Representations of fricatives in subcortical model responses: Comparisons with human consonant perception.

Fricatives are obstruent sound contrasts made by airflow constrictions in the vocal tract that produce turbulence across the constriction or at a site downstream from the constriction. Fricatives exhibit significant intra/intersubject and contextual variability. Yet, fricatives are perceived with high accuracy. The current study investigated modeled neural responses to fricatives in the auditory nerve (AN) and inferior colliculus (IC) with the hypothesis that response profiles across populations of neurons provide robust correlates to consonant perception. Stimuli were 270 intervocalic fricatives (10 speakers × 9 fricatives × 3 utterances). Computational model response profiles had characteristic frequencies that were log-spaced from 125 Hz to 8 or 20 kHz to explore the impact of high-frequency responses. Confusion matrices generated by k-nearest-neighbor subspace classifiers were based on the profiles of average rates across characteristic frequencies as feature vectors. Model confusion matrices were compared with published behavioral data. The modeled AN and IC neural responses provided better predictions of behavioral accuracy than the stimulus spectra, and IC showed better accuracy than AN. Behavioral fricative accuracy was explained by modeled neural response profiles, whereas confusions were only partially explained. Extended frequencies improved accuracy based on the model IC, corroborating the importance of extended high frequencies in speech perception.

Effects of sensorineural hearing loss on formant-frequency discrimination: Measurements and models.

This study concerns the effect of hearing loss on discrimination of formant frequencies in vowels. In the response of the healthy ear to a harmonic sound, auditory-nerve (AN) rate functions fluctuate at the fundamental frequency, F0. Responses of inner-hair-cells (IHCs) tuned near spectral peaks are captured (or dominated) by a single harmonic, resulting in lower fluctuation depths than responses of IHCs tuned between spectral peaks. Therefore, the depth of neural fluctuations (NFs) varies along the tonotopic axis and encodes spectral peaks, including formant frequencies of vowels. This NF code is robust across a wide range of sound levels and in background noise. The NF profile is converted into a rate-place representation in the auditory midbrain, wherein neurons are sensitive to low-frequency fluctuations. The NF code is vulnerable to sensorineural hearing loss (SNHL) because capture depends upon saturation of IHCs, and thus the interaction of cochlear gain with IHC transduction. In this study, formant-frequency discrimination limens (DLFFs) were estimated for listeners with normal hearing or mild to moderate SNHL. The F0 was fixed at 100 Hz, and formant peaks were either aligned with harmonic frequencies or placed between harmonics. Formant peak frequencies were 600 and 2000 Hz, in the range of first and second formants of several vowels. The difficulty of the task was varied by changing formant bandwidth to modulate the contrast in the NF profile. Results were compared to predictions from model auditory-nerve and inferior colliculus (IC) neurons, with listeners' audiograms used to individualize the AN model. Correlations between DLFFs, audiometric thresholds near the formant frequencies, age, and scores on the Quick speech-in-noise test are reported. SNHL had a strong effect on DLFF for the second formant frequency (F2), but relatively small effect on DLFF for the first formant (F1). The IC model appropriately predicted substantial threshold elevations for changes in F2 as a function of SNHL and little effect of SNHL on thresholds for changes in F1.

Sensorineural Hearing Loss Diminishes Use of Temporal Envelope Cues: Evidence From Roving-Level Tone-in-Noise Detection.

OBJECTIVES: The objective of our study is to understand how listeners with and without sensorineural hearing loss (SNHL) use energy and temporal envelope cues to detect tones in noise. Previous studies of low-frequency tone-in-noise detection have shown that when energy cues are made less reliable using a roving-level paradigm, thresholds of listeners with normal hearing (NH) are only slightly increased. This result is consistent with studies demonstrating the importance of temporal envelope cues for masked detection. In contrast, roving-level detection thresholds are more elevated in listeners with SNHL at the test frequency, suggesting stronger weighting of energy cues. The present study extended these tests to a wide range of frequencies and stimulus levels. The authors hypothesized that individual listeners with SNHL use energy and temporal envelope cues differently for masked detection at different frequencies and levels, depending on the degree of hearing loss. DESIGN: Twelve listeners with mild to moderate SNHL and 12 NH listeners participated. Tone-in-noise detection thresholds at 0.5, 1, 2, and 4 kHz in 1/3 octave bands of simultaneously gated Gaussian noise were obtained using a novel, two-part tracking paradigm. A track refers to the sequence of trials in an adaptive test procedure; the signal to noise ratio was the tracked variable. Each part of the track consisted of a two-alternative, two-interval, forced-choice procedure. The initial portion of the track estimated detection threshold using a fixed masker level. When the track continued, stimulus levels were randomly varied over a 20-dB rove range (±10 dB with respect to mean masker level), and a second threshold was estimated. Rove effect (RE) was defined as the difference between thresholds for the fixed- and roving-level tests. The size of the RE indicated how strongly a listener weighted energy-based cues for masked detection. Participants were tested at one to three masker levels per frequency, depending on audibility. RESULTS: Across all stimulus frequencies and levels, NH listeners had small REs (≈1 dB), whereas listeners with SNHL typically had larger REs. Some listeners with SNHL had larger REs at higher frequencies, where pure-tone audiometric thresholds were typically elevated. RE did not vary significantly with masker level for either group. Increased RE for the SNHL group was consistent with simulations in which energy cues were more heavily weighted than envelope cues. CONCLUSIONS: Tone-in-noise detection thresholds in NH listeners were typically elevated only slightly by the roving-level paradigm at any frequency or level tested, consistent with the primary use of level-independent cues, such as temporal envelope cues that are conveyed by fluctuations in neural responses. In comparison, thresholds of listeners with SNHL were more affected by the roving-level paradigm, suggesting stronger weighting of energy cues. For listeners with SNHL, the largest RE was observed at 4000 Hz, for which pure-tone audiometric thresholds were most elevated. Specifically, RE size at 4000 Hz was significantly correlated with higher pure-tone audiometric thresholds at the same frequency, after controlling for the effect of age. Future studies will explore strategies for restoring or enhancing neural fluctuation cues that may lead to improved hearing in noise for listeners with SNHL.

Detection thresholds for amplitude modulations of tones in budgerigar, rabbit, and human.

Envelope fluctuations of complex sounds carry information that is -essential for many types of discrimination and for detection in noise. To study the neural representation of envelope information and mechanisms for processing of this temporal aspect of sounds, it is useful to identify an animal model that can -sensitively detect amplitude modulations (AM). Low modulation frequencies, which dominate speech sounds, are of particular interest. Yet, most animal -models studied previously are relatively insensitive to AM at low modulation -frequencies. Rabbits have high thresholds for low-frequency modulations, -especially for tone carriers. Rhesus macaques are less sensitive than humans to low-frequency -modulations of wideband noise (O'Conner et al. Hear Res 277, 37-43, 2011). Rats and -chinchilla also have higher thresholds than humans for amplitude -modulations of noise (Kelly et al. J Comp Psychol 120, 98-105, 2006; Henderson et al. J Acoust Soc Am 75, -1177-1183, 1984). In contrast, the budgerigar has thresholds for AM detection of wideband noise similar to those of human listeners at low -modulation frequencies (Dooling and Searcy. Percept Psychophys 46, 65-71, 1981). A -one-interval, two-alternative operant conditioning procedure was used to estimate AM -detection thresholds for 4-kHz tone carriers at low modulation -frequencies (4-256 Hz). Budgerigar thresholds are comparable to those of human subjects in a comparable task. Implications of these comparative results for temporal coding of complex sounds are discussed. Comparative results for masked AM detection are also presented.

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.