Constructing Noise-Invariant Representations of Sound in the Auditory Pathway

Identifying behaviorally relevant sounds in the presence of background noise is one of the most important and poorly understood challenges faced by the auditory system. An elegant solution to this problem would be for the auditory system to represent sounds in a noise-invariant fashion. Since a major effect of background noise is to alter the statistics of the sounds reaching the ear, noise-invariant representations could be promoted by neurons adapting to stimulus statistics. Here we investigated the extent of neuronal adaptation to the mean and contrast of auditory stimulation as one ascends the auditory pathway. We measured these forms of adaptation by presenting complex synthetic and natural sounds, recording neuronal responses in the inferior colliculus and primary fields of the auditory cortex of anaesthetized ferrets, and comparing these responses with a sophisticated model of the auditory nerve. We find that the strength of both forms of adaptation increases as one ascends the auditory pathway. To investigate whether this adaptation to stimulus statistics contributes to the construction of noise-invariant sound representations, we also presented complex, natural sounds embedded in stationary noise, and used a decoding approach to assess the noise tolerance of the neuronal population code. We find that the code for complex sounds in the periphery is affected more by the addition of noise than the cortical code. We also find that noise tolerance is correlated with adaptation to stimulus statistics, so that populations that show the strongest adaptation to stimulus statistics are also the most noise-tolerant. This suggests that the increase in adaptation to sound statistics from auditory nerve to midbrain to cortex is an important stage in the construction of noise-invariant sound representations in the higher auditory brain.     

Studies of directional sensitivity of neurons in the cat primary auditory cortex

A set of impulsive transient signals has been synthesized for earphone delivery whose waveform and amplitude spectra, measured at the eardrum, mimic those of sounds arriving from a free-field source. The complete stimulus set forms a "virtual acoustic space" (VAS) for the cat. VAS stimuli are delivered via calibrated earphones sealed into the external meatus in cats under barbiturate anesthesia. Neurons recorded extracellularly in primary (AI) auditory cortex exhibit sensitivity to the direction of sound in VAS. The aggregation of effective sound directions forms a virtual space receptive field (VSRF). At about 20 dB above minimal threshold, VSRFs recorded in otherwise quiet and anechoic space fall into categories based on spatial dimension and location. The size, shape and location of VSRFs remain stable over many hours of recording and are found to be shaped by excitatory and inhibitory interactions of activity arriving from the two ears. Within the VSRF response latency and strength vary systematically with stimulus direction. In an ensemble of such neurons these functional gradients provide information about stimulus direction, which closely accounts for a human listener's spatial acuity. Raising stimulus intensity, introducing continuous background noise or presenting a conditioning stimulus all influence the extent of the VSRF but leave intact the gradient structure of the field. These and other findings suggest that such functional gradients in VSRFs of ensembles of AI neurons are instrumental in coding sound direction and robust enough to overcome interference from competing environmental sounds.     

Effects of urban noise on song and response behaviour in great tits

Acoustic communication is fundamental in avian territory defence and mate attraction. In urban environments where sound transmissions are more likely to be masked by low-frequency anthropogenic noise, acoustic adaptations may be advantageous. However, minor modifications to a signal could affect its efficacy. While recent research has shown that there is divergence between songs from noisy and quiet areas, it is unknown whether these differences affect the response to the signal by its receivers. Here, we show that there is a difference in spectral aspects of rural and urban song in a common passerine, the great tit Parus major, at 20 sites across the UK. We also provide, to our knowledge, the first demonstration that such environmentally induced differences in song influence the response of male territory holders. Males from quiet territories exhibited a significantly stronger response when hearing song from another territory holder with low background noise than from those with high background noise. The opposite distinction in response intensity to homotypic versus heterotypic song was observed in males from noisy territories. This behavioural difference may intensify further signal divergence between urban and rural populations and raises important questions concerning signal evolution.     

Intrinsic gain modulation and adaptive neural coding

In many cases, the computation of a neural system can be reduced to a receptive field, or a set of linear filters, and a thresholding function, or gain curve, which determines the firing probability; this is known as a linear/nonlinear model. In some forms of sensory adaptation, these linear filters and gain curve adjust very rapidly to changes in the variance of a randomly varying driving input. An apparently similar but previously unrelated issue is the observation of gain control by background noise in cortical neurons: the slope of the firing rate versus current (f-I) curve changes with the variance of background random input. Here, we show a direct correspondence between these two observations by relating variance-dependent changes in the gain of f-I curves to characteristics of the changing empirical linear/nonlinear model obtained by sampling. In the case that the underlying system is fixed, we derive relationships relating the change of the gain with respect to both mean and variance with the receptive fields derived from reverse correlation on a white noise stimulus. Using two conductance-based model neurons that display distinct gain modulation properties through a simple change in parameters, we show that coding properties of both these models quantitatively satisfy the predicted relationships. Our results describe how both variance-dependent gain modulation and adaptive neural computation result from intrinsic nonlinearity.     

Coding of time-dependent stimuli in homogeneous and heterogeneous neural populations

We compare the information transmission of a time-dependent signal by two types of uncoupled neuron populations that differ in their sources of variability: i) a homogeneous population whose units receive independent noise and ii) a deterministic heterogeneous population, where each unit exhibits a different baseline firing rate (’disorder’). Our criterion for making both sources of variability quantitatively comparable is that the interspike-interval distributions are identical for both systems. Numerical simulations using leaky integrate-and-fire neurons unveil that a non-zero amount of both noise or disorder maximizes the encoding efficiency of the homogeneous and heterogeneous system, respectively, as a particular case of suprathreshold stochastic resonance. Our findings thus illustrate that heterogeneity can render similarly profitable effects for neuronal populations as dynamic noise. The optimal noise/disorder depends on the system size and the properties of the stimulus such as its intensity or cutoff frequency. We find that weak stimuli are better encoded by a noiseless heterogeneous population, whereas for strong stimuli a homogeneous population outperforms an equivalent heterogeneous system up to a moderate noise level. Furthermore, we derive analytical expressions of the coherence function for the cases of very strong noise and of vanishing intrinsic noise or heterogeneity, which predict the existence of an optimal noise intensity. Our results show that, depending on the type of signal, noise as well as heterogeneity can enhance the encoding performance of neuronal populations.     

Representation of spectrotemporal sound information in the ascending auditory pathway

The representation of sound information in the central nervous system relies on the analysis of time-varying features in communication and other environmental sounds. How are auditory physiologists and theoreticians to choose an appropriate method for characterizing spectral and temporal acoustic feature representations in single neurons and neural populations? A brief survey of currently available scientific methods and their potential usefulness is given, with a focus on the strengths and weaknesses of using noise analysis techniques for approximating spectrotemporal response fields (STRFs). Noise analysis has been used to foster several conceptual advances in describing neural acoustic feature representation in a variety of species and auditory nuclei. STRFs have been used to quantitatively assess spectral and temporal transformations across mutually connected auditory nuclei, to identify neuronal interactions between spectral and temporal sound dimensions, and to compare linear vs. nonlinear response properties through state-dependent comparisons. We propose that noise analysis techniques used in combination with novel stimulus paradigms and parametric experiment designs will provide powerful means of exploring acoustic feature representations in the central nervous system.     

Prepulse Inhibition of Acoustic Startle Reflex as a Function of the Frequency Difference between Prepulse and Background Sounds in Mice

Background

Prepulse inhibition (PPI) depicts the effects of a weak sound preceding strong acoustic stimulus on acoustic startle response (ASR). Previous studies suggest that PPI is influenced by physical parameters of prepulse sound such as intensity and preceding time. The present study characterizes the impact of prepulse tone frequency on PPI.

Methods

Seven female C57BL mice were used in the present study. ASR was induced by a 100 dB SPL white noise burst. After assessing the effect of background sounds (white noise and pure tones) on ASR, PPI was tested by using prepulse pure tones with the background tone of either 10 or 18 kHz. The inhibitory effect was assessed by measuring and analyzing the changes in the first peak-to-peak magnitude, root mean square value, duration and latency of the ASR as the function of frequency difference between prepulse and background tones.

Results

Our data showed that ASR magnitude with pure tone background varied with tone frequency and was smaller than that with white noise background. Prepulse tone systematically reduced ASR as the function of the difference in frequency between prepulse and background tone. The 0.5 kHz difference appeared to be a prerequisite for inducing substantial ASR inhibition. The frequency dependence of PPI was similar under either a 10 or 18 kHz background tone.

Conclusion

PPI is sensitive to frequency information of the prepulse sound. However, the critical factor is not tone frequency itself, but the frequency difference between the prepulse and background tones.     

Individual right whales call louder in increased environmental noise

The ability to modify vocalizations to compensate for environmental noise is critical for successful communication in a dynamic acoustic environment. Many marine species rely on sound for vital life functions including communication, navigation and feeding. The impacts of significant increases in ocean noise levels from human activities are a current area of concern for the conservation of marine mammals. Here, we document changes in calling behaviour by individual endangered North Atlantic right whales (Eubalaena glacialis) in increased background noise. Right whales, like several bird and primate species, respond to periods of increased noise by increasing the amplitude of their calls. This behaviour may help maintain the communication range with conspecifics during periods of increased noise. These call modifications have implications for conservation efforts for right whales, affecting both the way whales use sound to communicate and our ability to detect them with passive acoustic monitoring systems.     

Lizards speed up visual displays in noisy motion habitats

Extensive research over the last few decades has revealed that many acoustically communicating animals compensate for the masking effect of background noise by changing the structure of their signals. Familiar examples include birds using acoustic properties that enhance the transmission of vocalizations in noisy habitats. Here, we show that the effects of background noise on communication signals are not limited to the acoustic modality, and that visual noise from windblown vegetation has an equally important influence on the production of dynamic visual displays. We found that two species of Puerto Rican lizard, Anolis cristatellus and A. gundlachi, increase the speed of body movements used in territorial signalling to apparently improve communication in visually 'noisy' environments of rapidly moving vegetation. This is the first evidence that animals change how they produce dynamic visual signals when communicating in noisy motion habitats. Taken together with previous work on acoustic communication, our results show that animals with very different sensory ecologies can face similar environmental constraints and adopt remarkably similar strategies to overcome these constraints.     

Communication in a noisy environment: short-term acoustic adjustments and the underlying acoustic niche of a Neotropical stream-breeding frog

Acoustically active animals may show long- and short-term adaptations in acoustic traits for coping with ambient noise. Given the key role of calls in anurans’ life history, long- and short-term adaptations are expected in species inhabiting noisy habitats. However, to disentangle such adaptations is a difficult task, incipiently addressed for Neotropical frogs. We investigated if males of a stream-breeding frog (Crossodactylus schmidti) adjust call traits according to the background noise, and if the signal-to-noise ratio (SNR) varies between call harmonics and along call notes. We measured sound pressure levels of calls and noise in the field and used a fine-scale acoustic analysis to describe the signal and noise structure and test for noise-related call adjustments. The multi-note harmonic call of C. schmidti greatly varied in the spectral structure, including a trend for increasing note amplitude along the call, a wide frequency bandwidth of the 2nd harmonic, a minor call frequency modulation due to a trend for increasing note frequency within the same harmonic, and a major call frequency modulation due to the variable location of the dominant harmonic along the call. Calls had significantly higher frequencies than the noise at the range of the 1st and the 2nd call harmonics, and significantly louder sound pressure than the noise at the range of all harmonics. Males emitted the majority of call notes showing positive SNR, and though males also emitted some notes with negative SNR, when a given harmonic was negative the other harmonics in the same note did not tend to be SNR-negative. Our results indicate that male C. schmidti show short-term acoustic adjustments that make the advertisement call effective for coping with the interference of the stream-generated noise. We suggest that the call spectral plasticity serves for coping with temporary changes in the background noise, whilst we also discuss the possibility that the redundant, harmonic-structured call may have evolved to diminish masking interference on the acoustic signal by the background noise. This is the first study to uncouple noise-related acoustic adjustments and putative long-term acoustic adaptations for a Hylodidae, providing insights on behavioral plasticity and signal evolution of stream-breeding frogs.     

Multimodal communication in a noisy environment: a case study of the Bornean rock frog Staurois parvus

High background noise is an impediment to signal detection and perception. We report the use of multiple solutions to improve signal perception in the acoustic and visual modality by the Bornean rock frog, Staurois parvus. We discovered that vocal communication was not impaired by continuous abiotic background noise characterised by fast-flowing water. Males modified amplitude, pitch, repetition rate and duration of notes within their advertisement call. The difference in sound pressure between advertisement calls and background noise at the call dominant frequency of 5578 Hz was 8 dB, a difference sufficient for receiver detection. In addition, males used several visual signals to communicate with conspecifics with foot flagging and foot flashing being the most common and conspicuous visual displays, followed by arm waving, upright posture, crouching, and an open-mouth display. We used acoustic playback experiments to test the efficacy-based alerting signal hypothesis of multimodal communication. In support of the alerting hypothesis, we found that acoustic signals and foot flagging are functionally linked with advertisement calling preceding foot flagging. We conclude that S. parvus has solved the problem of continuous broadband low-frequency noise by both modifying its advertisement call in multiple ways and by using numerous visual signals. This is the first example of a frog using multiple acoustic and visual solutions to communicate in an environment characterised by continuous noise.     

Information filtering by synchronous spikes in a neural population

Information about time-dependent sensory stimuli is encoded by the spike trains of neurons. Here we consider a population of uncoupled but noisy neurons (each subject to some intrinsic noise) that are driven by a common broadband signal. We ask specifically how much information is encoded in the synchronous activity of the population and how this information transfer is distributed with respect to frequency bands. In order to obtain some insight into the mechanism of information filtering effects found previously in the literature, we develop a mathematical framework to calculate the coherence of the synchronous output with the common stimulus for populations of simple neuron models. Within this frame, the synchronous activity is treated as the product of filtered versions of the spike trains of a subset of neurons. We compare our results for the simple cases of (1) a Poisson neuron with a rate modulation and (2) an LIF neuron with intrinsic white current noise and a current stimulus. For the Poisson neuron, formulas are particularly simple but show only a low-pass behavior of the coherence of synchronous activity. For the LIF model, in contrast, the coherence function of the synchronous activity shows a clear peak at high frequencies, comparable to recent experimental findings. We uncover the mechanism for this shift in the maximum of the coherence and discuss some biological implications of our findings.     

An overview of fish bioacoustics and the impacts of anthropogenic sounds on fishes

Fishes use a variety of sensory systems to learn about their environments and to communicate. Of the various senses, hearing plays a particularly important role for fishes in providing information, often from great distances, from all around these animals. This information is in all three spatial dimensions, often overcoming the limitations of other senses such as vision, touch, taste and smell. Sound is used for communication between fishes, mating behaviour, the detection of prey and predators, orientation and migration and habitat selection. Thus, anything that interferes with the ability of a fish to detect and respond to biologically relevant sounds can decrease survival and fitness of individuals and populations. Since the onset of the Industrial Revolution, there has been a growing increase in the noise that humans put into the water. These anthropogenic sounds are from a wide range of sources that include shipping, sonars, construction activities (e.g., wind farms, harbours), trawling, dredging and exploration for oil and gas. Anthropogenic sounds may be sufficiently intense to result in death or mortal injury. However, anthropogenic sounds at lower levels may result in temporary hearing impairment, physiological changes including stress effects, changes in behaviour or the masking of biologically important sounds. The intent of this paper is to review the potential effects of anthropogenic sounds upon fishes, the potential consequences for populations and ecosystems and the need to develop sound exposure criteria and relevant regulations. However, assuming that many readers may not have a background in fish bioacoustics, the paper first provides information on underwater acoustics, with a focus on introducing the very important concept of particle motion, the primary acoustic stimulus for all fishes, including elasmobranchs. The paper then provides background material on fish hearing, sound production and acoustic behaviour. This is followed by an overview of what is known about effects of anthropogenic sounds on fishes and considers the current guidelines and criteria being used world-wide to assess potential effects on fishes. Most importantly, the paper provides the most complete summary of the effects of anthropogenic noise on fishes to date. It is also made clear that there are currently so many information gaps that it is almost impossible to reach clear conclusions on the nature and levels of anthropogenic sounds that have potential to cause changes in animal behaviour, or even result in physical harm. Further research is required on the responses of a range of fish species to different sound sources, under different conditions. There is a need both to examine the immediate effects of sound exposure and the longer-term effects, in terms of fitness and likely impacts upon populations.     

Latency represents sound frequency in mouse IC

Frequency is one of the fundamental parameters of sound.The frequency of an acoustic stimulus can be represented by a neural response such as spike rate,and/or first spike latency(FSL)of a given neuron.The spike rates/frequency function of most neurons changes with different acoustic ampli-tudes,whereas FSL/frequency function is highly stable.This implies that FSL might represent the fre-quency of a sound stimulus more efficiently than spike rate.This study involved representations of acoustic frequency by spike rate and FSL of central inferior colliculus(IC)neurons responding to free-field pure-tone stimuli.We found that the FSLs of neurons responding to characteristic frequency(CF)of sound stimulus were usually the shortest,regardless of sound intensity,and that spike rates of most neurons showed a variety of function according to sound frequency,especially at high intensities.These results strongly suggest that FSL of auditory IC neurons can represent sound frequency more precisely than spike rate.     

Latency represents sound frequency in mouse IC

Frequency is one of the fundamental parameters of sound. The frequency of an acoustic stimulus can be represented by a neural response such as spike rate, and/or first spike latency (FSL) of a given neuron. The spike rates/frequency function of most neurons changes with different acoustic amplitudes, whereas FSL/frequency function is highly stable. This implies that FSL might represent the frequency of a sound stimulus more efficiently than spike rate. This study involved representations of acoustic frequency by spike rate and FSL of central inferior colliculus (IC) neurons responding to free-field pure-tone stimuli. We found that the FSLs of neurons responding to characteristic frequency (CF) of sound stimulus were usually the shortest, regardless of sound intensity, and that spike rates of most neurons showed a variety of function according to sound frequency, especially at high intensities.These results strongly suggest that FSL of auditory IC neurons can represent sound frequency more precisely than spike rate.     

Correlated neuronal discharges that increase coding efficiency during perceptual discrimination

During a sensory discrimination task, the responses of multiple sensory neurons must be combined to generate a choice. The optimal combination of responses is determined both by their dependence on the sensory stimulus and by their cofluctuations across trials-that is, the noise correlations. Positively correlated noise is considered deleterious, because it limits the coding accuracy of populations of similarly tuned neurons. However, positively correlated fluctuations between differently tuned neurons actually increase coding accuracy, because they allow the common noise to be subtracted without signal loss. This is demonstrated with data recorded from the secondary somatosensory cortex of monkeys performing a vibrotactile discrimination task. The results indicate that positive correlations are not always harmful and may be exploited by cortical networks to enhance the neural representation of features to be discriminated.     

The effect of surface wave propagation on neural responses to vibration in primate glabrous skin

Because tactile perception relies on the response of large populations of receptors distributed across the skin, we seek to characterize how a mechanical deformation of the skin at one location affects the skin at another. To this end, we introduce a novel non-contact method to characterize the surface waves produced in the skin under a variety of stimulation conditions. Specifically, we deliver vibrations to the fingertip using a vibratory actuator and measure, using a laser Doppler vibrometer, the surface waves at different distances from the locus of stimulation. First, we show that a vibration applied to the fingertip travels at least the length of the finger and that the rate at which it decays is dependent on stimulus frequency. Furthermore, the resonant frequency of the skin matches the frequency at which a subpopulation of afferents, namely Pacinian afferents, is most sensitive. We show that this skin resonance can lead to a two-fold increase in the strength of the response of a simulated afferent population. Second, the rate at which vibrations propagate across the skin is dependent on the stimulus frequency and plateaus at 7 m/s. The resulting delay in neural activation across locations does not substantially blur the temporal patterning in simulated populations of afferents for frequencies less than 200 Hz, which has important implications about how vibratory frequency is encoded in the responses of somatosensory neurons. Third, we show that, despite the dependence of decay rate and propagation speed on frequency, the waveform of a complex vibration is well preserved as it travels across the skin. Our results suggest, then, that the propagation of surface waves promotes the encoding of spectrally complex vibrations as the entire neural population is exposed to essentially the same stimulus. We also discuss the implications of our results for biomechanical models of the skin.     

Lip-Reading Aids Word Recognition Most in Moderate Noise: A Bayesian Explanation Using High-Dimensional Feature Space

Watching a speaker''s facial movements can dramatically enhance our ability to comprehend words, especially in noisy environments. From a general doctrine of combining information from different sensory modalities (the principle of inverse effectiveness), one would expect that the visual signals would be most effective at the highest levels of auditory noise. In contrast, we find, in accord with a recent paper, that visual information improves performance more at intermediate levels of auditory noise than at the highest levels, and we show that a novel visual stimulus containing only temporal information does the same. We present a Bayesian model of optimal cue integration that can explain these conflicts. In this model, words are regarded as points in a multidimensional space and word recognition is a probabilistic inference process. When the dimensionality of the feature space is low, the Bayesian model predicts inverse effectiveness; when the dimensionality is high, the enhancement is maximal at intermediate auditory noise levels. When the auditory and visual stimuli differ slightly in high noise, the model makes a counterintuitive prediction: as sound quality increases, the proportion of reported words corresponding to the visual stimulus should first increase and then decrease. We confirm this prediction in a behavioral experiment. We conclude that auditory-visual speech perception obeys the same notion of optimality previously observed only for simple multisensory stimuli.     

Your attention please: increasing ambient noise levels elicits a change in communication behaviour in humpback whales (Megaptera novaeangliae)

High background noise is an important obstacle in successful signal detection and perception of an intended acoustic signal. To overcome this problem, many animals modify their acoustic signal by increasing the repetition rate, duration, amplitude or frequency range of the signal. An alternative method to ensure successful signal reception, yet to be tested in animals, involves the use of two different types of signal, where one signal type may enhance the other in periods of high background noise. Humpback whale communication signals comprise two different types: vocal signals, and surface-generated signals such as ‘breaching’ or ‘pectoral slapping’. We found that humpback whales gradually switched from primarily vocal to primarily surface-generated communication in increasing wind speeds and background noise levels, though kept both signal types in their repertoire. Vocal signals have the advantage of having higher information content but may have the disadvantage of loosing this information in a noisy environment. Surface-generated sounds have energy distributed over a greater frequency range and may be less likely to become confused in periods of high wind-generated noise but have less information content when compared with vocal sounds. Therefore, surface-generated sounds may improve detection or enhance the perception of vocal signals in a noisy environment.     

Abiotic noise,call frequency and stream-breeding anuran assemblages

Environmental noise can be an important selective force modulating signal evolution in species with acoustic communication. Many anuran species breed alongside streams; hence, the sound produced by the flowing water is an important source of noise for acoustic communication. Since calling is physiologically very expensive in anurans, and communication is essential for reproduction, we expected adaptations that reduce environmental masking effects and allow acoustic communication in streamside breeders. This basic assumption of the acoustic adaptation hypothesis has not been yet evaluated at a large phylogenetic scale. We combined ahistorical and phylogenetic methods to test whether anuran species that breed alongside streams call at higher frequencies than species that breed away from streams. We compiled primary and secondary data on body size, breeding habitat, and the dominant frequency of the advertisement call for 110 species; 40 of them breed alongside streams and 70 away from streams. Call frequency was slightly higher and body size was significantly smaller in streamside breeding species. After controlling for the effects of body size and phylogenetic signal, only differences in body size persisted between species breeding at both kinds of habitats. Our data suggest that habitat filtering rather than acoustic adaptation explains the high call frequency of stream breeders. Species with large body size, pleiotropically constrained to utter low-frequency calls, would have succeeded less often in establishing viable populations alongside streams, due to the masking effect of low-frequency noise. Thus, small species calling at relatively high frequencies would be more common there. Although our data do not preclude adaptations to noisy habitats in some anuran species, they do not provide support for the acoustic adaptation hypothesis at a wider phylogenetic scale.