An intrusive method for estimating speech intelligibility from noisy and distorted signals.

An objective metric that predicts speech intelligibility under different types of noise and distortion would be desirable in voice communication. To date, the majority of studies concerning speech intelligibility metrics have focused on predicting the effects of individual noise or distortion mechanisms. This study proposes an objective metric, the spectrogram orthogonal polynomial measure (SOPM), that attempts to predict speech intelligibility for people with normal hearing under adverse conditions. The SOPM metric is developed by extracting features from the spectrogram using Krawtchouk moments. The metric's performance is evaluated for several types of noise (steady-state and fluctuating noise), distortions (peak clipping, center clipping, and phase jitters), ideal time-frequency segregation, and reverberation conditions both in quiet and noisy environments. High correlation (0.97-0.996) is achieved with the proposed metric when evaluated with subjective scores by normal-hearing subjects under various conditions.

Suboptimal use of neural information in a mammalian auditory system.

Establishing neural determinants of psychophysical performance requires both behavioral and neurophysiological metrics amenable to correlative analyses. It is often assumed that organisms use neural information optimally, such that any information available in a neural code that could improve behavioral performance is used. Studies have shown that detection of amplitude-modulated (AM) auditory tones by humans is correlated to neural synchrony thresholds, as recorded in rabbit at the level of the inferior colliculus, the first level of the ascending auditory pathway where neurons are tuned to AM stimuli. Behavioral thresholds in rabbit, however, are ∼10 dB higher (i.e., 3 times less sensitive) than in humans, and are better correlated to rate-based than temporal coding schemes in the auditory midbrain. The behavioral and physiological results shown here illustrate an unexpected, suboptimal utilization of available neural information that could provide new insights into the mechanisms that link neuronal function to behavior.

Updated parameters and expanded simulation options for a model of the auditory periphery.

A phenomenological model of the auditory periphery in cats was previously developed by Zilany and colleagues [J. Acoust. Soc. Am. 126, 2390-2412 (2009)] to examine the detailed transformation of acoustic signals into the auditory-nerve representation. In this paper, a few issues arising from the responses of the previous version have been addressed. The parameters of the synapse model have been readjusted to better simulate reported physiological discharge rates at saturation for higher characteristic frequencies [Liberman, J. Acoust. Soc. Am. 63, 442-455 (1978)]. This modification also corrects the responses of higher-characteristic frequency (CF) model fibers to low-frequency tones that were erroneously much higher than the responses of low-CF model fibers in the previous version. In addition, an analytical method has been implemented to compute the mean discharge rate and variance from the model's synapse output that takes into account the effects of absolute refractoriness.

Power-law dynamics in an auditory-nerve model can account for neural adaptation to sound-level statistics.

Neurons in the auditory system respond to recent stimulus-level history by adapting their response functions according to the statistics of the stimulus, partially alleviating the so-called "dynamic-range problem." However, the mechanism and source of this adaptation along the auditory pathway remain unknown. Inclusion of power-law dynamics in a phenomenological model of the inner hair cell (IHC)-auditory nerve (AN) synapse successfully explained neural adaptation to sound-level statistics, including the time course of adaptation of the mean firing rate and changes in the dynamic range observed in AN responses. A direct comparison between model responses to a dynamic stimulus and to an "inversely gated" static background suggested that AN dynamic-range adaptation largely results from the adaptation produced by the response history. These results support the hypothesis that the potential mechanism underlying the dynamic-range adaptation observed at the level of the auditory nerve is located peripheral to the spike generation mechanism and central to the IHC receptor potential.

A phenomenological model of the synapse between the inner hair cell and auditory nerve: long-term adaptation with power-law dynamics.

There is growing evidence that the dynamics of biological systems that appear to be exponential over short time courses are in some cases better described over the long-term by power-law dynamics. A model of rate adaptation at the synapse between inner hair cells and auditory-nerve (AN) fibers that includes both exponential and power-law dynamics is presented here. Exponentially adapting components with rapid and short-term time constants, which are mainly responsible for shaping onset responses, are followed by two parallel paths with power-law adaptation that provide slowly and rapidly adapting responses. The slowly adapting power-law component significantly improves predictions of the recovery of the AN response after stimulus offset. The faster power-law adaptation is necessary to account for the "additivity" of rate in response to stimuli with amplitude increments. The proposed model is capable of accurately predicting several sets of AN data, including amplitude-modulation transfer functions, long-term adaptation, forward masking, and adaptation to increments and decrements in the amplitude of an ongoing stimulus.

A phenomenological model of the synapse between the inner hair cell and auditory nerve: Implications of limited neurotransmitter release sites.

Peterson and Heil [Hear. Res., In Press] have argued that the statistics of spontaneous spiking in auditory nerve fibers (ANFs) can be best explained by a model with a limited number of synaptic vesicle docking (release) sites (∼4) and a relatively-long average redocking time (∼16-17 ms) for each of the sites. In this paper we demonstrate how their model can be: i) generalized to also describe sound-driven ANF responses and ii) incorporated into a well-established and widely-used model of the entire auditory periphery [Zilany et al., J. Acoust. Soc. Am. 135, 283-286, 2014]. The responses of the new model exhibit substantial improvement in several measures of ANF spiking statistics, and predicted physiological forward-masking and rate-level functions from the new model structure are shown to also better match published physiological data.

Representation of the vowel /epsilon/ in normal and impaired auditory nerve fibers: model predictions of responses in cats.

The temporal response of auditory-nerve (AN) fibers to a steady-state vowel is investigated using a computational auditory-periphery model. The model predictions are validated against a wide range of physiological data for both normal and impaired fibers in cats. The model incorporates two parallel filter paths, component 1 (C1) and component 2 (C2), which correspond to the active and passive modes of basilar membrane vibration, respectively, in the cochlea. The outputs of the two filters are subsequently transduced by two separate functions, added together, and then low-pass filtered by the inner hair cell (IHC) membrane, which is followed by the IHC-AN synapse and discharge generator. The C1 response dominates at low and moderate levels and is responsible for synchrony capture and multiformant responses seen in the vowel responses. The C2 response dominates at high levels and contributes to the loss of synchrony capture observed in normal and impaired fibers. The interaction between C1 and C2 responses explains the behavior of AN fibers in the transition region, which is characterized by two important observations in the vowel responses: First, all components of the vowel undergo the C1/C2 transition simultaneously, and second, the responses to the nonformant components of the vowel become substantial.

Modeling Responses in the Superior Paraolivary Nucleus: Implications for Forward Masking in the Inferior Colliculus.

A phenomenological model of the responses of neurons in the superior paraolivary nucleus (SPON) of the rodent is presented in this study. Pure tones at the characteristic frequency (CF) and broadband noise stimuli evoke offset-type responses in these neurons. SPON neurons also phase-lock to the envelope of sinusoidally amplitude-modulated (SAM) stimuli for a range of modulation frequencies. Model SPON neuron received inhibitory input that was relayed by the ipsilateral medial nucleus of the trapezoid body from the contralateral model ventral cochlear nucleus neuron. The SPON model response was simulated by detecting the slope of its inhibitory postsynaptic potential. Responses of the proposed model to pure tones at CF and broadband noise were offset-type independent of the duration of the input stimulus. SPON model responses were also synchronized to the envelope of SAM stimuli with precise timing for a range of modulation frequencies. Modulation transfer functions (MTFs) obtained from the model response to SAM stimuli resemble the physiological MTFs. The output of the proposed SPON model provides an input for models of physiological responses at higher levels of the ascending auditory pathway and can also be utilized to infer possible mechanisms underlying gap detection and duration encoding as well as forward masking at the level of the auditory midbrain.

Reference-Free Assessment of Speech Intelligibility Using Bispectrum of an Auditory Neurogram.

Sensorineural hearing loss occurs due to damage to the inner and outer hair cells of the peripheral auditory system. Hearing loss can cause decreases in audibility, dynamic range, frequency and temporal resolution of the auditory system, and all of these effects are known to affect speech intelligibility. In this study, a new reference-free speech intelligibility metric is proposed using 2-D neurograms constructed from the output of a computational model of the auditory periphery. The responses of the auditory-nerve fibers with a wide range of characteristic frequencies were simulated to construct neurograms. The features of the neurograms were extracted using third-order statistics referred to as bispectrum. The phase coupling of neurogram bispectrum provides a unique insight for the presence (or deficit) of supra-threshold nonlinearities beyond audibility for listeners with normal hearing (or hearing loss). The speech intelligibility scores predicted by the proposed method were compared to the behavioral scores for listeners with normal hearing and hearing loss both in quiet and under noisy background conditions. The results were also compared to the performance of some existing methods. The predicted results showed a good fit with a small error suggesting that the subjective scores can be estimated reliably using the proposed neural-response-based metric. The proposed metric also had a wide dynamic range, and the predicted scores were well-separated as a function of hearing loss. The proposed metric successfully captures the effects of hearing loss and supra-threshold nonlinearities on speech intelligibility. This metric could be applied to evaluate the performance of various speech-processing algorithms designed for hearing aids and cochlear implants.

Semi-supervised spike sorting using pattern matching and a scaled Mahalanobis distance metric.

Sorting action potentials (spikes) from tetrode recordings can be time consuming, labor intensive, and inconsistent, depending on the methods used and the experience of the operator. The techniques presented here were designed to address these issues. A feature related to the slope of the spike during repolarization is computed. A small subsample of the features obtained from the tetrode (ca. 10,000-20,000 events) is clustered using a modified version of k-means that uses Mahalanobis distance and a scaling factor related to the cluster size. The cluster-size-based scaling improves the clustering by increasing the separability of close clusters, especially when they are of disparate size. The full data set is then classified from the statistics of the clusters. The technique yields consistent results for a chosen number of clusters. A MATLAB implementation is able to classify more than 5000 spikes per second on a modern workstation.

Modeling auditory-nerve responses for high sound pressure levels in the normal and impaired auditory periphery.

This paper presents a computational model to simulate normal and impaired auditory-nerve (AN) fiber responses in cats. The model responses match physiological data over a wider dynamic range than previous auditory models. This is achieved by providing two modes of basilar membrane excitation to the inner hair cell (IHC) rather than one. The two modes are generated by two parallel filters, component 1 (C1) and component 2 (C2), and the outputs are subsequently transduced by two separate functions. The responses are then added and passed through the IHC low-pass filter followed by the IHC-AN synapse model and discharge generator. The C1 filter is a narrow-band, chirp filter with the gain and bandwidth controlled by a nonlinear feed-forward control path. This filter is responsible for low and moderate level responses. A linear, static, and broadly tuned C2 filter followed by a nonlinear, inverted and nonrectifying C2 transduction function is critical for producing transition region and high-level effects. Consistent with Kiang's two-factor cancellation hypothesis, the interaction between the two paths produces effects such as the C1/C2 transition and peak splitting in the period histogram. The model responses are consistent with a wide range of physiological data from both normal and impaired ears for stimuli presented at levels spanning the dynamic range of hearing.