Auditory Filter Research Articles

Sensory Processing Dysfunction and Mealtime Behavior Problems in Children With Autism

To compare sensory processing and mealtime problem behaviors among children with autism spectrum disorder (ASD) and typically developing controls, and to examine the relationship between atypical sensory processing and eating problems in children with ASD. 50 children (4-10 years) with a diagnosis of ASD as per DSM-5 were recruited from the pediatric psychology clinic of a tertiary care center in India. The Brief Assessment of Mealtime Behavior in Children (BAMBIC) and the Short Sensory Profile (SSP) were administered to measure feeding and sensory processing problems, respectively. Parents were interviewed about their child's dietary intake using a 3-day dietary recall. The ASD group showed greater mealtime behavior problems than the control group and had significantly higher total scores on the BAMBIC (P<0.001), and on two of the three subscales including food refusal (P<0.001) and disruptive behavior (P<0.001). The ASD group, relative to the neurotypical children, showed atypical response on majority of the subscales of the short sensory profile including tactile sensitivity (P<0.001), taste sensitivity (P<0.001), movement sensitivity (P<0.001), under responsiveness (P<0.001), auditory filtering (P<0.001), low weak/energy (P=0.02), and visual/auditory sensitivity (P<0.001). The study underscores the need for detailed evaluation of sensory processing and feeding problems of children with ASD so that the interventions can be tailored to address their unique sensory characteristics.

Filter algorithm based on cochlear mechanics and neuron filter mechanism and application on enhancement of audio signals

A filter algorithm based on cochlear mechanics and neuron filter mechanism is proposed from the view point of vibration. It helps to solve the problem that the non-linear amplification is rarely considered in studying the auditory filters. A cochlear mechanical transduction model is built to illustrate the audio signals processing procedure in cochlea, and then the neuron filter mechanism is modeled to indirectly obtain the outputs with the cochlear properties of frequency tuning and non-linear amplification. The mathematic description of the proposed algorithm is derived by the two models. The parameter space, the parameter selection rules and the error correction of the proposed algorithm are discussed. The unit impulse responses in the time domain and the frequency domain are simulated and compared to probe into the characteristics of the proposed algorithm. Then a 24-channel filter bank is built based on the proposed algorithm and applied to the enhancements of the audio signals. The experiments and comparisons verify that, the proposed algorithm can effectively divide the audio signals into different frequencies, significantly enhance the high frequency parts, and provide positive impacts on the performance of speech enhancement in different noise environments, especially for the babble noise and the volvo noise.

A Bayesian active-learning test for remote hearing-aid fitting

A major problem of remote or online hearing tests is the calibration of the equipment and control of the acoustic environment: For example, an audiogram requires calibrated headphones with a wide dynamic range and a silent room. Instead, a notched-noise test alleviates these problems and gives valuable information about auditory filters that can be used to fit a hearing aid. Although absolute levels still play a role, relative levels between signal and noise are more robust as the environment does not need to be quieter than the noise. Using traditional methods, determination of auditory-filter shapes across the whole frequency range would take hours. We developed a notched-noise test that uses Gaussian Processes to maintain a probabilistic estimate of the outcome, and to select stimulus parameters based on mutual information. This active-learning method gives an estimate of auditory-filter shapes in about 30 minutes across a wide range of frequency. We demonstrate that these estimates, even when obtained with headphones for which the frequency response and exact calibration is unknown, are sufficient to give a good initial fit of a hearing aid.

Study of sensory processing deficits in autism spectrum disorder symptom triad: an Egyptian sample

BackgroundAutism spectrum disorder (ASD) is one of the most common neurodevelopmental disorders. In DSM-IV, the diagnostic criteria of autism consisted of three domains: impairment in social interaction, communication deficits, and stereotypic behavior, while in DSM-5 they were condensed into two domains: social communication deficits and restricted patterns of behavior to which sensory processing deficits (SPD) were added, manifested by hypo- or hyper-reactivity to sensory stimuli or uncommon interests in sensory aspects of the surrounding environment. The purpose of this study is to determine the relation between SPD and the symptom triad in ASD namely social interaction, communication deficits, and stereotypic behavior. To our knowledge, this issue was not studied before in Egyptian literature.ResultsThere was a significant negative correlation between SPD (assessed by short sensory profile: SSP) and symptom triad in ASD. As regards social interaction, the ADI-R (A), there was significant negative correlation with the SSP total scores and all subscales except for low energy/weak and visual/auditory sensitivity, whereas under-responsive/seeks sensation held the highest negative correlation (p ˂ 0.008). As regards communication deficits, the ADI-R (B), there was significant negative correlation with the SSP total scores and the auditory filtering held the highest negative correlation (p ˂ 0.008), and as regards stereotypic behavior, the ADI-R (C), there was significant negative correlation with the SSP total scores and all subscales except for low energy/weak and visual/auditory sensitivity whereas taste/smell sensitivity and auditory filtering held the highest negative correlation (p ˂ 0.001). There was a high significant negative correlation between the severity of autistic symptoms (assessed by CARS) and the SSP total scores, and for all subscales (p ˂ 0.001) except for low energy/weak, the correlation was significant (not highly significant) (p ˂ 0.05). Finally, there was high percentage of parental consanguinity among the participants (80%).ConclusionsSPD negatively affected the ASD symptom triad which highlights the importance of sensory integration therapy (SIT) as a major core of ASD treatment alongside the other treatment modalities. Early ASD screening is mandatory in families with parental consanguinity.

Sensory Processing Disorder in Normally Developing Egyptian Children

Background: The aim of this study was to assess the sensory processing among typically developing children between 3-8 years old to follow any negative impact on their social and behavioral development.Objective: The purpose of this study was to Develop and standardize Arabic QuickSIN test that measures the signal-to-noise ratio loss.Patients and Methods: Sensory processing abilities were assessed for 190 typically developing children by using the Short Sensory Profile (SSP)caregiver questionnaire. A typical sensory performance was defined as an SSP total score of: 190-155 and the total scores of typical sensory performances for tactile sensitivity:35-30; taste/smell sensitivity:20-15; movement sensitivity: 15-13; under-responsive/seeks sensation:35-27; auditory filtering:30-23; low energy/weak:30-26; and visual/auditory sensitivity: 25-19. Below these scores shows atypical sensory performance.Results: The typically developed children showed some degree of sensory processing disorderon Under-responsive/seek sensitivity as well as Taste and smell sensitivity. Some children in this study (50.5 %) showed a degree of sensory processing impairment in the SSP total score. An increase in the tactile sensitivity was significantly associated with higher prevalence of the movement sensitivity (r=0.745, p value= <0.001). As for the Pearson correlation between the low energy/weak and the visual/auditory sensitivity is about r=0.698, which indicates that there is a moderate positive relationship between the 2 variables.Conclusion: These findings showed the types of sensory processing disorder in typically developing children and the difference between female and male children. Future studies are required on larger samples to clarify more the types of sensory impairment in typically developing children and to confirm these results.

The relationship between preschoolers’ sensory regulation and temperament: implications for parents and early childhood caregivers

ABSTRACT With a non-clinical sample of pre-schoolers, we used partial least squares regression to assess the extent to which children's seven sensory regulatory factors (X) account for variations in the three temperament factors (Y) and Y explains variations in X. The results suggested predicting temperament factors from sensory processing factors receive better outcomes. Finally, multiple regression analyses using the sensory processing measures to predict each of the three temperamental variables showed that (1) significant predictors of negative affectivity included difficulties in tactile sensitivity, taste/smell sensitivity, and low energy; (2) surgency was predicted by underresponsive/seeks sensation, auditory filtering, tactile sensitivity, and movement sensitivity; and (3) effortful control was predicted by low energy and auditory filtering. The hypothesis supported children's sensory processing preferences form the basis for the manifestation of temperament. Parents and educators need to notice environmental stimuli can cause challenges for children with sensory processing difficulties and manifesti difficult temperament.

Association between Body Mass Index and Sensory Processing in Childhood: InProS Study.

We assessed the association between body mass index (BMI) and sensory processing in 445 Spanish children aged 3–7 from the InProS project. Child sensory processing was measured using the short sensory profile (SSP); an atypical sensory performance was defined as an SSP total score <155 and scores of tactile sensitivity <30; taste/smell sensitivity <15; movement sensitivity <13; under-responsive/seeks sensation <27; auditory filtering <23; low energy/weak <26; and visual/auditory sensitivity <19. The BMI was calculated according to the cutoffs by the World Health Organization for children aged 0–5 and 5–19 years. We used multiple Poisson regression models with robust variance to obtain prevalence ratios (PR). No associations between children’s overweight and obesity and the prevalence of atypical sensory outcomes were observed. A one-point increase in BMI was significantly associated with a higher prevalence of atypical tactile sensitivity (PR = 1.07, 95% CI: 1.02; 1.12). A statistically marginal association was also observed for atypical total SSP (PR = 1.03, 95% CI: 1.00; 1.07) and atypical movement sensitivity (PR = 1.05, 95% CI: 1.00; 1.10). To our knowledge, this is the first time the association between children’s BMI and sensory processing has been reported. Our findings suggest that sensory processing issues may play a part in the complex context of childhood obesity. Further research is required to confirm these findings.

Effect of envelope signal-to-noise ratio on the intelligibility of speech in speech-spectrum shaped noise

In attempts to improve speech intelligibility, the envelope of speech in noise is commonly processed in the time domain and used to amplitude modulate the corrupted speech. In this study the envelope was constructed from the noisy speech, but separately and independently of the SNR of the noisy speech. Envelopes corresponding to 0, −3, −6, and −9 dB SNR were applied to the noisy speech when the latter had an SNR of either −9, −6, −3, or 0 dB, to evaluate the benefit to intelligibility of improving the SNR in the modulation domain. Signals from 200 Hz to 6 kHz were processed by MATLAB into sixteen contiguous subbands with bandwidth approximately 1.5×ERB of an auditory filter. The subband envelopes were formed from the absolute value of the signals and low-pass filtered. Eleven subjects aged 29 ± 8 years (mean and range) with normal hearing underwent the Modified Rhyme Test to assess speech intelligibility. The stimuli were presented diotically over earphones with the subject seated in an audiometric room. Statistically significant increases in mean word score of up to 35% could be obtained by improving the envelope SNR, suggesting this processing may benefit speech intelligibility. [Work supported by NIOSH.]

Investigation of a temporal modulation based method on the intelligibility of speech in speech-spectrum shaped noise

Temporal modulation is frequently considered for improving the intelligibility of noisy speech owing to the importance of the envelope in conveying speech information. In this study, we evaluated the performance of temporal modulation on word scores obtained with speech-in-noise. We firstly divided the noisy speech into 16 contiguous subbands from 200 Hz to 6 kHz with bandwidth approximately 1.5xERB of an auditory filter. The temporal envelope was constructed for each subband from the absolute value. It was then low-pass filtered at 16 Hz and used as the instantaneous gain for the subband's noisy speech. We established a psychophysical test (Modified Rhyme Test) with different speech-in-noise SNR values of 0, −3, −6, and −9 dB. Eleven native speakers (age 29 ± 8) with normal hearing were recruited for the study. The signals were processed by MATLAB and presented over earphones to participants seated in an audiometric room. Comparing the processed and unprocessed noisy speech, the mean differences in word scores for SNRs of 0, −3, −6, and −9 dB were −0.4%, 1.5%, −10.5%, and 1.4%, respectively. Using ANOVA, we concluded that the temporal modulation-based algorithm does not produce a statistically significant improvement in speech intelligibility. [Work supported by NIOSH.]

The Yost approach to pitch perception: What we can learn from noise-like stimuli

One of Bill's major contributions to psychoacoustics is his innovative work using rippled noise (RN) to understand pitch. Bill has argued that RN-pitch percepts can be accounted for by models based on autocorrelation. Working with Bill, we showed that behavioral responses obtained from chinchillas to RNs generally paralleled those obtained from humans, suggesting the underlying processing is similar between the two species. Recent work from my lab using noise-vocoded harmonic tone complexes (NV-HTCs) was inspired by Bill's approach with RNs. NV-HTCs can also evoke pitch percepts like RNs, but present a challenge to Bill's autocorrelation approach, because they can have strong harmonic structures with weak or no periodicities. However, when NV-HTCs are passed through a gammatone filterbank model, weak stimulus periodicities become augmented in summary correlograms. An analytical model based on summary correlograms for NV-HTC responses from (1) humans in a magnitude estimation task and (2) chinchillas in a stimulus generalization task suggests that the underlying processing does indeed differ between the two species. Specifically, chinchillas appear to process the envelope whereas humans appear to process the fine structure. This talk will explore why this apparent processing difference exists, despite the similarity in chinchilla and human auditory filter bandwidths.

Changes in audiometric threshold and frequency selectivity correlate with cochlear histopathology in macaque monkeys with permanent noise-induced hearing loss

Exposure to loud noise causes damage to the inner ear, including but not limited to outer and inner hair cells (OHCs and IHCs) and IHC ribbon synapses. This cochlear damage impairs auditory processing and increases audiometric thresholds (noise-induced hearing loss, NIHL). However, the exact relationship between the perceptual consequences of NIHL and its underlying cochlear pathology are poorly understood. This study used a nonhuman primate model of NIHL to relate changes in frequency selectivity and audiometric thresholds to indices of cochlear histopathology. Three macaques (one Macaca mulatta and two Macaca radiata) were trained to detect tones in quiet and in noises that were spectrally notched around the tone frequency. Audiograms were derived from tone thresholds in quiet; perceptual auditory filters were derived from tone thresholds in notched-noise maskers using the rounded-exponential fit. Data were obtained before and after a four-hour exposure to a 50-Hz noise centered at 2 kHz at 141 or 146 dB SPL. Noise exposure caused permanent audiometric threshold shifts and broadening of auditory filters at and above 2 kHz, with greater changes observed for the 146-dB-exposed monkeys. The normalized bandwidth of the perceptual auditory filters was strongly correlated with audiometric threshold at each tone frequency. While changes in audiometric threshold and perceptual auditory filter widths were primarily determined by the extent of OHC survival, additional variability was explained by including interactions among OHC, IHC, and ribbon synapse survival. This is the first study to provide within-subject comparisons of auditory filter bandwidths in an animal model of NIHL and correlate these NIHL-related perceptual changes with cochlear histopathology. These results expand the foundations for ongoing investigations of the neural correlates of NIHL-related perceptual changes.

Assessing the Quality of Low-Frequency Acoustic Hearing: Implications for Combined Electroacoustic Stimulation With Cochlear Implants.

There are many potential advantages to combined electric and acoustic stimulation (EAS) with a cochlear implant (CI), including benefits for hearing in noise, localization, frequency selectivity, and music enjoyment. However, performance on these outcome measures is variable, and the residual acoustic hearing may not be beneficial for all patients. As such, we propose a measure of spectral resolution that might be more predictive of the usefulness of the residual hearing than the audiogram alone. In the following experiments, we measured performance on spectral resolution and speech perception tasks in individuals with normal hearing (NH) using low-pass filters to simulate steeply sloping audiograms of typical EAS candidates and compared it with performance on these tasks for individuals with sensorineural hearing loss with similar audiometric configurations. Because listeners with NH had similar levels of audibility and bandwidth to listeners with hearing loss, differences between the groups could be attributed to distortions due to hearing loss. Listeners with NH (n = 12) and those with hearing loss (n = 23) with steeply sloping audiograms participated in this study. The group with hearing loss consisted of 7 EAS users, 14 hearing aid users, and 3 who did not use amplification in the test ear. Spectral resolution was measured with the spectral-temporal modulated ripple test (SMRT), and speech perception was measured with AzBio sentences in quiet and noise. Listeners with NH listened to stimuli through low-pass filters and at two levels (40 and 60 dBA) to simulate low and high audibility. Listeners with hearing loss listened to SMRT stimuli unaided at their most comfortable listening level and speech stimuli at 60 dBA. Results suggest that performance with SMRT is significantly worse for listeners with hearing loss than for listeners with NH and is not related to audibility. Performance on the speech perception task declined with decreasing frequency information for both listeners with NH and hearing loss. Significant correlations were observed between speech perception, SMRT scores, and mid-frequency audiometric thresholds for listeners with hearing loss. NH simulations describe a "best case scenario" for hearing loss where audibility is the only deficit. For listeners with hearing loss, the likely broadening of auditory filters, loss of cochlear nonlinearities, and possible cochlear dead regions may have contributed to distorted spectral resolution and thus deviations from the NH simulations. Measures of spectral resolution may capture an aspect of hearing loss not evident from the audiogram and be a useful tool for assessing the contributions of residual hearing post-cochlear implantation.

Comodulation masking release with random variations of flanking-band center frequencies.

Comodulation masking release (CMR) is an effect that is associated with auditory sensitivity to coherent amplitude modulations in different frequency regions. The present study investigated if this comodulation is detected by a direct comparison of auditory filter outputs, or if common masker fluctuations are first extracted by a broadly tuned stage that integrates information across a large spectral range. To this end, a modified flanking-band experiment with a narrowband noise masker at the signal frequency (on-frequency masker), and two flanking bands (FBs), one centered below and one above the signal frequency, were used. The center frequencies (CFs) of FBs changed whenever the masker had a local envelope minimum. The center frequencies were randomly chosen from a range of frequencies around the average CF of each FB. A CMR was measured even for large CF variations of FBs, where the envelopes at the off-frequency auditory filters were no longer the same as the masker envelope at the on-frequency auditory filter. This supports the hypothesis of a broadly tuned stage to determine masker comodulation. For two experimental settings, CMR deteriorated for very large variations of CFs of FBs, suggesting a spectral weighting of the off-frequency auditory filters in this broadly tuned stage.

Comodulation masking release effect in children with and without dyslexia

&lt;p class="abstract"&gt;&lt;strong&gt;Background:&lt;/strong&gt; Speech recognition in a modulating noise background can be facilitated by a process attributable to comodulation masking release (CMR). CMR is usually assumed to depend on comparisons of the outputs of different auditory filters. There was an immense importance to study to find the CMR effect in children with and without dyslexia.&lt;/p&gt;&lt;p class="abstract"&gt;&lt;strong&gt;Methods:&lt;/strong&gt; The study was to find the CMR effect in children with and without dyslexia. The current research was carried out through five steps i.e. auditory attention task stimuli preparation, auditory performance test, CMR stimuli preparation, CMR task and statistical analysis. Through these processes for measuring the CMR was done for the children with and without dyslexia. All the data were tabulated and statistically computed for the analysis of the data. SPSS software version 16 was used for the statistical analysis of the data. &lt;/p&gt;&lt;p class="abstract"&gt;&lt;strong&gt;Results:&lt;/strong&gt; Independent t-test was used for the statistical analysis while the comparison between groups. Paired t- test was used for the statistical analysis while the comparison within the group at 95% confidence interval. These results indicated that the amount effect of CMR is greater in children with dyslexia. There was not a significant difference of CMR between the children with and without dyslexia to the effect of CMR was not significantly different between the ears in children without dyslexia and with dyslexia.&lt;/p&gt;&lt;p class="abstract"&gt;&lt;strong&gt;Conclusions:&lt;/strong&gt; The present study indicates that children with dyslexia have selective inability to use the temporal and spectral cues necessary for signal extraction in CMR.&lt;/p&gt;

Effects of rehabilitation training on an elderly population with mild to moderate hearing loss: study protocol for a randomised clinical trial

Background: Age-related hearing loss (presbycusis) is a form of hearing loss in over 60-years-olds and has a negative impact on quality of life. Presbycusis is multifactorial and is predominately characterised with a loss of speech perception in noise. In the cochlea, auditory filters decompose broadband sound into a series of narrowband output signals, which contains two kinds of temporal information: slow changes in overall amplitude envelope (ENV) and faster variations in temporal fine structure (TFS). TFS is important for recognition of target speech in noise. The main aim of the study is to evaluate the effect of TFS rehabilitation training in participants over the age of 60 years with mild to moderate hearing loss. Methods: A randomised clinical trial conducted on 30 participants with mild (loss of 20-39dB HL) to moderate (40-69dB HL) hearing loss, aged between 60 and 75 years old. Participants with conductive hearing loss, abnormal middle ear pathology and central nervous system disease were excluded. Participants were selected randomly to an intervention and control group with a 1:1 ratio. Rehabilitation for the intervention Group are 30-minute sessions three times a week for a total five weeks of vowel consonant vowel words that are used to eliminate ENV and keep only TFS. Word in noise test, binaural TFS test, and Speech, Spatial and Qualities of Hearing Scale scores are performed at the beginning and end of study to evaluate the effect of rehabilitation training. Conclusion: Life expectancy in the elderly has improved, leading to an increased prevalence of age-related diseases including presbycusis. A literature review highlighted that TFS damage is permanent; however, in this study we will attempt to prove that TFS training may lead to speech in noise perception improvement. Trial registration: Registry of Clinical Trials, IRCT2019625044006N1 (7 th August 2019).

Neural fluctuation cues for simultaneous notched-noise masking and profile-analysis tasks: Insights from model midbrain responses.

Results of simultaneous notched-noise masking are commonly interpreted as reflecting the bandwidth of underlying auditory filters. This interpretation assumes that listeners detect a tone added to notched-noise based on an increase in energy at the output of an auditory filter. Previous work challenged this assumption by showing that randomly and independently varying (roving) the levels of each stimulus interval does not substantially worsen listener thresholds [Lentz, Richards, and Matiasek (1999). J. Acoust. Soc. Am. 106, 2779-2792]. Lentz et al. further challenged this assumption by showing that filter bandwidths based on notched-noise results were different from those based on a profile-analysis task [Green (1983). Am. Psychol. 38, 133-142; (1988). (Oxford University Press, New York)], although these estimates were later reconciled by emphasizing spectral peaks of the profile-analysis stimulus [Lentz (2006). J. Acoust. Soc. Am. 120, 945-956]. Here, a single physiological model is shown to account for performance in fixed- and roving-level notched-noise tasks and the Lentz et al. profile-analysis task. This model depends on peripheral neural fluctuation cues that are transformed into the average rates of model inferior colliculus neurons. Neural fluctuations are influenced by peripheral filters, synaptic adaptation, cochlear amplification, and saturation of inner hair cells, an element not included in previous theories of envelope-based cues for these tasks. Results suggest reevaluation of the interpretation of performance in these paradigms.

ANALYSIS OF SENSORY PROCESSING ABILITIES AMONG AUTISM SPECTRUM DISORDERS CHILDREN

BACKGROUND AND AIMS Autism is characterized by the complex paradigm of intellectual disabilities that leads to impairment in sensory integration, communication barriers and social interaction. Therefore, the purpose of this study is to analyze the sensory processing abilities among autism children. Further the study will be helpful in identifying difficulties faced by autistic children in multidimensional aspects. METHODOLOGY A cross-sectional study was conducted at special education institute/schools of Karachi among the Autism Spectrum Children to evaluate the patterns of sensory integration by using Short Sensory Profile. RESULTS A total of 50 children, 23 males and 27 females with mean age of 9.5±1.84 years included in the study. On SSP, the mean scores of participants were 17.89±7.61 in tactile sensitivity, 16.66±6.90 in Under-responsive/Seeks Sensation, 14.45±5.75 in Visual/Auditory Sensitivity, 17.77±7.04 in Auditory Filtering, 19.57±7.91 in Low Energy/Weak while lowest scores were reported in Taste/Smell Sensitivity and Movement Sensitivity. CONCLUSION It was concluded that definite difference is more prevalent among children with autism whereas varying percentages are found on different items on Short sensory profile. Further studies are required to assess the sensory processing issues for more useful outcomes.

Computational models for predicting sound quality

The quality of an audio device, such as a microphone, amplifier, or headphone, depends on how accurately the device transmits the properties of the sound source to the ear(s) of the listener. Two types of “distortion” can occur in this transmission: (1) “Linear” distortion, which may be described as a deviation of the frequency response from the “target” response; (2) Nonlinear distortion, which is characterised by frequency components in the output of the device that were not present in the input. These two forms of distortion have different perceptual effects. Their effects on sound quality can be predicted using a model of auditory processing with the following stages: (1) A filter to take into account the transmission of sound from the device to the ear of the listener; (2) A filter to simulate the effects of transmission through the middle ear; (3) An array of bandpass filters to simulate the auditory filters that exist in the cochlea of the inner ear. For predicting the perceptual effects of linear distortion, a model operating in the frequency domain can be used. For predicting the perceptual effects of nonlinear distortion, a model operating in the time domain is required, since the detailed waveforms at the outputs of the auditory filters need to be considered. The models described have been shown to give accurate predictions for a wide range of “artificial” and “real” linear and nonlinear distortions.

Ship Type Classification by Convolutional Neural Networks with Auditory-like Mechanisms.

Ship type classification with radiated noise helps monitor the noise of shipping around the hydrophone deployment site. This paper introduces a convolutional neural network with several auditory-like mechanisms for ship type classification. The proposed model mainly includes a cochlea model and an auditory center model. In cochlea model, acoustic signal decomposition at basement membrane is implemented by time convolutional layer with auditory filters and dilated convolutions. The transformation of neural patterns at hair cells is modeled by a time frequency conversion layer to extract auditory features. In the auditory center model, auditory features are first selectively emphasized in a supervised manner. Then, spectro-temporal patterns are extracted by deep architecture with multistage auditory mechanisms. The whole model is optimized with an objective function of ship type classification to form the plasticity of the auditory system. The contributions compared with an auditory inspired convolutional neural network include the improvements in dilated convolutions, deep architecture and target layer. The proposed model can extract auditory features from a raw hydrophone signal and identify types of ships under different working conditions. The model achieved a classification accuracy of 87.2% on four ship types and ocean background noise.

The gammachirp auditory filter and its application to speech perception

The gammachirp auditory filter and its application to speech perception