Processing In Auditory Cortex Research Articles

Multiple processes of vocal sensory-motor interaction in primate auditory cortex

Sensory-motor interactions in the auditory system play an important role in vocal self-monitoring and control. These result from top-down corollary discharges, relaying predictions about vocal timing and acoustics. Recent evidence suggests such signals may be two distinct processes, one suppressing neural activity during vocalization and another enhancing sensitivity to sensory feedback, rather than a single mechanism. Single-neuron recordings have been unable to disambiguate due to overlap of motor signals with sensory inputs. Here, we sought to disentangle these processes in marmoset auditory cortex during production of multi-phrased ‘twitter’ vocalizations. Temporal responses revealed two timescales of vocal suppression: temporally-precise phasic suppression during phrases and sustained tonic suppression. Both components were present within individual neurons, however, phasic suppression presented broadly regardless of frequency tuning (gating), while tonic was selective for vocal frequencies and feedback (prediction). This suggests that auditory cortex is modulated by concurrent corollary discharges during vocalization, with different computational mechanisms.

Attention-Driven Modulation of Auditory Cortex Activity during Selective Listening in a Multispeaker Setting.

Real-world listening settings often consist of multiple concurrent sound streams. To limit perceptual interference during selective listening, the auditory system segregates and filters the relevant sensory input. Previous work provided evidence that the auditory cortex is critically involved in this process and selectively gates attended input toward subsequent processing stages. We studied at which level of auditory cortex processing this filtering of attended information occurs using functional magnetic resonance imaging (fMRI) and a naturalistic selective listening task. Forty-five human listeners (of either sex) attended to one of two continuous speech streams, presented either concurrently or in isolation. Functional data were analyzed using an inter-subject analysis to assess stimulus-specific components of ongoing auditory cortex activity. Our results suggest that stimulus-related activity in the primary auditory cortex and the adjacent planum temporale are hardly affected by attention, whereas brain responses at higher stages of the auditory cortex processing hierarchy become progressively more selective for the attended input. Consistent with these findings, a complementary analysis of stimulus-driven functional connectivity further demonstrated that information on the to-be-ignored speech stream is shared between the primary auditory cortex and the planum temporale but largely fails to reach higher processing stages. Our findings suggest that the neural processing of ignored speech cannot be effectively suppressed at the level of early cortical processing of acoustic features but is gradually attenuated once the competing speech streams are fully segregated.

Neuromagnetic representation of melodic contour processing in human auditory cortex.

The pattern of ups and downs in a sequence with varying pitch can be heard as a melodic contour. Contrary to single pitch, the neural representation of melodic contour information in the auditory cortex is rarely investigated, and it is not clear whether the processing entails a hemispheric asymmetry. The present magnetoencephalography study assessed the neuromagnetic responses of N = 18 normal-hearing adults to four-note sequences with fixed vs. varying pitch that were presented either monaurally or diotically; data were analyzed using minimum-norm reconstructions. The first note of the sequences elicited prominent transient activity in posterior auditory regions (Planum temporale), especially contralateral to the ear of entry. In contrast, the response to the subsequent notes originated from more anterior areas (Planum polare) and was larger for melodic contours than for fixed pitch sequences, independent from the ear of entry and without hemispheric asymmetry. Together, the results point to a gradient in the early cortical processing of melodic contours, both in spatial and functional terms, where posterior auditory activity reflects the onset of a pitch sequence and anterior activity reflects its subsequent notes, including the difference between sequences with fixed pitch and melodic contours.

Auditory processing remains sensitive to environmental experience during adolescence in a rodent model

Elevated neural plasticity during development contributes to dramatic improvements in perceptual, motor, and cognitive skills. However, malleable neural circuits are vulnerable to environmental influences that may disrupt behavioral maturation. While these risks are well-established prior to sexual maturity (i.e., critical periods), the degree of neural vulnerability during adolescence remains uncertain. Here, we induce transient hearing loss (HL) spanning adolescence in gerbils, and ask whether behavioral and neural maturation are disrupted. We find that adolescent HL causes a significant perceptual deficit that can be attributed to degraded auditory cortex processing, as assessed with wireless single neuron recordings and within-session population-level analyses. Finally, auditory cortex brain slices from adolescent HL animals reveal synaptic deficits that are distinct from those typically observed after critical period deprivation. Taken together, these results show that diminished adolescent sensory experience can cause long-lasting behavioral deficits that originate, in part, from a dysfunctional cortical circuit.

Unified Coding of Spectral and Temporal Phonetic Cues: Electrophysiological Evidence for Abstract Phonological Features.

Spoken word recognition models and phonological theory propose that abstract features play a central role in speech processing. It remains unknown, however, whether auditory cortex encodes linguistic features in a manner beyond the phonetic properties of the speech sounds themselves. We took advantage of the fact that English phonology functionally codes stops and fricatives as voiced or voiceless with two distinct phonetic cues: Fricatives use a spectral cue, whereas stops use a temporal cue. Evidence that these cues can be grouped together would indicate the disjunctive coding of distinct phonetic cues into a functionally defined abstract phonological feature. In English, the voicing feature, which distinguishes the consonants [s] and [t] from [z] and [d], respectively, is hypothesized to be specified only for voiceless consonants (e.g., [s t]). Here, participants listened to syllables in a many-to-one oddball design, while their EEG was recorded. In one block, both voiceless stops and fricatives were the standards. In the other block, both voiced stops and fricatives were the standards. A critical design element was the presence of intercategory variation within the standards. Therefore, a many-to-one relationship, which is necessary to elicit an MMN, existed only if the stop and fricative standards were grouped together. In addition to the ERPs, event-related spectral power was also analyzed. Results showed an MMN effect in the voiceless standards block-an asymmetric MMN-in a time window consistent with processing in auditory cortex, as well as increased prestimulus beta-band oscillatory power to voiceless standards. These findings suggest that (i) there is an auditory memory trace of the standards based on the shared [voiceless] feature, which is only functionally defined; (ii) voiced consonants are underspecified; and (iii) features can serve as a basis for predictive processing. Taken together, these results point toward auditory cortex's ability to functionally code distinct phonetic cues together and suggest that abstract features can be used to parse the continuous acoustic signal.

Effects of Cortical Cooling on Sound Processing in Auditory Cortex and Thalamus of Awake Marmosets.

The auditory thalamus is the central nexus of bottom-up connections from the inferior colliculus and top-down connections from auditory cortical areas. While considerable efforts have been made to investigate feedforward processing of sounds in the auditory thalamus (medial geniculate body, MGB) of non-human primates, little is known about the role of corticofugal feedback in the MGB of awake non-human primates. Therefore, we developed a small, repositionable cooling probe to manipulate corticofugal feedback and studied neural responses in both auditory cortex and thalamus to sounds under conditions of normal and reduced cortical temperature. Cooling-induced increases in the width of extracellularly recorded spikes in auditory cortex were observed over the distance of several hundred micrometers away from the cooling probe. Cortical neurons displayed reduction in both spontaneous and stimulus driven firing rates with decreased cortical temperatures. In thalamus, cortical cooling led to increased spontaneous firing and either increased or decreased stimulus driven activity. Furthermore, response tuning to modulation frequencies of temporally modulated sounds and spatial tuning to sound source location could be altered (increased or decreased) by cortical cooling. Specifically, best modulation frequencies of individual MGB neurons could shift either toward higher or lower frequencies based on the vector strength or the firing rate. The tuning of MGB neurons for spatial location could both sharpen or widen. Elevation preference could shift toward higher or lower elevations and azimuth tuning could move toward ipsilateral or contralateral locations. Such bidirectional changes were observed in many parameters which suggests that the auditory thalamus acts as a filter that could be adjusted according to behaviorally driven signals from auditory cortex. Future work will have to delineate the circuit elements responsible for the observed effects.

Unilateral Conductive Hearing Loss Disrupts the Developmental Refinement of Binaural Processing in the Rat Primary Auditory Cortex.

Binaural hearing is critically important for the perception of sound spatial locations. The primary auditory cortex (AI) has been demonstrated to be necessary for sound localization. However, after hearing onset, how the processing of binaural cues by AI neurons develops, and how the binaural processing of AI neurons is affected by reversible unilateral conductive hearing loss (RUCHL), are not fully elucidated. Here, we determined the binaural processing of AI neurons in four groups of rats: postnatal day (P) 14–18 rats, P19–30 rats, P57–70 adult rats, and RUCHL rats (P57–70) with RUCHL during P14–30. We recorded the responses of AI neurons to both monaural and binaural stimuli with variations in interaural level differences (ILDs) and average binaural levels. We found that the monaural response types, the binaural interaction types, and the distributions of the best ILDs of AI neurons in P14–18 rats are already adult-like. However, after hearing onset, there exist developmental refinements in the binaural processing of AI neurons, which are exhibited by the increase in the degree of binaural interaction, and the increase in the sensitivity and selectivity to ILDs. RUCHL during early hearing development affects monaural response types, decreases the degree of binaural interactions, and decreases both the selectivity and sensitivity to ILDs of AI neurons in adulthood. These new evidences help us to understand the refinements and plasticity in the binaural processing of AI neurons during hearing development, and might enhance our understanding in the neuronal mechanism of developmental changes in auditory spatial perception.

Expectation of self-generated sounds drives predictive processing in mouse auditory cortex

Behavior exerts a strong influence over sensory responses in the brain. In the auditory cortex, neural responses to self-generated sounds are suppressed, suggesting that prediction may play a critical role in local sensory processing. However, it is unclear whether this phenomenon derives from a precise movement-based prediction or how it affects the neural representation of incoming stimuli. We address these questions by designing a behavioral paradigm where mice learn to expect the predictable acoustic consequences of a simple forelimb movement. Dense neuronal recordings from auditory cortex revealed suppression of neural responses that was strongest for the expected tone during a narrow time window surrounding the specific sound-associated movement. Predictive suppression was concentrated in layers 2/3 and 5, preceded by the arrival of movement information in deep layers, and driven by motor-sensory coupling rather than task engagement, behavioral relevance, or reward association. Despite large-scale suppression at the population level, the auditory cortex retained a sparse representation of the expected self-generated sounds, while a substantial population of auditory cortex neurons in L2/3 and L5 encoded prediction errors. These findings illustrate that expectation, learned through motor-sensory experience, drives layer-specific predictive processing in the mouse auditory cortex.

Laser-Induced Apoptosis of Corticothalamic Neurons in Layer VI of Auditory Cortex Impact on Cortical Frequency Processing.

Corticofugal projections outnumber subcortical input projections by far. However, the specific role for signal processing of corticofugal feedback is still less well understood in comparisonto the feedforward projection. Here, we lesioned corticothalamic (CT) neurons in layers V and/or VI of the auditory cortex of Mongolian gerbils by laser-induced photolysis to investigate their contribution to cortical activation patterns. We have used laminar current-source density (CSD) recordings of tone-evoked responses and could show that, particularly, lesion of CT neurons in layer VI affected cortical frequency processing. Specifically, we found a decreased gain of best-frequency input in thalamocortical (TC)-recipient input layers that correlated with the relative lesion of layer VI neurons, but not layer V neurons. Using cortical silencing with the GABAa-agonist muscimol and layer-specific intracortical microstimulation (ICMS), we found that direct activation of infragranular layers recruited a local recurrent cortico-thalamo-cortical loop of synaptic input. This recurrent feedback was also only interrupted when lesioning layer VI neurons, but not cells in layer V. Our study thereby shows distinct roles of these two types of CT neurons suggesting a particular impact of CT feedback from layer VI to affect the local feedforward frequency processing in auditory cortex.

Plasticity of Multidimensional Receptive Fields in Core Rat Auditory Cortex Directed by Sound Statistics

Sensory cortical neurons can nonlinearly integrate a wide range of inputs. The outcome of this nonlinear process can be approximated by more than one receptive field component or filter to characterize the ensuing stimulus preference. The functional properties of multidimensional filters are, however, not well understood. Here we estimated two spectrotemporal receptive fields (STRFs) per neuron using maximally informative dimension analysis. We compared their temporal and spectral modulation properties and determined the stimulus information captured by the two STRFs in core rat auditory cortical fields, primary auditory cortex (A1) and ventral auditory field (VAF). The first STRF is the dominant filter and acts as a sound feature detector in both fields. The second STRF is less feature specific, preferred lower modulations, and had less spike information compared to the first STRF. The information jointly captured by the two STRFs was larger than that captured by the sum of the individual STRFs, reflecting nonlinear interactions of two filters. This information gain was larger in A1. We next determined how the acoustic environment affects the structure and relationship of these two STRFs. Rats were exposed to moderate levels of spectrotemporally modulated noise during development. Noise exposure strongly altered the spectrotemporal preference of the first STRF in both cortical fields. The interaction between the two STRFs was reduced by noise exposure in A1 but not in VAF. The results reveal new functional distinctions between A1 and VAF indicating that (i) A1 has stronger interactions of the two STRFs than VAF, (ii) noise exposure diminishes modulation parameter representation contained in the noise more strongly for the first STRF in both fields, and (iii) plasticity induced by noise exposure can affect the strength of filter interactions in A1. Taken together, ascertaining two STRFs per neuron enhances the understanding of cortical information processing and plasticity effects in core auditory cortex.

Auditory Mismatch Negativity in Clinical High Risk for Psychosis and Schizophrenia, and Association With Symptoms and Cognition

Primary Auditory Cortex processing deficits are a core feature of schizophrenia, reflected in impaired generation of auditory mismatch negativity (MMN), an EEG-derived event-related potential. Auditory MMN impairments in response to deviant stimuli can predict outcome among clinical high risk (CHR) for psychosis subjects. We hypothesized that auditory MMN deficits in CHR patients would be associated with symptom severity and cognitive dysfunction

Nicotine Enhances Amplitude and Consistency of Timing of Responses to Acoustic Trains in A1.

Systemic nicotine enhances neural processing in primary auditory cortex (A1) as determined using tone-evoked, current-source density (CSD) measurements. For example, nicotine enhances the characteristic frequency (CF)-evoked current sink in layer 4 of A1, increasing amplitude and decreasing latency. However, since presenting auditory stimuli within a stream of stimuli increases the complexity of response dynamics, we sought to determine the effects of nicotine on CSD responses to trains of CF stimuli (one-second trains at 2–40 Hz; each train repeated 25 times). CSD recordings were obtained using a 16-channel multiprobe inserted in A1 of urethane/xylazine-anesthetized mice, and analysis focused on two current sinks in the middle (layer 4) and deep (layers 5/6) layers. CF trains produced adaptation of the layer 4 response that was weak at 2 Hz, stronger at 5–10 Hz and complete at 20–40 Hz. In contrast, the layer 5/6 current sink exhibited less adaptation at 2–10 Hz, and simultaneously recorded auditory brainstem responses (ABRs) showed no adaptation even at 40 Hz. Systemic nicotine (2.1 mg/kg) enhanced layer 4 responses throughout the one-second stimulus train at rates ≤10 Hz. Nicotine enhanced both response amplitude within each train and the consistency of response timing across 25 trials. Nicotine did not alter the degree of adaptation over one-second trials, but its effect to increase amplitudes revealed a novel, slower form of adaptation that developed over multiple trials. Nicotine did not affect responses that were fully adapted (20–40 Hz trains), nor did nicotine affect any aspect of the layer 5/6 current sink or ABRs. The overall effect of nicotine in layer 4 was to enhance all responses within each train, to emphasize earlier trials across multiple trials, and to improve the consistency of timing across all trials. These effects may improve processing of complex acoustic streams, including speech, that contain information in the 2–10 Hz range.

Linking central auditory deficits to cognitive decline in mice with amyloid deposition

AbstractBackgroundAlzheimer’s disease (AD) is a form of dementia that is characterized by the progressive loss of cognitive capacity, including the loss of executive functions and memory. There is growing evidence suggesting that hearing loss and dementia are tightly linked, with recent studies showing that hearing loss is independently associated with accelerated cognitive decline. However, the neural mechanisms linking hearing loss to AD‐related pathologies and subsequent cognitive decline remain unknown.MethodWe will test the hypothesis that amyloid pathology impacts feedforward sensory processing in primary auditory cortex (A1). We performed in vivo two‐photon calcium imaging in awake, head‐fixed APPswe/PS1dE9 (APP/PS1) mice to investigate how AD pathology impacts central auditory processing. APP/PS1 mice express chimeric mouse/human amyloid precursor protein (APP) and mutant presinilin‐1 (PS1). These transgenes target neurons of the central nervous system and lead to the rapid accumulation of soluble amyloid beta and subsequent plaque deposition in cortical and hippocampal areas. Our study uses the fine‐scale tonotopic arrangement of excitatory neurons in A1 as an assay for feedforward sensory integration. With two‐photon calcium imaging, we have the spatial and temporal resolution to monitor the activity of hundreds of neurons simultaneously with single‐cell resolution.ResultPreliminary data collected from 11 mice (7 APP+, 4 WT) suggests clear differences in the processing of feedforward sensory information, namely differences in the tonotopic arrangement of neurons in A1 between WT and APP/PS1 mice.ConclusionFuture work will investigate how AD pathology affects sensorimotor learning. Learning rates will be monitored in APP/PS1 mice trained on a go/no‐go stimulus recognition task. To probe the neural computations that are critical for the selection of auditory cues during behavior, excitatory networks in A1 will be monitored using two‐photon calcium imaging while animals perform the task. This study hopes to elucidate neural mechanisms of how amyloid beta pathology impacts central auditory processing, and thus, mechanistically link hearing loss, AD‐related pathologies and cognitive impairments.

Difficulties with Speech-in-Noise Perception Related to Fundamental Grouping Processes in Auditory Cortex.

In our everyday lives, we are often required to follow a conversation when background noise is present (“speech-in-noise” [SPIN] perception). SPIN perception varies widely—and people who are worse at SPIN perception are also worse at fundamental auditory grouping, as assessed by figure-ground tasks. Here, we examined the cortical processes that link difficulties with SPIN perception to difficulties with figure-ground perception using functional magnetic resonance imaging. We found strong evidence that the earliest stages of the auditory cortical hierarchy (left core and belt areas) are similarly disinhibited when SPIN and figure-ground tasks are more difficult (i.e., at target-to-masker ratios corresponding to 60% rather than 90% performance)—consistent with increased cortical gain at lower levels of the auditory hierarchy. Overall, our results reveal a common neural substrate for these basic (figure-ground) and naturally relevant (SPIN) tasks—which provides a common computational basis for the link between SPIN perception and fundamental auditory grouping.

Neuronal Avalanches in Input and Associative Layers of Auditory Cortex.

The primary auditory cortex processes acoustic sequences for the perception of behaviorally meaningful sounds such as speech. Sound information arrives at its input layer four from where activity propagates to associative layer 2/3. It is currently not known whether there is a characteristic organization of neuronal population activity across layers and sound levels during sound processing. Here, we identify neuronal avalanches, which in theory and experiments have been shown to maximize dynamic range and optimize information transfer within and across networks, in primary auditory cortex. We used in vivo 2-photon imaging of pyramidal neurons in cortical layers L4 and L2/3 of mouse A1 to characterize the populations of neurons that were active spontaneously, i.e., in the absence of a sound stimulus, and those recruited by single-frequency tonal stimuli at different sound levels. Single-frequency sounds recruited neurons of widely ranging frequency selectivity in both layers. We defined neuronal ensembles as neurons being active within or during successive temporal windows at the temporal resolution of our imaging. For both layers, neuronal ensembles were highly variable in size during spontaneous activity as well as during sound presentation. Ensemble sizes distributed according to power laws, the hallmark of neuronal avalanches, and were similar across sound levels. Avalanches activated by sound were composed of neurons with diverse tuning preference, yet with selectivity independent of avalanche size. Our results suggest that optimization principles identified for avalanches guide population activity in L4 and L2/3 of auditory cortex during and in-between stimulus processing.

Delayed M50/M100 evoked response component latency in minimally verbal/nonverbal children who have autism spectrum disorder

Abnormal auditory neuromagnetic M50 and M100 responses, reflecting primary/secondary auditory cortex processing, have been reported in children who have autism spectrum disorder (ASD). Some studies have reported an association between delays in these responses and language impairment. However, as most prior research has focused on verbal individuals with ASD without cognitive impairment, rather little is known about neural activity during auditory processing in minimally verbal or nonverbal children who have ASD (ASD-MVNV)—children with little or no speech and often significant cognitive impairment. To understand the neurophysiological mechanisms underlying auditory processing in ASD-MVNV children, magnetoencephalography (MEG) measured M50 and M100 responses arising from left and right superior temporal gyri during tone stimuli in three cohorts: (1) MVNV children who have ASD (ASD-MVNV), (2) verbal children who have ASD and no intellectual disability (ASD-V), and (3) typically developing (TD) children. One hundred and five participants (8–12 years) were included in the final analyses (ASD-MVNV: n = 16, 9.85 ± 1.32 years; ASD-V: n = 55, 10.64 ± 1.31 years; TD: n = 34, 10.18 ± 1.36 years). ASD-MVNV children showed significantly delayed M50 and M100 latencies compared to TD. These delays tended to be greater than the corresponding delays in verbal children with ASD. Across cohorts, delayed latencies were associated with language and communication skills, assessed by the Vineland Adaptive Behavior Scale Communication Domain. Findings suggest that auditory cortex neural activity measures could be dimensional objective indices of language impairment in ASD for either diagnostic (e.g., via threshold or cutoff) or prognostic (considering the continuous variable) use.

Direct electrophysiological mapping of human pitch-related processing in auditory cortex

This work sought correlates of pitch perception, defined by neural activity above the lower limit of pitch (LLP), in auditory cortical neural ensembles, and examined their topographical distribution. Local field potentials (LFPs) were recorded in eight patients undergoing invasive recordings for pharmaco-resistant epilepsy. Stimuli consisted of bursts of broadband noise followed by regular interval noise (RIN). RIN was presented at rates below and above the LLP to distinguish responses related to the regularity of the stimulus and the presence of pitch itself. LFPs were recorded from human cortical homologues of auditory core, belt, and parabelt regions using multicontact depth electrodes implanted in Heschl’s gyrus (HG) and Planum Temporale (PT), and subdural grid electrodes implanted over lateral superior temporal gyrus (STG). Evoked responses corresponding to the temporal regularity of the stimulus were assessed using autocorrelation of the evoked responses, and occurred for stimuli below and above the LLP. Induced responses throughout the high gamma range (60–200 Hz) were present for pitch values above the LLP, with onset latencies of approximately 70 ms. Mapping of the induced responses onto a common brain space demonstrated variability in the topographical distribution of high gamma responses across subjects. Induced responses were present throughout the length of HG and on PT, which is consistent with previous functional neuroimaging studies. Moreover, in each subject, a region within lateral STG showed robust induced responses at pitch-evoking stimulus rates. This work suggests a distributed representation of pitch processing in neural ensembles in human homologues of core and non-core auditory cortex.

Speaker-normalized sound representations in the human auditory cortex

The acoustic dimensions that distinguish speech sounds (like the vowel differences in “boot” and “boat”) also differentiate speakers’ voices. Therefore, listeners must normalize across speakers without losing linguistic information. Past behavioral work suggests an important role for auditory contrast enhancement in normalization: preceding context affects listeners’ perception of subsequent speech sounds. Here, using intracranial electrocorticography in humans, we investigate whether and how such context effects arise in auditory cortex. Participants identified speech sounds that were preceded by phrases from two different speakers whose voices differed along the same acoustic dimension as target words (the lowest resonance of the vocal tract). In every participant, target vowels evoke a speaker-dependent neural response that is consistent with the listener’s perception, and which follows from a contrast enhancement model. Auditory cortex processing thus displays a critical feature of normalization, allowing listeners to extract meaningful content from the voices of diverse speakers.

Inner speech is accompanied by a temporally-precise and content-specific corollary discharge

When we move our articulator organs to produce overt speech, the brain generates a corollary discharge that acts to suppress the neural and perceptual responses to our speech sounds. Recent research suggests that inner speech – the silent production of words in one's mind – is also accompanied by a corollary discharge. Here, we show that this corollary discharge contains information about the temporal and physical properties of inner speech. In two experiments, participants produced an inner phoneme at a precisely-defined moment in time. An audible phoneme was presented 300 ms before, concurrently with, or 300 ms after participants produced the inner phoneme. We found that producing the inner phoneme attenuated the N1 component of the event-related potential – an index of auditory cortex processing – but only when the inner and audible phonemes occurred concurrently and matched on content. If the audible phoneme was presented before or after the production of the inner phoneme, or if the inner phoneme did not match the content of the audible phoneme, there was no attenuation of the N1. These results suggest that inner speech is accompanied by a temporally-precise and content-specific corollary discharge. We conclude that these results support the notion of a functional equivalence between the neural processes that underlie the production of inner and overt speech, and may provide a platform for identifying inner speech abnormalities in disorders in which they have been putatively associated, such as schizophrenia.

Signature Patterns for Top-Down and Bottom-Up Information Processing via Cross-Frequency Coupling in Macaque Auditory Cortex.

Predictive coding is a theoretical framework that provides a functional interpretation of top-down and bottom-up interactions in sensory processing. The theory suggests there are differences in message passing up versus down the cortical hierarchy. These differences result from the linear feedforward of prediction errors, and the nonlinear feedback of predictions. This implies that cross-frequency interactions should predominate top-down. But it remains unknown whether these differences are expressed in cross-frequency interactions in the brain. Here we examined bidirectional cross-frequency coupling across four sectors of the auditory hierarchy in the macaque. We computed two measures of cross-frequency coupling, phase–amplitude coupling (PAC) and amplitude–amplitude coupling (AAC). Our findings revealed distinct patterns for bottom-up and top-down information processing among cross-frequency interactions. Both top-down and bottom-up interactions made prominent use of low frequencies: low-to-low-frequency (theta, alpha, beta) and low-frequency-to-high- gamma couplings were predominant top-down, while low-frequency-to-low-gamma couplings were predominant bottom-up. These patterns were largely preserved across coupling types (PAC and AAC) and across stimulus types (natural and synthetic auditory stimuli), suggesting that they are a general feature of information processing in auditory cortex. Our findings suggest the modulatory effect of low frequencies on gamma-rhythms in distant regions is important for bidirectional information transfer. The finding of low-frequency-to-low-gamma interactions in the bottom-up direction suggest that nonlinearities may also play a role in feedforward message passing. Altogether, the patterns of cross-frequency interaction we observed across the auditory hierarchy are largely consistent with the predictive coding framework.