow listeners to exploit structured variation for speech perception, voice recognition, and memory of spoken language. Collectively, the results (1) identify principles that govern how listeners modify the mapping to speech sounds to reflect cumulative experience with talkers’ phonetic input, (2) show that sensitivity to structured phonetic input facilitates identification of a talker’s voice in addition to the linguistic message, and (3) demonstrate that talker identity can provide structure for integration of spoken language in memory. These findings help explicate a theoretical framework that accounts for tension in a linguistic architecture that uses both abstract and instance-specific representational knowledge to guide spoken language processing.
March 2019
| HRC Seminar with Mathieu Lavandier | March 23rd |
Mathieu Lavandier
Title: Binaural modelling of speech intelligibility in noisy rooms
Abstract: A model will be presented that allows predicting binaural speech intelligibility in noise for normal-hearing (NH) and hearing-impaired (HI) listeners. It is based on a previous NH model by Lavandier and Culling (2010) that has been validated in various conditions, involving different levels of reverberation and spatial configurations for multiple stationary speech-shaped noise maskers.The new model inputs are the masker and target signals at the listener’s ears, as well as his/her audiogram that is used to implement an internal noise. The model computes per time frame and frequency band a binaural masking level difference (BMLD) when all signal levels are higher than the internal noise level, and the signal-to-noise ratio (SNR) at the better ear using the maximum between the masker and internal noise levels. The BMLD and better-ear SNR are integrated across time and frequency and summed to obtain a binaural ratio, the relative differences of which can be compared to speech reception threshold differences measured in listening tests. This model was validated using data involving NH and HI listeners. Stimuli were anechoic and played using headphones. The target, in front of the listener, was presented simultaneously with two speech-shaped noises or noise-vocoded speech either collocated with the target or artificially spatially separated (by presenting each masker on a different single ear). Different amplifications were applied to the stimuli, involving changes in presentation levels and audibility. The model was able to predict the speech reception thresholds, the spatial release from masking and the difference between HI and NH thresholds with the same accuracy as previous models limited to NH predictions.
| HRC Seminar with Huanping Dai | March 30th |
Huanping Dai, University of Arizona
Title: N/A
Abstract: N/A
April 2019
| HRC Seminar with Mounya Elhilali | April 5th |
Mounya Elhilali
Title: Statistical inference in auditory perception
Abstract: The brain’s ability to extract statistical regularities from sound is an important tool in auditory scene analysis, necessary for object recognition, structural learning (in speech or music), texture perception, and novel event detection. In this talk, we explore perceptual, neural and computational principles underlying the brain’s ability to collect complex and non-deterministic patterns in echoic memory. These processes underlie statistical inference in auditory perception and guide our ability to delineate salient changes in our acoustic environment. We will explore the acoustic dimensionality and statistical granularity of this inference space using complex soundscapes; and discuss theoretical interpretation of perceptual and neural findings in terms of coding mechanisms in the brain.
| HRC Seminar with Matthew Leonard | April 26th |
Matthew Leonard
Title: Dynamic Brain Networks for Speech Comprehension
Abstract: Speech is an acoustically and cognitively dynamic signal, unfolding rapidly and requiring listeners to integrate multiple sources of information in real-time for comprehension. Recent advances in human neuroscience, and particularly direct electrocorticography (ECoG), have revealed key neural representations underlying the coherent perception of speech. At a fundamental level, local neural populations in the superior temporal gyrus (STG) encode acoustic-phonetic features in speech. However, the representations of these features are strongly modulated by linguistic predictions, learned knowledge, and perceptual goals, suggesting that STG feature detectors are highly context-dependent. These results will be discussed in the context of a novel framework for understanding how temporal integration and binding across auditory and cognitive representations leads to core linguistic units like syllables and words.
September 2019
| HRC Seminar with Kelly Whiteford | September 13th |
Kelly Whiteford, University of Minnesota
Title: Perception of Frequency Modulation is Mediated by Cochlear Place Coding
Abstract: Humans have acute sensitivity to frequency modulation (FM) at carriers and rates important for speech and music (fc < ~4-5 kHz and fm < ~ 10 Hz). Such fine-grained sensitivity is believed to be related to the fidelity of precise stimulus-driven spike timing (time code). Alternatively, FM sensitivity could be explained by a unitary code for FM at all rates and carrier frequencies, based on cochlear conversion of FM to amplitude modulation (AM) (place code). Experiment 1 tested this hypothesis directly by assessing sensitivity to low-carrier (fc = 1 kHz) FM and AM at slow (fm = 1 Hz) and fast rates (fm = 20 Hz) in a large group of listeners varying in degree of sensorineural hearing loss, resulting in widely varying fidelity of place-based (tonotopic) coding. Results supported a unitary code for FM, as both slow and fast FM were related to estimates of place coding fidelity. Experiment 2 was designed to test if enhanced FM sensitivity for low carriers at slow rates could potentially be explained by central, rather than peripheral, limitations. We simulated FM by simultaneously presenting two AM envelopes out of phase with carriers spaced far apart on the tonotopic axis, thereby providing informative place but not timing cues. Sensitivity for simulated FM mirrored that found for traditional FM detection, with performance best at slow rates with low-frequency carriers. Results suggest that the deterioration in performance at high carrier frequencies may reflect central, rather than peripheral, limitations.
| HRC Seminar with Dan Sanes | September 27th |
Dan Sanes, New York University
Title: Identifying CNS contributions to hearing loss-induced perceptual deficits
Abstract: A heightened period of central nervous system (CNS) plasticity permits early sensory experience to influence functional maturation, thereby shaping adult perceptual skills. Support for this idea is based largely on the neural consequences of sensory deprivation, such as developmental hearing loss. However, hearing loss commonly involves permanent damage to the cochlea, making it difficult to attribute perceptual deficits to CNS dysfunction. By inducing a reversible form of hearing loss during a developmental critical period, we subsequently find that performance on an auditory psychometric task is degraded even though cochlear function is normal at the time of testing. Furthermore, the same manipulation leads to a permanent disruption of CNS synaptic inhibition. To test whether this CNS deficit is causally related to degraded perception, we have restored synaptic inhibition with a pharmacological treatment, and find that auditory perceptual skill is completely rescued. Although these results are encouraging, many children with mild-to-moderate hearing loss are not identified until later in development. Therefore, we have gone on to study transient hearing loss that occurs after the developmental critical period, and again find significant perceptual deficits despite a normally functioning cochlea. However, our preliminary results suggest that the CNS contribution is quite unique. Taken together, our results imply that specific CNS deficits contribute to hearing loss-induced perceptual deficits, and might be targeted for remediation.
October 2019
| HRC Seminar with Skyler Jennings | October 11th |
Skyler Jennings
Title: Does Olivocochlear Efferent Feedback Improve Perception of the Temporal Envelope?
Abstract: Adults with cochlear hearing loss struggle to understand speech in noisy backgrounds even after treatment with modern hearing devices, such as hearing aids and cochlear implants. This difficulty may stem from dysfunction of the medial olivocochlear (MOC) reflex, which is hypothesized to facilitate perceptual performance in background noise based on studies from laboratory animals. Perceptual performance is mediated, in part, by decoding the gross fluctuations in an acoustic temporal envelope. Our recent research shows that perception of the temporal envelope is improved by a priming (or “precursor”) stimulus in normal- hearing, but not hearing-impaired individuals. We show that manipulations of stimulus properties and the influence of hearing impairment on this “precursor effect” are consistent with feedback from the MOC reflex. We evaluate the neural coding of the temporal envelope using methods that emphasize cochlear or brainstem/cortical neural generators, and that control for potential middle ear muscle activity. This method shows promise as an assay of MOC reflex activity and as a vehicle to test whether the MOC reflex is responsible for improving perception of the temporal envelope.
| HRC Seminar with Christian Stilp | October 18th |
Christian Stilp, University of Louisville
Title: (Acoustic) Context is Everything in Speech Perception
Abstract: All perception takes place in context. Objects and events in the environment are never perceived in isolation, but relative to other surrounding stimuli. This is true across all sensory modalities, and particularly significant in speech perception. Acoustic characteristics of surrounding sounds have a powerful influence on perception of speech sounds. In this talk, I will review and extend what we know about two classic effects of surrounding spectral context on perception: spectral contrast effects and auditory enhancement effects. Speech sound categorization is exquisitely sensitive to contrast effects and enhancement effects, which I will show are related to each other at the individual differences level. I will also link these questions and draw connections with another widely-studied phenomenon in speech research, talker normalization. Similarities, differences, and interrelationships among these phenomena will be discussed.
November 2019
| HRC Seminar with Alex Francis | November 8th |
Alex Francis, Purdue University
Title: Noise, Listening Effort, Emotion, and Health
Abstract: Noise is a significant source of annoyance and distress and is increasingly recognized as a major public health issue. Workplace noise impairs cognitive performance and increases fatigue and susceptibility to chronic disease. Certain sounds may be particularly troublesome for individuals with tinnitus, hyperacusis, or misophonia. And people with hearing loss find listening in noise to be disproportionately difficult and effortful. In this talk I will present a preliminary overview of a broadly focused research program addressing multiple aspects of the question of why background noise can be so problematic for so many people. In particular, I will focus on the likely contribution of emotional response to the short- and long-term consequences of noise exposure. I hypothesize that noise becomes annoying or frustrating when irrelevant sounds interfere with task performance, e.g. through distraction and/or increased listening effort. This response may be stronger when the noise is not avoidable or controllable, and/or when exerted effort feels ineffective or unrewarding. Chronic annoyance, in turn, may induce physiological stress responses that damage long-term health. This implies that the short- and long-term effect of noise on job performance and health may vary depending on individual differences in noise tolerance, information processing (cognitive capacity), susceptibility to distraction (selective attention), and emotional responsivity (affective psychophysiology).
| HRC Seminar with Julius Golstein | November 22nd |
Julius Golstein
Title: Application of the Classical Phon-Loudness Model to Reveal Middle-Ear and Acoustic-Reflex Functions
Abstract: Equal Loudness-level Contours (Fletcher and Munson, 1933) measured by Lydolf and H. Moller (1997) show increases in ELC growth rates at each frequency with loudness above 60 phons and below 1 kHz, suggesting the influence of acoustic reflex attenuation. Classical psychophysical power-law models (Stevens, 1966) were fit to ELC data (20 – 80 phons) judged to be free of reflex attenuation. The higher level ELC data exceed extrapolated predictions from the low-level models. These ELC differences are modeled as reductions in middle ear input. The majority of these differences represent acoustic-reflex attenuation of middle-ear input, with log-linear growth rates <1>1 dB/dB are interpreted as middle-ear attenuation from unmonitored static eardrum pressure assumed to accompany low-frequency pressure-field stimuli. Both mechanisms add middle ear attenuation by increasing the stiffness of stapes motion, and can be modeled as mass-spring-dashpot systems (Kringlebotn and Gundersen, 1985; Kringlebotn, 2000). A model of reflex attenuation dynamics (Goldstein, 2014) based on the acoustic-reflex power-law behavior, represents the salient features of ipsilateral eardrum impedance modulation (A. Moller, 1962). Acoustic reflex properties from this study and Rabinowitz (1977) are compatible.
