Acoustic Input Research Articles

Systematic changes in neural selectivity reflect the acquired salience of category-diagnostic dimensions.

Humans and other animals develop remarkable behavioral specializations for identifying, differentiating, and acting on classes of ecologically important signals. Ultimately, this expertise is flexible enough to support diverse perceptual judgments: a voice, for example, simultaneously conveys what a talker says as well as myriad cues about her identity and state. Mature perception across complex signals thus involves both discovering and learning regularities that best inform diverse perceptual judgments, and weighting this information flexibly as task demands change. Here, we test whether this flexibility may involve endogenous attentional gain to task-relevant dimensions. We use two prospective auditory category learning tasks to relate a complex, entirely novel soundscape to four classes of "alien identity" and two classes of "alien size." Identity, but not size, categorization requires discovery and learning of patterned acoustic input situated in one of two simultaneous, frequency-delimited bands. This allows us to capitalize on the coarsely segregated frequency-band-specific channels of auditory tonotopic maps using fMRI to ask whether category-relevant perceptual information is prioritized relative to simultaneous, uninformative information. Among participants expert at alien identity categorization, we observe prioritization of the diagnostic frequency band that persists even when the diagnostic information becomes irrelevant in the size categorization task. Tellingly, the neural selectivity evoked implicitly in categorization aligns closely with activation driven by explicit, sustained selective attention to other sounds presented in the same frequency band. Additionally, we observe fingerprints of individual differences in the learning trajectories taken to achieve expert-level categorization in patterns of neural activity associated with the diagnostic dimension. In all, this indicates that acquiring categories can drive the emergence of acquired attentional salience to dimensions of acoustic input.

Fast frequency modulation is encoded according to the listener expectations in the human subcortical auditory pathway

Abstract Expectations aid and bias our perception. For instance, expected words are easier to recognise than unexpected words, particularly in noisy environments, and incorrect expectations can make us misunderstand our conversational partner. Expectations are combined with the output from the sensory pathways to form representations of auditory objects in the cerebral cortex. Previous literature has shown that expectations propagate further down to subcortical stations during the encoding of static pure tones. However, it is unclear whether expectations also drive the subcortical encoding of subtle dynamic elements of the acoustic signal that are not represented in the tonotopic axis. Here, we tested the hypothesis that subjective expectations drive the encoding of fast frequency modulation (FM) in the human subcortical auditory pathway. We used fMRI to measure neural responses in the human auditory midbrain (inferior colliculus) and thalamus (medial geniculate body). Participants listened to sequences of FM-sweeps for which they held different expectations based on the task instructions. We found robust evidence that the responses in auditory midbrain and thalamus encode the difference between the acoustic input and the subjective expectations of the listener. The results indicate that FM-sweeps are already encoded at the level of the human auditory midbrain and that encoding is mainly driven by subjective expectations. We conclude that the subcortical auditory pathway is integrated in the cortical network of predictive processing and that expectations are used to optimise the encoding of fast dynamic elements of the acoustic signal.

Coronal underspecification as an emerging property in the development of speech processing.

Is the developing lexicon phonologically detailed or are representations underspecified? Experimental results from toddlers suggest phonological specificity. By contrast, the featurally underspecified lexicon theory (Lahiri, 2018; Lahiri & Reetz, 2010), motivated by evidence such as the cross-linguistic prevalence of phenomena such as coronal assimilation (rainbow → rai[m]bow), proposes that coronal sounds are unspecified for place of articulation even in the adult lexicon. The featurally underspecified lexicon, therefore, predicts that asymmetries in mispronunciation sensitivity are also present in the developing lexicon. Recent research (Ren et al., 2019) has rejected this, reporting similar sensitivity to mispronunciation of coronals and noncoronals at 19 months. Using a more sensitive experimental paradigm, we provide new evidence demonstrating a lack of asymmetries at 18 months, but mispronunciation sensitivity for coronals disappears by 24 months. In an intermodal preferential looking study, growth curve analysis shows that 18-month-olds are sensitive to mispronunciations of words with a coronal (e.g., duck vs. *buck) and noncoronal (e.g., bird vs. *dird) onset. At 24 months, mispronunciations of coronal-onset words were treated just like the accurate pronunciations. We conclude that coronals are underspecified in the developing lexicon at 24 months. We propose a model under which initial representations are phonetic in nature and require exact acoustic input, whereas phonological coronal underspecification at the lexical level emerges gradually as a result of exposure to variation in the input such as coronal assimilations that only become detectable patterns with growing lexical and segmentation skills. (PsycInfo Database Record (c) 2024 APA, all rights reserved).

Simulation of Piezoelectric Transducer Microphone Diaphragm Based on Different Materials

Piezoelectric microphone which utilizes MEMS technology is a type of transducer that converts an input acoustic signal into an output electrical signal. The characteristics of the microphone diaphragm such as the diaphragm design features and the type of piezoelectric materials used will affect the performance of the microphone in terms of sensitivity. It is hard to control the stress of the diaphragm used in the MEMS transducer microphone. A modification of the diaphragm is done in this project to reduce the residual stress of the piezoelectric transducer. In addition, finite element analysis namely structural, modal and harmonic were carried out using Ansys 15.0 to simulate the mechanical and dynamic behaviour of the microphone diaphragm. Two types of diaphragm structure were designed, namely square and circular, while three types of piezoelectric material which are AlN, PZT and ZnO were used as the diaphragm material. The structural analysis findings of the diaphragm subjected to 1 Pa pressure revealed that the circular diaphragm made of AlN material exhibited the highest stress, reaching 43.05 GPa, surpassing the stresses observed in the other two materials. On the contrary, the square diaphragm composed of PZT material demonstrated the lowest stress, with only 1.55 GPa. In terms of resonance frequency, the circular AlN diaphragm achieved the highest resonant frequency, reaching 449.84 kHz, whereas the square PZT diaphragm exhibited the lowest frequency at 200.25 kHz. In general, the circular diaphragm design consistently yielded higher first resonant frequencies compared to the square design.The results show that the circular diaphragm with AlN piezoelectric materials is the ideal diaphragm in the microphone because of the highest stress generated and the first resonant frequency. The stress is related to the sensitivity of a microphone while the high resonant frequency can lead to the better optimization of signal to noise ratio control.

An introduction to acoustic location Template-based positioning model

Indoor location information, which is an indispensable label for linking the vast amount of data circulating in the world of IoT, 5G, and AI, is essential. This paper proposes a novel template pattern matching-based acoustic positioning method, Location Template-based Positioning Model, for indoor acoustic source localisation. This innovative technology, similar to the passive SONAR systems on submarines, uses audible acoustic waves to locate acoustic sources in complex indoor environments. The proposed LTPM is a statistical model, and it calculates the coordinates of acoustic sources by matching input acoustic signals with template acoustic signals that are pre-collected from preset locations. A comprehensive overview of the principles and implementations of LTPM is provided, and its feasibility is verified through preliminary positioning tests in this paper. It is hoped to provide research references for the advancement of pattern matching-based acoustic indoor positioning technologies.

Neural processing in the primary auditory cortex following cholinergic lesions of the basal forebrain in ferrets

Cortical acetylcholine (ACh) release has been linked to various cognitive functions, including perceptual learning. We have previously shown that cortical cholinergic innervation is necessary for accurate sound localization in ferrets, as well as for their ability to adapt with training to altered spatial cues. To explore whether these behavioral deficits are associated with changes in the response properties of cortical neurons, we recorded neural activity in the primary auditory cortex (A1) of anesthetized ferrets in which cholinergic inputs had been reduced by making bilateral injections of the immunotoxin ME20.4-SAP in the nucleus basalis (NB) prior to training the animals. The pattern of spontaneous activity of A1 units recorded in the ferrets with cholinergic lesions (NB ACh–) was similar to that in controls, although the proportion of burst-type units was significantly lower. Depletion of ACh also resulted in more synchronous activity in A1. No changes in thresholds, frequency tuning or in the distribution of characteristic frequencies were found in these animals. When tested with normal acoustic inputs, the spatial sensitivity of A1 neurons in the NB ACh– ferrets and the distribution of their preferred interaural level differences also closely resembled those found in control animals, indicating that these properties had not been altered by sound localization training with one ear occluded. Simulating the animals’ previous experience with a virtual earplug in one ear reduced the contralateral preference of A1 units in both groups, but caused azimuth sensitivity to change in slightly different ways, which may reflect the modest adaptation observed in the NB ACh– group. These results show that while ACh is required for behavioral adaptation to altered spatial cues, it is not required for maintenance of the spectral and spatial response properties of A1 neurons.

Multi-Modal Attentive Prompt Learning for Few-shot Emotion Recognition in Conversations

Emotion recognition in conversations (ERC) has emerged as an important research area in Natural Language Processing and Affective Computing, focusing on accurately identifying emotions within the conversational utterance. Conventional approaches typically rely on labeled training samples for fine-tuning pre-trained language models (PLMs) to enhance classification performance. However, the limited availability of labeled data in real-world scenarios poses a significant challenge, potentially resulting in diminished model performance. In response to this challenge, we present the Multi-modal Attentive Prompt (MAP) learning framework, tailored specifically for few-shot emotion recognition in conversations. The MAP framework consists of four integral modules: multi-modal feature extraction for the sequential embedding of text, visual, and acoustic inputs; a multi-modal prompt generation module that creates six manually-designed multi-modal prompts; an attention mechanism for prompt aggregation; and an emotion inference module for emotion prediction. To evaluate our proposed model’s efficacy, we conducted extensive experiments on two widely recognized benchmark datasets, MELD and IEMOCAP. Our results demonstrate that the MAP framework outperforms state-of-the-art ERC models, yielding notable improvements of 3.5% and 0.4% in micro F1 scores. These findings highlight the MAP learning framework’s ability to effectively address the challenge of limited labeled data in emotion recognition, offering a promising strategy for improving ERC model performance.

Acute Aromatase Inhibition Impairs Neural and Behavioral Auditory Scene Analysis in Zebra Finches.

Auditory perception can be significantly disrupted by noise. To discriminate sounds from noise, auditory scene analysis (ASA) extracts the functionally relevant sounds from acoustic input. The zebra finch communicates in noisy environments. Neurons in their secondary auditory pallial cortex (caudomedial nidopallium, NCM) can encode song from background chorus, or scenes, and this capacity may aid behavioral ASA. Furthermore, song processing is modulated by the rapid synthesis of neuroestrogens when hearing conspecific song. To examine whether neuroestrogens support neural and behavioral ASA in both sexes, we retrodialyzed fadrozole (aromatase inhibitor, FAD) and recorded in vivo awake extracellular NCM responses to songs and scenes. We found that FAD affected neural encoding of songs by decreasing responsiveness and timing reliability in inhibitory (narrow-spiking), but not in excitatory (broad-spiking) neurons. Congruently, FAD decreased neural encoding of songs in scenes for both cell types, particularly in females. Behaviorally, we trained birds using operant conditioning and tested their ability to detect songs in scenes after administering FAD orally or injected bilaterally into NCM. Oral FAD increased response bias and decreased correct rejections in females, but not in males. FAD in NCM did not affect performance. Thus, FAD in the NCM impaired neuronal ASA but that did not lead to behavioral disruption suggesting the existence of resilience or compensatory responses. Moreover, impaired performance after systemic FAD suggests involvement of other aromatase-rich networks outside the auditory pathway in ASA. This work highlights how transient estrogen synthesis disruption can modulate higher-order processing in an animal model of vocal communication.

A robotic platform for integrating biomimetic sonar and AI

To gain a first-hand view of the biosonar inputs that a bat receives when operating in a complex natural environment, a bioinspired sonar head to mimic the auditory periphery of the bats' biosonar system is currently under development. The sonar head seeks to replicate the static geometric complexity of the baffles that emit the biosonar pulses and receive the returning echoes in horseshoe bats. Furthermore, some of the dynamic deformations of these structures are mimicked using tension-driven soft-robotics actuated by a set of 12 motors (two for the noseleaf and five for the each pinna). The periphery has been interfaced with acoustic emission and reception systems as well as a modular system control system for the acoustic input and output as well as the motors. Besides low-level control, the sonar head has onboard computing resources that can support deep-learning inference in the loop with acoustic data acquisition and pulse generation. All functional components have been integrated into a custom-designed shell that allows operating the system in natural outdoor environments. A wireless user interface allows remote control over the system so that the tasks of carrying and orienting the sonar head and controlling its function can be shared between two individuals.

Effect of reverberation and informational masking on speech perception by adolescents and young adults with cochlear implants

Children with severe to profound hearing loss fitted with cochlear implants (CIs) can develop speech with degraded auditory inputs. Little is known about how speech perception by CI users who developed speech through electrical hearing is affected by realistic reverberation—a source of additional degradation in acoustic inputs to the CI speech processor. Moreover, how these young CI users handle competing maskers with high informational masking, such as two-talker babble, in speech-in-noise perception is not well characterized. Both reverberation and competing maskers are relevant factors in everyday classroom listening scenarios. In this work, we studied a cohort of adolescents and young adults with early fitting of CIs. Speech reception thresholds (SRTs) were measured in five types of maskers: speech-shaped noise (SSN), SSN modulated by a speech envelope, two-talker babble in English and in French, and time-reversed two-talker English babble. SRTs from each masker type were repeated in an anechoic condition and two reverberation conditions mimicking a classroom meeting classroom acoustics standard versus a lecture hall. When compared with a group of age-matched normal-hearing listeners, young CI users showed steeper increase of masked SRTs by increasing reverberation but comparable release from informational masking. [Work supported by the Hearing Health Foundation.]

Acoustic impedance measurement head design and evaluation

Acoustic input impedance measurements are a useful tool for characterizing the behavior of a wind instrument. Methods for measuring acoustic input impedances include the two-microphone three-calibration (TMTC) technique (Gibiat and Laloë, 1990) and the multi-microphone method (Jang and Ih, 1998). We present and evaluate the design of a new impedance measurement head based on the multi-microphone method. We will discuss practical aspects such as impedance head calibration, measurement stimulus, and leak detection.

Neural response properties predict perceived contents and locations elicited by intracranial electrical stimulation of human auditory cortex.

Intracranial electrical stimulation (iES) of auditory cortex can elicit sound experiences with a variety of perceived contents (hallucination or illusion) and locations (contralateral or bilateral side), independent of actual acoustic inputs. However, the neural mechanisms underlying this elicitation heterogeneity remain undiscovered. Here, we collected subjective reports following iES at 3062 intracranial sites in 28 patients (both sexes) and identified 113 auditory cortical sites with iES-elicited sound experiences. We then decomposed the sound-induced intracranial electroencephalogram (iEEG) signals recorded from all 113 sites into time-frequency features. We found that the iES-elicited perceived contents can be predicted by the early high-γ features extracted from sound-induced iEEG. In contrast, the perceived locations elicited by stimulating hallucination sites and illusion sites are determined by the late high-γ and long-lasting α features, respectively. Our study unveils the crucial neural signatures of iES-elicited sound experiences in human and presents a new strategy to hearing restoration for individuals suffering from deafness.

The Use of Deep Learning Software in the Detection of Voice Disorders: A Systematic Review.

To summarize the use of deep learning in the detection of voice disorders using acoustic and laryngoscopic input, compare specific neural networks in terms of accuracy, and assess their effectiveness compared to expert clinical visual examination. Embase, MEDLINE, and Cochrane Central. Databases were screened through November 11, 2023 for relevant studies. The inclusion criteria required studies to utilize a specified deep learning method, use laryngoscopy or acoustic input, and measure accuracy of binary classification between healthy patients and those with voice disorders. Thirty-four studies met the inclusion criteria, with 18 focusing on voice analysis, 15 on imaging analysis, and 1 both. Across the 18 acoustic studies, 21 programs were used for identification of organic and functional voice disorders. These technologies included 10 convolutional neural networks (CNNs), 6 multilayer perceptrons (MLPs), and 5 other neural networks. The binary classification systems yielded a mean accuracy of 89.0% overall, including 93.7% for MLP programs and 84.5% for CNNs. Among the 15 imaging analysis studies, a total of 23 programs were utilized, resulting in a mean accuracy of 91.3%. Specifically, the twenty CNNs achieved a mean accuracy of 92.6% compared to 83.0% for the 3 MLPs. Deep learning models were shown to be highly accurate in the detection of voice pathology, with CNNs most effective for assessing laryngoscopy images and MLPs most effective for assessing acoustic input. While deep learning methods outperformed expert clinical exam in limited comparisons, further studies integrating external validation are necessary.

Method of Underwater Acoustic Signal Denoising Based on Dual-Path Transformer Network

The presence of natural ambient noise interferes with the system for locating and identifying underwater targets. This paper suggests that a Dual-Path Transformation Network (DPTN) reduces ambient noise in underwater acoustic signals. First, the input acoustic signals’ higher-order non-linear features are extracted using a multi-scale convolutional encoder neural network. Second, sub-vectors with the same length are created according to the time dimension from the higher-order non-linear features. The sub-vectors are stitched together to form a three-dimensional tensor. Third, a neural network transformer based on the feed-forward network is constructed. Further, to capture long-term series features and separate the target signal from the noisy signals, the three-dimensional tensor is used as the input of the transformer-based masking network. Finally, overlap-add and transpose are used to obtain discernible target signals. The experimental results verify the effectiveness of the proposed underwater acoustic signal denoising algorithm and demonstrate that the proposed DPRN method can obtain higher output signal-to-noise ratio (SNR) and the scale-invariant signal-to-noise ratio (SI-SNR) compared with the other classical algorithms.

Quantifying the Impact of Auditory Deafferentation on Speech Perception.

The past decade has seen a wealth of research dedicated to determining which and how morphological changes in the auditory periphery contribute to people experiencing hearing difficulties in noise despite having clinically normal audiometric thresholds in quiet. Evidence from animal studies suggests that cochlear synaptopathy in the inner ear might lead to auditory nerve deafferentation, resulting in impoverished signal transmission to the brain. Here, we quantify the likely perceptual consequences of auditory deafferentation in humans via a physiologically inspired encoding-decoding model. The encoding stage simulates the processing of an acoustic input stimulus (e.g., speech) at the auditory periphery, while the decoding stage is trained to optimally regenerate the input stimulus from the simulated auditory nerve firing data. This allowed us to quantify the effect of different degrees of auditory deafferentation by measuring the extent to which the decoded signal supported the identification of speech in quiet and in noise. In a series of experiments, speech perception thresholds in quiet and in noise increased (worsened) significantly as a function of the degree of auditory deafferentation for modeled deafferentation greater than 90%. Importantly, this effect was significantly stronger in a noisy than in a quiet background. The encoding-decoding model thus captured the hallmark symptom of degraded speech perception in noise together with normal speech perception in quiet. As such, the model might function as a quantitative guide to evaluating the degree of auditory deafferentation in human listeners.

The role of L2 input in developing a novel L2 contrast phonetically and phonologically: Production evidence from a residence abroad context

This study analyses the production of French /y/ and /u/ by 42 native English learners of French (ELoF) at the start and end of a Residence Abroad (RA) in a French-speaking country. As an approximation of both phonological and phonetic development, categorical change is teased apart from gradient change using k-medoid clustering of acoustic data and different input measures are tested as predictors of both types of development. Results of the phonological analysis reveal that while no change occurs for /y/, the proportion that learners correctly use their back vowel in /u/ contexts increases over the RA. The quantity of French input declared over the RA is a significant predictor of this categorical change, especially the amount of auditory and visual engagement (e.g. listening and reading in French). Results of the phonetic analysis indicate, instead, that /y/ becomes more target-like over the RA, while /u/ makes less progress, suggesting a partial mismatch between the phonological and phonetic levels. Nevertheless, the phonetic development of /u/ is more substantial for individuals who: (1) have been learning French longer, (2) have yet to experience naturalistic exposure by the start of the RA, and (3) are English language teachers over the RA rather than students at a foreign institution. Taken together, these results have implications for the link between second language (L2) input and L2 production, the assumptions of L2 speech models, and the relationship between different levels of linguistic representation in L2 speech learning.

Rhythmic modulation of prediction errors: A top-down gating role for the beta-range in speech processing.

Natural speech perception requires processing the ongoing acoustic input while keeping in mind the preceding one and predicting the next. This complex computational problem could be handled by a dynamic multi-timescale hierarchical inferential process that coordinates the information flow up and down the language network hierarchy. Using a predictive coding computational model (Precoss-β) that identifies online individual syllables from continuous speech, we address the advantage of a rhythmic modulation of up and down information flows, and whether beta oscillations could be optimal for this. In the model, and consistent with experimental data, theta and low-gamma neural frequency scales ensure syllable-tracking and phoneme-level speech encoding, respectively, while the beta rhythm is associated with inferential processes. We show that a rhythmic alternation of bottom-up and top-down processing regimes improves syllable recognition, and that optimal efficacy is reached when the alternation of bottom-up and top-down regimes, via oscillating prediction error precisions, is in the beta range (around 20-30 Hz). These results not only demonstrate the advantage of a rhythmic alternation of up- and down-going information, but also that the low-beta range is optimal given sensory analysis at theta and low-gamma scales. While specific to speech processing, the notion of alternating bottom-up and top-down processes with frequency multiplexing might generalize to other cognitive architectures.

Acoustic Radiation Prediction Model Rationality and Mechanism of Steel-Spring Floating-Slab Tracks on Bridges

In the actual operation of urban rail transit (URT), the vibrations of steel-spring floating-slab tracks (SSFSTs) are amplified, and the track structure has strong low-frequency acoustic radiation; therefore, it is necessary to study the acoustic radiation of SSFSTs. In addition, multi-block short track structures are often laid within the URT lines; however, many researchers studying the reduction of vibration track service performance problems only select one or several block tracks to study. In reality, many short track structures will become sound sources when a train passes, and different sound sources will have various acoustic effects during the propagation process; therefore, it is necessary to study the rationality of any track acoustic model that analyzes the acoustic radiation problem. In order to more accurately predict the acoustic characteristics of steel-spring short floating-slab tracks (SSSFSTs) on a one-span bridge, train-track-bridge interaction theory and the acoustic boundary element method (BEM) were adopted to study the acoustical differences and mechanism of the float-slab number in the acoustic model. The results showed that with the increase in the floating-slab number in the acoustic model, the acoustic radiation ability of SSSFSTs and the sound pressure in the sound field increased; however, it was not a simple linear increase. Thus, the floating-slab number in the acoustic model not only affected the acoustic radiation ability but also caused acoustic effects during the propagation process, which affected the predicted results. The vibration characteristics of each floating-slab were different, and the acoustic input conditions of different numbers of floating-slabs in the acoustic model led to significant differences in the acoustic analysis. There was also obviously a different acoustic contribution of each floating-slab to the same sound field point, which led to the significant influence of the sound pressure at the sound field points when using different acoustic models. Therefore, using acoustic models with different floating-slab numbers had a significant effect on the acoustic analysis of SSSFSTs. In order to study the acoustic characteristics of SSSFSTs on a one-span bridge, it was necessary to establish a complete acoustic model.

Estimation of input power to track energy propagation through a small air handling unit

Statistical energy analysis (SEA) is dependent on having accurate inputs. The inputs to an SEA model are structural and acoustic input powers. In this work, blocked force analysis is utilized to determine the structural and acoustic input powers to a small air handling unit (AHU). An electromagnetic shaker connected to a small stiff plate bolted to a panel of the AHU is used as an input force to the unit. The acoustic input consists of a calibrated volume velocity source placed inside the air handler. The inversely determined input powers are then used as inputs to the SEA model to track the structure-borne and airborne energy propagation through the small air handler. The results demonstrate that the inversely determined input powers can be used to reliably estimate inputs for the SEA model.

Lightweight modular construction of school classrooms, reverberation time prediction versus field testing

Lightweight modular schools are in the early phases of being designed and constructed throughout NSW as part of Schools Infrastructure “Modern Methods of Construction (MMC).” This method proposes the use of CLT (Cross Laminated Timber) panels, timber floor and wall cassettes, and/or hybrid systems. Schools Infrastructure NSW has developed, along with PKA Acoustic Consulting’s acoustic input, the “Kit of Parts” catalogue which utilises predetermined set of components that can easily be assembled and constructed. The classroom layout is therefore defined and can be incorporated into the acoustic predictive model. Additionally, the absorption characteristics of lightweight constructions differ to conventional construction and therefore should be considered. This presentation discusses the significance of construction type and fitout finishes, by comparing predicted reverb times of modular lightweight classrooms with in-situ absorptive testing at various cold shell, warm shell, and completed fitout stages of a school classroom and open learning space.