Integration over multiple timescales in primary auditory cortex

Neural encoding of linguistic speech cues is unaffected by cognitive decline, but decreases with increasing hearing impairment

The multivariate temporal response function (mTRF) is an effective tool for investigating the neural encoding of acoustic and complex linguistic features in natural continuous speech. In this study, we investigated how neural representations of speech features derived from natural stimuli are related to early signs of cognitive decline in older adults, taking into account the effects of hearing. Participants without ( n = 25 ) and with ( n = 19 ) early signs of cognitive decline listened to an audiobook while their electroencephalography responses were recorded. Using the mTRF framework, we modeled the relationship between speech input and neural response via different acoustic, segmented and linguistic encoding models and examined the response functions in terms of encoding accuracy, signal power, peak amplitudes and latencies. Our results showed no significant effect of cognitive decline or hearing ability on the neural encoding of acoustic and linguistic speech features. However, we found a significant interaction between hearing ability and the word-level segmentation model, suggesting that hearing impairment specifically affects encoding accuracy for this model, while other features were not affected by hearing ability. These results suggest that while speech processing markers remain unaffected by cognitive decline and hearing loss per se, neural encoding of word-level segmented speech features in older adults is affected by hearing loss but not by cognitive decline. This study emphasises the effectiveness of mTRF analysis in studying the neural encoding of speech and argues for an extension of research to investigate its clinical impact on hearing loss and cognition.

Auditory discrimination learning differentially modulates neural representation in auditory cortex subregions and inter-areal connectivity

Changes in sound-evoked responses in the auditory cortex (ACtx) occur during learning, but how learning alters neural responses in different ACtx subregions and changes their interactions is unclear. To address these questions, we developed an automated training and widefield imaging system to longitudinally track the neural activity of all mouse ACtx subregions during a tone discrimination task. We find that responses in primary ACtx are highly informative of learned stimuli and behavioral outcomes throughout training. In contrast, representations of behavioral outcomes in the dorsal posterior auditory field, learned stimuli in the dorsal anterior auditory field, and inter-regional correlations between primary and higher-order areas are enhanced with training. Moreover, ACtx response changes vary between stimuli, and such differences display lag synchronization with the learning rate. These results indicate that learning alters functional connections between ACtx subregions, inducing region-specific modulations by propagating behavioral information from primary to higher-order areas.

Eelbrain, a Python toolkit for time-continuous analysis with temporal response functions

Even though human experience unfolds continuously in time, it is not strictly linear; instead, it entails cascading processes building hierarchical cognitive structures. For instance, during speech perception, humans transform a continuously varying acoustic signal into phonemes, words, and meaning, and these levels all have distinct but interdependent temporal structures. Time-lagged regression using temporal response functions (TRFs) has recently emerged as a promising tool for disentangling electrophysiological brain responses related to such complex models of perception. Here, we introduce the Eelbrain Python toolkit, which makes this kind of analysis easy and accessible. We demonstrate its use, using continuous speech as a sample paradigm, with a freely available EEG dataset of audiobook listening. A companion GitHub repository provides the complete source code for the analysis, from raw data to group-level statistics. More generally, we advocate a hypothesis-driven approach in which the experimenter specifies a hierarchy of time-continuous representations that are hypothesized to have contributed to brain responses, and uses those as predictor variables for the electrophysiological signal. This is analogous to a multiple regression problem, but with the addition of a time dimension. TRF analysis decomposes the brain signal into distinct responses associated with the different predictor variables by estimating a multivariate TRF (mTRF), quantifying the influence of each predictor on brain responses as a function of time(-lags). This allows asking two questions about the predictor variables: (1) Is there a significant neural representation corresponding to this predictor variable? And if so, (2) what are the temporal characteristics of the neural response associated with it? Thus, different predictor variables can be systematically combined and evaluated to jointly model neural processing at multiple hierarchical levels. We discuss applications of this approach, including the potential for linking algorithmic/representational theories at different cognitive levels to brain responses through computational models with appropriate linking hypotheses.

A sparse code for natural sound context in auditory cortex

Accurate sound perception can require integrating information over hundreds of milliseconds or even seconds. Spectro-temporal models of sound coding by single neurons in auditory cortex indicate that the majority of sound-evoked activity can be attributed to stimuli with a few tens of milliseconds. It remains uncertain how the auditory system integrates information about sensory context on a longer timescale. Here we characterized long-lasting contextual effects in auditory cortex (AC) using a diverse set of natural sound stimuli. We measured context effects as the difference in a neuron's response to a single probe sound following two different context sounds. Many AC neurons showed context effects lasting longer than the temporal window of a traditional spectro-temporal receptive field. The duration and magnitude of context effects varied substantially across neurons and stimuli. This diversity of context effects formed a sparse code across the neural population that encoded a wider range of contexts than any constituent neuron. Encoding model analysis indicates that context effects can be explained by activity in the local neural population, suggesting that recurrent local circuits support a long-lasting representation of sensory context in auditory cortex.

Short-term neuronal and synaptic plasticity act in synergy for deviance detection in spiking networks

Sensory areas of cortex respond more strongly to infrequent stimuli when these violate previously established regularities, a phenomenon known as deviance detection (DD). Previous modeling work has mainly attempted to explain DD on the basis of synaptic plasticity. However, a large fraction of cortical neurons also exhibit firing rate adaptation, an underexplored potential mechanism. Here, we investigate DD in a spiking neuronal network model with two types of short-term plasticity, fast synaptic short-term depression (STD) and slower threshold adaptation (TA). We probe the model with an oddball stimulation paradigm and assess DD by evaluating the network responses. We find that TA is sufficient to elicit DD. It achieves this by habituating neurons near the stimulation site that respond earliest to the frequently presented standard stimulus (local fatigue), which diminishes the response and promotes the recovery (global fatigue) of the wider network. Further, we find a synergy effect between STD and TA, where they interact with each other to achieve greater DD than the sum of their individual effects. We show that this synergy is caused by the local fatigue added by STD, which inhibits the global response to the frequently presented stimulus, allowing greater recovery of TA-mediated global fatigue and making the network more responsive to the deviant stimulus. Finally, we show that the magnitude of DD strongly depends on the timescale of stimulation. We conclude that highly predictable information can be encoded in strong local fatigue, which allows greater global recovery and subsequent heightened sensitivity for DD.

Time as a supervisor: temporal regularity and auditory object learning

Sensory systems appear to learn to transform incoming sensory information into perceptual representations, or "objects," that can inform and guide behavior with minimal explicit supervision. Here, we propose that the auditory system can achieve this goal by using time as a supervisor, i.e., by learning features of a stimulus that are temporally regular. We will show that this procedure generates a feature space sufficient to support fundamental computations of auditory perception. In detail, we consider the problem of discriminating between instances of a prototypical class of natural auditory objects, i.e., rhesus macaque vocalizations. We test discrimination in two ethologically relevant tasks: discrimination in a cluttered acoustic background and generalization to discriminate between novel exemplars. We show that an algorithm that learns these temporally regular features affords better or equivalent discrimination and generalization than conventional feature-selection algorithms, i.e., principal component analysis and independent component analysis. Our findings suggest that the slow temporal features of auditory stimuli may be sufficient for parsing auditory scenes and that the auditory brain could utilize these slowly changing temporal features.

Adaptive mechanisms facilitate robust performance in noise and in reverberation in an auditory categorization model

For robust vocalization perception, the auditory system must generalize over variability in vocalization production as well as variability arising from the listening environment (e.g., noise and reverberation). We previously demonstrated using guinea pig and marmoset vocalizations that a hierarchical model generalized over production variability by detecting sparse intermediate-complexity features that are maximally informative about vocalization category from a dense spectrotemporal input representation. Here, we explore three biologically feasible model extensions to generalize over environmental variability: (1) training in degraded conditions, (2) adaptation to sound statistics in the spectrotemporal stage and (3) sensitivity adjustment at the feature detection stage. All mechanisms improved vocalization categorization performance, but improvement trends varied across degradation type and vocalization type. One or more adaptive mechanisms were required for model performance to approach the behavioral performance of guinea pigs on a vocalization categorization task. These results highlight the contributions of adaptive mechanisms at multiple auditory processing stages to achieve robust auditory categorization.

Repetition plasticity in primary auditory cortex occurs across long timescales for spectrotemporally randomized pure-tones

Repetition plasticity is a ubiquitous property of sensory systems in which repetitive sensation causes either a decrease ("repetition suppression", i.e. "adaptation") or increase ("repetition enhancement", i.e. "facilitation") in the amplitude of neural responses. Timescales of repetition plasticity for sensory neurons typically span milliseconds to tens of seconds, with longer durations for cortical vs subcortical regions. Here, we used 2-photon (2P) imaging to study repetition plasticity in mouse primary auditory cortex (A1) layer 2/3 (L2/3) during the presentation of spectrotemporally randomized pure-tone frequencies. Our study revealed subpopulations of neurons with repetition plasticity for equiprobable frequencies spaced minutes apart over a 20-minute period. We found both repetition suppression and enhancement in individual neurons and on average across populations. Each neuron tended to show repetition plasticity for 1-2 pure-tone frequencies near the neuron's best frequency. Moreover, we found correlated changes in neural response amplitude and latency across stimulus repetitions. Together, our results highlight cortical specialization for pattern recognition over long timescales in complex acoustic sequences.

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.

Deep neural networks effectively model neural adaptation to changing background noise and suggest nonlinear noise filtering methods in auditory cortex

The human auditory system displays a robust capacity to adapt to sudden changes in background noise, allowing for continuous speech comprehension despite changes in background environments. However, despite comprehensive studies characterizing this ability, the computations that underly this process are not well understood. The first step towards understanding a complex system is to propose a suitable model, but the classical and easily interpreted model for the auditory system, the spectro-temporal receptive field (STRF), cannot match the nonlinear neural dynamics involved in noise adaptation. Here, we utilize a deep neural network (DNN) to model neural adaptation to noise, illustrating its effectiveness at reproducing the complex dynamics at the levels of both individual electrodes and the cortical population. By closely inspecting the model's STRF-like computations over time, we find that the model alters both the gain and shape of its receptive field when adapting to a sudden noise change. We show that the DNN model's gain changes allow it to perform adaptive gain control, while the spectro-temporal change creates noise filtering by altering the inhibitory region of the model's receptive field. Further, we find that models of electrodes in nonprimary auditory cortex also exhibit noise filtering changes in their excitatory regions, suggesting differences in noise filtering mechanisms along the cortical hierarchy. These findings demonstrate the capability of deep neural networks to model complex neural adaptation and offer new hypotheses about the computations the auditory cortex performs to enable noise-robust speech perception in real-world, dynamic environments.

Auditory cortex modelled as a dynamical network of oscillators: understanding event-related fields and their adaptation

Adaptation, the reduction of neuronal responses by repetitive stimulation, is a ubiquitous feature of auditory cortex (AC). It is not clear what causes adaptation, but short-term synaptic depression (STSD) is a potential candidate for the underlying mechanism. In such a case, adaptation can be directly linked with the way AC produces context-sensitive responses such as mismatch negativity and stimulus-specific adaptation observed on the single-unit level. We examined this hypothesis via a computational model based on AC anatomy, which includes serially connected core, belt, and parabelt areas. The model replicates the event-related field (ERF) of the magnetoencephalogram as well as ERF adaptation. The model dynamics are described by excitatory and inhibitory state variables of cell populations, with the excitatory connections modulated by STSD. We analysed the system dynamics by linearising the firing rates and solving the STSD equation using time-scale separation. This allows for characterisation of AC dynamics as a superposition of damped harmonic oscillators, so-called normal modes. We show that repetition suppression of the N1m is due to a mixture of causes, with stimulus repetition modifying both the amplitudes and the frequencies of the normal modes. In this view, adaptation results from a complete reorganisation of AC dynamics rather than a reduction of activity in discrete sources. Further, both the network structure and the balance between excitation and inhibition contribute significantly to the rate with which AC recovers from adaptation. This lifetime of adaptation is longer in the belt and parabelt than in the core area, despite the time constants of STSD being spatially homogeneous. Finally, we critically evaluate the use of a single exponential function to describe recovery from adaptation.

Encoding of acquired sound-sequence salience by auditory cortical offset responses

Behaviorally relevant sounds are often composed of distinct acoustic units organized into specific temporal sequences. The meaning of such sound sequences can therefore be fully recognized only when they have terminated. However, the neural mechanisms underlying the perception of sound sequences remain unclear. Here, we use two-photon calcium imaging in the auditory cortex of behaving mice to test the hypothesis that neural responses to termination of sound sequences ("Off-responses") encode their acoustic history and behavioral salience. We find that auditory cortical Off-responses encode preceding sound sequences and that learning to associate a sound sequence with a reward induces enhancement of Off-responses relative to responses during the sound sequence ("On-responses"). Furthermore, learning enhances network-level discriminability of sound sequences by Off-responses. Last, learning-induced plasticity of Off-responses but not On-responses lasts to the next day. These findings identify auditory cortical Off-responses as a key neural signature of acquired sound-sequence salience.

Learning Speech Production and Perception through Sensorimotor Interactions

Action and perception are closely linked in many behaviors necessitating a close coordination between sensory and motor neural processes so as to achieve a well-integrated smoothly evolving task performance. To investigate the detailed nature of these sensorimotor interactions, and their role in learning and executing the skilled motor task of speaking, we analyzed ECoG recordings of responses in the high-γ band (70-150 Hz) in human subjects while they listened to, spoke, or silently articulated speech. We found elaborate spectrotemporally modulated neural activity projecting in both "forward" (motor-to-sensory) and "inverse" directions between the higher-auditory and motor cortical regions engaged during speaking. Furthermore, mathematical simulations demonstrate a key role for the forward projection in "learning" to control the vocal tract, beyond its commonly postulated predictive role during execution. These results therefore offer a broader view of the functional role of the ubiquitous forward projection as an important ingredient in learning, rather than just control, of skilled sensorimotor tasks.

Complementary Effects of Adaptation and Gain Control on Sound Encoding in Primary Auditory Cortex

An important step toward understanding how the brain represents complex natural sounds is to develop accurate models of auditory coding by single neurons. A commonly used model is the linear-nonlinear spectro-temporal receptive field (STRF; LN model). The LN model accounts for many features of auditory tuning, but it cannot account for long-lasting effects of sensory context on sound-evoked activity. Two mechanisms that may support these contextual effects are short-term plasticity (STP) and contrast-dependent gain control (GC), which have inspired expanded versions of the LN model. Both models improve performance over the LN model, but they have never been compared directly. Thus, it is unclear whether they account for distinct processes or describe one phenomenon in different ways. To address this question, we recorded activity of neurons in primary auditory cortex (A1) of awake ferrets during presentation of natural sounds. We then fit models incorporating one nonlinear mechanism (GC or STP) or both (GC+STP) using this single dataset, and measured the correlation between the models' predictions and the recorded neural activity. Both the STP and GC models performed significantly better than the LN model, but the GC+STP model outperformed both individual models. We also quantified the equivalence of STP and GC model predictions and found only modest similarity. Consistent results were observed for a dataset collected in clean and noisy acoustic contexts. These results establish general methods for evaluating the equivalence of arbitrarily complex encoding models and suggest that the STP and GC models describe complementary processes in the auditory system.

Estimating and interpreting nonlinear receptive field of sensory neural responses with deep neural network models

Our understanding of nonlinear stimulus transformations by neural circuits is hindered by the lack of comprehensive yet interpretable computational modeling frameworks. Here, we propose a data-driven approach based on deep neural networks to directly model arbitrarily nonlinear stimulus-response mappings. Reformulating the exact function of a trained neural network as a collection of stimulus-dependent linear functions enables a locally linear receptive field interpretation of the neural network. Predicting the neural responses recorded invasively from the auditory cortex of neurosurgical patients as they listened to speech, this approach significantly improves the prediction accuracy of auditory cortical responses, particularly in nonprimary areas. Moreover, interpreting the functions learned by neural networks uncovered three distinct types of nonlinear transformations of speech that varied considerably from primary to nonprimary auditory regions. The ability of this framework to capture arbitrary stimulus-response mappings while maintaining model interpretability leads to a better understanding of cortical processing of sensory signals.

Streaming of Repeated Noise in Primary and Secondary Fields of Auditory Cortex

Statistical regularities in natural sounds facilitate the perceptual segregation of auditory sources, or streams. Repetition is one cue that drives stream segregation in humans, but the neural basis of this perceptual phenomenon remains unknown. We demonstrated a similar perceptual ability in animals by training ferrets of both sexes to detect a stream of repeating noise samples (foreground) embedded in a stream of random samples (background). During passive listening, we recorded neural activity in primary auditory cortex (A1) and secondary auditory cortex (posterior ectosylvian gyrus, PEG). We used two context-dependent encoding models to test for evidence of streaming of the repeating stimulus. The first was based on average evoked activity per noise sample and the second on the spectro-temporal receptive field. Both approaches tested whether differences in neural responses to repeating versus random stimuli were better modeled by scaling the response to both streams equally (global gain) or by separately scaling the response to the foreground versus background stream (stream-specific gain). Consistent with previous observations of adaptation, we found an overall reduction in global gain when the stimulus began to repeat. However, when we measured stream-specific changes in gain, responses to the foreground were enhanced relative to the background. This enhancement was stronger in PEG than A1. In A1, enhancement was strongest in units with low sparseness (i.e., broad sensory tuning) and with tuning selective for the repeated sample. Enhancement of responses to the foreground relative to the background provides evidence for stream segregation that emerges in A1 and is refined in PEG.SIGNIFICANCE STATEMENT To interact with the world successfully, the brain must parse behaviorally important information from a complex sensory environment. Complex mixtures of sounds often arrive at the ears simultaneously or in close succession, yet they are effortlessly segregated into distinct perceptual sources. This process breaks down in hearing-impaired individuals and speech recognition devices. By identifying the underlying neural mechanisms that facilitate perceptual segregation, we can develop strategies for ameliorating hearing loss and improving speech recognition technology in the presence of background noise. Here, we present evidence to support a hierarchical process, present in primary auditory cortex and refined in secondary auditory cortex, in which sound repetition facilitates segregation.

The role of adaptation in generating monotonic rate codes in auditory cortex

In primary auditory cortex, slowly repeated acoustic events are represented temporally by the stimulus-locked activity of single neurons. Single-unit studies in awake marmosets (Callithrix jacchus) have shown that a sub-population of these neurons also monotonically increase or decrease their average discharge rate during stimulus presentation for higher repetition rates. Building on a computational single-neuron model that generates stimulus-locked responses with stimulus evoked excitation followed by strong inhibition, we find that stimulus-evoked short-term depression is sufficient to produce synchronized monotonic positive and negative responses to slowly repeated stimuli. By exploring model robustness and comparing it to other models for adaptation to such stimuli, we conclude that short-term depression best explains our observations in single-unit recordings in awake marmosets. Together, our results show how a simple biophysical mechanism in single neurons can generate complementary neural codes for acoustic stimuli.

Decoding speech from spike-based neural population recordings in secondary auditory cortex of non-human primates

Direct electronic communication with sensory areas of the neocortex is a challenging ambition for brain-computer interfaces. Here, we report the first successful neural decoding of English words with high intelligibility from intracortical spike-based neural population activity recorded from the secondary auditory cortex of macaques. We acquired 96-channel full-broadband population recordings using intracortical microelectrode arrays in the rostral and caudal parabelt regions of the superior temporal gyrus (STG). We leveraged a new neural processing toolkit to investigate the choice of decoding algorithm, neural preprocessing, audio representation, channel count, and array location on neural decoding performance. The presented spike-based machine learning neural decoding approach may further be useful in informing future encoding strategies to deliver direct auditory percepts to the brain as specific patterns of microstimulation.

Spectral tuning of adaptation supports coding of sensory context in auditory cortex

Perception of vocalizations and other behaviorally relevant sounds requires integrating acoustic information over hundreds of milliseconds. Sound-evoked activity in auditory cortex typically has much shorter latency, but the acoustic context, i.e., sound history, can modulate sound evoked activity over longer periods. Contextual effects are attributed to modulatory phenomena, such as stimulus-specific adaption and contrast gain control. However, an encoding model that links context to natural sound processing has yet to be established. We tested whether a model in which spectrally tuned inputs undergo adaptation mimicking short-term synaptic plasticity (STP) can account for contextual effects during natural sound processing. Single-unit activity was recorded from primary auditory cortex of awake ferrets during presentation of noise with natural temporal dynamics and fully natural sounds. Encoding properties were characterized by a standard linear-nonlinear spectro-temporal receptive field (LN) model and variants that incorporated STP-like adaptation. In the adapting models, STP was applied either globally across all input spectral channels or locally to subsets of channels. For most neurons, models incorporating local STP predicted neural activity as well or better than LN and global STP models. The strength of nonlinear adaptation varied across neurons. Within neurons, adaptation was generally stronger for spectral channels with excitatory than inhibitory gain. Neurons showing improved STP model performance also tended to undergo stimulus-specific adaptation, suggesting a common mechanism for these phenomena. When STP models were compared between passive and active behavior conditions, response gain often changed, but average STP parameters were stable. Thus, spectrally and temporally heterogeneous adaptation, subserved by a mechanism with STP-like dynamics, may support representation of the complex spectro-temporal patterns that comprise natural sounds across wide-ranging sensory contexts.

Invariant neural responses for sensory categories revealed by the time-varying information for communication calls

Although information theoretic approaches have been used extensively in the analysis of the neural code, they have yet to be used to describe how information is accumulated in time while sensory systems are categorizing dynamic sensory stimuli such as speech sounds or visual objects. Here, we present a novel method to estimate the cumulative information for stimuli or categories. We further define a time-varying categorical information index that, by comparing the information obtained for stimuli versus categories of these same stimuli, quantifies invariant neural representations. We use these methods to investigate the dynamic properties of avian cortical auditory neurons recorded in zebra finches that were listening to a large set of call stimuli sampled from the complete vocal repertoire of this species. We found that the time-varying rates carry 5 times more information than the mean firing rates even in the first 100 ms. We also found that cumulative information has slow time constants (100–600 ms) relative to the typical integration time of single neurons, reflecting the fact that the behaviorally informative features of auditory objects are time-varying sound patterns. When we correlated firing rates and information values, we found that average information correlates with average firing rate but that higher-rates found at the onset response yielded similar information values as the lower-rates found in the sustained response: the onset and sustained response of avian cortical auditory neurons provide similar levels of independent information about call identity and call-type. Finally, our information measures allowed us to rigorously define categorical neurons; these categorical neurons show a high degree of invariance for vocalizations within a call-type. Peak invariance is found around 150 ms after stimulus onset. Surprisingly, call-type invariant neurons were found in both primary and secondary avian auditory areas.

Just as the recognition of faces requires neural representations that are invariant to scale and rotation, the recognition of behaviorally relevant auditory objects, such as spoken words, requires neural representations that are invariant to the speaker uttering the word and to his or her location. Here, we used information theory to investigate the time course of the neural representation of bird communication calls and of behaviorally relevant categories of these same calls: the call-types of the bird’s repertoire. We found that neurons in both the primary and secondary avian auditory cortex exhibit invariant responses to call renditions within a call-type, suggestive of a potential role for extracting the meaning of these communication calls. We also found that time plays an important role: first, neural responses carry significantly more information when represented by temporal patterns calculated at the small time scale of 10 ms than when measured as average rates and, second, this information accumulates in a non-redundant fashion up to long integration times of 600 ms. This rich temporal neural representation is matched to the temporal richness found in the communication calls of this species.

A dynamic network model of temporal receptive fields in primary auditory cortex

Auditory neurons encode stimulus history, which is often modelled using a span of time-delays in a spectro-temporal receptive field (STRF). We propose an alternative model for the encoding of stimulus history, which we apply to extracellular recordings of neurons in the primary auditory cortex of anaesthetized ferrets. For a linear-non-linear STRF model (LN model) to achieve a high level of performance in predicting single unit neural responses to natural sounds in the primary auditory cortex, we found that it is necessary to include time delays going back at least 200 ms in the past. This is an unrealistic time span for biological delay lines. We therefore asked how much of this dependence on stimulus history can instead be explained by dynamical aspects of neurons. We constructed a neural-network model whose output is the weighted sum of units whose responses are determined by a dynamic firing-rate equation. The dynamic aspect performs low-pass filtering on each unit's response, providing an exponentially decaying memory whose time constant is individual to each unit. We find that this dynamic network (DNet) model, when fitted to the neural data using STRFs of only 25 ms duration, can achieve prediction performance on a held-out dataset comparable to the best performing LN model with STRFs of 200 ms duration. These findings suggest that integration due to the membrane time constants or other exponentially-decaying memory processes may underlie linear temporal receptive fields of neurons beyond 25 ms.

Neural responses to natural and model-matched stimuli reveal distinct computations in primary and nonprimary auditory cortex

A central goal of sensory neuroscience is to construct models that can explain neural responses to natural stimuli. As a consequence, sensory models are often tested by comparing neural responses to natural stimuli with model responses to those stimuli. One challenge is that distinct model features are often correlated across natural stimuli, and thus model features can predict neural responses even if they do not in fact drive them. Here, we propose a simple alternative for testing a sensory model: we synthesize a stimulus that yields the same model response as each of a set of natural stimuli, and test whether the natural and "model-matched" stimuli elicit the same neural responses. We used this approach to test whether a common model of auditory cortex-in which spectrogram-like peripheral input is processed by linear spectrotemporal filters-can explain fMRI responses in humans to natural sounds. Prior studies have that shown that this model has good predictive power throughout auditory cortex, but this finding could reflect feature correlations in natural stimuli. We observed that fMRI responses to natural and model-matched stimuli were nearly equivalent in primary auditory cortex (PAC) but that nonprimary regions, including those selective for music or speech, showed highly divergent responses to the two sound sets. This dissociation between primary and nonprimary regions was less clear from model predictions due to the influence of feature correlations across natural stimuli. Our results provide a signature of hierarchical organization in human auditory cortex, and suggest that nonprimary regions compute higher-order stimulus properties that are not well captured by traditional models. Our methodology enables stronger tests of sensory models and could be broadly applied in other domains.

Recent advances in understanding the auditory cortex

Our ability to make sense of the auditory world results from neural processing that begins in the ear, goes through multiple subcortical areas, and continues in the cortex. The specific contribution of the auditory cortex to this chain of processing is far from understood. Although many of the properties of neurons in the auditory cortex resemble those of subcortical neurons, they show somewhat more complex selectivity for sound features, which is likely to be important for the analysis of natural sounds, such as speech, in real-life listening conditions. Furthermore, recent work has shown that auditory cortical processing is highly context-dependent, integrates auditory inputs with other sensory and motor signals, depends on experience, and is shaped by cognitive demands, such as attention. Thus, in addition to being the locus for more complex sound selectivity, the auditory cortex is increasingly understood to be an integral part of the network of brain regions responsible for prediction, auditory perceptual decision-making, and learning. In this review, we focus on three key areas that are contributing to this understanding: the sound features that are preferentially represented by cortical neurons, the spatial organization of those preferences, and the cognitive roles of the auditory cortex.

Focal Suppression of Distractor Sounds by Selective Attention in Auditory Cortex

Auditory selective attention is required for parsing crowded acoustic environments, but cortical systems mediating the influence of behavioral state on auditory perception are not well characterized. Previous neurophysiological studies suggest that attention produces a general enhancement of neural responses to important target sounds versus irrelevant distractors. However, behavioral studies suggest that in the presence of masking noise, attention provides a focal suppression of distractors that compete with targets. Here, we compared effects of attention on cortical responses to masking versus non-masking distractors, controlling for effects of listening effort and general task engagement. We recorded single-unit activity from primary auditory cortex (A1) of ferrets during behavior and found that selective attention decreased responses to distractors masking targets in the same spectral band, compared with spectrally distinct distractors. This suppression enhanced neural target detection thresholds, suggesting that limited attention resources serve to focally suppress responses to distractors that interfere with target detection. Changing effort by manipulating target salience consistently modulated spontaneous but not evoked activity. Task engagement and changing effort tended to affect the same neurons, while attention affected an independent population, suggesting that distinct feedback circuits mediate effects of attention and effort in A1.

Coordinated neuronal ensembles in primary auditory cortical columns

The synchronous activity of groups of neurons is increasingly thought to be important in cortical information processing and transmission. However, most studies of processing in the primary auditory cortex (AI) have viewed neurons as independent filters; little is known about how coordinated AI neuronal activity is expressed throughout cortical columns and how it might enhance the processing of auditory information. To address this, we recorded from populations of neurons in AI cortical columns of anesthetized rats and, using dimensionality reduction techniques, identified multiple coordinated neuronal ensembles (cNEs), which are groups of neurons with reliable synchronous activity. We show that cNEs reflect local network configurations with enhanced information encoding properties that cannot be accounted for by stimulus-driven synchronization alone. Furthermore, similar cNEs were identified in both spontaneous and evoked activity, indicating that columnar cNEs are stable functional constructs that may represent principal units of information processing in AI.

Incorporating behavioral and sensory context into spectro-temporal models of auditory encoding

For several decades, auditory neuroscientists have used spectro-temporal encoding models to understand how neurons in the auditory system represent sound. Derived from early applications of systems identification tools to the auditory periphery, the spectro-temporal receptive field (STRF) and more sophisticated variants have emerged as an efficient means of characterizing representation throughout the auditory system. Most of these encoding models describe neurons as static sensory filters. However, auditory neural coding is not static. Sensory context, reflecting the acoustic environment, and behavioral context, reflecting the internal state of the listener, can both influence sound-evoked activity, particularly in central auditory areas. This review explores recent efforts to integrate context into spectro-temporal encoding models. It begins with a brief tutorial on the basics of estimating and interpreting STRFs. Then it describes three recent studies that have characterized contextual effects on STRFs, emerging over a range of timescales, from many minutes to tens of milliseconds. An important theme of this work is not simply that context influences auditory coding, but also that contextual effects span a large continuum of internal states. The added complexity of these context-dependent models introduces new experimental and theoretical challenges that must be addressed in order to be used effectively. Several new methodological advances promise to address these limitations and allow the development of more comprehensive context-dependent models in the future.

Encoding and Decoding Models in Cognitive Electrophysiology

Cognitive neuroscience has seen rapid growth in the size and complexity of data recorded from the human brain as well as in the computational tools available to analyze this data. This data explosion has resulted in an increased use of multivariate, model-based methods for asking neuroscience questions, allowing scientists to investigate multiple hypotheses with a single dataset, to use complex, time-varying stimuli, and to study the human brain under more naturalistic conditions. These tools come in the form of "Encoding" models, in which stimulus features are used to model brain activity, and "Decoding" models, in which neural features are used to generated a stimulus output. Here we review the current state of encoding and decoding models in cognitive electrophysiology and provide a practical guide toward conducting experiments and analyses in this emerging field. Our examples focus on using linear models in the study of human language and audition. We show how to calculate auditory receptive fields from natural sounds as well as how to decode neural recordings to predict speech. The paper aims to be a useful tutorial to these approaches, and a practical introduction to using machine learning and applied statistics to build models of neural activity. The data analytic approaches we discuss may also be applied to other sensory modalities, motor systems, and cognitive systems, and we cover some examples in these areas. In addition, a collection of Jupyter notebooks is publicly available as a complement to the material covered in this paper, providing code examples and tutorials for predictive modeling in python. The aim is to provide a practical understanding of predictive modeling of human brain data and to propose best-practices in conducting these analyses.

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.

Variable Temporal Integration of Stimulus Patterns in the Mouse Barrel Cortex

Making sense of the world requires distinguishing temporal patterns and sequences lasting hundreds of milliseconds or more. How cortical circuits integrate over time to represent specific sensory sequences remains elusive. Here we assessed whether neurons in the barrel cortex (BC) integrate information about temporal patterns of whisker movements. We performed cell-attached recordings in anesthetized mice while delivering whisker deflections at variable intervals and compared the information carried by neurons about the latest interstimulus interval (reflecting sensitivity to instantaneous frequency) and earlier intervals (reflecting integration over timescales up to several hundred milliseconds). Neurons carried more information about the latest interval than earlier ones. The amount of temporal integration varied with neuronal responsiveness and with the cortical depth of the recording site, that is, with laminar location. A subset of neurons in the upper layers displayed the strongest integration. Highly responsive neurons in the deeper layers encoded the latest interval but integrated particularly weakly. Under these conditions, BC neurons act primarily as encoders of current stimulation parameters; however, our results suggest that temporal integration over hundreds of milliseconds can emerge in some neurons within BC.

Predictive Ensemble Decoding of Acoustical Features Explains Context-Dependent Receptive Fields

A primary goal of auditory neuroscience is to identify the sound features extracted and represented by auditory neurons. Linear encoding models, which describe neural responses as a function of the stimulus, have been primarily used for this purpose. Here, we provide theoretical arguments and experimental evidence in support of an alternative approach, based on decoding the stimulus from the neural response. We used a Bayesian normative approach to predict the responses of neurons detecting relevant auditory features, despite ambiguities and noise. We compared the model predictions to recordings from the primary auditory cortex of ferrets and found that: (1) the decoding filters of auditory neurons resemble the filters learned from the statistics of speech sounds; (2) the decoding model captures the dynamics of responses better than a linear encoding model of similar complexity; and (3) the decoding model accounts for the accuracy with which the stimulus is represented in neural activity, whereas linear encoding model performs very poorly. Most importantly, our model predicts that neuronal responses are fundamentally shaped by "explaining away," a divisive competition between alternative interpretations of the auditory scene.

SIGNIFICANCE STATEMENT

Neural responses in the auditory cortex are dynamic, nonlinear, and hard to predict. Traditionally, encoding models have been used to describe neural responses as a function of the stimulus. However, in addition to external stimulation, neural activity is strongly modulated by the responses of other neurons in the network. We hypothesized that auditory neurons aim to collectively decode their stimulus. In particular, a stimulus feature that is decoded (or explained away) by one neuron is not explained by another. We demonstrated that this novel Bayesian decoding model is better at capturing the dynamic responses of cortical neurons in ferrets. Whereas the linear encoding model poorly reflects selectivity of neurons, the decoding model can account for the strong nonlinearities observed in neural data.

Network Receptive Field Modeling Reveals Extensive Integration and Multi-feature Selectivity in Auditory Cortical Neurons

Cortical sensory neurons are commonly characterized using the receptive field, the linear dependence of their response on the stimulus. In primary auditory cortex neurons can be characterized by their spectrotemporal receptive fields, the spectral and temporal features of a sound that linearly drive a neuron. However, receptive fields do not capture the fact that the response of a cortical neuron results from the complex nonlinear network in which it is embedded. By fitting a nonlinear feedforward network model (a network receptive field) to cortical responses to natural sounds, we reveal that primary auditory cortical neurons are sensitive over a substantially larger spectrotemporal domain than is seen in their standard spectrotemporal receptive fields. Furthermore, the network receptive field, a parsimonious network consisting of 1-7 sub-receptive fields that interact nonlinearly, consistently better predicts neural responses to auditory stimuli than the standard receptive fields. The network receptive field reveals separate excitatory and inhibitory sub-fields with different nonlinear properties, and interaction of the sub-fields gives rise to important operations such as gain control and conjunctive feature detection. The conjunctive effects, where neurons respond only if several specific features are present together, enable increased selectivity for particular complex spectrotemporal structures, and may constitute an important stage in sound recognition. In conclusion, we demonstrate that fitting auditory cortical neural responses with feedforward network models expands on simple linear receptive field models in a manner that yields substantially improved predictive power and reveals key nonlinear aspects of cortical processing, while remaining easy to interpret in a physiological context.

Distinct Spatiotemporal Response Properties of Excitatory Versus Inhibitory Neurons in the Mouse Auditory Cortex

In the auditory system, early neural stations such as brain stem are characterized by strict tonotopy, which is used to deconstruct sounds to their basic frequencies. But higher along the auditory hierarchy, as early as primary auditory cortex (A1), tonotopy starts breaking down at local circuits. Here, we studied the response properties of both excitatory and inhibitory neurons in the auditory cortex of anesthetized mice. We used in vivo two photon-targeted cell-attached recordings from identified parvalbumin-positive neurons (PVNs) and their excitatory pyramidal neighbors (PyrNs). We show that PyrNs are locally heterogeneous as characterized by diverse best frequencies, pairwise signal correlations, and response timing. In marked contrast, neighboring PVNs exhibited homogenous response properties in pairwise signal correlations and temporal responses. The distinct physiological microarchitecture of different cell types is maintained qualitatively in response to natural sounds. Excitatory heterogeneity and inhibitory homogeneity within the same circuit suggest different roles for each population in coding natural stimuli.

The Representation of Prediction Error in Auditory Cortex

To survive, organisms must extract information from the past that is relevant for their future. How this process is expressed at the neural level remains unclear. We address this problem by developing a novel approach from first principles. We show here how to generate low-complexity representations of the past that produce optimal predictions of future events. We then illustrate this framework by studying the coding of ‘oddball’ sequences in auditory cortex. We find that for many neurons in primary auditory cortex, trial-by-trial fluctuations of neuronal responses correlate with the theoretical prediction error calculated from the short-term past of the stimulation sequence, under constraints on the complexity of the representation of this past sequence. In some neurons, the effect of prediction error accounted for more than 50% of response variability. Reliable predictions often depended on a representation of the sequence of the last ten or more stimuli, although the representation kept only few details of that sequence.

A crucial aspect of all life is the ability to use past events in order to guide future behavior. To do that, creatures need the ability to predict future events. Indeed, predictability has been shown to affect neuronal responses in many animals and under many conditions. Clearly, the quality of predictions should depend on the amount and detail of the past information used to generate them. Here, by using a basic principle from information theory, we show how to derive explicitly the tradeoff between quality of prediction and complexity of the representation of past information. We then apply these ideas to a concrete case–neuronal responses recorded in auditory cortex during the presentation of oddball sequences, consisting of two tones with varying probabilities. We show that the neuronal responses fit quantitatively the prediction errors of optimal predictors derived from our theory, and use that result in order to deduce the properties of the representations of the past in the auditory system. We conclude that these memory representations have surprisingly long duration (10 stimuli back or more), but keep relatively little detail about this past. Our theory can be applied widely to other sensory systems.

Capturing contextual effects in spectro-temporal receptive fields

Spectro-temporal receptive fields (STRFs) are thought to provide descriptive images of the computations performed by neurons along the auditory pathway. However, their validity can be questioned because they rely on a set of assumptions that are probably not fulfilled by real neurons exhibiting contextual effects, that is, nonlinear interactions in the time or frequency dimension that cannot be described with a linear filter. We used a novel approach to investigate how a variety of contextual effects, due to facilitating nonlinear interactions and synaptic depression, affect different STRF models, and if these effects can be captured with a context field (CF). Contextual effects were incorporated in simulated networks of spiking neurons, allowing one to define the true STRFs of the neurons. This, in turn, made it possible to evaluate the performance of each STRF model by comparing the estimations with the true STRFs. We found that currently used STRF models are particularly poor at estimating inhibitory regions. Specifically, contextual effects make estimated STRFs dependent on stimulus density in a contrasting fashion: inhibitory regions are underestimated at lower densities while artificial inhibitory regions emerge at higher densities. The CF was found to provide a solution to this dilemma, but only when it is used together with a generalized linear model. Our results therefore highlight the limitations of the traditional STRF approach and provide useful recipes for how different STRF models and stimuli can be used to arrive at reliable quantifications of neural computations in the presence of contextual effects. The results therefore push the purpose of STRF analysis from simply finding an optimal stimulus toward describing context-dependent computations of neurons along the auditory pathway.

Input-Specific Gain Modulation by Local Sensory Context Shapes Cortical and Thalamic Responses to Complex Sounds

Sensory neurons are customarily characterized by one or more linearly weighted receptive fields describing sensitivity in sensory space and time. We show that in auditory cortical and thalamic neurons, the weight of each receptive field element depends on the pattern of sound falling within a local neighborhood surrounding it in time and frequency. Accounting for this change in effective receptive field with spectrotemporal context improves predictions of both cortical and thalamic responses to stationary complex sounds. Although context dependence varies among neurons and across brain areas, there are strong shared qualitative characteristics. In a spectrotemporally rich soundscape, sound elements modulate neuronal responsiveness more effectively when they coincide with sounds at other frequencies, and less effectively when they are preceded by sounds at similar frequencies. This local-context-driven lability in the representation of complex sounds—a modulation of “input-specific gain” rather than “output gain”—may be a widespread motif in sensory processing.

•

Gain of neuronal responses to sound components varies with immediate acoustic context

•

“Contextual gain fields” can be estimated from neuronal responses to complex sounds

•

Coincident sound at different frequencies boosts gain in cortex and thalamus

•

Preceding sound at similar frequency reduces gain for longer in cortex than thalamus

A Model of Synaptic Reconsolidation

Reconsolidation of memories has mostly been studied at the behavioral and molecular level. Here, we put forward a simple extension of existing computational models of synaptic consolidation to capture hippocampal slice experiments that have been interpreted as reconsolidation at the synaptic level. The model implements reconsolidation through stabilization of consolidated synapses by stabilizing entities combined with an activity-dependent reservoir of stabilizing entities that are immune to protein synthesis inhibition (PSI). We derive a reduced version of our model to explore the conditions under which synaptic reconsolidation does or does not occur, often referred to as the boundary conditions of reconsolidation. We find that our computational model of synaptic reconsolidation displays complex boundary conditions. Our results suggest that a limited resource of hypothetical stabilizing molecules or complexes, which may be implemented by protein phosphorylation or different receptor subtypes, can underlie the phenomenon of synaptic reconsolidation.

Presynaptic Adenosine Receptor-Mediated Regulation of Diverse Thalamocortical Short-Term Plasticity in the Mouse Whisker Pathway

Short-term synaptic plasticity (STP) sets the sensitivity of a synapse to incoming activity and determines the temporal patterns that it best transmits. In “driver” thalamocortical (TC) synaptic populations, STP is dominated by depression during stimulation from rest. However, during ongoing stimulation, lemniscal TC connections onto layer 4 neurons in mouse barrel cortex express variable STP. Each synapse responds to input trains with a distinct pattern of depression or facilitation around its mean steady-state response. As a result, in common with other synaptic populations, lemniscal TC synapses express diverse rather than uniform dynamics, allowing for a rich representation of temporally varying stimuli. Here, we show that this STP diversity is regulated presynaptically. Presynaptic adenosine receptors of the A1R type, but not kainate receptors (KARs), modulate STP behavior. Blocking the receptors does not eliminate diversity, indicating that diversity is related to heterogeneous expression of multiple mechanisms in the pathway from presynaptic calcium influx to neurotransmitter release.

Measuring the Performance of Neural Models

Good metrics of the performance of a statistical or computational model are essential for model comparison and selection. Here, we address the design of performance metrics for models that aim to predict neural responses to sensory inputs. This is particularly difficult because the responses of sensory neurons are inherently variable, even in response to repeated presentations of identical stimuli. In this situation, standard metrics (such as the correlation coefficient) fail because they do not distinguish between explainable variance (the part of the neural response that is systematically dependent on the stimulus) and response variability (the part of the neural response that is not systematically dependent on the stimulus, and cannot be explained by modeling the stimulus-response relationship). As a result, models which perfectly describe the systematic stimulus-response relationship may appear to perform poorly. Two metrics have previously been proposed which account for this inherent variability: Signal Power Explained (SPE, Sahani and Linden, 2003), and the normalized correlation coefficient (CC norm , Hsu et al., 2004). Here, we analyze these metrics, and show that they are intimately related. However, SPE has no lower bound, and we show that, even for good models, SPE can yield negative values that are difficult to interpret. CC norm is better behaved in that it is effectively bounded between -1 and 1, and values below zero are very rare in practice and easy to interpret. However, it was hitherto not possible to calculate CC norm directly; instead, it was estimated using imprecise and laborious resampling techniques. Here, we identify a new approach that can calculate CC norm quickly and accurately. As a result, we argue that it is now a better choice of metric than SPE to accurately evaluate the performance of neural models.

Memory Stacking in Hierarchical Networks

Robust representations of sounds with a complex spectrotemporal structure are thought to emerge in hierarchically organized auditory cortex, but the computational advantage of this hierarchy remains unknown. Here, we used computational models to study how such hierarchical structures affect temporal binding in neural networks. We equipped individual units in different types of feedforward networks with local memory mechanisms storing recent inputs and observed how this affected the ability of the networks to process stimuli context dependently. Our findings illustrate that these local memories stack up in hierarchical structures and hence allow network units to exhibit selectivity to spectral sequences longer than the time spans of the local memories. We also illustrate that short-term synaptic plasticity is a potential local memory mechanism within the auditory cortex, and we show that it can bring robustness to context dependence against variation in the temporal rate of stimuli, while introducing nonlinearities to response profiles that are not well captured by standard linear spectrotemporal receptive field models. The results therefore indicate that short-term synaptic plasticity might provide hierarchically structured auditory cortex with computational capabilities important for robust representations of spectrotemporal patterns.

Incorporating Midbrain Adaptation to Mean Sound Level Improves Models of Auditory Cortical Processing

Adaptation to stimulus statistics, such as the mean level and contrast of recently heard sounds, has been demonstrated at various levels of the auditory pathway. It allows the nervous system to operate over the wide range of intensities and contrasts found in the natural world. Yet current standard models of the response properties of auditory neurons do not incorporate such adaptation. Here we present a model of neural responses in the ferret auditory cortex (the IC Adaptation model), which takes into account adaptation to mean sound level at a lower level of processing: the inferior colliculus (IC). The model performs high-pass filtering with frequency-dependent time constants on the sound spectrogram, followed by half-wave rectification, and passes the output to a standard linear-nonlinear (LN) model. We find that the IC Adaptation model consistently predicts cortical responses better than the standard LN model for a range of synthetic and natural stimuli. The IC Adaptation model introduces no extra free parameters, so it improves predictions without sacrificing parsimony. Furthermore, the time constants of adaptation in the IC appear to be matched to the statistics of natural sounds, suggesting that neurons in the auditory midbrain predict the mean level of future sounds and adapt their responses appropriately.

SIGNIFICANCE STATEMENT

An ability to accurately predict how sensory neurons respond to novel stimuli is critical if we are to fully characterize their response properties. Attempts to model these responses have had a distinguished history, but it has proven difficult to improve their predictive power significantly beyond that of simple, mostly linear receptive field models. Here we show that auditory cortex receptive field models benefit from a nonlinear preprocessing stage that replicates known adaptation properties of the auditory midbrain. This improves their predictive power across a wide range of stimuli but keeps model complexity low as it introduces no new free parameters. Incorporating the adaptive coding properties of neurons will likely improve receptive field models in other sensory modalities too.

The Essential Complexity of Auditory Receptive Fields

Encoding properties of sensory neurons are commonly modeled using linear finite impulse response (FIR) filters. For the auditory system, the FIR filter is instantiated in the spectro-temporal receptive field (STRF), often in the framework of the generalized linear model. Despite widespread use of the FIR STRF, numerous formulations for linear filters are possible that require many fewer parameters, potentially permitting more efficient and accurate model estimates. To explore these alternative STRF architectures, we recorded single-unit neural activity from auditory cortex of awake ferrets during presentation of natural sound stimuli. We compared performance of > 1000 linear STRF architectures, evaluating their ability to predict neural responses to a novel natural stimulus. Many were able to outperform the FIR filter. Two basic constraints on the architecture lead to the improved performance: (1) factorization of the STRF matrix into a small number of spectral and temporal filters and (2) low-dimensional parameterization of the factorized filters. The best parameterized model was able to outperform the full FIR filter in both primary and secondary auditory cortex, despite requiring fewer than 30 parameters, about 10% of the number required by the FIR filter. After accounting for noise from finite data sampling, these STRFs were able to explain an average of 40% of A1 response variance. The simpler models permitted more straightforward interpretation of sensory tuning properties. They also showed greater benefit from incorporating nonlinear terms, such as short term plasticity, that provide theoretical advances over the linear model. Architectures that minimize parameter count while maintaining maximum predictive power provide insight into the essential degrees of freedom governing auditory cortical function. They also maximize statistical power available for characterizing additional nonlinear properties that limit current auditory models.

Understanding how the brain solves sensory problems can provide useful insight for the development of automated systems such as speech recognizers and image classifiers. Recent developments in nonlinear regression and machine learning have produced powerful algorithms for characterizing the input-output relationship of complex systems. However, the complexity of sensory neural systems, combined with practical limitations on experimental data, make it difficult to apply arbitrarily complex analyses to neural data. In this study we pushed analysis in the opposite direction, toward simpler models. We asked how simple a model can be while still capturing the essential sensory properties of neurons in auditory cortex. We found that substantially simpler formulations of the widely-used spectro-temporal receptive field are able to perform as well as the best current models. These simpler formulations define new basis sets that can be incorporated into state-of-the-art machine learning algorithms for a more exhaustive exploration of sensory processing.

Synaptic plasticity as a cortical coding scheme

Processing of auditory information requires constant adjustment due to alterations of the environment and changing conditions in the nervous system with age, health, and experience. Consequently, patterns of activity in cortical networks have complex dynamics over a wide range of timescales, from milliseconds to days and longer. In the primary auditory cortex (AI), multiple forms of adaptation and plasticity shape synaptic input and action potential output. However, the variance of neuronal responses has made it difficult to characterize AI receptive fields and to determine the function of AI in processing auditory information such as vocalizations. Here we describe recent studies on the temporal modulation of cortical responses and consider the relation of synaptic plasticity to neural coding.

Spike timing precision changes with spike rate adaptation in the owl's auditory space map

Spike rate adaptation (SRA) is a continuing change of responsiveness to ongoing stimuli, which is ubiquitous across species and levels of sensory systems. Under SRA, auditory responses to constant stimuli change over time, relaxing toward a long-term rate often over multiple timescales. With more variable stimuli, SRA causes the dependence of spike rate on sound pressure level to shift toward the mean level of recent stimulus history. A model based on subtractive adaptation (Benda J, Hennig RM. J Comput Neurosci 24: 113-136, 2008) shows that changes in spike rate and level dependence are mechanistically linked. Space-specific neurons in the barn owl's midbrain, when recorded under ketamine-diazepam anesthesia, showed these classical characteristics of SRA, while at the same time exhibiting changes in spike timing precision. Abrupt level increases of sinusoidally amplitude-modulated (SAM) noise initially led to spiking at higher rates with lower temporal precision. Spike rate and precision relaxed toward their long-term values with a time course similar to SRA, results that were also replicated by the subtractive model. Stimuli whose amplitude modulations (AMs) were not synchronous across carrier frequency evoked spikes in response to stimulus envelopes of a particular shape, characterized by the spectrotemporal receptive field (STRF). Again, abrupt stimulus level changes initially disrupted the temporal precision of spiking, which then relaxed along with SRA. We suggest that shifts in latency associated with stimulus level changes may differ between carrier frequency bands and underlie decreased spike precision. Thus SRA is manifest not simply as a change in spike rate but also as a change in the temporal precision of spiking.

Connecting Neural Codes with Behavior in the Auditory System of Drosophila

Brains are optimized for processing ethologically relevant sensory signals. However, few studies have characterized the neural coding mechanisms that underlie the transformation from natural sensory information to behavior. Here, we focus on acoustic communication in Drosophila melanogaster and use computational modeling to link natural courtship song, neuronal codes, and female behavioral responses to song. We show that melanogaster females are sensitive to long timescale song structure (on the order of tens of seconds). From intracellular recordings, we generate models that recapitulate neural responses to acoustic stimuli. We link these neural codes with female behavior by generating model neural responses to natural courtship song. Using a simple decoder, we predict female behavioral responses to the same song stimuli with high accuracy. Our modeling approach reveals how long timescale song features are represented by the Drosophila brain and how neural representations can be decoded to generate behavioral selectivity for acoustic communication signals.

Rhythmic auditory cortex activity at multiple timescales shapes stimulus-response gain and background firing

The phase of low-frequency network activity in the auditory cortex captures changes in neural excitability, entrains to the temporal structure of natural sounds, and correlates with the perceptual performance in acoustic tasks. Although these observations suggest a causal link between network rhythms and perception, it remains unknown how precisely they affect the processes by which neural populations encode sounds. We addressed this question by analyzing neural responses in the auditory cortex of anesthetized rats using stimulus-response models. These models included a parametric dependence on the phase of local field potential rhythms in both stimulus-unrelated background activity and the stimulus-response transfer function. We found that phase-dependent models better reproduced the observed responses than static models, during both stimulation with a series of natural sounds and epochs of silence. This was attributable to two factors: (1) phase-dependent variations in background firing (most prominent for delta; 1-4 Hz); and (2) modulations of response gain that rhythmically amplify and attenuate the responses at specific phases of the rhythm (prominent for frequencies between 2 and 12 Hz). These results provide a quantitative characterization of how slow auditory cortical rhythms shape sound encoding and suggest a differential contribution of network activity at different timescales. In addition, they highlight a putative mechanism that may implement the selective amplification of appropriately timed sound tokens relative to the phase of rhythmic auditory cortex activity.

Dynamic encoding of speech sequence probability in human temporal cortex

Sensory processing involves identification of stimulus features, but also integration with the surrounding sensory and cognitive context. Previous work in animals and humans has shown fine-scale sensitivity to context in the form of learned knowledge about the statistics of the sensory environment, including relative probabilities of discrete units in a stream of sequential auditory input. These statistics are a defining characteristic of one of the most important sequential signals humans encounter: speech. For speech, extensive exposure to a language tunes listeners to the statistics of sound sequences. To address how speech sequence statistics are neurally encoded, we used high-resolution direct cortical recordings from human lateral superior temporal cortex as subjects listened to words and nonwords with varying transition probabilities between sound segments. In addition to their sensitivity to acoustic features (including contextual features, such as coarticulation), we found that neural responses dynamically encoded the language-level probability of both preceding and upcoming speech sounds. Transition probability first negatively modulated neural responses, followed by positive modulation of neural responses, consistent with coordinated predictive and retrospective recognition processes, respectively. Furthermore, transition probability encoding was different for real English words compared with nonwords, providing evidence for online interactions with high-order linguistic knowledge. These results demonstrate that sensory processing of deeply learned stimuli involves integrating physical stimulus features with their contextual sequential structure. Despite not being consciously aware of phoneme sequence statistics, listeners use this information to process spoken input and to link low-level acoustic representations with linguistic information about word identity and meaning.

One hundred ways to process time, frequency, rate and scale in the central auditory system: a pattern-recognition meta-analysis

The mammalian auditory system extracts features from the acoustic environment based on the responses of spatially distributed sets of neurons in the subcortical and cortical auditory structures. The characteristic responses of these neurons (linearly approximated by their spectro-temporal receptive fields, or STRFs) suggest that auditory representations are formed, as early as in the inferior colliculi, on the basis of a time, frequency, rate (temporal modulations) and scale (spectral modulations) analysis of sound. However, how these four dimensions are integrated and processed in subsequent neural networks remains unclear. In this work, we present a new methodology to generate computational insights into the functional organization of such processes. We first propose a systematic framework to explore more than a hundred different computational strategies proposed in the literature to process the output of a generic STRF model. We then evaluate these strategies on their ability to compute perceptual distances between pairs of environmental sounds. Finally, we conduct a meta-analysis of the dataset of all these algorithms' accuracies to examine whether certain combinations of dimensions and certain ways to treat such dimensions are, on the whole, more computationally effective than others. We present an application of this methodology to a dataset of ten environmental sound categories, in which the analysis reveals that (1) models are most effective when they organize STRF data into frequency groupings—which is consistent with the known tonotopic organization of receptive fields in auditory structures -, and that (2) models that treat STRF data as time series are no more effective than models that rely only on summary statistics along time—which corroborates recent experimental evidence on texture discrimination by summary statistics.