Cochlear Filter Research Articles

Wave-based signal-to-noise enhancement and the shape of cochlear filters

Cochlear transfer functions measured at one location—a.k.a. “cochlear filters”—are characterized by a steep roll-off as the frequency increases above the local best (responding) frequency (BF). The functional role of the sharp “cut-off” is not well understood, although it was previously hypothesized that it plays a major role in encoding sound frequency. Less hypothetical in nature, our work elucidates the role of the steep cut-off in improving signal detection in cochlear-like amplification strategies. We study signal amplification in generic active gain media models and physics-based cochlear models as well. We demonstrate that the optimal strategy for boosting the signal-to-noise ratio (SNR) at a given location “x” requires high gain basal to “x,” followed by rapid attenuation (i.e., sharp cut-off) beyond it. This strategy of amplification followed by a sharp cut-off is precisely how the cochlea processes traveling waves of given frequencies; it boils down to a peculiar form of spatial filtering where waves coming from the “signal side” are amplified, while waves coming from the direction where there is noise but no signal are squelched. This noise “squelching” action manifests in the cochlear filters as the steep high-frequency roll-off.

Noise within: Signal-to-noise enhancement via coherent wave amplification in the mammalian cochlea.

The extraordinary sensitivity of the mammalian inner ear has captivated scientists for decades, largely due to the crucial role played by the outer hair cells (OHCs) and their unique electromotile properties. Typically arranged in three rows along the sensory epithelium, the OHCs work in concert via mechanisms collectively referred to as the "cochlear amplifier" to boost the cochlear response to faint sounds. While simplistic views attribute this enhancement solely to the OHC-based increase in cochlear gain, the inevitable presence of internal noise requires a more rigorous analysis. Achieving a genuine boost in sensitivity through amplification requires that signals be amplified more than internal noise, and this requirement presents the cochlea with an intriguing challenge. Here we analyze the effects of spatially distributed cochlear-like amplification on both signals and internal noise. By combining a straightforward mathematical analysis with a simplified model of cochlear mechanics designed to capture the essential physics, we generalize previous results about the impact of spatially coherent amplification on signal degradation in active gain media. We identify and describe the strategy employed by the cochlea to amplify signals more than internal noise and thereby enhance the sensitivity of hearing. For narrow-band signals, this effective, wave-based strategy consists of spatially amplifying the signal within a localized cochlear region, followed by rapid attenuation. Location-dependent wave amplification and attenuation meet the necessary conditions for amplifying near-characteristic frequency (CF) signals more than internal noise components of the same frequency. Our analysis reveals that the sharp wave cutoff past the CF location greatly reduces noise contamination. The distinctive asymmetric shape of the "cochlear filters" thus underlies a crucial but previously unrecognized mechanism of cochlear noise reduction.

The Shape of Noise to Come: Signal vs. Noise Amplification in the Active Cochlea.

According to the dominant view, the mammalian cochlea spatially amplifies signals by actively pumping energy into the traveling wave. That is, signals are amplified as they propagate through a region where the medium's resistance is effectively negative. While signal amplification has been extensively studied in active cochlear models, the same cannot be said for amplification of internal noise. According to transmission-line theory, signals are amplified more than internal noise in regions where the net resistance is negative. Here we generalize this finding by showing that a distributed system composed of cascaded "noisy" amplifiers boosts signals more rapidly than the internal noise; the larger the amplifier gain, the larger the signal-to-noise ratio (SNR) of the amplified signal. We further show that this mechanism operates in existing active cochlear models: the cochlear amplifier increases the SNR of cochlear responses, and thus enhances cochlear sensitivity. When considering also that the cochlear amplifier narrows the bandwidth of the "cochlear filters", activation of the cochlear amplifiers dramatically increases the SNR (by about one order of magnitude in our simulations) from the tail to the peak of the traveling wave. We further demonstrate that the tapered ear-horn-like cochlear geometry significantly improves the SNR of basilar-membrane responses.

Problem Analysis and Legal Protection of the Exercise of Teachers’ Educational Disciplinary Rights Based on the Background of Big Data

Abstract The right of education and discipline is an important way of school education and teaching management, teachers to fulfill the teaching and educating people, the implementation of the fundamental task of moral education. This paper firstly discusses the dilemma of exercising the right to discipline teachers in education, and also analyzes the legal nature of the right to discipline in education and the impact on the emotional performance of teachers and students in the process of exercising the right. Secondly, cochlear filtering combined with CNN and LSTM network is introduced to extract the speech characteristics of teachers in the process of exercising the right of education and discipline, and a hybrid neural network model is used to realize the recognition and prediction of students’ auditory emotions. Finally, in order to verify the effectiveness of the method of this paper, experimental test analysis was carried out, and a comprehensive rule of law guarantee proposal was given in the process of exercising the right of teachers’ educational discipline. The results show that the maximum value of the intensity of the teacher’s speech signal after processing using the cochlear filter is 78.28dB, and the difference with the original signal is only 0.32%. The accuracy of recognizing students’ auditory emotions reached 90.48% after over 50 iterations. Under the background of big data, the right to discipline teachers in education needs to be analyzed with the help of technology for the data analysis of the appropriateness of exercise, and it is united in a number of aspects, such as strengthening the legislation, standardizing the implementation, strengthening the supervision, and perfecting the relief, as a way to help the comprehensive rule of law operation of the right to discipline teachers in education.

Case reopened: A temporal basis for harmonic pitch templates in the early auditory system?a).

A fundamental assumption of rate-place models of pitch is the existence of harmonic templates in the central nervous system (CNS). Shamma and Klein [(2000). J. Acoust. Soc. Am. 107, 2631-2644] hypothesized that these templates have a temporal basis. Coincidences in the temporal fine-structure of neural spike trains, even in response to nonharmonic, stochastic stimuli, would be sufficient for the development of harmonic templates. The physiological plausibility of this hypothesis is tested. Responses to pure tones, low-pass noise, and broadband noise from auditory nerve fibers and brainstem "high-sync" neurons are studied. Responses to tones simulate the output of fibers with infinitely sharp filters: for these responses, harmonic structure in a coincidence matrix comparing pairs of spike trains is indeed found. However, harmonic template structure is not observed in coincidences across responses to broadband noise, which are obtained from nerve fibers or neurons with enhanced synchronization. Using a computer model based on that of Shamma and Klein, it is shown that harmonic templates only emerge when consecutive processing steps (cochlear filtering, lateral inhibition, and temporal enhancement) are implemented in extreme, physiologically implausible form. It is concluded that current physiological knowledge does not support the hypothesis of Shamma and Klein (2000).

Spectral and temporal resolution in listeners with age-dependent hearing loss

Two basic mechanisms perform the frequency analysis of sounds: (i) the spectral mechanism presented by a bank of frequency-tuned filters and (ii) The temporal mechanism that uses the temporal structure of the signal. The temporal structure manifests in the autocorrelation function. The roles of the spectral and temporal mechanisms in the age-dependent hearing loss were investigated using rippled-spectrum signals. It has been shown previously that discrimination of two rippled signals with different positions of ripples at the frequency scale is mostly performed the spectral processing; for discrimination between a rippled and “flat” signal, the temporal processing dominated. In cases of hearing loss, the decrease of the acuteness of frequency-tuned cochlear filters resulted in reduction of the ripple-pattern resolution by the spectral mechanism. This effect did not influence the capabilities of the temporal mechanism. However, listeners with the age-dependenthearing loss did not perceive high sound frequencies. High frequencies were more favorable for temporal processing, so the inability to perceive these frequencies negatively influenced the ripple resolution. Thus, resolution of rippled signals may reduce at any discrimination task; however, the mechanisms of the reduction differed depending on the task. [Work supported by The Russian Science Foundation, Grant 23-25-00148.]

Optimal feature selection and detection of sickle cell anemia detection using enhanced whale optimization with clustering based boosted C5.0 algorithm for tribes of Nilgiris

Degradation of red blood cells (RBC) causes many diseases, like sickle cell anemia. Diagnosing this disease takes more time because peripheral blood samples must be examined under the microscope. Since the isolated RBC observation is subjective and high error rate leads to reduce the accuracy, the technology is needed to perform this approach. To fulfil these requirements, optimal feature selection and detection of sickle cell anemia using enhanced whale optimization with clustering based boosted C5.0 algorithm in tribes of Nilgiris is proposed in this manuscript. The input dataset is taken from real data set via non-government organization named NAWA (situated in Kotagiri). The reason behind of Nilgiri region is considered in this work is a collection of sickle cell anemia test results of the tribe people who lives in different areas of Nilgiri region. These images are pre-processed using wavelet packet transform cochlear filter bank method to eradicate the noises and to improve the superiority of image. After that, the features are extracted using force-invariant improved feature extraction method. To select the optimal features, enhanced whale optimization (EWO) algorithm is used. These optimal features are classified utilizing clustering based boosted C5.0 algorithm. The proposed method is activated on PYTHON. The proposed method shows 52.32%, 43.78%, 32.78% and 45.90% higher accuracy compared with the existing methods, such as RGSA-MLP-SCA, CRFA-SVM-SCA, AO-LSTM-SCA and BOA-CNN-SCA.

Hearing as adaptive cascaded envelope interpolation

The human auditory system is designed to capture and encode sounds from our surroundings and conspecifics. However, the precise mechanisms by which it adaptively extracts the most important spectro-temporal information from sounds are still not fully understood. Previous auditory models have explained sound encoding at the cochlear level using static filter banks, but this vision is incompatible with the nonlinear and adaptive properties of the auditory system. Here we propose an approach that considers the cochlear processes as envelope interpolations inspired by cochlear physiology. It unifies linear and nonlinear adaptive behaviors into a single comprehensive framework that provides a data-driven understanding of auditory coding. It allows simulating a broad range of psychophysical phenomena from virtual pitches and combination tones to consonance and dissonance of harmonic sounds. It further predicts the properties of the cochlear filters such as frequency selectivity. Here we propose a possible link between the parameters of the model and the density of hair cells on the basilar membrane. Cascaded Envelope Interpolation may lead to improvements in sound processing for hearing aids by providing a non-linear, data-driven, way to preprocessing of acoustic signals consistent with peripheral processes.

Sensitivity to Frequency Modulation is Limited Centrally.

Modulations in both amplitude and frequency are prevalent in natural sounds and are critical in defining their properties. Humans are exquisitely sensitive to frequency modulation (FM) at the slow modulation rates and low carrier frequencies that are common in speech and music. This enhanced sensitivity to slow-rate and low-frequency FM has been widely believed to reflect precise, stimulus-driven phase locking to temporal fine structure in the auditory nerve. At faster modulation rates and/or higher carrier frequencies, FM is instead thought to be coded by coarser frequency-to-place mapping, where FM is converted to amplitude modulation (AM) via cochlear filtering. Here, we show that patterns of human FM perception that have classically been explained by limits in peripheral temporal coding are instead better accounted for by constraints in the central processing of fundamental frequency (F0) or pitch. We measured FM detection in male and female humans using harmonic complex tones with an F0 within the range of musical pitch but with resolved harmonic components that were all above the putative limits of temporal phase locking (>8 kHz). Listeners were more sensitive to slow than fast FM rates, even though all components were beyond the limits of phase locking. In contrast, AM sensitivity remained better at faster than slower rates, regardless of carrier frequency. These findings demonstrate that classic trends in human FM sensitivity, previously attributed to auditory nerve phase locking, may instead reflect the constraints of a unitary code that operates at a more central level of processing.SIGNIFICANCE STATEMENT Natural sounds involve dynamic frequency and amplitude fluctuations. Humans are particularly sensitive to frequency modulation (FM) at slow rates and low carrier frequencies, which are prevalent in speech and music. This sensitivity has been ascribed to encoding of stimulus temporal fine structure (TFS) via phase-locked auditory nerve activity. To test this long-standing theory, we measured FM sensitivity using complex tones with a low F0 but only high-frequency harmonics beyond the limits of phase locking. Dissociating the F0 from TFS showed that FM sensitivity is limited not by peripheral encoding of TFS but rather by central processing of F0, or pitch. The results suggest a unitary code for FM detection limited by more central constraints.

Perceptual simultaneity range for two dichotically presented pure tones with frequency separation below 0.5 Bark.

The perceptual simultaneity range for two diotically presented tones increases with decreasing frequency separation of the two tones from approximately 0.5 Bark. As the present study of two frequency regions shows, this effect is not observed when the two tones are not presented to the same ear, i.e., presented dichotically. Since the increase in simultaneity is only observed when the tones are presented to the same ear, it is possible that it reflects the time-frequency uncertainty within a cochlear filter.

Masking effects of amplitude modulation on frequency-modulated tones

Human sensitivity to frequency modulation (FM) is best for low carrier frequencies (<∼4–5 kHz) and slow modulation rates (<5–10 Hz). This high sensitivity is thought to be afforded by neural phase locking to temporal fine structure (TFS). At faster rates and higher carriers, TFS cues may no longer be available, with sensitivity to FM relying instead on conversion to amplitude modulation (AM) via cochlear filtering. One way to test this hypothesis is to measure the masking produced by additional AM, with the prediction that AM masking will be more pronounced when FM is encoded via AM than via TFS cues. This study tested AM masking of FM and of “simulated FM,” which involves out-of-phase AM dyads. FM detection was measured for carrier frequencies of 1 and 6 kHz and modulation rates of 2 and 20 Hz with and without AM masking. Preliminary results suggest that FM and simulated FM sensitivity is more affected by AM at fast than slow rates at both low and high carrier frequencies. Overall, the results do not provide support for the idea that AM interference can be used to distinguish between TFS- and envelope-based codes for FM. [Work supported by NIH grant R21 DC019409.]

Testing the generation of harmonic templates from coincidence patterns in response to noise: Physiological and modeling considerations

Harmonic templates are a critical component of spectral pitch theories. Shamma and Klein (2000) hypothesize that such templates can arise from neural coincidence patterns in response to stochastic stimuli without pitch.The key feature of their computational model is a higher number of coincidences between neurons with harmonically related characteristic frequencies (CFs). We tested the hypothesis with physiological data and a computational model. We recorded neural responses in chinchilla to various periodic and non-periodic stimuli, both from the auditory nerve and the trapezoid body. We then examined counts of spike coincidences across fibers. The key prediction of a higher number of coincidences between pairs of fibers with harmonically related CFs than for other pairs was observed for responses to narrowband signals (tones at CF), but not to broadband noise. We then tested a model similar to that of Shamma and Klein. Only with cochlear filters sharper than plausible even in humans, followed by additional spectral sharpening through lateral inhibition and extreme temporal sharpening, do we obtain harmonic templates similar to their model. Although our results do not speak to the existence of harmonic templates per se, we conclude that the specific mechanism of this hypothesis has little physiological plausibility.

A comparative study of eight human auditory models of monaural processing.

A number of auditory models have been developed using diverging approaches, either physiological or perceptual, but they share comparable stages of signal processing, as they are inspired by the same constitutive parts of the auditory system. We compare eight monaural models that are openly accessible in the Auditory Modelling Toolbox. We discuss the considerations required to make the model outputs comparable to each other, as well as the results for the following model processing stages or their equivalents: Outer and middle ear, cochlear filter bank, inner hair cell, auditory nerve synapse, cochlear nucleus, and inferior colliculus. The discussion includes a list of recommendations for future applications of auditory models.

Testing the plausibility of harmonic templates based on temporal fine-structure

The existence of harmonic templates in the central nervous system is a basic tenet of place models of pitch. Shamma & Klein (2000) propose that coincidences in temporal fine-structure across auditory nerve fibers in response to broadband stimuli suffice to set up harmonic templates for all fundamental frequencies in the phase-locking range. However, their computational model uses cochlear filters with unphysiologically steep slopes, calling into question the plausibility of the proposed scheme. We obtained responses from chinchilla auditory nerve fibers and “high-sync” neurons to pure tones, low-pass noise, and broadband noise, including “warped” noise (Heinz, 2006; Cedolin & Delgutte, 2010). Coincidences were computed using shuffled auto- or crosscorrelograms. We find that the maximum number of coincidences in response to tones of different frequencies, simulating the output of fibers with infinite sharp filters, indeed shows the template structure modeled by Shamma and Klein. This was not the case for responses to low-pass or broadband noise. However, if coincidences were only counted at zero delay, higher numbers of coincidences were indeed found for harmonically-related characteristic frequencies, even in response to noise. We conclude that steep-sloped cochlear filtering is not a necessary condition to obtain harmonic templates based on fine-structure.

Effect of Musical Experience on Cochlear Frequency Resolution: An Estimation of PTCs, DLF and SOAEs.

Background: There is a lot of debate on whether musical experience/training can affect cochlear filter characteristics and thereby enhance frequency resolution skills in musicians (Ms). Thus, the objective of the study was to compare between Ms and non-musicians (NMs) and correlate the frequency resolution skills with years of musical experience, the difference limen for frequency (DLF), and the Q10 values of psychophysical tuning curves (PTCs) measured at 2 characteristic frequencies (CFs), 1000 and 4000 Hz. The secondary objective was to find out whether spontaneous otoacoustic emissions (SOAEs) can be recorded among Ms more frequently than among NMs. Methods:Thirty-six listeners with normal hearing participated in the study. They were 18 Ms with a minimum of 8 years of musical experience, and 18 NMs. Forward-masked PTCs and DLF tests were assessed for each listener in the right ear at 2 CFs, 1 and 4 kHz, and SOAEs were measured using standard protocol. Results:The results obtained indicated that values of the psychophysical measures, DLF and Q10, were better among the Ms group. Statistically significant differences were seen when measurements were done at 1 KHz and 4 KHz. SOAEs were recorded in a higher number of Ms than NMs. Musical experience had a moderate positive correlation with PTC Q10 values at both 1 and 4 kHz, but not DLF values. Conclusion:This study indicates that musical experience enhances peripheral filtering, and thereby betters cochlear frequency selectivity in Ms.

Mechanisms of Spectrotemporal Modulation Detection for Normal- and Hearing-Impaired Listeners.

Spectrotemporal modulations (STM) are essential features of speech signals that make them intelligible. While their encoding has been widely investigated in neurophysiology, we still lack a full understanding of how STMs are processed at the behavioral level and how cochlear hearing loss impacts this processing. Here, we introduce a novel methodological framework based on psychophysical reverse correlation deployed in the modulation space to characterize the mechanisms underlying STM detection in noise. We derive perceptual filters for young normal-hearing and older hearing-impaired individuals performing a detection task of an elementary target STM (a given product of temporal and spectral modulations) embedded in other masking STMs. Analyzed with computational tools, our data show that both groups rely on a comparable linear (band-pass)–nonlinear processing cascade, which can be well accounted for by a temporal modulation filter bank model combined with cross-correlation against the target representation. Our results also suggest that the modulation mistuning observed for the hearing-impaired group results primarily from broader cochlear filters. Yet, we find idiosyncratic behaviors that cannot be captured by cochlear tuning alone, highlighting the need to consider variability originating from additional mechanisms. Overall, this integrated experimental-computational approach offers a principled way to assess suprathreshold processing distortions in each individual and could thus be used to further investigate interindividual differences in speech intelligibility.

Nonlinear reflection as a cause of the short-latency component in stimulus-frequency otoacoustic emissions simulated by the methods of compression and suppression.

Stimulus-frequency otoacoustic emissions (SFOAEs) are generated by coherent reflection of forward traveling waves by perturbations along the basilar membrane. The strongest wavelets are backscattered near the place where the traveling wave reaches its maximal amplitude (tonotopic place). Therefore, the SFOAE group delay might be expected to be twice the group delay estimated in the cochlear filters. However, experimental data have yielded steady-state SFOAE components with near-zero latency. A cochlear model is used to show that short-latency SFOAE components can be generated due to nonlinear reflection of the compressor or suppressor tones used in SFOAE measurements. The simulations indicate that suppressors produce more pronounced short-latency components than compressors. The existence of nonlinear reflection components due to suppressors can also explain why SFOAEs can still be detected when suppressors are presented more than half an octave above the probe-tone frequency. Simulations of the SFOAE suppression tuning curves showed that phase changes in the SFOAE residual as the suppressor frequency increases are mostly determined by phase changes of the nonlinear reflection component.

Characterizing amplitude and frequency modulation cues in natural soundscapes: A pilot study on four habitats of a biosphere reserve.

Natural soundscapes correspond to the acoustical patterns produced by biological and geophysical sound sources at different spatial and temporal scales for a given habitat. This pilot study aims to characterize the temporal-modulation information available to humans when perceiving variations in soundscapes within and across natural habitats. This is addressed by processing soundscapes from a previous study [Krause, Gage, and Joo. (2011). Landscape Ecol. 26, 1247] via models of human auditory processing extracting modulation at the output of cochlear filters. The soundscapes represent combinations of elevation, animal, and vegetation diversity in four habitats of the biosphere reserve in the Sequoia National Park (Sierra Nevada, USA). Bayesian statistical analysis and support vector machine classifiers indicate that: (i) amplitude-modulation (AM) and frequency-modulation (FM) spectra distinguish the soundscapes associated with each habitat; and (ii) for each habitat, diurnal and seasonal variations are associated with salient changes in AM and FM cues at rates between about 1 and 100 Hz in the low (<0.5 kHz) and high (>1-3 kHz) audio-frequency range. Support vector machine classifications further indicate that soundscape variations can be classified accurately based on these perceptually inspired representations.

Fine-grained statistical structure of speech.

In spite of its acoustic diversity, the speech signal presents statistical regularities that can be exploited by biological or artificial systems for efficient coding. Independent Component Analysis (ICA) revealed that on small time scales (∼ 10 ms), the overall structure of speech is well captured by a time-frequency representation whose frequency selectivity follows the same power law in the high frequency range 1–8 kHz as cochlear frequency selectivity in mammals. Variations in the power-law exponent, i.e. different time-frequency trade-offs, have been shown to provide additional adaptation to phonetic categories. Here, we adopt a parametric approach to investigate the variations of the exponent at a finer level of speech. The estimation procedure is based on a measure that reflects the sparsity of decompositions in a set of Gabor dictionaries whose atoms are Gaussian-modulated sinusoids. We examine the variations of the exponent associated with the best decomposition, first at the level of phonemes, then at an intra-phonemic level. We show that this analysis offers a rich interpretation of the fine-grained statistical structure of speech, and that the exponent values can be related to key acoustic properties. Two main results are: i) for plosives, the exponent is lowered by the release bursts, concealing higher values during the opening phases; ii) for vowels, the exponent is bound to formant bandwidths and decreases with the degree of acoustic radiation at the lips. This work further suggests that an efficient coding strategy is to reduce frequency selectivity with sound intensity level, congruent with the nonlinear behavior of cochlear filtering.

An Efficient and Perceptually Motivated Auditory Neural Encoding and Decoding Algorithm for Spiking Neural Networks.

The auditory front-end is an integral part of a spiking neural network (SNN) when performing auditory cognitive tasks. It encodes the temporal dynamic stimulus, such as speech and audio, into an efficient, effective and reconstructable spike pattern to facilitate the subsequent processing. However, most of the auditory front-ends in current studies have not made use of recent findings in psychoacoustics and physiology concerning human listening. In this paper, we propose a neural encoding and decoding scheme that is optimized for audio processing. The neural encoding scheme, that we call Biologically plausible Auditory Encoding (BAE), emulates the functions of the perceptual components of the human auditory system, that include the cochlear filter bank, the inner hair cells, auditory masking effects from psychoacoustic models, and the spike neural encoding by the auditory nerve. We evaluate the perceptual quality of the BAE scheme using PESQ; the performance of the BAE based on sound classification and speech recognition experiments. Finally, we also built and published two spike-version of speech datasets: the Spike-TIDIGITS and the Spike-TIMIT, for researchers to use and benchmarking of future SNN research.