Spectrotemporal receptive fields in the inferior colliculus revealing selectivity for spectral motion in conspecific vocalizations

A computational model of auditory chirp-velocity sensitivity and amplitude-modulation tuning in inferior colliculus neurons

We demonstrate a model of chirp-velocity sensitivity in the inferior colliculus (IC) that retains the tuning to amplitude modulation (AM) that was established in earlier models. The mechanism of velocity sensitivity is sequence detection by octopus cells of the posteroventral cochlear nucleus, which have been proposed in physiological studies to respond preferentially to the order of arrival of cross-frequency inputs of different amplitudes. Model architecture is based on coincidence detection of a combination of excitatory and inhibitory inputs. Chirp-sensitivity of the IC output is largely controlled by the strength and timing of the chirp-sensitive octopus-cell inhibitory input. AM tuning is controlled by inhibition and excitation that are tuned to the same frequency. We present several example neurons that demonstrate the feasibility of the model in simulating realistic chirp-sensitivity and AM tuning for a wide range of characteristic frequencies. Additionally, we explore the systematic impact of varying parameters on model responses. The proposed model can be used to assess the contribution of IC chirp-velocity sensitivity to responses to complex sounds, such as speech.

Humans use local spectrotemporal correlations to detect rising and falling pitch

To discern speech or appreciate music, the human auditory system detects how pitch increases or decreases over time. However, the algorithms used to detect changes in pitch, or pitch motion, are incompletely understood. Here, using psychophysics, computational modeling, functional neuroimaging, and analysis of recorded speech, we ask if humans detect pitch motion using computations analogous to those used by the visual system. We adapted stimuli from studies of vision to create novel auditory correlated noise stimuli that elicited robust pitch motion percepts. Crucially, these stimuli possess no persistent features across frequency or time, but do possess positive or negative local spectrotemporal correlations in intensity. In psychophysical experiments, we found clear evidence that humans judge pitch direction based on both positive and negative spectrotemporal correlations. The observed sensitivity to negative correlations is a direct analogue of illusory "reverse-phi" motion in vision, and thus constitutes a new auditory illusion. Our behavioral results and computational modeling led us to hypothesize that human auditory processing employs pitch direction opponency. fMRI measurements in auditory cortex supported this hypothesis. To link our psychophysical findings to real-world pitch perception, we analyzed recordings of English and Mandarin speech and discovered that pitch direction was robustly signaled by the same positive and negative spectrotemporal correlations used in our psychophysical tests, suggesting that sensitivity to both positive and negative correlations confers ecological benefits. Overall, this work reveals that motion detection algorithms sensitive to local correlations are deployed by the central nervous system across disparate modalities (vision and audition) and dimensions (space and frequency).

A Computational Model of Auditory Chirp-Velocity Sensitivity and Amplitude-Modulation Tuning in Inferior Colliculus Neurons

We demonstrate a model of chirp-velocity sensitivity in the inferior colliculus (IC) that retains the tuning to amplitude modulation (AM) that was established in earlier models. The mechanism of velocity sensitivity is sequence detection by octopus cells of the posteroventral cochlear nucleus, which have been proposed in physiological studies to respond preferentially to the order of arrival of cross-frequency inputs of different amplitudes. Model architecture is based on coincidence detection of a combination of excitatory and inhibitory inputs. Chirp-sensitivity of the IC output is largely controlled by the strength and timing of the chirp-sensitive octopus-cell inhibitory input. AM tuning is controlled by inhibition and excitation that are tuned to the same frequency. We present several example neurons that demonstrate the feasibility of the model in simulating realistic chirp-sensitivity and AM tuning for a wide range of characteristic frequencies. Additionally, we explore the systematic impact of varying parameters on model responses. The proposed model can be used to assess the contribution of IC chirp-velocity sensitivity to responses to complex sounds, such as speech.

Identifying neuron types and circuit mechanisms in the auditory midbrain

The inferior colliculus (IC) is a critical computational hub in the central auditory pathway. From its position in the midbrain, the IC receives nearly all the ascending output from the lower auditory brainstem and provides the main source of auditory information to the thalamocortical system. In addition to being a crossroads for auditory circuits, the IC is rich with local circuits and contains more than five times as many neurons as the nuclei of the lower auditory brainstem combined. These results hint at the enormous computational power of the IC, and indeed, systems-level studies have identified numerous important transformations in sound coding that occur in the IC. However, despite decades of effort, the cellular mechanisms underlying IC computations and how these computations change following hearing loss have remained largely impenetrable. In this review, we argue that this challenge persists due to the surprisingly difficult problem of identifying the neuron types and circuit motifs that comprise the IC. After summarizing the extensive evidence pointing to a diversity of neuron types in the IC, we highlight the successes of recent efforts to parse this complexity using molecular markers to define neuron types. We conclude by arguing that the discovery of molecularly identifiable neuron types ushers in a new era for IC research marked by molecularly targeted recordings and manipulations. We propose that the ability to reproducibly investigate IC circuits at the neuronal level will lead to rapid advances in understanding the fundamental mechanisms driving IC computations and how these mechanisms shift following hearing loss.

Sensitivity to direction and velocity of fast frequency chirps in the inferior colliculus of awake rabbit

Neurons in the mammalian inferior colliculus (IC) are sensitive to the velocity (speed and direction) of fast frequency chirps contained in Schroeder-phase harmonic complexes (SCHR). However, IC neurons are also sensitive to stimulus periodicity, a prominent feature of SCHR stimuli. Here, to disentangle velocity sensitivity from periodicity tuning, we introduced a novel stimulus consisting of aperiodic random chirps. Extracellular, single-unit recordings were made in the IC of Dutch-belted rabbits in response to both SCHR and aperiodic chirps. Rate-velocity functions were constructed from aperiodic-chirp responses and compared to SCHR rate profiles, revealing interactions between stimulus periodicity and neural velocity sensitivity. A generalized linear model analysis demonstrated that periodicity tuning influences SCHR response rates more strongly than velocity sensitivity. Principal component analysis of rate-velocity functions revealed that neurons were more often sensitive to the direction of lower-velocity chirps and were less often sensitive to the direction of higher-velocity chirps. Overall, these results demonstrate that sensitivity to chirp velocity is common in the IC. Harmonic sounds with complex phase spectra, such as speech and music, contain chirps, and velocity sensitivity would shape IC responses to these sounds.

A time-causal and time-recursive scale-covariant scale-space representation of temporal signals and past time

This article presents an overview of a theory for performing temporal smoothing on temporal signals in such a way that: (i) temporally smoothed signals at coarser temporal scales are guaranteed to constitute simplifications of corresponding temporally smoothed signals at any finer temporal scale (including the original signal) and (ii) the temporal smoothing process is both time-causal and time-recursive, in the sense that it does not require access to future information and can be performed with no other temporal memory buffer of the past than the resulting smoothed temporal scale-space representations themselves. For specific subsets of parameter settings for the classes of linear and shift-invariant temporal smoothing operators that obey this property, it is shown how temporal scale covariance can be additionally obtained, guaranteeing that if the temporal input signal is rescaled by a uniform temporal scaling factor, then also the resulting temporal scale-space representations of the rescaled temporal signal will constitute mere rescalings of the temporal scale-space representations of the original input signal, complemented by a shift along the temporal scale dimension. The resulting time-causal limit kernel that obeys this property constitutes a canonical temporal kernel for processing temporal signals in real-time scenarios when the regular Gaussian kernel cannot be used, because of its non-causal access to information from the future, and we cannot additionally require the temporal smoothing process to comprise a complementary memory of the past beyond the information contained in the temporal smoothing process itself, which in this way also serves as a multi-scale temporal memory of the past. We describe how the time-causal limit kernel relates to previously used temporal models, such as Koenderink's scale-time kernels and the ex-Gaussian kernel. We do also give an overview of how the time-causal limit kernel can be used for modelling the temporal processing in models for spatio-temporal and spectro-temporal receptive fields, and how it more generally has a high potential for modelling neural temporal response functions in a purely time-causal and time-recursive way, that can also handle phenomena at multiple temporal scales in a theoretically well-founded manner. We detail how this theory can be efficiently implemented for discrete data, in terms of a set of recursive filters coupled in cascade. Hence, the theory is generally applicable for both: (i) modelling continuous temporal phenomena over multiple temporal scales and (ii) digital processing of measured temporal signals in real time. We conclude by stating implications of the theory for modelling temporal phenomena in biological, perceptual, neural and memory processes by mathematical models, as well as implications regarding the philosophy of time and perceptual agents. Specifically, we propose that for A-type theories of time, as well as for perceptual agents, the notion of a non-infinitesimal inner temporal scale of the temporal receptive fields has to be included in representations of the present, where the inherent nonzero temporal delay of such time-causal receptive fields implies a need for incorporating predictions from the actual time-delayed present in the layers of a perceptual hierarchy, to make it possible for a representation of the perceptual present to constitute a representation of the environment with timing properties closer to the actual present.

Quantitative models of auditory cortical processing

To generate insight from experimental data, it is critical to understand the inter-relationships between individual data points and place them in context within a structured framework. Quantitative modeling can provide the scaffolding for such an endeavor. Our main objective in this review is to provide a primer on the range of quantitative tools available to experimental auditory neuroscientists. Quantitative modeling is advantageous because it can provide a compact summary of observed data, make underlying assumptions explicit, and generate predictions for future experiments. Quantitative models may be developed to characterize or fit observed data, to test theories of how a task may be solved by neural circuits, to determine how observed biophysical details might contribute to measured activity patterns, or to predict how an experimental manipulation would affect neural activity. In complexity, quantitative models can range from those that are highly biophysically realistic and that include detailed simulations at the level of individual synapses, to those that use abstract and simplified neuron models to simulate entire networks. Here, we survey the landscape of recently developed models of auditory cortical processing, highlighting a small selection of models to demonstrate how they help generate insight into the mechanisms of auditory processing. We discuss examples ranging from models that use details of synaptic properties to explain the temporal pattern of cortical responses to those that use modern deep neural networks to gain insight into human fMRI data. We conclude by discussing a biologically realistic and interpretable model that our laboratory has developed to explore aspects of vocalization categorization in the auditory pathway.

Two stages of bandwidth scaling drives efficient neural coding of natural sounds

Theories of efficient coding propose that the auditory system is optimized for the statistical structure of natural sounds, yet the transformations underlying optimal acoustic representations are not well understood. Using a database of natural sounds including human speech and a physiologically-inspired auditory model, we explore the consequences of peripheral (cochlear) and mid-level (auditory midbrain) filter tuning transformations on the representation of natural sound spectra and modulation statistics. Whereas Fourier-based sound decompositions have constant time-frequency resolution at all frequencies, cochlear and auditory midbrain filters bandwidths increase proportional to the filter center frequency. This form of bandwidth scaling produces a systematic decrease in spectral resolution and increase in temporal resolution with increasing frequency. Here we demonstrate that cochlear bandwidth scaling produces a frequency-dependent gain that counteracts the tendency of natural sound power to decrease with frequency, resulting in a whitened output representation. Similarly, bandwidth scaling in mid-level auditory filters further enhances the representation of natural sounds by producing a whitened modulation power spectrum (MPS) with higher modulation entropy than both the cochlear outputs and the conventional Fourier MPS. These findings suggest that the tuning characteristics of the peripheral and mid-level auditory system together produce a whitened output representation in three dimensions (frequency, temporal and spectral modulation) that reduces redundancies and allows for a more efficient use of neural resources. This hierarchical multi-stage tuning strategy is thus likely optimized to extract available information and may underlies perceptual sensitivity to natural sounds.

The prefrontal cortex of the Mexican free-tailed bat is more selective to communication calls than primary auditory cortex

In this study, we examined the auditory responses of a prefrontal area, the frontal auditory field (FAF), of an echolocating bat (Tadarida brasiliensis) and presented a comparative analysis of the neuronal response properties between the FAF and the primary auditory cortex (A1). We compared single-unit responses from the A1 and the FAF elicited by pure tones, downward frequency-modulated sweeps (dFMs), and species-specific vocalizations. Unlike the A1, FAFs were not frequency tuned. However, progressive increases in dFM sweep rate elicited a systematic increase of response precision, a phenomenon that does not take place in the A1. Call selectivity was higher in the FAF versus A1. We calculated the neuronal spectrotemporal receptive fields (STRFs) and spike-triggered averages (STAs) to predict responses to the communication calls and provide an explanation for the differences in call selectivity between the FAF and A1. In the A1, we found a high correlation between predicted and evoked responses. However, we did not generate reasonable STRFs in the FAF, and the prediction based on the STAs showed lower correlation coefficient than that of the A1. This suggests nonlinear response properties in the FAF that are stronger than the linear response properties in the A1. Stimulating with a call sequence increased call selectivity in the A1, but it remained unchanged in the FAF. These data are consistent with a role for the FAF in assessing distinctive acoustic features downstream of A1, similar to the role proposed for primate ventrolateral prefrontal cortex.NEW & NOTEWORTHY In this study, we examined the neuronal responses of a frontal cortical area in an echolocating bat to behaviorally relevant acoustic stimuli and compared them with those in the primary auditory cortex (A1). In contrast to the A1, neurons in the bat frontal auditory field are not frequency tuned but showed a higher selectivity for social signals such as communication calls. The results presented here indicate that the frontal auditory field may represent an additional processing center for behaviorally relevant sounds.

Tuned in to communication sounds: Neuronal sensitivity in the túngara frog midbrain to frequency modulated signals

For complex communication signals, it is often difficult to identify the information-bearing elements and their parameters necessary to elicit functional behavior. Consequently, it may be difficult to design stimuli that test how neurons contribute to communicative processing. For túngara frogs (Physalaemus pustulosus), however, previous behavioral testing with numerous stimuli showed that a particular frequency modulated (FM) transition in the male call is required to elicit phonotaxis and vocal responses. Modeled on such behavioral experiments, we used awake in vivo recordings of single units in the midbrain to determine if their excitation was biased to behaviorally important FM parameters. Comparisons of stimulus driven action potentials revealed greatest excitation to the behaviorally important FM transition: a downward FM sweep or step that crosses ~600 Hz. Previous studies using long-duration acoustic exposure found immediate early gene expression in many midbrain neurons to be most sensitive to similar FM. However, those data could not determine if FM coding was accomplished by the population and/or individual neurons. Our data suggest both coding schemes could operate, as 1) individual neurons are more sensitive to the behaviorally significant FM transition and 2) when single unit recordings are analytically combined across cells, the combined code can produce high stimulus discrimination (FM vs. noise driven excitation), approaching that found in behavioral discrimination of call vs. noise.

5-HT1A Receptors Alter Temporal Responses to Broadband Vocalizations in the Mouse Inferior Colliculus Through Response Suppression

Neuromodulatory systems may provide information on social context to auditory brain regions, but relatively few studies have assessed the effects of neuromodulation on auditory responses to acoustic social signals. To address this issue, we measured the influence of the serotonergic system on the responses of neurons in a mouse auditory midbrain nucleus, the inferior colliculus (IC), to vocal signals. Broadband vocalizations (BBVs) are human-audible signals produced by mice in distress as well as by female mice in opposite-sex interactions. The production of BBVs is context-dependent in that they are produced both at early stages of interactions as females physically reject males and at later stages as males mount females. Serotonin in the IC of males corresponds to these events, and is elevated more in males that experience less female rejection. We measured the responses of single IC neurons to five recorded examples of BBVs in anesthetized mice. We then locally activated the 5-HT1A receptor through iontophoretic application of 8-OH-DPAT. IC neurons showed little selectivity for different BBVs, but spike trains were characterized by local regions of high spike probability, which we called "response features." Response features varied across neurons and also across calls for individual neurons, ranging from 1 to 7 response features for responses of single neurons to single calls. 8-OH-DPAT suppressed spikes and also reduced the numbers of response features. The weakest response features were the most likely to disappear, suggestive of an "iceberg"-like effect in which activation of the 5-HT1A receptor suppressed weakly suprathreshold response features below the spiking threshold. Because serotonin in the IC is more likely to be elevated for mounting-associated BBVs than for rejection-associated BBVs, these effects of the 5-HT1A receptor could contribute to the differential auditory processing of BBVs in different behavioral subcontexts.

Efficient encoding of spectrotemporal information for bat echolocation

In most animals, natural stimuli are characterized by a high degree of redundancy, limiting the ensemble of ecologically valid stimuli to a significantly reduced subspace of the representation space. Neural encodings can exploit this redundancy and increase sensing efficiency by generating low-dimensional representations that retain all information essential to support behavior. In this study, we investigate whether such an efficient encoding can be found to support a broad range of echolocation tasks in bats. Starting from an ensemble of echo signals collected with a biomimetic sonar system in natural indoor and outdoor environments, we use independent component analysis to derive a low-dimensional encoding of the output of a cochlear model. We show that this compressive encoding retains all essential information. To this end, we simulate a range of psycho-acoustic experiments with bats. In these simulations, we train a set of neural networks to use the encoded echoes as input while performing the experiments. The results show that the neural networks’ performance is at least as good as that of the bats. We conclude that our results indicate that efficient encoding of echo information is feasible and, given its many advantages, very likely to be employed by bats. Previous studies have demonstrated that low-dimensional encodings allow for task resolution at a relatively high level. In contrast to previous work in this area, we show that high performance can also be achieved when low-dimensional filters are derived from a data set of realistic echo signals, not tailored to specific experimental conditions.

We show that complex (and simple) echoes from real environments can be efficiently and effectively represented using a small set of filters. Critically, we show that high performance across a range of tasks can be achieved when low-dimensional filters are derived from a data set of realistic echo signals, not tailored to specific experimental conditions. The redundancy in echoic information opens up the opportunity for efficient encoding, reducing the computational load of echo processing as well as the memory load for storing the information. Therefore, we predict the auditory system of bats to capitalize on this opportunity for efficient coding by implementing filters with spectrotemporal properties akin to those hypothesized here. Indeed, the filters we obtain here are similar to those found in other animals and other sensing capabilities. Our results indicate that bats could exploit the redundancy in sonar signals to implement an efficient neural encoding of the relevant information.

Natural Statistics as Inference Principles of Auditory Tuning in Biological and Artificial Midbrain Networks

Bats provide a powerful mammalian model to explore the neural representation of complex sounds, as they rely on hearing to survive in their environment. The inferior colliculus (IC) is a central hub of the auditory system that receives converging projections from the ascending pathway and descending inputs from auditory cortex. In this work, we build an artificial neural network to replicate auditory characteristics in IC neurons of the big brown bat. We first test the hypothesis that spectro-temporal tuning of IC neurons is optimized to represent the natural statistics of conspecific vocalizations. We estimate spectro-temporal receptive fields (STRFs) of IC neurons and compare tuning characteristics to statistics of bat calls. The results indicate that the FM tuning of IC neurons is matched with the statistics. Then, we investigate this hypothesis on the network optimized to represent natural sound statistics and to compare its output with biological responses. We also estimate biomimetic STRFs from the artificial network and correlate their characteristics to those of biological neurons. Tuning properties of both biological and artificial neurons reveal strong agreement along both spectral and temporal dimensions, and suggest the presence of nonlinearity, sparsity, and complexity constraints that underlie the neural representation in the auditory midbrain. Additionally, the artificial neurons replicate IC neural activities in discrimination of social calls, and provide simulated results for a noise robust discrimination. In this way, the biomimetic network allows us to infer the neural mechanisms by which the bat’s IC processes natural sounds used to construct the auditory scene.

Recurrent network dynamics shape direction selectivity in primary auditory cortex

Detecting the direction of frequency modulation (FM) is essential for vocal communication in both animals and humans. Direction-selective firing of neurons in the primary auditory cortex (A1) has been classically attributed to temporal offsets between feedforward excitatory and inhibitory inputs. However, it remains unclear how cortical recurrent circuitry contributes to this computation. Here, we used two-photon calcium imaging and whole-cell recordings in awake mice to demonstrate that direction selectivity is not caused by temporal offsets between synaptic currents, but by an asymmetry in total synaptic charge between preferred and non-preferred directions. Inactivation of cortical somatostatin-expressing interneurons (SOM cells) reduced direction selectivity, revealing its cortical contribution. Our theoretical models showed that charge asymmetry arises due to broad spatial topography of SOM cell-mediated inhibition which regulates signal amplification in strongly recurrent circuitry. Together, our findings reveal a major contribution of recurrent network dynamics in shaping cortical tuning to behaviorally relevant complex sounds.

Auditory processing in the zebra finch midbrain: single unit responses and effect of rearing experience

In birds the auditory system plays a key role in providing the sensory input used to discriminate between conspecific and heterospecific vocal signals. In those species that are known to learn their vocalizations, for example, songbirds, it is generally considered that this ability arises and is manifest in the forebrain, although there is no a priori reason why brainstem components of the auditory system could not also play an important part. To test this assumption, we used groups of normal reared and cross-fostered zebra finches that had previously been shown in behavioural experiments to reduce their preference for conspecific songs subsequent to cross fostering experience with Bengalese finches, a related species with a distinctly different song. The question we asked, therefore, is whether this experiential change also changes the bias in favour of conspecific song displayed by auditory midbrain units of normally raised zebra finches. By recording the responses of single units in MLd to a variety of zebra finch and Bengalese finch songs in both normally reared and cross-fostered zebra finches, we provide a positive answer to this question. That is, the difference in response to conspecific and heterospecific songs seen in normal reared zebra finches is reduced following cross-fostering. In birds the virtual absence of mammalian-like cortical projections upon auditory brainstem nuclei argues against the interpretation that MLd units change, as observed in the present experiments, as a result of top-down influences on sensory processing. Instead, it appears that MLd units can be influenced significantly by sensory inputs arising directly from a change in auditory experience during development.

Functional organization of the primary auditory cortex of the free-tailed bat Tadarida brasiliensis

The Mexican free-tailed bat, Tadarida brasiliensis, is a fast-flying bat that hunts by biosonar at high altitudes in open space. The auditory periphery and ascending auditory pathways have been described in great detail for this species, but nothing is yet known about its auditory cortex. Here we describe the topographical organization of response properties in the primary auditory cortex (AC) of the Mexican free-tailed bat with emphasis on the sensitivity for FM sweeps and echo-delay tuning. Responses of 716 units to pure tones and of 373 units to FM sweeps and FM-FM pairs were recorded extracellularly using multielectrode arrays in anesthetized bats. A general tonotopy was confirmed with low frequencies represented caudally and high frequencies represented rostrally. Characteristic frequencies (CF) ranged from 15 to 70 kHz, and fifty percent of CFs fell between 20 and 30 kHz, reflecting a hyper-representation of a bandwidth corresponding to search-phase echolocation pulses. Most units showed a stronger response to downward rather than upward FM sweeps and forty percent of the neurons interspersed throughout AC (150/371) showed echo-delay sensitivity to FM-FM pairs. Overall, the results illustrate that the free-tailed bat auditory cortex is organized similarly to that of other FM-type insectivorous bats.

GABAA receptors contribute more to rate than temporal coding in the IC of awake mice

Speech is our most important form of communication, yet we have a poor understanding of how communication sounds are processed by the brain. Mice make great model organisms to study neural processing of communication sounds because of their rich repertoire of social vocalizations and because they have brain structures analogous to humans, such as the auditory midbrain nucleus inferior colliculus (IC). Although the combined roles of GABAergic and glycinergic inhibition on vocalization selectivity in the IC have been studied to a limited degree, the discrete contributions of GABAergic inhibition have only rarely been examined. In this study, we examined how GABAergic inhibition contributes to shaping responses to pure tones as well as selectivity to complex sounds in the IC of awake mice. In our set of long-latency neurons, we found that GABAergic inhibition extends the evoked firing rate range of IC neurons by lowering the baseline firing rate but maintaining the highest probability of firing rate. GABAergic inhibition also prevented IC neurons from bursting in a spontaneous state. Finally, we found that although GABAergic inhibition shaped the spectrotemporal response to vocalizations in a nonlinear fashion, it did not affect the neural code needed to discriminate vocalizations, based either on spiking patterns or on firing rate. Overall, our results emphasize that even if GABAergic inhibition generally decreases the firing rate, it does so while maintaining or extending the abilities of neurons in the IC to code the wide variety of sounds that mammals are exposed to in their daily lives.NEW & NOTEWORTHY GABAergic inhibition adds nonlinearity to neuronal response curves. This increases the neuronal range of evoked firing rate by reducing baseline firing. GABAergic inhibition prevents bursting responses from neurons in a spontaneous state, reducing noise in the temporal coding of the neuron. This could result in improved signal transmission to the cortex.

Multisensory perceptual interactions between higher-order temporal frequency signals

Naturally occurring signals in audition and touch can be complex and marked by temporal variations in frequency and amplitude. Auditory frequency sweep processing has been studied extensively; however, much less is known about sweep processing in touch because studies have primarily focused on the perception of simple sinusoidal vibrations. Given the extensive interactions between audition and touch in the frequency processing of pure tone signals, we reasoned that these senses might also interact in the processing of higher-order frequency representations like sweeps. In a series of psychophysical experiments, we characterized the influence of auditory distractors on the ability of participants to discriminate tactile frequency sweeps. Auditory frequency sweeps systematically biased the tactile perception of sweep direction. Importantly, auditory cues exerted little influence on tactile sweep direction perception when the sounds and vibrations occupied different absolute frequency ranges or when the sounds consisted of intensity sweeps. Thus, audition and touch interact in frequency sweep perception in a frequency- and feature-specific manner. Our results demonstrate that audio-tactile interactions are not constrained to the processing of simple sinusoids. Because higher-order frequency representations may be synthesized from simpler representations, our findings imply that multisensory interactions in the temporal frequency domain span multiple hierarchical levels in sensory processing. (PsycINFO Database Record (c) 2019 APA, all rights reserved).

A temporal integration mechanism enhances frequency selectivity of broadband inputs to inferior colliculus

Accurately resolving frequency components in sounds is essential for sound recognition, yet there is little direct evidence for how frequency selectivity is preserved or newly created across auditory structures. We demonstrate that prepotentials (PPs) with physiological properties resembling presynaptic potentials from broadly tuned brainstem inputs can be recorded concurrently with postsynaptic action potentials in inferior colliculus (IC). These putative brainstem inputs (PBIs) are broadly tuned and exhibit delayed and spectrally interleaved excitation and inhibition not present in the simultaneously recorded IC neurons (ICNs). A sharpening of tuning is accomplished locally at the expense of spike-timing precision through nonlinear temporal integration of broadband inputs. A neuron model replicates the finding and demonstrates that temporal integration alone can degrade timing precision while enhancing frequency tuning through interference of spectrally in- and out-of-phase inputs. These findings suggest that, in contrast to current models that require local inhibition, frequency selectivity can be sharpened through temporal integration, thus supporting an alternative computational strategy to quickly refine frequency selectivity.

Processing of fast amplitude modulations in bat auditory cortex matches communication call-specific sound features

Bats use a large repertoire of calls for social communication. In the bat Phyllostomus discolor, social communication calls are often characterized by sinusoidal amplitude and frequency modulations with modulation frequencies in the range of 100-130 Hz. However, peaks in mammalian auditory cortical modulation transfer functions are typically limited to modulation frequencies below 100 Hz. We investigated the coding of sinusoidally amplitude modulated sounds in auditory cortical neurons in P. discolor by constructing rate and temporal modulation transfer functions. Neuronal responses to playbacks of various communication calls were additionally recorded and compared with the neurons' responses to sinusoidally amplitude-modulated sounds. Cortical neurons in the posterior dorsal field of the auditory cortex were tuned to unusually high modulation frequencies: rate modulation transfer functions often peaked around 130 Hz (median: 87 Hz), and the median of the highest modulation frequency that evoked significant phase-locking was also 130 Hz. Both values are much higher than reported from the auditory cortex of other mammals, with more than 51% of the units preferring modulation frequencies exceeding 100 Hz. Conspicuously, the fast modulations preferred by the neurons match the fast amplitude and frequency modulations of prosocial, and mostly of aggressive, communication calls in P. discolor. We suggest that the preference for fast amplitude modulations in the P. discolor dorsal auditory cortex serves to reliably encode the fast modulations seen in their communication calls. NEW & NOTEWORTHY Neural processing of temporal sound features is crucial for the analysis of communication calls. In bats, these calls are often characterized by fast temporal envelope modulations. Because auditory cortex neurons typically encode only low modulation frequencies, it is unclear how species-specific vocalizations are cortically processed. We show that auditory cortex neurons in the bat Phyllostomus discolor encode fast temporal envelope modulations. This property improves response specificity to communication calls and thus might support species-specific communication.

A hierarchical sparse coding model predicts acoustic feature encoding in both auditory midbrain and cortex

The auditory pathway consists of multiple stages, from the cochlear nucleus to the auditory cortex. Neurons acting at different stages have different functions and exhibit different response properties. It is unclear whether these stages share a common encoding mechanism. We trained an unsupervised deep learning model consisting of alternating sparse coding and max pooling layers on cochleogram-filtered human speech. Evaluation of the response properties revealed that computing units in lower layers exhibited spectro-temporal receptive fields (STRFs) similar to those of inferior colliculus neurons measured in physiological experiments, including properties such as sound onset and termination, checkerboard pattern, and spectral motion. Units in upper layers tended to be tuned to phonetic features such as plosivity and nasality, resembling the results of field recording in human auditory cortex. Variation of the sparseness level of the units in each higher layer revealed a positive correlation between the sparseness level and the strength of phonetic feature encoding. The activities of the units in the top layer, but not other layers, correlated with the dynamics of the first two formants (F1, F2) of all phonemes, indicating the encoding of phoneme dynamics in these units. These results suggest that the principles of sparse coding and max pooling may be universal in the human auditory pathway.

When speech enters the ear, it is subjected to a series of processing stages prior to arriving at the auditory cortex. Neurons acting at different processing stages have different response properties. For example, at the auditory midbrain, a neuron may specifically detect the onsets of a frequency component in the speech, whereas in the auditory cortex, a neuron may specifically detect phonetic features. The encoding mechanisms underlying these neuronal functions remain unclear. To address this issue, we designed a hierarchical sparse coding model, inspired by the sparse activity of neurons in the sensory system, to learn features in speech signals. We found that the computing units in different layers exhibited hierarchical extraction of speech sound features, similar to those of neurons in the auditory midbrain and auditory cortex, although the computational principles in these layers were the same. The results suggest that sparse coding and max pooling represent universal computational principles throughout the auditory pathway.

Theta band oscillations reflect more than entrainment: behavioral and neural evidence demonstrates an active chunking process

Parsing continuous acoustic streams into perceptual units is fundamental to auditory perception. Previous studies have uncovered a cortical entrainment mechanism in the delta and theta bands (~1-8 Hz) that correlates with formation of perceptual units in speech, music, and other quasi-rhythmic stimuli. Whether cortical oscillations in the delta-theta bands are passively entrained by regular acoustic patterns or play an active role in parsing the acoustic stream is debated. Here, we investigate cortical oscillations using novel stimuli with 1/f modulation spectra. These 1/f signals have no rhythmic structure but contain information over many timescales because of their broadband modulation characteristics. We chose 1/f modulation spectra with varying exponents of f, which simulate the dynamics of environmental noise, speech, vocalizations, and music. While undergoing magnetoencephalography (MEG) recording, participants listened to 1/f stimuli and detected embedded target tones. Tone detection performance varied across stimuli of different exponents and can be explained by local signal-to-noise ratio computed using a temporal window around 200 ms. Furthermore, theta band oscillations, surprisingly, were observed for all stimuli, but robust phase coherence was preferentially displayed by stimuli with exponents 1 and 1.5. We constructed an auditory processing model to quantify acoustic information on various timescales and correlated the model outputs with the neural results. We show that cortical oscillations reflect a chunking of segments, > 200 ms. These results suggest an active auditory segmentation mechanism, complementary to entrainment, operating on a timescale of ~200 ms to organize acoustic information.

Active acoustic interference elicits echolocation changes in heterospecific bats

Echolocating bats often forage in the presence of both conspecific and heterospecific individuals, which have the potential to produce acoustic interference. Recent studies have shown that at least one bat species, the Brazilian free-tailed bat (Tadarida brasiliensis), produces specialized social signals that disrupt the sonar of conspecific competitors. We herein discuss the differences between passive and active jamming signals and test whether heterospecific jamming occurs in species overlapping spatiotemporally, as well as whether such interference elicits a jamming avoidance response. We compare the capture rates of tethered moths and the echolocation parameters of big brown bats (Eptesicus fuscus) challenged with the playback of the jamming signal normally produced by Brazilian free-tailed bats and playback of deconstructed versions of this signal. There were no differences in the capture rates of targets with and without the jamming signal, although significant changes in both spectral and temporal features of the bats' echolocation were observed. These changes are consistent with improvements of the signal-to-noise ratio in the presence of acoustic interference. Accordingly, we propose to expand the traditional definition of the jamming avoidance response, stating that echolocation changes in response to interference should decrease similarity between the two signals, to include any change that increases the ability to separate returning echoes from active jamming stimuli originating from conspecific and heterospecific organisms. Flexibility in echolocation is an important characteristic for overcoming various forms of acoustic interference and may serve a purpose in interspecific interactions as well as intraspecific ones.

Temporal Coding of Voice Pitch Contours in Mandarin Tones

Accurate perception of time-variant pitch is important for speech recognition, particularly for tonal languages with different lexical tones such as Mandarin, in which different tones convey different semantic information. Previous studies reported that the auditory nerve and cochlear nucleus can encode different pitches through phase-locked neural activities. However, little is known about how the inferior colliculus (IC) encodes the time-variant periodicity pitch of natural speech. In this study, the Mandarin syllable /ba/ pronounced with four lexical tones (flat, rising, falling then rising and falling) were used as stimuli. Local field potentials (LFPs) and single neuron activity were simultaneously recorded from 90 sites within contralateral IC of six urethane-anesthetized and decerebrate guinea pigs in response to the four stimuli. Analysis of the temporal information of LFPs showed that 93% of the LFPs exhibited robust encoding of periodicity pitch. Pitch strength of LFPs derived from the autocorrelogram was significantly (p < 0.001) stronger for rising tones than flat and falling tones. Pitch strength are also significantly increased (p < 0.05) with the characteristic frequency (CF). On the other hand, only 47% (42 or 90) of single neuron activities were significantly synchronized to the fundamental frequency of the stimulus suggesting that the temporal spiking pattern of single IC neuron could encode the time variant periodicity pitch of speech robustly. The difference between the number of LFPs and single neurons that encode the time-variant F0 voice pitch supports the notion of a transition at the level of IC from direct temporal coding in the spike trains of individual neurons to other form of neural representation.

Dual Coding of Frequency Modulation in the Ventral Cochlear Nucleus

Frequency modulation (FM) is a common acoustic feature of natural sounds and is known to play a role in robust sound source recognition. Auditory neurons show precise stimulus-synchronized discharge patterns that may be used for the representation of low-rate FM. However, it remains unclear whether this representation is based on synchronization to slow temporal envelope (ENV) cues resulting from cochlear filtering or phase locking to faster temporal fine structure (TFS) cues. To investigate the plausibility of those encoding schemes, single units of the ventral cochlear nucleus of guinea pigs of either sex were recorded in response to sine FM tones centered at the unit's best frequency (BF). The results show that, in contrast to high-BF units, for modulation depths within the receptive field, low-BF units (<4 kHz) demonstrate good phase locking to TFS. For modulation depths extending beyond the receptive field, the discharge patterns follow the ENV and fluctuate at the modulation rate. The receptive field proved to be a good predictor of the ENV responses for most primary-like and chopper units. The current in vivo data also reveal a high level of diversity in responses across unit types. TFS cues are mainly conveyed by low-frequency and primary-like units and ENV cues by chopper and onset units. The diversity of responses exhibited by cochlear nucleus neurons provides a neural basis for a dual-coding scheme of FM in the brainstem based on both ENV and TFS cues.SIGNIFICANCE STATEMENT Natural sounds, including speech, convey informative temporal modulations in frequency. Understanding how the auditory system represents those frequency modulations (FM) has important implications as robust sound source recognition depends crucially on the reception of low-rate FM cues. Here, we recorded 115 single-unit responses from the ventral cochlear nucleus in response to FM and provide the first physiological evidence of a dual-coding mechanism of FM via synchronization to temporal envelope cues and phase locking to temporal fine structure cues. We also demonstrate a diversity of neural responses with different coding specializations. These results support the dual-coding scheme proposed by psychophysicists to account for FM sensitivity in humans and provide new insights on how this might be implemented in the early stages of the auditory pathway.

Cluster-based analysis improves predictive validity of spike-triggered receptive field estimates

Spectrotemporal receptive field (STRF) characterization is a central goal of auditory physiology. STRFs are often approximated by the spike-triggered average (STA), which reflects the average stimulus preceding a spike. In many cases, the raw STA is subjected to a threshold defined by gain values expected by chance. However, such correction methods have not been universally adopted, and the consequences of specific gain-thresholding approaches have not been investigated systematically. Here, we evaluate two classes of statistical correction techniques, using the resulting STRF estimates to predict responses to a novel validation stimulus. The first, more traditional technique eliminated STRF pixels (time-frequency bins) with gain values expected by chance. This correction method yielded significant increases in prediction accuracy, including when the threshold setting was optimized for each unit. The second technique was a two-step thresholding procedure wherein clusters of contiguous pixels surviving an initial gain threshold were then subjected to a cluster mass threshold based on summed pixel values. This approach significantly improved upon even the best gain-thresholding techniques. Additional analyses suggested that allowing threshold settings to vary independently for excitatory and inhibitory subfields of the STRF resulted in only marginal additional gains, at best. In summary, augmenting reverse correlation techniques with principled statistical correction choices increased prediction accuracy by over 80% for multi-unit STRFs and by over 40% for single-unit STRFs, furthering the interpretational relevance of the recovered spectrotemporal filters for auditory systems analysis.

Human Brainstem Exhibits higher Sensitivity and Specificity than Auditory-Related Cortex to Short-Term Phonetic Discrimination Learning

Phonetic discrimination learning is an active perceptual process that operates under the influence of cognitive control mechanisms by increasing the sensitivity of the auditory system to the trained stimulus attributes. It is assumed that the auditory cortex and the brainstem interact in order to refine how sounds are transcribed into neural codes. Here, we evaluated whether these two computational entities are prone to short-term functional changes, whether there is a chronological difference in malleability, and whether short-term training suffices to alter reciprocal interactions. We performed repeated cortical (i.e., mismatch negativity responses, MMN) and subcortical (i.e., frequency-following response, FFR) EEG measurements in two groups of participants who underwent one hour of phonetic discrimination training or were passively exposed to the same stimulus material. The training group showed a distinctive brainstem energy reduction in the trained frequency-range (i.e., first formant), whereas the passive group did not show any response modulation. Notably, brainstem signal change correlated with the behavioral improvement during training, this result indicating a close relationship between behavior and underlying brainstem physiology. Since we did not reveal group differences in MMN responses, results point to specific short-term brainstem changes that precede functional alterations in the auditory cortex.

Responses to Social Vocalizations in the Dorsal Cochlear Nucleus of Mice

Identifying sounds is critical for an animal to make appropriate behavioral responses to environmental stimuli, including vocalizations from conspecifics. Identification of vocalizations may be supported by neuronal selectivity in the auditory pathway. The first place in the ascending auditory pathway where neuronal selectivity to vocalizations has been found is in the inferior colliculus (IC), but very few brainstem nuclei have been evaluated. Here, we tested whether selectivity to vocalizations is present in the dorsal cochlear nucleus (DCN). We recorded extracellular neural responses in the DCN of mice and found that fusiform cells responded in a heterogeneous and selective manner to mouse ultrasonic vocalizations. Most fusiform cells responded to vocalizations that contained spectral energy at much higher frequencies than the characteristic frequencies of the cells. To understand this mismatch of stimulus properties and frequency tuning of the cells, we developed a dynamic, nonlinear model of the cochlea that simulates cochlear distortion products on the basilar membrane. We preprocessed the vocalization stimuli through this model and compared responses to these distorted vocalizations with responses to the original vocalizations. We found that fusiform cells in the DCN respond in a heterogeneous manner to vocalizations, and that these neurons can use distortion products as a mechanism for encoding ultrasonic vocalizations. In addition, the selective neuronal responses were dependent on the presence of inhibitory sidebands that modulated the response depending on the temporal structure of the distortion product. These findings suggest that important processing of complex sounds occurs at a very early stage of central auditory processing and is not strictly a function of the cortex.

Modulation of auditory brainstem responses by serotonin and specific serotonin receptors

The neuromodulator serotonin is found throughout the auditory system from the cochlea to the cortex. Although effects of serotonin have been reported at the level of single neurons in many brainstem nuclei, how these effects correspond to more integrated measures of auditory processing has not been well-explored. In the present study, we aimed to characterize the effects of serotonin on far-field auditory brainstem responses (ABR) across a wide range of stimulus frequencies and intensities. Using a mouse model, we investigated the consequences of systemic serotonin depletion, as well as the selective stimulation and suppression of the 5-HT1 and 5-HT2 receptors, on ABR latency and amplitude. Stimuli included tone pips spanning four octaves presented over a forty dB range. Depletion of serotonin reduced the ABR latencies in Wave II and later waves, suggesting that serotonergic effects occur as early as the cochlear nucleus. Further, agonists and antagonists of specific serotonergic receptors had different profiles of effects on ABR latencies and amplitudes across waves and frequencies, suggestive of distinct effects of these agents on auditory processing. Finally, most serotonergic effects were more pronounced at lower ABR frequencies, suggesting larger or more directional modulation of low-frequency processing. This is the first study to describe the effects of serotonin on ABR responses across a wide range of stimulus frequencies and amplitudes, and it presents an important step in understanding how serotonergic modulation of auditory brainstem processing may contribute to modulation of auditory perception.

Balance or imbalance: inhibitory circuits for direction selectivity in the auditory system

The auditory system detects and processes dynamic sound information transmitted in the environment. Other than the basic acoustic parameters, such as frequency, amplitude and phase, the time-varying changes of these parameters must also be encoded in our brain. Frequency-modulated (FM) sound is socially and environmentally significant, and the direction of FM sweeps is essential for animal communication and human speech. Many auditory neurons selectively respond to the directional change of such FM signals. In the past half century, our knowledge of auditory representation and processing has been updated frequently, due to technological advancement. Recently, in vivo whole-cell voltage clamp recordings have been applied to different brain regions in sensory systems. These recordings illustrate the synaptic mechanisms underlying basic sensory information processing and provide profound insights toward our understanding of neural circuits for complex signal analysis. In this review, we summarize the major findings of direction selectivity at several key auditory regions and emphasize on the recent discoveries on the synaptic mechanisms for direction selectivity in the auditory system. We conclude this review by describing promising technical developments in dissecting neural circuits and future directions in the study of complex sound analysis.

Idealized computational models for auditory receptive fields

We present a theory by which idealized models of auditory receptive fields can be derived in a principled axiomatic manner, from a set of structural properties to (i) enable invariance of receptive field responses under natural sound transformations and (ii) ensure internal consistency between spectro-temporal receptive fields at different temporal and spectral scales. For defining a time-frequency transformation of a purely temporal sound signal, it is shown that the framework allows for a new way of deriving the Gabor and Gammatone filters as well as a novel family of generalized Gammatone filters, with additional degrees of freedom to obtain different trade-offs between the spectral selectivity and the temporal delay of time-causal temporal window functions. When applied to the definition of a second-layer of receptive fields from a spectrogram, it is shown that the framework leads to two canonical families of spectro-temporal receptive fields, in terms of spectro-temporal derivatives of either spectro-temporal Gaussian kernels for non-causal time or a cascade of time-causal first-order integrators over the temporal domain and a Gaussian filter over the logspectral domain. For each filter family, the spectro-temporal receptive fields can be either separable over the time-frequency domain or be adapted to local glissando transformations that represent variations in logarithmic frequencies over time. Within each domain of either non-causal or time-causal time, these receptive field families are derived by uniqueness from the assumptions. It is demonstrated how the presented framework allows for computation of basic auditory features for audio processing and that it leads to predictions about auditory receptive fields with good qualitative similarity to biological receptive fields measured in the inferior colliculus (ICC) and primary auditory cortex (A1) of mammals.

Spectrotemporal response properties of core auditory cortex neurons in awake monkey

So far, most studies of core auditory cortex (AC) have characterized the spectral and temporal tuning properties of cells in non-awake, anesthetized preparations. As experiments in awake animals are scarce, we here used dynamic spectral-temporal broadband ripples to study the properties of the spectrotemporal receptive fields (STRFs) of AC cells in awake monkeys. We show that AC neurons were typically most sensitive to low ripple densities (spectral) and low velocities (temporal), and that most cells were not selective for a particular spectrotemporal sweep direction. A substantial proportion of neurons preferred amplitude-modulated sounds (at zero ripple density) to dynamic ripples (at non-zero densities). The vast majority (>93%) of modulation transfer functions were separable with respect to spectral and temporal modulations, indicating that time and spectrum are independently processed in AC neurons. We also analyzed the linear predictability of AC responses to natural vocalizations on the basis of the STRF. We discuss our findings in the light of results obtained from the monkey midbrain inferior colliculus by comparing the spectrotemporal tuning properties and linear predictability of these two important auditory stages.

Discriminative learning of receptive fields from responses to non-Gaussian stimulus ensembles

Analysis of sensory neurons' processing characteristics requires simultaneous measurement of presented stimuli and concurrent spike responses. The functional transformation from high-dimensional stimulus space to the binary space of spike and non-spike responses is commonly described with linear-nonlinear models, whose linear filter component describes the neuron's receptive field. From a machine learning perspective, this corresponds to the binary classification problem of discriminating spike-eliciting from non-spike-eliciting stimulus examples. The classification-based receptive field (CbRF) estimation method proposed here adapts a linear large-margin classifier to optimally predict experimental stimulus-response data and subsequently interprets learned classifier weights as the neuron's receptive field filter. Computational learning theory provides a theoretical framework for learning from data and guarantees optimality in the sense that the risk of erroneously assigning a spike-eliciting stimulus example to the non-spike class (and vice versa) is minimized. Efficacy of the CbRF method is validated with simulations and for auditory spectro-temporal receptive field (STRF) estimation from experimental recordings in the auditory midbrain of Mongolian gerbils. Acoustic stimulation is performed with frequency-modulated tone complexes that mimic properties of natural stimuli, specifically non-Gaussian amplitude distribution and higher-order correlations. Results demonstrate that the proposed approach successfully identifies correct underlying STRFs, even in cases where second-order methods based on the spike-triggered average (STA) do not. Applied to small data samples, the method is shown to converge on smaller amounts of experimental recordings and with lower estimation variance than the generalized linear model and recent information theoretic methods. Thus, CbRF estimation may prove useful for investigation of neuronal processes in response to natural stimuli and in settings where rapid adaptation is induced by experimental design.

Binaural gain modulation of spectrotemporal tuning in the interaural level difference-coding pathway

In the brainstem, the auditory system diverges into two pathways that process different sound localization cues, interaural time differences (ITDs) and level differences (ILDs). We investigated the site where ILD is detected in the auditory system of barn owls, the posterior part of the lateral lemniscus (LLDp). This structure is equivalent to the lateral superior olive in mammals. The LLDp is unique in that it is the first place of binaural convergence in the brainstem where monaural excitatory and inhibitory inputs converge. Using binaurally uncorrelated noise and a generalized linear model, we were able to estimate the spectrotemporal tuning of excitatory and inhibitory inputs to these cells. We show that the response of LLDp neurons is highly locked to the stimulus envelope. Our data demonstrate that spectrotemporally tuned, temporally delayed inhibition enhances the reliability of envelope locking by modulating the gain of LLDp neurons' responses. The dependence of gain modulation on ILD shown here constitutes a means for space-dependent coding of stimulus identity by the initial stages of the auditory pathway.

Auditory cortical areas activated by slow frequency-modulated sounds in mice

Species-specific vocalizations in mice have frequency-modulated (FM) components slower than the lower limit of FM direction selectivity in the core region of the mouse auditory cortex. To identify cortical areas selective to slow frequency modulation, we investigated tonal responses in the mouse auditory cortex using transcranial flavoprotein fluorescence imaging. For differentiating responses to frequency modulation from those to stimuli at constant frequencies, we focused on transient fluorescence changes after direction reversal of temporally repeated and superimposed FM sweeps. We found that the ultrasonic field (UF) in the belt cortical region selectively responded to the direction reversal. The dorsoposterior field (DP) also responded weakly to the reversal. Regarding the responses in UF, no apparent tonotopic map was found, and the right UF responses were significantly larger in amplitude than the left UF responses. The half-max latency in responses to FM sweeps was shorter in UF compared with that in the primary auditory cortex (A1) or anterior auditory field (AAF). Tracer injection experiments in the functionally identified UF and DP confirmed that these two areas receive afferent inputs from the dorsal part of the medial geniculate nucleus (MG). Calcium imaging of UF neurons stained with fura-2 were performed using a two-photon microscope, and the presence of UF neurons that were selective to both direction and direction reversal of slow frequency modulation was demonstrated. These results strongly suggest a role for UF, and possibly DP, as cortical areas specialized for processing slow frequency modulation in mice.

Conserved mechanisms of vocalization coding in mammalian and songbird auditory midbrain

The ubiquity of social vocalizations among animals provides the opportunity to identify conserved mechanisms of auditory processing that subserve communication. Identifying auditory coding properties that are shared across vocal communicators will provide insight into how human auditory processing leads to speech perception. Here, we compare auditory response properties and neural coding of social vocalizations in auditory midbrain neurons of mammalian and avian vocal communicators. The auditory midbrain is a nexus of auditory processing because it receives and integrates information from multiple parallel pathways and provides the ascending auditory input to the thalamus. The auditory midbrain is also the first region in the ascending auditory system where neurons show complex tuning properties that are correlated with the acoustics of social vocalizations. Single unit studies in mice, bats and zebra finches reveal shared principles of auditory coding including tonotopy, excitatory and inhibitory interactions that shape responses to vocal signals, nonlinear response properties that are important for auditory coding of social vocalizations and modulation tuning. Additionally, single neuron responses in the mouse and songbird midbrain are reliable, selective for specific syllables, and rely on spike timing for neural discrimination of distinct vocalizations. We propose that future research on auditory coding of vocalizations in mouse and songbird midbrain neurons adopt similar experimental and analytical approaches so that conserved principles of vocalization coding may be distinguished from those that are specialized for each species. This article is part of a Special Issue entitled "Communication Sounds and the Brain: New Directions and Perspectives".

The dominant role of inhibition in creating response selectivities for communication calls in the brainstem auditory system

This review is concerned with how communication calls are processed and represented by populations of neurons in both the inferior colliculus (IC), the auditory midbrain nucleus, and the dorsal nucleus of the lateral lemniscus (DNLL), the nucleus just caudal to the IC. The review has five sections where focus in each section is on inhibition and its role in shaping response selectivity for communication calls. In the first section, the lack of response selectivity for calls in DNLL neurons is presented and discusses why inhibition plays virtually no role in shaping selectivity. In the second section, the lack of selectivity in the DNLL is contrasted with the high degree of response selectivity in the IC. The third section then reviews how inhibition in the IC shapes response selectivities for calls, and how those selectivities can create a population response with a distinctive response profile to a particular call, which differs from the population profile evoked by any other call. The fourth section is concerned with the specifics of inhibition in the IC, and how the interaction of excitation and inhibition creates directional selectivities for frequency modulations, one of the principal acoustic features of communication signals. The two major hypotheses for directional selectivity are presented. One is the timing hypothesis, which holds that the precise timing of excitation relative to inhibition is the feature that shapes directionality. The other hypothesis is that the relative magnitudes of excitation and inhibition are the dominant features that shape directionality, where timing is relatively unimportant. The final section then turns to the role of serotonin, a neuromodulator that can markedly change responses to calls in the IC. Serotonin provides a linkage between behavioral states and processing. This linkage is discussed in the final section together with the hypothesis that serotonin acts to enhances the contrast in the population responses to various calls over and above the distinctive population responses that were created by inhibition. This article is part of a Special Issue entitled "Communication Sounds and the Brain: New Directions and Perspectives".

Development of on-off spiking in superior paraolivary nucleus neurons of the mouse

The superior paraolivary nucleus (SPON) is a prominent cell group in the auditory brain stem that has been increasingly implicated in representing temporal sound structure. Although SPON neurons selectively respond to acoustic signals important for sound periodicity, the underlying physiological specializations enabling these responses are poorly understood. We used in vitro and in vivo recordings to investigate how SPON neurons develop intrinsic cellular properties that make them well suited for encoding temporal sound features. In addition to their hallmark rebound spiking at the stimulus offset, SPON neurons were characterized by spiking patterns termed onset, adapting, and burst in response to depolarizing stimuli in vitro. Cells with burst spiking had some morphological differences compared with other SPON neurons and were localized to the dorsolateral region of the nucleus. Both membrane and spiking properties underwent strong developmental regulation, becoming more temporally precise with age for both onset and offset spiking. Single-unit recordings obtained in young mice demonstrated that SPON neurons respond with temporally precise onset spiking upon tone stimulation in vivo, in addition to the typical offset spiking. Taken together, the results of the present study demonstrate that SPON neurons develop sharp on-off spiking, which may confer sensitivity to sound amplitude modulations or abrupt sound transients. These findings are consistent with the proposed involvement of the SPON in the processing of temporal sound structure, relevant for encoding communication cues.

Intracellular responses to frequency modulated tones in the dorsal cortex of the mouse inferior colliculus

Frequency modulations occur in many natural sounds, including vocalizations. The neuronal response to frequency modulated (FM) stimuli has been studied extensively in different brain areas, with an emphasis on the auditory cortex and the central nucleus of the inferior colliculus. Here, we measured the responses to FM sweeps in whole-cell recordings from neurons in the dorsal cortex of the mouse inferior colliculus. Both up- and downward logarithmic FM sweeps were presented at two different speeds to both the ipsi- and the contralateral ear. Based on the number of action potentials that were fired, between 10 and 24% of cells were selective for rate or direction of the FM sweeps. A somewhat lower percentage of cells, 6–21%, showed selectivity based on EPSP size. To study the mechanisms underlying the generation of FM selectivity, we compared FM responses with responses to simple tones in the same cells. We found that if pairs of neurons responded in a similar way to simple tones, they generally also responded in a similar way to FM sweeps. Further evidence that FM selectivity can be generated within the dorsal cortex was obtained by reconstructing FM sweeps from the response to simple tones using three different models. In about half of the direction selective neurons the selectivity was generated by spectrally asymmetric synaptic inhibition. In addition, evidence for direction selectivity based on the timing of excitatory responses was also obtained in some cells. No clear evidence for the local generation of rate selectivity was obtained. We conclude that FM direction selectivity can be generated within the dorsal cortex of the mouse inferior colliculus by multiple mechanisms.

Differences between spectro-temporal receptive fields derived from artificial and natural stimuli in the auditory cortex

Spectro-temporal properties of auditory cortex neurons have been extensively studied with artificial sounds but it is still unclear whether they help in understanding neuronal responses to communication sounds. Here, we directly compared spectro-temporal receptive fields (STRFs) obtained from the same neurons using both artificial stimuli (dynamic moving ripples, DMRs) and natural stimuli (conspecific vocalizations) that were matched in terms of spectral content, average power and modulation spectrum. On a population of auditory cortex neurons exhibiting reliable tuning curves when tested with pure tones, significant STRFs were obtained for 62% of the cells with vocalizations and 68% with DMR. However, for many cells with significant vocalization-derived STRFs (STRF_voc) and DMR-derived STRFs (STRF_dmr), the BF, latency, bandwidth and global STRFs shape differed more than what would be predicted by spiking responses simulated by a linear model based on a non-homogenous Poisson process. Moreover STRF_voc predicted neural responses to vocalizations more accurately than STRF_dmr predicted neural response to DMRs, despite similar spike-timing reliability for both sets of stimuli. Cortical bursts, which potentially introduce nonlinearities in evoked responses, did not explain the differences between STRF_voc and STRF_dmr. Altogether, these results suggest that the nonlinearity of auditory cortical responses makes it difficult to predict responses to communication sounds from STRFs computed from artificial stimuli.

Local inhibition modulates learning-dependent song encoding in the songbird auditory cortex

Changes in inhibition during development are well documented, but the role of inhibition in adult learning-related plasticity is not understood. In songbirds, vocal recognition learning alters the neural representation of songs across the auditory forebrain, including the caudomedial nidopallium (NCM), a region analogous to mammalian secondary auditory cortices. Here, we block local inhibition with the iontophoretic application of gabazine, while simultaneously measuring song-evoked spiking activity in NCM of European starlings trained to recognize sets of conspecific songs. We find that local inhibition differentially suppresses the responses to learned and unfamiliar songs and enhances spike-rate differences between learned categories of songs. These learning-dependent response patterns emerge, in part, through inhibitory modulation of selectivity for song components and the masking of responses to specific acoustic features without altering spectrotemporal tuning. The results describe a novel form of inhibitory modulation of the encoding of learned categories and demonstrate that inhibition plays a central role in shaping the responses of neurons to learned, natural signals.

Inhibition shapes response selectivity in the inferior colliculus by gain modulation

Pharmacological block of inhibition is often used to determine if inhibition contributes to spike selectivity, in which a preferred stimulus evokes more spikes than a null stimulus. When inhibitory block reduces spike selectivity, a common interpretation is that differences between the preferred- and null-evoked inhibitions created the selectivity from less-selective excitatory inputs. In models based on empirical properties of cells from the inferior colliculus (IC) of awake bats, we show that inhibitory differences are not required. Instead, inhibition can enhance spike selectivity by changing the gain, the ratio of output spikes to input current. Within the model, we made preferred stimuli that evoked more spikes than null stimuli using five distinct synaptic mechanisms. In two cases, synaptic selectivity (the differences between the preferred and null inputs) was entirely excitatory, and in two it was entirely inhibitory. In each case, blocking inhibition eliminated spike selectivity. Thus, observing spike rates following inhibitory block did not distinguish among the cases where synaptic selectivity was entirely excitatory or inhibitory. We then did the same modeling experiment using empirical synaptic conductances derived from responses to preferred and null sounds. In most cases, inhibition in the model enhanced spike selectivity mainly by gain modulation and firing rate reduction. Sometimes, inhibition reduced the null gain to zero, eliminating null-evoked spikes. In some cases, inhibition increased the preferred gain more than the null gain, enhancing the difference between the preferred- and null-evoked spikes. Finally, inhibition kept firing rates low. When selectivity is quantified by the selectivity index (SI, the ratio of the difference to the sum of the spikes evoked by the preferred and null stimuli), inhibitory block reduced the SI by increasing overall firing rates. These results are consistent with inhibition shaping spike selectivity by gain control.

From behavioral context to receptors: serotonergic modulatory pathways in the IC

In addition to ascending, descending, and lateral auditory projections, inputs extrinsic to the auditory system also influence neural processing in the inferior colliculus (IC). These types of inputs often have an important role in signaling salient factors such as behavioral context or internal state. One route for such extrinsic information is through centralized neuromodulatory networks like the serotonergic system. Serotonergic inputs to the IC originate from centralized raphe nuclei, release serotonin in the IC, and activate serotonin receptors expressed by auditory neurons. Different types of serotonin receptors act as parallel pathways regulating specific features of circuitry within the IC. This results from variation in subcellular localizations and effector pathways of different receptors, which consequently influence auditory responses in distinct ways. Serotonin receptors may regulate GABAergic inhibition, influence response gain, alter spike timing, or have effects that are dependent on the level of activity. Serotonin receptor types additionally interact in nonadditive ways to produce distinct combinatorial effects. This array of effects of serotonin is likely to depend on behavioral context, since the levels of serotonin in the IC transiently increase during behavioral events including stressful situations and social interaction. These studies support a broad model of serotonin receptors as a link between behavioral context and reconfiguration of circuitry in the IC, and the resulting possibility that plasticity at the level of specific receptor types could alter the relationship between context and circuit function.

Sparse codes for speech predict spectrotemporal receptive fields in the inferior colliculus

We have developed a sparse mathematical representation of speech that minimizes the number of active model neurons needed to represent typical speech sounds. The model learns several well-known acoustic features of speech such as harmonic stacks, formants, onsets and terminations, but we also find more exotic structures in the spectrogram representation of sound such as localized checkerboard patterns and frequency-modulated excitatory subregions flanked by suppressive sidebands. Moreover, several of these novel features resemble neuronal receptive fields reported in the Inferior Colliculus (IC), as well as auditory thalamus and cortex, and our model neurons exhibit the same tradeoff in spectrotemporal resolution as has been observed in IC. To our knowledge, this is the first demonstration that receptive fields of neurons in the ascending mammalian auditory pathway beyond the auditory nerve can be predicted based on coding principles and the statistical properties of recorded sounds.

FM-selective networks in human auditory cortex revealed using fMRI and multivariate pattern classification

Frequency modulation (FM) is an acoustic feature of nearly all complex sounds. Directional FM sweeps are especially pervasive in speech, music, animal vocalizations, and other natural sounds. Although the existence of FM-selective cells in the auditory cortex of animals has been documented, evidence in humans remains equivocal. Here we used multivariate pattern analysis to identify cortical selectivity for direction of a multitone FM sweep. This method distinguishes one pattern of neural activity from another within the same ROI, even when overall level of activity is similar, allowing for direct identification of FM-specialized networks. Standard contrast analysis showed that despite robust activity in auditory cortex, no clusters of activity were associated with up versus down sweeps. Multivariate pattern analysis classification, however, identified two brain regions as selective for FM direction, the right primary auditory cortex on the supratemporal plane and the left anterior region of the superior temporal gyrus. These findings are the first to directly demonstrate existence of FM direction selectivity in the human auditory cortex.

The generation of direction selectivity in the auditory system

Both human speech and animal vocal signals contain frequency-modulated (FM) sounds. Although central auditory neurons that selectively respond to the direction of frequency modulation are known, the synaptic mechanisms underlying the generation of direction selectivity (DS) remain elusive. Here we show the emergence of DS neurons in the inferior colliculus by mapping the three major subcortical auditory nuclei. Cell-attached recordings reveal a highly reliable and precise firing of DS neurons to FM sweeps in a preferred direction. By using in vivo whole-cell current-clamp and voltage-clamp recordings, we found that the synaptic inputs to DS neurons are not direction selective, but temporally reversed excitatory and inhibitory synaptic inputs are evoked in response to opposing directions of FM sweeps. The construction of such temporal asymmetry, resulting DS, and its topography can be attributed to the spectral disparity of the excitatory and the inhibitory synaptic tonal receptive fields.

Receptive field dimensionality increases from the auditory midbrain to cortex

In the primary auditory cortex, spectrotemporal receptive fields (STRFs) are composed of multiple independent components that capture the processing of disparate stimulus aspects by any given neuron. The origin of these multidimensional stimulus filters in the central auditory system is unknown. To determine whether multicomponent STRFs emerge prior to the forebrain, we recorded from single neurons in the main obligatory station of the auditory midbrain, the inferior colliculus. By comparing results of different spike-triggered techniques, we found that the neural responses in the inferior colliculus can be accounted for by a single stimulus filter. This was observed for all temporal response patterns, from strongly phasic to tonic. Our results reveal that spectrotemporal stimulus encoding undergoes a fundamental transformation along the auditory neuraxis, with the emergence of multidimensional receptive fields beyond the auditory midbrain.

Selectivity for spectral motion as a neural computation for encoding natural communication signals in bat inferior colliculus

This study examines the neural computations performed by neurons in the auditory system to be selective for the direction and velocity of signals sweeping upward or downward in frequency, termed spectral motion. We show that neurons in the auditory midbrain of Mexican free-tailed bats encode multiple spectrotemporal features of natural communication sounds. These features to which each neuron is tuned are nonlinearly combined to produce selectivity for spectral motion cues present in their conspecific calls, such as direction and velocity. We find that the neural computations resulting in selectivity for spectral motion are analogous to models of motion selectivity studied in vision. Our analysis revealed that auditory neurons in the inferior colliculus (IC) are avoiding spectrotemporal modulations that are redundant across different bat communication signals and are specifically tuned for modulations that distinguish each call from another by their frequency-modulated direction and velocity, suggesting that spectral motion is the neural computation through which IC neurons are encoding specific features of conspecific vocalizations.