Low-dimensional interference of mid-level sound statistics predicts human speech recognition in natural environmental noise.

Recognizing speech in noise, such as in a busy street or restaurant, is an essential listening task where the task difficulty varies across acoustic environments and noise levels. Yet, current cognitive models are unable to account for changing real-world hearing sensitivity. Here, using natural and perturbed background sounds we demonstrate that spectrum and modulations statistics of environmental backgrounds drastically impact human word recognition accuracy and they do so independently of the noise level. These sound statistics can facilitate or hinder recognition - at the same noise level accuracy can range from 0% to 100%, depending on the background. To explain this perceptual variability, we optimized a biologically grounded hierarchical model, consisting of frequency-tuned cochlear filters and subsequent mid-level modulation-tuned filters that account for central auditory tuning. Low-dimensional summary statistics from the mid-level model accurately predict single trial perceptual judgments, accounting for more than 90% of the perceptual variance across backgrounds and noise levels, and substantially outperforming a cochlear model. Furthermore, perceptual transfer functions in the mid-level auditory space identify multi-dimensional natural sound features that impact recognition. Thus speech recognition in natural backgrounds involves interference of multiple summary statistics that are well described by an interpretable, low-dimensional auditory model. Since this framework relates salient natural sound cues to single trial perceptual judgements, it may improve outcomes for auditory prosthetics and clinical measurements of real-world hearing sensitivity.

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.

Distinct neural ensemble response statistics are associated with recognition and discrimination of natural sound textures.

The perception of sound textures, a class of natural sounds defined by statistical sound structure such as fire, wind, and rain, has been proposed to arise through the integration of time-averaged summary statistics. Where and how the auditory system might encode these summary statistics to create internal representations of these stationary sounds, however, is unknown. Here, using natural textures and synthetic variants with reduced statistics, we show that summary statistics modulate the correlations between frequency organized neuron ensembles in the awake rabbit inferior colliculus (IC). These neural ensemble correlation statistics capture high-order sound structure and allow for accurate neural decoding in a single trial recognition task with evidence accumulation times approaching 1 s. In contrast, the average activity across the neural ensemble (neural spectrum) provides a fast (tens of milliseconds) and salient signal that contributes primarily to texture discrimination. Intriguingly, perceptual studies in human listeners reveal analogous trends: the sound spectrum is integrated quickly and serves as a salient discrimination cue while high-order sound statistics are integrated slowly and contribute substantially more toward recognition. The findings suggest statistical sound cues such as the sound spectrum and correlation structure are represented by distinct response statistics in auditory midbrain ensembles, and that these neural response statistics may have dissociable roles and time scales for the recognition and discrimination of natural sounds.

Optimal Multichannel Artifact Prediction and Removal for Neural Stimulation and Brain Machine Interfaces.

Neural implants that deliver multi-site electrical stimulation to the nervous systems are no longer the last resort but routine treatment options for various neurological disorders. Multi-site electrical stimulation is also widely used to study nervous system function and neural circuit transformations. These technologies increasingly demand dynamic electrical stimulation and closed-loop feedback control for real-time assessment of neural function, which is technically challenging since stimulus-evoked artifacts overwhelm the small neural signals of interest. We report a novel and versatile artifact removal method that can be applied in a variety of settings, from single- to multi-site stimulation and recording and for current waveforms of arbitrary shape and size. The method capitalizes on linear electrical coupling between stimulating currents and recording artifacts, which allows us to estimate a multi-channel linear Wiener filter to predict and subsequently remove artifacts via subtraction. We confirm and verify the linearity assumption and demonstrate feasibility in a variety of recording modalities, including in vitro sciatic nerve stimulation, bilateral cochlear implant stimulation, and multi-channel stimulation and recording between the auditory midbrain and cortex. We demonstrate a vast enhancement in the recording quality with a typical artifact reduction of 25-40 dB. The method is efficient and can be scaled to arbitrary number of stimulus and recording sites, making it ideal for applications in large-scale arrays, closed-loop implants, and high-resolution multi-channel brain-machine interfaces.

Spiking network optimized for word recognition in noise predicts auditory system hierarchy.

The auditory neural code is resilient to acoustic variability and capable of recognizing sounds amongst competing sound sources, yet, the transformations enabling noise robust abilities are largely unknown. We report that a hierarchical spiking neural network (HSNN) optimized to maximize word recognition accuracy in noise and multiple talkers predicts organizational hierarchy of the ascending auditory pathway. Comparisons with data from auditory nerve, midbrain, thalamus and cortex reveals that the optimal HSNN predicts several transformations of the ascending auditory pathway including a sequential loss of temporal resolution and synchronization ability, increasing sparseness, and selectivity. The optimal organizational scheme enhances performance by selectively filtering out noise and fast temporal cues such as voicing periodicity, that are not directly relevant to the word recognition task. An identical network arranged to enable high information transfer fails to predict auditory pathway organization and has substantially poorer performance. Furthermore, conventional single-layer linear and nonlinear receptive field networks that capture the overall feature extraction of the HSNN fail to achieve similar performance. The findings suggest that the auditory pathway hierarchy and its sequential nonlinear feature extraction computations enhance relevant cues while removing non-informative sources of noise, thus enhancing the representation of sounds in noise impoverished conditions.

A neural ensemble correlation code for sound category identification.

Humans and other animals effortlessly identify natural sounds and categorize them into behaviorally relevant categories. Yet, the acoustic features and neural transformations that enable sound recognition and the formation of perceptual categories are largely unknown. Here, using multichannel neural recordings in the auditory midbrain of unanesthetized female rabbits, we first demonstrate that neural ensemble activity in the auditory midbrain displays highly structured correlations that vary with distinct natural sound stimuli. These stimulus-driven correlations can be used to accurately identify individual sounds using single-response trials, even when the sounds do not differ in their spectral content. Combining neural recordings and an auditory model, we then show how correlations between frequency-organized auditory channels can contribute to discrimination of not just individual sounds but sound categories. For both the model and neural data, spectral and temporal correlations achieved similar categorization performance and appear to contribute equally. Moreover, both the neural and model classifiers achieve their best task performance when they accumulate evidence over a time frame of approximately 1-2 seconds, mirroring human perceptual trends. These results together suggest that time-frequency correlations in sounds may be reflected in the correlations between auditory midbrain ensembles and that these correlations may play an important role in the identification and categorization of natural sounds.

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.

A Hierarchy of Time Scales for Discriminating and Classifying the Temporal Shape of Sound in Three Auditory Cortical Fields.

Auditory cortex is essential for mammals, including rodents, to detect temporal "shape" cues in the sound envelope but it remains unclear how different cortical fields may contribute to this ability (Lomber and Malhotra, 2008; Threlkeld et al., 2008). Previously, we found that precise spiking patterns provide a potential neural code for temporal shape cues in the sound envelope in the primary auditory (A1), and ventral auditory field (VAF) and caudal suprarhinal auditory field (cSRAF) of the rat (Lee et al., 2016). Here, we extend these findings and characterize the time course of the temporally precise output of auditory cortical neurons in male rats. A pairwise sound discrimination index and a Naive Bayesian classifier are used to determine how these spiking patterns could provide brain signals for behavioral discrimination and classification of sounds. We find response durations and optimal time constants for discriminating sound envelope shape increase in rank order with: A1 < VAF < cSRAF. Accordingly, sustained spiking is more prominent and results in more robust sound discrimination in non-primary cortex versus A1. Spike-timing patterns classify 10 different sound envelope shape sequences and there is a twofold increase in maximal performance when pooling output across the neuron population indicating a robust distributed neural code in all three cortical fields. Together, these results support the idea that temporally precise spiking patterns from primary and non-primary auditory cortical fields provide the necessary signals for animals to discriminate and classify a large range of temporal shapes in the sound envelope.SIGNIFICANCE STATEMENT Functional hierarchies in the visual cortices support the concept that classification of visual objects requires successive cortical stages of processing including a progressive increase in classical receptive field size. The present study is significant as it supports the idea that a similar progression exists in auditory cortices in the time domain. We demonstrate for the first time that three cortices provide temporal spiking patterns for robust temporal envelope shape discrimination but only the ventral non-primary cortices do so on long time scales. This study raises the possibility that primary and non-primary cortices provide unique temporal spiking patterns and time scales for perception of sound envelope shape.

Origins of scale invariance in vocalization sequences and speech.

To communicate effectively animals need to detect temporal vocalization cues that vary over several orders of magnitude in their amplitude and frequency content. This large range of temporal cues is evident in the power-law scale-invariant relationship between the power of temporal fluctuations in sounds and the sound modulation frequency (f). Though various forms of scale invariance have been described for natural sounds, the origins and implications of scale invariant phenomenon remain unknown. Using animal vocalization sequences, including continuous human speech, and a stochastic model of temporal amplitude fluctuations we demonstrate that temporal acoustic edges are the primary acoustic cue accounting for the scale invariant phenomenon. The modulation spectrum of vocalization sequences and the model both exhibit a dual regime lowpass structure with a flat region at low modulation frequencies and scale invariant 1/f2 trend for high modulation frequencies. Moreover, we find a time-frequency tradeoff between the average vocalization duration of each vocalization sequence and the cutoff frequency beyond which scale invariant behavior is observed. These results indicate that temporal edges are universal features responsible for scale invariance in vocalized sounds. This is significant since temporal acoustic edges are salient perceptually and the auditory system could exploit such statistical regularities to minimize redundancies and generate compact neural representations of vocalized sounds.

Neural spike-timing patterns vary with sound shape and periodicity in three auditory cortical fields.

Mammals perceive a wide range of temporal cues in natural sounds, and the auditory cortex is essential for their detection and discrimination. The rat primary (A1), ventral (VAF), and caudal suprarhinal (cSRAF) auditory cortical fields have separate thalamocortical pathways that may support unique temporal cue sensitivities. To explore this, we record responses of single neurons in the three fields to variations in envelope shape and modulation frequency of periodic noise sequences. Spike rate, relative synchrony, and first-spike latency metrics have previously been used to quantify neural sensitivities to temporal sound cues; however, such metrics do not measure absolute spike timing of sustained responses to sound shape. To address this, in this study we quantify two forms of spike-timing precision, jitter, and reliability. In all three fields, we find that jitter decreases logarithmically with increase in the basis spline (B-spline) cutoff frequency used to shape the sound envelope. In contrast, reliability decreases logarithmically with increase in sound envelope modulation frequency. In A1, jitter and reliability vary independently, whereas in ventral cortical fields, jitter and reliability covary. Jitter time scales increase (A1 < VAF < cSRAF) and modulation frequency upper cutoffs decrease (A1 > VAF > cSRAF) with ventral progression from A1. These results suggest a transition from independent encoding of shape and periodicity sound cues on short time scales in A1 to a joint encoding of these same cues on longer time scales in ventral nonprimary cortices.

Age-related decrease in theta and gamma coherence across dorsal ca1 pyramidale and radiatum layers.

In both humans and rodents, aging is linked to impairments in hippocampus dependent learning. Given such deficits, one would expect corresponding changes in hippocampal local field potentials, which represent the integration of multiple inputs onto a given dendritic field within the hippocampus. The current experiment examined coherence of theta and gamma in young and aged rats at sub-millimeter and millimeter distant locations both within and across layers in CA1 of the dorsal hippocampus. The degree to which different dendritic layers show coherent oscillations indicates the uniformity of the inputs and local circuitry, and may form an important element for processing information. Aged rats had lower coherence in all frequency ranges; this was most marked within a layer as the distance between electrodes increased. Notably, unlike younger rats, in the aged rats coherence was not affected by running on the maze. Furthermore, despite the previously reported effects of cholinergic activation on theta frequency and power, there was no effect of the cholinomimetic physostigmine on coherence. These data indicate an age related fragmentation in hippocampal processing that may underlie some of the observed learning and memory deficits.

A high-density, high-channel count, multiplexed µECoG array for auditory-cortex recordings.

Our understanding of the large-scale population dynamics of neural activity is limited, in part, by our inability to record simultaneously from large regions of the cortex. Here, we validated the use of a large-scale active microelectrode array that simultaneously records 196 multiplexed micro-electrocortigraphical (µECoG) signals from the cortical surface at a very high density (1,600 electrodes/cm(2)). We compared µECoG measurements in auditory cortex using a custom "active" electrode array to those recorded using a conventional "passive" µECoG array. Both of these array responses were also compared with data recorded via intrinsic optical imaging, which is a standard methodology for recording sound-evoked cortical activity. Custom active µECoG arrays generated more veridical representations of the tonotopic organization of the auditory cortex than current commercially available passive µECoG arrays. Furthermore, the cortical representation could be measured efficiently with the active arrays, requiring as little as 13.5 s of neural data acquisition. Next, we generated spectrotemporal receptive fields from the recorded neural activity on the active µECoG array and identified functional organizational principles comparable to those observed using intrinsic metabolic imaging and single-neuron recordings. This new electrode array technology has the potential for large-scale, temporally precise monitoring and mapping of the cortex, without the use of invasive penetrating electrodes.

Theta dynamics in rat: speed and acceleration across the Septotemporal axis.

Theta (6-12 Hz) rhythmicity in the local field potential (LFP) reflects a clocking mechanism that brings physically isolated neurons together in time, allowing for the integration and segregation of distributed cell assemblies. Variation in the theta signal has been linked to locomotor speed, sensorimotor integration as well as cognitive processing. Previously, we have characterized the relationship between locomotor speed and theta power and how that relationship varies across the septotemporal (long) axis of the hippocampus (HPC). The current study investigated the relationship between whole body acceleration, deceleration and theta indices at CA1 and dentate gyrus (DG) sites along the septotemporal axis of the HPC in rats. Results indicate that whole body acceleration and deceleration predicts a significant amount of variability in the theta signal beyond variation in locomotor speed. Furthermore, deceleration was more predictive of variation in theta amplitude as compared to acceleration as rats traversed a linear track. Such findings highlight key variables that systematically predict the variability in the theta signal across the long axis of the HPC. A better understanding of the relative contribution of these quantifiable variables and their variation as a function of experience and environmental conditions should facilitate our understanding of the relationship between theta and sensorimotor/cognitive functions.

Novel acoustic stimuli can alter locomotor speed to hippocampal theta relationship.

Hippocampal theta (6-12 Hz) plays a critical role in synchronizing the discharge of action potentials, ultimately orchestrating individual neurons into large-scale ensembles. Alterations in theta dynamics may reflect variations in sensorimotor integration, the flow of sensory input, and/or cognitive processing. Previously we have investigated septotemporal variation in the locomotor speed to theta amplitude relationship as well as how that relationship is systematically altered as a function of novel, physical space. In the present study, we ask, beyond physical space, whether persistent and passive sound delivery can alter septal theta local field potential rhythm dynamics. Results indicate pronounced alterations in the slope of the speed to theta amplitude relationship as a function of sound presentation and location. Further, this reduction in slope habituates across days. The current findings highlight that moment-to-moment alterations in theta amplitude is a rich dynamic index that is quantitatively related to both alterations in motor behavior and sensory experience. The implications of these phenomena are discussed with respect to emergent cognitive functions subserved by hippocampal circuits.

Novel space alters theta and gamma synchrony across the longitudinal axis of the hippocampus.

Hippocampal theta (6-10 Hz) and gamma (25-50 Hz and 65-100 Hz) local field potentials (LFPs) reflect the dynamic synchronization evoked by inputs impinging upon hippocampal neurons. Novel experience is known to engage hippocampal physiology and promote successful encoding. Does novelty synchronize or desynchronize theta and/or gamma frequency inputs across the septotemporal (long) axis of the hippocampus (HPC)? The present study tested the hypothesis that a novel spatial environment would alter theta power and coherence across the long axis. We compared theta and gamma LFP signals at individual (power) and millimeter distant electrode pairs (coherence) within the dentate gyrus (DG) and CA1 region while rats navigated a runway (1) in a familiar environment, (2) with a modified path in the same environment and (3) in a novel space. Locomotion in novel space was related to increases in theta and gamma power at most CA1 and DG sites. The increase in theta and gamma power was concurrent with an increase in theta and gamma coherence across the long axis of CA1; however, there was a significant decrease in theta coherence across the long axis of the DG. These findings illustrate significant shifts in the synchrony of entorhinal, CA3 and/or neuromodulatory afferents conveying novel spatial information to the dendritic fields of CA1 and DG targets across the long axis of the HPC. This shift suggests that the entire theta/gamma-related input to the CA1 network, and likely output, receives and conveys a more coherent message in response to novel sensory experience. Such may contribute to the successful encoding of novel sensory experience.

Dissociation between dorsal and ventral hippocampal theta oscillations during decision-making.

Hippocampal theta oscillations are postulated to support mnemonic processes in humans and rodents. Theta oscillations facilitate encoding and spatial navigation, but to date, it has been difficult to dissociate the effects of volitional movement from the cognitive demands of a task. Therefore, we examined whether volitional movement or cognitive demands exerted a greater modulating factor over theta oscillations during decision-making. Given the anatomical, electrophysiological, and functional dissociations along the dorsal-ventral axis, theta oscillations were simultaneously recorded in the dorsal and ventral hippocampus in rats trained to switch between place and motor-response strategies. Stark differences in theta characteristics were found between the dorsal and ventral hippocampus in frequency, power, and coherence. Theta power increased in the dorsal, but decreased in the ventral hippocampus, during the decision-making epoch. Interestingly, the relationship between running speed and theta power was uncoupled during the decision-making epoch, a phenomenon limited to the dorsal hippocampus. Theta frequency increased in both the dorsal and ventral hippocampus during the decision epoch, although this effect was greater in the dorsal hippocampus. Despite these differences, ventral hippocampal theta was responsive to the navigation task; theta frequency, power, and coherence were all affected by cognitive demands. Theta coherence increased within the dorsal hippocampus during the decision-making epoch on all three tasks. However, coherence selectively increased throughout the hippocampus (dorsal to ventral) on the task with new hippocampal learning. Interestingly, most results were consistent across tasks, regardless of hippocampal-dependent learning. These data indicate increased integration and cooperation throughout the hippocampus during information processing.

Hippocampal theta, gamma, and theta-gamma coupling: effects of aging, environmental change, and cholinergic activation.

Hippocampal theta and gamma oscillations coordinate the timing of multiple inputs to hippocampal neurons and have been linked to information processing and the dynamics of encoding and retrieval. One major influence on hippocampal rhythmicity is from cholinergic afferents. In both humans and rodents, aging is linked to impairments in hippocampus-dependent function along with degradation of cholinergic function. Cholinomimetics can reverse some age-related memory impairments and modulate oscillations in the hippocampus. Therefore, one would expect corresponding changes in these oscillations and possible rescue with the cholinomimetic physostigmine. Hippocampal activity was recorded while animals explored a familiar or a novel maze configuration. Reexposure to a familiar situation resulted in minimal aging effects or changes in theta or gamma oscillations. In contrast, exploration of a novel maze configuration increased theta power; this was greater in adult than old animals, although the deficit was reversed with physostigmine. In contrast to the theta results, the effects of novelty, age, and/or physostigmine on gamma were relatively weak. Unrelated to the behavioral situation were an age-related decrease in the degree of theta-gamma coupling and the fact that physostigmine lowered the frequency of theta in both adult and old animals. The results indicate that age-related changes in gamma and theta modulation of gamma, while reflecting aging changes in hippocampal circuitry, seem less related to aging changes in information processing. In contrast, the data support a role for theta and the cholinergic system in encoding and that hippocampal aging is related to impaired encoding of new information.

Ketamine disrupts theta synchrony across the septotemporal axis of the CA1 region of hippocampus.

The hippocampal theta signal reflects moment-to-moment variation in the synchrony of synaptic input to hippocampal neurons. Consistent with the topography of hippocampal afferents, the synchrony (coherence) of the theta signal varies across the septotemporal axis. Septotemporal variation in the theta signal can also be observed in relation to ongoing and past experience. Thus there is a systematic decrease in the relationship between locomotor speed and theta power across the septotemporal axis, septal hippocampus exhibiting the strongest relationship. Conversely, theta in temporal hippocampus decrements over repeated behavioral experience (running episodes), while theta in the septal hippocampus does not. Ketamine is an N-methyl-D-aspartate (NMDA) antagonist that can decrease theta power. The present study examined whether ketamine treatment could alter theta coherence across the long axis independent of changes in locomotor behavior. Rats were well trained to navigate a linear runway and outfitted with electrodes at different septotemporal positions within CA1. Locomotor behavior and theta coherence and power were examined after administration of 2.5 and 10 mg/kg ketamine. Ketamine (2.5 mg/kg) decreased theta coherence between distant CA1 electrode sites without altering running speed or theta power. Both doses of ketamine also blunted and reversed the decrement in theta power observed at midseptotemporal and temporal electrodes over repeated run sessions. The results demonstrate the sensitivity of global network synchronization to relatively low doses of ketamine and septotemporal differences in the influence of ketamine on hippocampal dynamics in relation to past experience.

Spectrotemporal sound preferences of neighboring inferior colliculus neurons: implications for local circuitry and processing.

How do local circuits in the inferior colliculus (IC) process and transform spectral and temporal sound information? Using a four-tetrode array we examined the functional properties of the IC and metrics of its micro circuitry by recording neural activity from neighboring single neurons in the cat. Spectral and temporal response preferences were compared for neurons found on the same and adjacent tetrodes (ATs), as well as across distant recording sites. We found that neighboring neurons had similar preferences while neurons recorded across distant sites were less similar. Best frequency (BF) was the most correlated parameter between neighboring neurons and BF differences exhibited unique clustering at ~0.3 octave intervals, indicative of the frequency band lamina. Other spectral and temporal parameters of the receptive fields were more similar for neighboring neurons than for those at distant sites and the receptive field similarity was larger for neurons with small differences in BF. Furthermore, correlated firing was stronger for neighboring neuron pairs and increased with proximity and decreasing BF difference. Thus, although response selectivities are quite diverse in the IC, spectral, and temporal preference within a local microcircuit are functionally quite similar. This suggests a scheme where local circuits are organized into zones that are specialized for processing distinct spectrotemporal cues.

Precise feature based time scales and frequency decorrelation lead to a sparse auditory code.

Sparse redundancy reducing codes have been proposed as efficient strategies for representing sensory stimuli. A prevailing hypothesis suggests that sensory representations shift from dense redundant codes in the periphery to selective sparse codes in cortex. We propose an alternative framework where sparseness and redundancy depend on sensory integration time scales and demonstrate that the central nucleus of the inferior colliculus (ICC) of cats encodes sound features by precise sparse spike trains. Direct comparisons with auditory cortical neurons demonstrate that ICC responses were sparse and uncorrelated as long as the spike train time scales were matched to the sensory integration time scales relevant to ICC neurons. Intriguingly, correlated spiking in the ICC was substantially lower than predicted by linear or nonlinear models and strictly observed for neurons with best frequencies within a "critical band," the hallmark of perceptual frequency resolution in mammals. This is consistent with a sparse asynchronous code throughout much of the ICC and a complementary correlation code within a critical band that may allow grouping of perceptually relevant cues.