- Home
- Previous APAN Programs
- 2020 Session Browser
2020 Session Browser
Note: all sessions are listed in Eastern Daylight Time (UTC−04:00).
Israel Nelken
Topic areas: hierarchical sensory organization neural coding
Thursday, 10/22 | 10:00AM - 11:00AM EDT | Keynote
Jennifer Bizley, Edward Chang, Lori Holt
Topic areas: brain processing of speech and language correlates of auditory behavior/perception neural coding
Friday, 10/23 | 9:45AM - 10:30AM EDT | Panel Discussion
Benjamin Morillon
Topic areas: brain processing of speech and language correlates of auditory behavior/perception
Friday, 10/23 | 10:30AM - 11:00AM EDT | Young Investigator Spotlight
Abstract
A major debate in cognitive neuroscience concerns whether brain asymmetry for speech and music emerges from differential sensitivity to acoustical cues or from domain-specific neural networks. In a first study, we took advantage of the spectro-temporal modulation framework applied on a unique corpus of sung speech stimuli in which melodic and verbal content was crossed and balanced. We show that perception of speech content decreases selectively with degradation of temporal information, whereas perception of melodic content decreases only with spectral degradation. fMRI data show that the neural decoding of speech and music depends on activity patterns in left and right auditory regions, respectively. This asymmetry is supported by specific sensitivity of decoding accuracy to spectra-temporal modulation rates within each region. Crucially, the effects of degradation on perception were paralleled by their effects on neural classification. Capitalizing on intracranial data from 96 epileptic patients, we also investigated the asymmetric sampling in time (AST) hypothesis, which predicts that (1) the auditory system employs a two-timescales processing mode, (2) present both hemispheres but with a different ratio of fast and slow timescales, (3) that emerges outside of primary cortical regions. We sensitively validated each of these predictions and provide a precise estimate of the processing timescales in the auditory cortical pathway. In particular, we reveal that asymmetric sampling in associative areas is subtended by distinct two-timescales processing modes. Overall, these results suggest a match between acoustical properties of human communicative signals and neural specializations adapted to that purpose. Crowdcast Session
Jose Peña, Catherine Carr
Topic areas: cross-species comparisons neural coding neuroethology and communication
Friday, 10/23 | 1:15PM - 1:30PM EDT | Special Presentation
Agnès Landemard, Célian Bimbard, Sam Norman-Haigneré and Yves Boubenec
Topic areas: brain processing of speech and language cross-species comparisons hierarchical sensory organization neural coding
auditory cortex natural sounds cross-species speech encoding vocalizationsThursday, 10/22 | 11:15AM - 12:15PM EDT | Virtual poster
Abstract
Little is known about how neural representations of natural stimuli differ across species. Speech and music for example play a unique role in human hearing, but it is unclear how cortical responses to natural sounds differ between humans and other animals. Using functional Ultrasound imaging, we measured cortical responses in the ferret to a set of natural and spectrotemporally-matched synthetic sounds previously tested in humans. We found that tuning for spectrotemporal modulations present in both the natural and synthetic sounds was similar between humans and ferrets and could be quantitatively predicted across species. By contrast, only humans showed non-primary neural populations that responded selectively to the natural vs. synthetic sounds. This finding suggests that the unique demands of speech and music have substantially altered higher-order representations of natural sounds in the human brain, while largely preserving tuning for lower-level acoustic features based on frequency and modulation.
Michelle Moerel, Agustin Lage-Castellanos, Omer Faruk Gulban and Federico De Martino
Topic areas: auditory memory and cognition correlates of auditory behavior/perception subcortical auditory processing
auditory pathway attention ultra-high field fMRI population receptive fields auditory cortex subcortical auditory processingThursday, 10/22 | 11:15AM - 12:15PM EDT | Virtual poster
Abstract
Auditory attention enables selecting behaviourally relevant information from a complex auditory scene. While invasive electrophysiological studies in animals suggest that attention acts through rapid changes in auditory neuronal processing properties[1], the exact nature of these modulations along the human auditory pathway is not yet understood. Here we use ultra-high field functional Magnetic Resonance Imaging (fMRI) to examine auditory processing while participants (N = 10) listen to ripple sounds and natural sounds. Participants perform a detection task on the ripple sounds. By manipulating the chance of target occurrence, participants alternatively attend to low frequency (300 Hz) or high frequency (4 kHz) ripple sounds. Instead, responses to natural sounds are used to compute separate neuronal population receptive fields (PRFs)[2,3] for the attentional task condition in order to test for changes in processing properties (i.e., in PRFs) with attention. Preliminary results show reliable measurements of responses to sounds in the auditory cortex and small subcortical structures (i.e., the cochlear nucleus [CN], superior olivary body [SOC], inferior colliculus [IC], and medial geniculate body [MGB]; Figure 1). In the auditory cortex, a shift in preferred frequency with attention could be observed (Figure 2). Frequency specificity (i.e., tuning width) did not change with attention. Future work will extend these analyses to subcortical structures in order to examine attentional effects in earlier hierarchical processing stages. 1. Fritz, et al. R. Nat. Neurosci. 6, 1216–1223 (2003). 2. Moerel, et al. J. Neurosci. 32, 14205–14216 (2012). 3. Dumoulin & Wandell. Neuroimage 39, 647–660 (2008).
Gioia De Franceschi and Tania Barkat
Topic areas: correlates of auditory behavior/perception hierarchical sensory organization neural coding thalamocortical circuitry and function
Sensory perception task-engagement attention arousal reward electrophysiology auditory cortex medial geniculate body inferior colliculus mouseFriday, 10/23 | 1:30PM - 2:30PM EDT | Virtual poster
Abstract
Sensory processing varies depending on behavioral context. Here, we asked how task-engagement modulates neurons in the auditory system. We trained mice in a simple tone-detection task, and compared their neuronal activity during passive hearing and active listening. Electrophysiological extracellular recordings in the inferior colliculus, medial geniculate body, primary auditory cortex and anterior auditory field revealed widespread modulations across all regions and cortical layers, and in both putative regular and fast-spiking cortical neurons. Clustering analysis unveiled ten distinct modulation patterns that could either enhance or suppress neuronal activity. Task-engagement changed the tone-onset response in most neurons. Such modulations first emerged in subcortical areas, ruling out cortical feedback from primary auditory areas as the only mechanism underlying subcortical modulations. Half the neurons additionally displayed late modulations associated with licking, arousal or reward. Our results reveal the presence of functionally distinct subclasses of neurons, differentially sensitive to specific task-related variables but anatomically distributed along the auditory pathway.
Malinda McPherson, River Grace and Josh McDermott
Topic areas: auditory memory and cognition correlates of auditory behavior/perception
Noise Pitch Discrimination Harmonicity Auditory Scene Analysis Speech in Noise Auditory GroupingFriday, 10/23 | 11:15AM - 12:15PM EDT | Virtual poster
Abstract
Hearing in noise is a core problem in audition, and a challenge for hearing impaired listeners, yet the underlying mechanisms are poorly understood. We explored whether harmonic frequency relations, one of the primary cues used to segregate concurrent sound sources, also aids hearing in noise. We measured detection thresholds for vowels and tones embedded in noise. In order to test the role of harmonic frequency structure in detection, sounds were resynthesized to be either harmonic or inharmonic. We found that harmonic signals were consistently easier to detect in noise than otherwise identical inharmonic signals. Harmonicity also improved pitch discrimination in noise. Musicians and non-musicians showed a comparable harmonic advantage for detecting tones in noise. The results suggest that harmonicity helps to detect and discriminate signals in noise, revealing a previously unappreciated aspect of auditory scene analysis.
Elena Rotondo and Kasia Bieszczad
Topic areas: auditory memory and cognition correlates of auditory behavior/perception subcortical auditory processing
memory auditory cortex auditory brain stem response specificityThursday, 10/22 | 1:45PM - 2:45PM EDT | Virtual poster
Abstract
Individuals differ in the sensory specificity with which memory is formed along a continuum from highly cue-specific to -general. To discover sensory neural codes for memory specificity, we take advantage of two different memories in an auditory model of learning: (1) tone-reward memory, where a pure tone is associated with reward and evokes the learned response, and (2) extinction memory, where a pure tone is explicitly associated with no reward and produces the inhibition of a learned response. If a form of auditory system plasticity codes for memory specificity per se, then the form and direction of plasticity should be shared among animals with signal-specific tone-reward and signal-specific extinction memory. Results reveal that only animals that form signal-specific (vs. -general) memory develop signal-specific auditory system plasticity. Moreover, the magnitude of plasticity correlates with behavioral measures of memory specificity on an individual basis. We find that changes in the amplitude of peak 5 in the auditory brainstem response encode signal specificity in both memory types, but also appear to be sensitive to association with reward. Amplitude changes occur in opposite directions for tone-reward vs. extinction memory. We find that plasticity of auditory cortical bandwidth tuning is also signal specific. However, it occurs in the same direction in both memory types. Therefore, sharpening of cortical tuning bandwidth that is selective for the representation of the learned cue may encode for specificity of memory per se, while other forms of plasticity may encode both signal specificity and associative links with reward.
Nicole Angenstein, Jörg Stadler and André Brechmann
Topic areas: auditory memory and cognition correlates of auditory behavior/perception
aging auditory cortex categorization functional magnetic resonance imaging hemispheric interaction hemispheric specialization sequential comparisonFriday, 10/23 | 11:15AM - 12:15PM EDT | Virtual poster + podium teaser
Abstract
The processing of complex acoustic stimuli like speech requires the involvement of the auditory cortex in both hemispheres and therefore hemispheric interaction. In older adults, this hemispheric interaction seems to be reduced. The present study shows how younger (18-38 years) and older adults (56-75 years) differ in auditory processing when tasks require different levels of hemispheric interaction. One task was the categorization of frequency modulated (FM) tones according to their FM direction which mainly involves the right auditory cortex. The second tasks was the sequential comparison of the same tones according to their FM direction which requires the involvement of both hemispheres and thus a higher degree of hemispheric interaction. The level of difficulty of the tasks was adjusted to achieve equal individual performance. Using functional magnetic resonance imaging with the contralateral noise procedure, we observed that the auditory cortex is stronger involved in the processing of the tasks in older than in younger adults. This increased involvement is particularly evident when the task requires efficient hemispheric interaction. Diffusion tensor imaging data suggests reduced hemispheric interaction in the older adults due to changes in microanatomy, particularly of the posterior corpus callosum. This reduced interaction probably makes it more difficult for the older persons to process complex auditory stimuli. Stronger activation of frontal and parietal regions in the older persons may indicate compensatory processes especially during the task that requires strong hemispheric interaction. This suggests that older participants require more cognitive resources to achieve the same performance as younger participants.
Patricia Valerio, Mari Nakamura, Stitipragyan Bhumika and Tania Barkat
Topic areas: brain processing of speech and language hierarchical sensory organization neural coding
Developmental plasticity Critical period Sensory processing Sensory feature Auditory cortex Pure tone Frequency modulated sweep Electrophysiology GABA White noiseFriday, 10/23 | 1:30PM - 2:30PM EDT | Virtual poster
Abstract
Critical periods are time windows of heightened plasticity during postnatal brain development. They are specific to sensory features and do not all happen at the same time. For example, the critical period for pure tone precedes the critical period for frequency modulated sweep (FMS) by about two weeks. Whether such critical periods are timed by a temporally precise developmental program or sequentially organized, where the closing of one window triggers the opening of the next, is not known. We used in vivo electrophysiological recordings in combination with molecular and sensory manipulations to elucidate the biological constraints on critical period timing in the mouse auditory system. Passive sound exposure during development shows that the cortical representation of the two sound features pure tone and FMS are not influencing each other. Enhancing gamma-aminobutyric acid (GABA) function before the critical period for pure tone accelerated it without changing the critical period for FMS. Similarly, delaying by ten days the critical period for pure tone by rearing mice in white noise (WN) had no effect on the critical period for FMS. However, the critical period for FMS started only if the critical period for pure tone had occurred. Together, these results indicate that distinct critical periods, although sequentially organized, can be temporally shifted independently of each other and are therefore timed by a temporally precise development program. Our findings shed new light on the dependence of sensory features on each other and on the mechanisms at play in developmental plasticity.
Prachi Patel, Stephan Bickel, Jose Herrero, Ashesh Mehta and Nima Mesgarani
Topic areas: correlates of auditory behavior/perception hierarchical sensory organization
ECoG Speech processing sound localization human auditory cortexThursday, 10/22 | 12:30PM - 12:45PM EDT | Short talk
Abstract
Humans can easily attend to the speech of a single talker in a mixture when the talkers are spatially separated. However, the neural mechanisms in the human brain that enable this ability are unclear. Specifically, it is not known how spatial separation affects the neural representation of the speech mixture, and how attention to either the location or the identity of a talker modulates the cortical response. We recorded intracranially from the human auditory cortex while subjects listened to spatial talkers either presented stand alone or in a mixture. We demonstrate that the neural encoding of phonemes is unaffected by the location of a talker presented in isolation, but narrows to the speech of the contralateral talker when presented in a mixture. Further, we show that attention to the spatial location and the talker identity have differential neural effects: attention to location modulates the overall baseline of neural response level while attention to talker identity modulates spectrotemporal selectivity. Our results reveal how the human auditory cortex adjusts its neural tuning to enable listeners to attend to a talker in a spatially separated mixture, with implications for more accurate neurophysiological models of speech processing. Crowdcast Session
Johannes Wetekam
Topic areas: auditory memory and cognition subcortical auditory processing
Auditory brainstem responses Frequency following responses Auditory system Distortion products Cochlea BatsFriday, 10/23 | 1:30PM - 2:30PM EDT | Virtual poster
Abstract
Bats are a widely used experimental animal model for studying audition. Most neurophysiology studies in this animal group relied on invasive techniques to capture neural responses. Here, we used almost non-invasive recording techniques, namely: auditory brainstem potentials (ABR) and frequency following responses, to study the auditory system of the frugivorous bat species Carollia perspicillata. Based on the recorded ABR signals, objective hearing thresholds of the animals were determined, using a bootstrap method based on the RMS of the measured potentials. We observed that the hearing threshold’s sensitivity depended strongly on steps taken for signal preprocessing (such as filtering of the signal). The lowest significant threshold value obtained by any of the three tested filter settings produced hearing sensitivity values similar to those reported in studies using DPOAEs, invasive neurophysiology and behavioral measurements. Frequency following responses to amplitude modulated stimuli could be recorded up to the highest tested frequency of 1280 Hz, revealing an interesting change in response pattern with increasing modulation frequency. From an “event-locked” coding of lower AM-frequencies to a “frequency-locked” coding of higher AM-frequencies. Altogether, we present quantitative data on the use of classic non-invasive electrophysiology techniques in a hearing specialist (bats).
Kameron Clayton, Ross Williamson, Yurika Watanabe, Kenneth Hancock, Gen-Ichi Tasaka, Adi Mizrahi, Troy Hackett and Daniel Polley
Topic areas: correlates of auditory behavior/perception neural coding neuroethology and communication thalamocortical circuitry and function
motor preparatory corticothalamic sensorimotor integration globus pallidus FoxP2Thursday, 10/22 | 1:45PM - 2:45PM EDT | Virtual poster
Abstract
During active sensing, neural responses to sensory inputs directly generated by our own movements are suppressed. In the auditory cortex (ACtx), self-initiated movements elicit corollary discharge from secondary motor cortex (M2) that suppresses pyramidal neuron (PyrN) spiking via recruitment of local inhibitory neurons. Using single unit electrophysiology, we observed that ACtx layer (L)6 PyrNs were also activated hundreds of milliseconds prior to movement onset, at approximately the same time as fast spiking inhibitory neurons. Most L6 PyrNs were corticothalamic (CT) cells, which all expressed FoxP2, a protein marker enriched in brain areas that integrate sensory inputs to control vocal motor behaviors. Using a combination of two-photon calcium imaging and optogenetically targeted single unit recordings, we found that L6 CTs were strongly activated prior to orofacial movements, but not locomotion. The specific recruitment by facial movements suggests that L6 CTs may receive a different source of motor input, beyond M2. Monosynaptic rabies tracing revealed that L6 CTs received ten times more direct inputs from the basal ganglia than M2. Our prior work found that optogenetic activation of L6 CTs reset the phase of low-frequency ACtx oscillations, modified sensory tuning in the thalamus and cortex, and switched perception into modes of enhanced detection or enhanced discrimination. Here, we show that preparatory facial movements can act like an optogenetic pulse, possibly priming cortical circuits for vocal feedback monitoring. These findings identify new pathways and local circuits for motor modulation of sound processing and suggest a new role for CT neurons in active sensing.
Aleena Garner and Georg Keller
Topic areas: auditory memory and cognition multisensory processes neural coding
audio-visual associative learning predictive processing cortico-cortical communicationFriday, 10/23 | 12:15PM - 12:30PM EDT | Short talk
Abstract
Learned associations between stimuli in different sensory modalities can shape the way we perceive these stimuli (Mcgurk and Macdonald, 1976). During audio-visual associative learning, auditory cortex has been shown to underlie multi-modal plasticity in visual cortex (McIntosh et al., 1998; Zangenehpour and Zatorre, 2010). However, how processing in visual cortex is altered when an auditory stimulus signals a visual event and what the neural mechanisms are that mediate such experience- dependent audio-visual associations is not well understood. Here we describe a neural mechanism that contributes to shaping visual representations of behaviorally relevant stimuli through direct interactions between auditory and visual cortices. We show that auditory association with a visual stimulus leads to an experience-dependent suppression of visual responses in visual cortex. This suppression of the predictable visual stimulus response is driven in part by input from auditory cortex. By recording from auditory cortex axons in visual cortex, we find that these axons carry a mixture of auditory and retinotopically matched visual input. Moreover, optogenetic stimulation of auditory cortex axons in visual cortex selectively suppresses the neurons responsive to the associated visual stimulus after, but not before, learning. Our results are consistent with the interpretation that cross-modal associations can be stored in long-range cortical connections and that with learning these cross-modal connections function to suppress the responses to predictable input. Crowdcast Session
Karolina Ignatiadis, Diane Baier, Brigitta Tóth and Robert Baumgartner
Topic areas: auditory memory and cognition correlates of auditory behavior/perception
auditory looming bias spectral cues electroencephalography functional connectivity parietal cortexThursday, 10/22 | 11:15AM - 12:15PM EDT | Virtual poster
Abstract
Our auditory system constantly monitors our environment, informing us about changes in our surroundings and warning us of potential threats. Human listeners are more sensitive to sounds signaling approaching objects rather than receding ones, an effect known as auditory looming bias. Possibly reflecting an evolutionary trait, this fast perceptual bias has been extensively studied using stimuli varying in overall sound intensity to induce the sensation of objects moving in distance. A recent electroencephalographic (EEG) study revealed a top-down cortical interaction in the brain networks involved, and assigned a crucial role to the prefrontal cortex in prioritizing looming versus receding stimuli. We here tested the generalizability of these findings on an existing EEG data set, originally used to prove that the looming effect is also elicited without changes in overall intensity, by using spectral modifications instead. Our functional source connectivity results suggest a critical involvement of the parietal cortex, thereby extending the cortical network associated with prioritizing approaching sounds.
Ana Polterovich, Maciej M Jankowski, Alex Kazakov, Johannes Niediek and Israel Nelken
Topic areas: correlates of auditory behavior/perception
Auditory cortex Freely moving Behavior Electrophysiology Sound localizationFriday, 10/23 | 1:30PM - 2:30PM EDT | Virtual poster
Abstract
We explore rat behavior and electrophysiology in a complex setting (the Rat Interactive Fantasy Facility, RIFF). The RIFF consists of a large circular arena (160 cm diameter) with 6 interaction areas (IAs) that each have a water port, a food port, and two loudspeakers. Rat behavior is monitored online using video tracking and nose-poke identification. Neural responses are recorded using telemetry (a 64-channel TBSI transmitter integrated in an Alpha-Omega SNR data acquisition system) or a logger on the head of the animal (RatLog-64, Deuteron Technologies). We trained rats to perform a sound localization task in the RIFF. Auditory cues consisted of 6 different modified human words that were played from each IA separately. When a rat reached the center of the arena, one of the sounds started playing (once every 2 seconds) and the rat had to identify the correct location and collect a reward (food or water) within 20 seconds. Trials were terminated by access to a wrong port or by time out. The rats were able to learn the task rapidly without guidance. Different rats showed different strategies of exploration and reward collection. Recording neural activity in primary auditory cortex during behavior, we found significant correlations between neuronal activity and a range of non-auditory behaviorally-related variables. Interestingly, behavioral outcomes were correlated with the neural responses before sound presentation, but not with the neuronal responses to sounds.
Blaise Robert, Tatenda Chakoma, Eyal Y. Kimchi and Daniel B. Polley
Topic areas: correlates of auditory behavior/perception neural coding
cholinergic basal forebrain plasticity learning fiber photometry neuromodulator behavior pupillometryThursday, 10/22 | 11:15AM - 12:15PM EDT | Virtual poster
Abstract
The cholinergic basal forebrain (CBF) can reshape cortical sound processing, yet prior studies have treated “the basal forebrain” as a single monolithic structure and have mostly studied its involvement with the auditory cortex through artificial stimulation or pharmacology protocols. Here, we performed genetically targeted deep brain fiber imaging from cholinergic cells in two distinct regions of the basal forebrain that both project to the auditory cortex, but have independent inputs: the nuclei of the diagonal band (NDB) and the caudal tail of the globus pallidus/substantia innominata (GP/SI). We investigated sensory, motor and cognitive activators of NDB and GP/SI using daily bulk GCaMP imaging throughout an extended auditory learning protocol. We found that activity in NDB closely follows pupil size, while responses to auditory stimuli were stronger in GP/SI and modulated by novelty. NDB and GP/SI responses were both changed over the course of auditory learning, but had different selectivity for reinforcement valence and reinforcement prediction error. Responses to rewarded sounds were selectively potentiated in NDB, whereas GP/SI responded more robustly to punished stimuli. To address how CBF cell body imaging translated into cortical acetylcholine levels, we performed dual fiber imaging from auditory and visual cortex neurons that expressed a genetically encoded fluorescent acetylcholine sensor. These findings confirm our cell body imaging experiments and underscore the stronger functional coupling of the CBF to the auditory cortex, not visual cortex. These findings provide us with the design principles for novel behavioral tasks that are tailored to recruit specific neuromodulatory brain centers.
Sarineh Keshishzadeh, Markus Garrett and Sarah Verhulst
Topic areas: auditory disorders neuroethology and communication
individualized hearing profile auditory evoked potential cochlear synaptopathy envelope following response auditory brainstem response personalized auditory periphery modelThursday, 10/22 | 11:15AM - 12:15PM EDT | Virtual poster
Abstract
Estimating individual auditory nerve (AN) damage profiles in patients with mixed cochlear- and AN-damage (cochlear synaptopathy, CS) is difficult as a direct quantification of AN-damage in live humans is impossible. We hence rely on the non-invasive and indirect markers of AN-damage, established in animal studies. However, auditory evoked potential (AEP) markers of CS are sensitive to both outer-hair-cell (OHC) and AN aspects of sensorineural hearing loss (SNHL) and the impaired functionality of these elements affects AEP-derived metrics differently. Accordingly, auditory models which incorporate different sources of SNHL, can be used in combination with experimentally recorded AEPs to develop personalized SNHL profiles. Here, we investigate which markers of auditory brainstem response (ABR) and envelope following response (EFR) are best suited to develop individualized models. First, we determined the cochlear-gain-loss parameters associated with OHC damage, based on individual audiometric thresholds. Then, we simulated AEPs for different degrees of AN-damage by reducing the population of different AN fiber types in a CF-dependent manner. Using a classification technique, we determined which AN-damage profile resulted in the best simulated-recorded AEP match. A validation procedure was adopted to test whether different AEP metrics recorded from the same listener could be explained by the same SNHL pattern. The validation showed that the combination of ABR wave-I latency growth-slope with the EFR strength yields the best prediction of individual SNHL profiles. The proposed method can be used in the development of individualized hearing aid algorithms which compensate for degraded sound processing as a consequence of coexisting SNHL pathologies.
Pradeep Dheerendra, Sukhbinder Kumar, Yoshihito Shigihara and Timothy D Griffiths
Topic areas: auditory memory and cognition
Auditory Working memory MEG Functional ConnectivityThursday, 10/22 | 11:15AM - 12:15PM EDT | Virtual poster
Abstract
We aim to understand the dynamics underlying auditory working memory for maintaining 'simple' tones. We recorded magnetoencephalography (MEG) in 17 subjects while they maintained one of the two presented tones (or ignore both in the control condition). After 12s, subjects compared the pitch of a test tone with the maintained tone. Analysis of evoked responses showed persistent activity throughout maintenance compared to the pre-stimulus silent baseline but only at the start of maintenance when compared to the control condition. The evoked response during maintenance was source localised against baseline to bilateral auditory cortex. Analysis of induced responses showed suppressed alpha in the left auditory cortex, enhanced theta in medial prefrontal cortex, and enhanced beta in cerebellum. In a second experiment, 19 new subjects were presented with a tone and a Gabor patch and a retro-cue indicating whether to maintain auditory or visual information for 12s. Analysis of the induced responses in auditory condition yielded results similar to those observed in the first experiment. Connectivity analysis showed that the theta activity in medial prefrontal was phase-locked to activity in the left hippocampus and left auditory cortex. The beta activity in cerebellum was phase-locked to left Inferior Frontal Gyrus (IFG) activity and correlated to subject’s task accuracy. Our data clearly shows a network of brain areas involving pre-frontal and hippocampus, IFG and cerebellum for maintaining sounds in the auditory cortex, consistent with previous fMRI and ECoG experiments [1, 2]. [1] Kumar, et al., J Neurosci, 2016 [2] Kumar, et al., bioRxiv, 2020
Felix Bröhl and Christoph Kayser
Topic areas: brain processing of speech and language correlates of auditory behavior/perception
speech entrainment brain rhythms hearing EEGFriday, 10/23 | 11:15AM - 12:15PM EDT | Virtual poster
Abstract
The representation of speech in the brain is often examined by measuring the alignment of rhythmic brain activity to the speech envelope. To conveniently quantify this alignment (termed ‘speech tracking’) many studies consider the overall speech envelope, which combines acoustic fluctuations across the spectral range. Using EEG recordings, we show that using this overall envelope provides a distorted picture on speech encoding. We systematically investigated the encoding of spectrally-limited speech envelopes presented by individual and multiple noise carriers in the human brain. Tracking in the 1 to 6 Hz EEG bands differentially reflected low (0.2 - 0.83 kHz) and high (2.66 - 8 kHz) frequency envelopes. This was independent of the specific carrier frequency but sensitive to attentional manipulations, and reflects the context-dependent emphasis of information from distinct spectral ranges of the speech envelope in low frequency brain activity. As low and high frequency speech envelopes relate to distinct phonemic features, our results suggest that functionally distinct processes contribute to speech tracking in the same EEG bands, and are easily confounded when considering the overall speech envelope.
Sophie Bagur, Jacques Bourg, Alexandre Kempf, Thibault Tarpin, Khalil Bergaoui, Jean-Luc Puel, Jerome Bourien, Sebastian Ceballo and Brice Bathellier
Topic areas: hierarchical sensory organization neural coding subcortical auditory processing thalamocortical circuitry and function
Population coding Auditory hierarchy Two photon imaging Decorrelation Medial Geniculate Body Inferior Colliculus Auxitory Cortex CochleaFriday, 10/23 | 11:15AM - 12:15PM EDT | Virtual poster + podium teaser
Abstract
Decades of auditory system exploration have shown that sound representations by single neurons evolve along the auditory hierarchy. However, the lack of population-wide comparisons specific levels of auditory system, makes it difficult to infer the computational roles of this evolution. Here, we used two-photon imaging in awake, head-fixed mice to broadly sample, at single cell resolution, neural population responses to a wide set of sounds across the auditory cortex (~60.000 neurons), the medial geniculate nucleus (~60.000 axon boutons in A1) and the inferior colliculus (~15.000 neurons). Cochlear responses were simulated by a detailed biophysical model. Using a noise-corrected unbiased estimator of the correlation between population responses for all sound pairs, we observed that at each successive step of the auditory hierarchy sound representations become increasingly dissimilar, especially for sounds with overlapping frequencies. Closer analysis of this global trend shows that different sound types are decorrelated at different levels. For example, periodic amplitude modulations are decorrelated in the midbrain whereas amplitude and frequency ramps decorrelate in the cortex. Interestingly, the same decorrelation process is observed in deep convolutional networks trained to classify multiple classes of sounds. In contrast, single layer linear non-linear models fail to reproduce the gradual decorrelation, even if applied to fit data of a given level based on the responses of the immediately preceding level. Hence, each step of the auditory hierarchy implements multi-layered non-linear computations that progressively transform initially overlapping input information into better separated representations that support distinct perceptions of sounds that differ by few aspects.
Justin Yao and Dan Sanes
Topic areas: correlates of auditory behavior/perception neural coding
auditory cortex temporal coding rate coding awake-behaving animal psychophysicsFriday, 10/23 | 11:15AM - 12:15PM EDT | Virtual poster
Abstract
Core auditory cortex (AC) neurons encode slow acoustic modulations with temporally patterned activity. However, whether temporal information is necessary to explain auditory perceptual skills remains uncertain. Here, we recorded from gerbil AC neurons telemetrically while subjects performed a Go-Nogo amplitude modulation (AM) rate discrimination task. Animals were trained to discriminate between a 4 Hz AM broadband noise (Nogo), and AM rates >4 Hz (Go). Neurometric thresholds based on spike pattern or firing rate from individual units were compared to an animal’s behavioral threshold during a single test session. For both metrics, a proportion of units possessed neural thresholds that were better than the behavioral threshold, suggesting that firing rate could provide sufficient information for this perceptual task. However, when many neurons were pooled, a population decoder that relied on temporal information outperformed a decoder that relied on firing rate alone. The number of neurons required to explain best behavioral performance was 3-6 times greater for the firing rate decoder. Nonetheless, the firing rate decoder remained sufficient to explain average behavioral performance. This leaves open the possibility that more demanding perceptual judgements may require temporal information. To test this idea, we asked whether accurate classification of six different AM rates, between 4 to 12 Hz, required AC temporal discharge patterns. In fact, accurate classification of these AM stimuli depended on the inclusion of temporal information. Overall, our results suggest that a AC rate code is sufficient to explain fine discrimination, but a temporal code is likely needed for categorization task performance.
Manaswini Kar, Pilar Montes-Lourido and Srivatsun Sadagopan
Topic areas: hierarchical sensory organization neural coding neuroethology and communication
Vocalizations Auditory cortex Hierarchical processing Information theory SelectivityThursday, 10/22 | 11:15AM - 12:15PM EDT | Virtual poster
Abstract
In early auditory processing stages, neural representations are likely optimized for efficient and complete representation of sounds. Theoretical work from our lab indicates that at higher processing stages, representations may be optimized for supporting specific behavioral tasks. Supporting this model, for the task of call categorization, neurons in auditory cortex of marmosets and guinea pigs showed high selectivity for intermediate call features. Here we begin to address how early dense representations are transformed into higher task-dependent representations. We recorded neural activity from the thalamus (vMGB), thalamorecipient (A1 L4), and superficial (A1 L2/3) layers of auditory cortex in unanesthetized guinea pigs passively listening to an extensive range of conspecific calls. Neurons in vMGB and A1 L4 responded densely to most call categories and throughout the call durations. In contrast, a large fraction of A1 L2/3 neurons responded sparsely and selectively to one or two call categories, and only for short durations within a call. Information theoretic analyses revealed that average information was high in the A1 L4 population as most neurons contained similar information across multiple stimuli. In contrast, individual A1 L2/3 neurons were highly informative about few stimuli, and conveyed high levels of information per spike. The dimensionality of L4 and L2/3 representations showed differences consistent with these transformations. These results suggest that an abrupt transformation in neural representation occurs between A1 L4 and Al L2/3, leading to the emergence of call-selective responses in A1 L2/3. We are currently investigating biophysical and circuit mechanisms that could underlie this transformation.
Aravindakshan Parthasarathy, Kenneth Hancock and Daniel Polley
Topic areas: brain processing of speech and language correlates of auditory behavior/perception neural coding subcortical auditory processing
frequency following response envelope following response EEG temporal processing evoked potentials pattern detection psychophysics amplitude modulation hearing diagnosticsFriday, 10/23 | 1:30PM - 2:30PM EDT | Virtual poster
Abstract
The auditory system has the remarkable ability to represent temporal information on timescales that span multiple orders of magnitude. Effective neural representation of multiple temporal elements is critical for speech comprehension, where complementary information occurs simultaneously at differing temporal resolutions– pitch and speaker identity at millisecond resolution, phonemes and envelope fluctuations over hundreds of milliseconds, words and syllables over seconds and linguistic and statistical context over longer timescales. Objective assessment of these varying temporal representations has traditionally been carried out by independently measuring the neural entrainment to these stimuli – frequency following responses (FFRs) to low frequency tones fluctuating over milliseconds, envelope following responses (EFRs) to stimulus envelopes over tens of milliseconds and cortically evoked potentials to syllabic events over seconds. In this study, we describe a new response that is entrained to the change in stimulus amplitude modulation (AM) over hundreds of milliseconds, the AM change response (or AMCR). The AMCR is sensitive to changes in modulation frequency in an ongoing stimulus, and proportional to the difference in AM rates. By presenting these varying AM rates in random arrangements, or repeating them in patterns that last over seconds, neural entrainment can be assessed to all individual elements of the stimuli occurring over multiple timescales. Ongoing work is assessing whether the FFR, the EFR, the AMCFR and EEG responses on longer timescales are sensitive to context, allowing for the simultaneous assessment of multiple temporal elements.
Min Wu, Hans Bosker, Rocha Macarena and Lars Riecke
Topic areas: brain processing of speech and language
Sentence processing Attention Frequency taggingFriday, 10/23 | 11:15AM - 12:15PM EDT | Virtual poster
Abstract
Speech studies have indicated that behavioral detection of acoustic/linguistic events is more rapid during words occurring late in sentences. The aim of this ongoing study is to characterize how sentence structure affects the sequential auditory processing of individual words and interacts with listeners’ attention. Auditory stimuli consist of an isochronous sequence of amplitude-modulated monosyllabic words (presented to the participants’ right ear) and a continuous amplitude-modulated tone (presented to the left ear). Words are ordered either randomly (unstructured speech) or as quartets forming sentences (structured speech). Selective attention is manipulated by having participants detect either a target word (speech task) or a brief loudness change in the tone (distractor task). Preliminary word-detection data suggest a word-position effect in structured speech (i.e., shorter reaction times to target words occurring late in sentences). No such effect is observed when identical target words are presented in unstructured speech (i.e., without sentence structure), which is in line with previous behavioral studies. Similar to these behavioral observations, cortical responses to individual words (assessed with EEG-frequency tagging) suggest a word-position effect in structured speech (i.e., weaker frequency-following responses to target words occurring late in sentences), while this trend is not observed in the distractor task or unstructured speech. A tentative interpretation is that sentence structure may regulate the gain of auditory cortical responses at the time scale of individual words. This putative effect might be mediated by semantic predictability and require selective attention to the words.
Julia Erb, Jens Kreitewolf, Ana Pinheiro and Jonas Obleser
Topic areas: auditory memory and cognition brain processing of speech and language correlates of auditory behavior/perception
speech perception reverse correlation spectro-temporal modulations hallucinations schizotypyFriday, 10/23 | 11:15AM - 12:15PM EDT | Virtual poster
Abstract
Hallucinations constitute an intriguing model of how percepts are generated and how perception can fail. Here, we investigate the hypothesis that an altered perceptual weighting of the spectro-temporal modulations that characterize speech contributes to the emergence of auditory verbal hallucinations. Healthy adults (N=168) varying in their predisposition for hallucinations had to choose the ‘more speech-like’ of two presented ambiguous sound textures and give a confidence judgement. Using psychophysical reverse correlation, we quantified the contribution of different acoustic features to a listener’s perceptual decisions. Higher hallucination proneness covaried with lower perceptual weighting of speech-typical, low-frequency acoustic energy. Remarkably, higher confidence judgements in single trials depended not only on acoustic evidence but also on an individual’s hallucination proneness and schizotypy score. In line with an account of altered perceptual priors and differential weighting of sensory evidence, these results show that hallucination-prone individuals exhibit qualitative and quantitative changes in their perception of the modulations typical for speech.
Emina Alickovic, Carina Graversen, Lorenz Fiedler, Dorothea Wendt, Hamish Innes-Brown, Louis Villejouberts, Sébastien Santurette, Elaine Hoi Ning Ng and Thomas Behrens
Topic areas: brain processing of speech and language correlates of auditory behavior/perception
hearing aids neural speech processing EEG noise reduction amplificationThursday, 10/22 | 1:45PM - 2:45PM EDT | Virtual poster
Abstract
Natural listening situations that require listeners to selectively attend to a talker of interest in noisy environments with multiple competing talkers are among the most challenging situations encountered by hearing impaired (HI) listeners. Such challenges become even more pronounced with increasing background noise level and may partially be overcome by adequate hearing aid (HA) amplification and noise reduction (NR) support. Using Electroencephalography (EEG), it has been demonstrated that the auditory cortex selectively represents the target talker with significantly higher fidelity than other competing talkers in normal-hearing and HI listeners. An NR scheme in commercial HAs was also found to enhance the neural representation of the foreground and suppress the neural representation of the background noise. Motivated by these findings, the objective of this study was to investigate whether neural speech processing was affected by different adaptive amplification and NR strategies in commercial HAs. We recorded EEG responses in thirty HI listeners fitted with HAs in which two different adaptive amplification and NR strategies were implemented, resulting in a total of four conditions. The participants were instructed to attend to one of two simultaneous talkers in the foreground mixed with multi-talker babble noise in the background (+3 dB SNR). Preliminary results indicate that both the choice of amplification and NR strategies can affect the neural speech processing. These results suggest that the neural representation of attended and ignored speech, in line with the behavioral performance, may be sensitive to the choice of HA signal processing strategy in HA users.
Sébastien Santurette, Martha Larsen, Pernille Aaby Gade, Lu Xia, Jens-Christian Britze Kijne and Josefine Juul Jensen
Topic areas: auditory disorders correlates of auditory behavior/perception
Selective attention Auditory streaming Hearing aids Psychoacoustics Noise reductionThursday, 10/22 | 1:45PM - 2:45PM EDT | Virtual poster
Abstract
With hearing loss (HL), the challenge of selectively attending to one of multiple talkers may be exacerbated by the presence of background noise and alleviated by adequate hearing aid (HA) amplification and noise reduction support. When attempting to focus on one of several speech streams, listeners with HL typically show more confusion errors than their normal-hearing peers, while still showing a buildup of auditory streaming over time. We investigated whether aided performance in such a task was affected by background noise and compensation with different HA processing schemes. Thirty HA users were asked to repeat four-digit sequences randomly located at -30, 0, or +30 degrees azimuth, while competing digits were simultaneously presented from the two other locations, either in quiet or in babble noise played from -100, 180, and +100 degrees. Two HA amplification strategies and two noise reduction schemes were tested. While background noise negatively affected performance, the listeners always scored above chance. Mean scores increased with digit position, suggesting a buildup of auditory streaming. Confusion errors were relatively higher than random errors, confirming that the ability to focus on the stream of interest and ignore competing streams drove performance. Scores differed between the two HA amplification strategies in both quiet and noise conditions. In noise, the activation of both noise reduction schemes increased performance. Overall, these results validate the applicability of a competing digits task to behaviorally assess selective attention in HA users and indicate that performance in this task is sensitive to differences in HA signal processing.
Alejandro Tabas and Katharina von Kriegstein
Topic areas: hierarchical sensory organization neural coding subcortical auditory processing
predictive coding subcortical pathway SSA computational modelling fMRI MGB ICThursday, 10/22 | 1:45PM - 2:45PM EDT | Virtual poster
Abstract
The predictive coding framework suggests that sensory neurons adapt their responses by constantly matching incoming stimuli against an internal prediction derived from a generative model of the sensory input. Although predictive coding is generally accepted to underlay cortical sensory processing, the role of predictability in subcortical sensory coding is still unclear. Single neurons and neuronal ensembles of the mammal subcortical sensory pathway nuclei exhibit stimulus specific adaptation (SSA), a phenomenon where neurons adapt to regularly occurring stimuli yet show restored responses to a stimulus with a divergent pitch. Although SSA is often interpreted as prediction error, computational models to date have successfully explained it in terms of synaptic fatigue. Mesoscopic data in humans is also ambiguous: BOLD responses mimic the single-cell SSA phenomenology, but the frequency-following-response (FFR), a continuous evoked potential partly elicited by subcortical sources, shows repetition and expectation enhancement. Here, we first introduce novel experimental data showing that human BOLD responses in the subcortical pathway reflect predictability, even when the participant expectations are modulated by abstract rules. These data suggests that predictive coding is the main normative mechanism underlying SSA in the human subcortical pathway. Next, we develop a new model of pitch encoding following the main directives of predictive coding that explains: 1) single-neuron SSA as recorded in mammals, 2) the repetition/expectation suppression of the BOLD responses, and 3) the repetition/expectation enhancement of the FFR. Together, our results suggest that the subcortical auditory pathway is fully integrated in the cortical predictive processing network.
Caitlin Price and Gavin Bidelman
Topic areas: brain processing of speech and language correlates of auditory behavior/perception hierarchical sensory organization subcortical auditory processing
attentional modulation hierarchical speech encoding cognitive factors frequency-following response functional connectivity EEGFriday, 10/23 | 1:30PM - 2:30PM EDT | Virtual poster
Abstract
Communication in noise is an incredibly complex process that requires efficient neural encoding throughout the entire auditory pathway as well as contributions from higher-order cognitive processes (i.e., attention) to effectively extract speech cues for perception. Attention enables the selection and prioritization of task-relevant inputs over competing background noise. It is thought to enhance speech processing by fine-tuning neural encoding across the pathway via top-down, or corticofugal, projections. However, whether attention actively influences early sensory processing within auditory brainstem and how it contributes to speech understanding remains unclear. The current study evaluated top-down, attentional modulation of SIN processing and how attention interacts with early sensory encoding. Hierarchical speech processing in brainstem and cortex was assessed by measuring source-resolved frequency-following responses (FFRs) and event-related potentials (ERPs) in normal hearing, young adults during active SIN perceptual tasks. Functional connectivity measures were used to assess the strength and direction of neural signaling between these responses and identify “bottom-up” vs. “top-down” (corticofugal) communication within the auditory brainstem-cortical pathway. Comparisons between (i) active and passive SIN tasks and (ii) clean and noise responses evaluated attentional modulation of this circuit and whether added cognitive demands of noise altered hierarchical neural processing. We found that attention modulates SIN processing at both subcortical and cortical levels and strengthens bidirectional neural signaling within the central auditory pathway. A relative disengagement of corticofugal transmission was observed in noise but only for passive listening suggesting attention aids SIN perception by maintaining top-down reinforcement of acoustic feature encoding within the primary auditory pathways.
Matthew Leonard, Emily Stephen, Lingyun Zhao and Edward Chang
Topic areas: brain processing of speech and language correlates of auditory behavior/perception neural coding
Speech ECoG Perception Binding SequencingThursday, 10/22 | 11:15AM - 12:15PM EDT | Virtual poster
Abstract
As natural speech unfolds rapidly, listeners must extract meaningful linguistic units like words, which are composed of specific sequences of phonemes and syllables. Often, there are no clear acoustic markers indicating where words begin and end. How does the brain segment continuous speech into perceptually-meaningful whole word chunks? We recorded neural activity directly from the human cortex while listeners heard continuous speech that could be perceived as multiple words (e.g., “/…seɪseɪseɪseɪ…/”, which can be heard as “say” or “ace”). They indicated which word they were hearing at every point during the experiment, allowing us to examine neural activity associated with distinct percepts from an identical sound. Some neural populations, primarily in superior temporal gyrus, encoded specific words (e.g., “say” vs “ace”), while others encoded a more general phonological sequence order (e.g., CV vs VC; “say/day” vs “ace/aid”). Furthermore, many populations encoded perception for both 1- and 2-syllable sequences (e.g., “/…sɜrtɛnsɜrtɛnsɜrtɛn…/, which can be heard as “certain” or “tensor”), demonstrating representation of length-independent word-level units. Perceptual encoding occurred most strongly in populations that were weakly tuned to acoustic-phonetic features like vowel formants, demonstrating that flexible, context-specific representations of sub-word speech content are a mechanism for binding input across time. In addition, populations tuned to speech envelope features encoded word percepts by varying the timing of the response, suggesting that both amplitude and temporal codes are used to represent words. Together, these results provide a direct demonstration of neural computations that underlie the perception of word-level phonological sequences during continuous, connected speech.
Mark Saddler, Ray Gonzalez and Josh McDermott
Topic areas: correlates of auditory behavior/perception hierarchical sensory organization neural coding
pitch perception artificial neural network peripheral coding sound statisticsThursday, 10/22 | 1:45PM - 2:45PM EDT | Virtual poster
Abstract
Behavior results from computations on sensory receptor responses that enable interaction with the environment. Behavior is plausibly shaped by both receptors and the environment for which organisms are optimized, but these constraints are often opaque. One classic example is pitch perception, whose properties are commonly linked to peripheral neural coding limits rather than environmental acoustic constraints. We trained artificial neural networks to estimate fundamental frequency from simulated cochlear representations of natural sounds. The best-performing networks replicated many characteristics of human pitch judgments. To probe how our ears and environment shape these characteristics, we optimized networks given altered cochleae or sound statistics. Human-like behavior depended critically on the fidelity of phase-locked auditory nerve spiking, but also on training with natural sounds. The results suggest that human pitch perception is optimized for natural sounds heard through a cochlea, illustrating the use of contemporary neural networks to reveal underpinnings of behavior.
Alexander Pei, Winko An and Barbara Shinn-Cunningham
Topic areas: brain processing of speech and language correlates of auditory behavior/perception hierarchical sensory organization neural coding
auditory attention electroencephalography frontoparietal connectivity alpha oscillations gamma oscillationsFriday, 10/23 | 1:30PM - 2:30PM EDT | Virtual poster
Abstract
Parietal alpha oscillations play an inhibitory role in audiospatial attention, presumably by suppressing auditory object representations that dominantly reflect events from contralateral space. Spatial attention manifests as increases in alpha ipsilateral to the side of attention. Alpha oscillations also functionally integrate frontal and parietal cortices to perform working memory and attention tasks. We investigated both the inhibitory and functional integrative roles of alpha oscillations in audiospatial attention by recording EEG during a cued-location attention task. Phase-amplitude coupling analysis demonstrates that in both preparatory and stimuli periods, the phase of right parietal alpha oscillations modulates the power of right parietal gamma oscillations during audiospatial attention. Weighted phase-lag index measures showed that alpha oscillations also functionally coupled right parietal and right frontal cortices during left spatial attention. Using dynamic causal modeling for induced responses, we found that models with frontal to parietal directionality have higher evidence than models with parietal to frontal directionality, implicating top-down frontal to parietal interactions. Dynamic causal models showed that frontal alpha increases parietal alpha and simultaneously decreases parietal gamma. Together, these results provided evidence that during top-down audiospatial attention, frontoparietal cortices are synchronized through alpha oscillations; these alpha oscillations simultaneously reduce cortical excitability and parietal gamma oscillations.
Jian Carlo Nocon, Howard Gritton, Kenny Chou, Xue Han and Kamal Sen
Topic areas: neural coding
Sound localization Cocktail party problem Auditory neuroscience Computational modelingFriday, 10/23 | 11:15AM - 12:15PM EDT | Virtual poster
Abstract
An impressive feature of the human and animal auditory system is its ability to segregate one sound from competing sounds in a complex auditory scene (i.e., the Cocktail Party Problem (CPP)). While mechanisms to extract cues in the auditory periphery are well understood, little remains known about how cortical circuits contribute to resolving the CPP. Previously, we constructed a computational model of cortical responses to spatially distributed sound mixtures in songbirds. This model explained experimentally observed broad spatial tuning in the presence of single target stimuli, and the sharpening of spatial tuning in the presence of competing masker stimuli. Specifically, it accurately predicted the position of “hotspots” of high discrimination performance on a spatial grid of all possible configurations of target and masker locations. The model predicted that hotspots arise from inhibiting neurons tuned to specific spatial locations. Recently, we discovered similar spatial grids in mouse A1 neurons. Here, we present a computational model to explain the structure of mouse A1 grids. While model structure is similar to that for songbirds, inputs are modified to match spatial tuning curves in the mouse midbrain. First, we catalog archetypes of grids that could arise from distinct patterns of input convergence onto a cortical unit. We then model the effects of cross-spatial inhibition on hotspots and spatial grids. Finally, we present a quantitative method for fitting spatial grids in mouse A1 based on experimental data and propose a series of future experiments in mice to probe cortical circuitry underlying the CPP with optogenetics.
Jacques Pesnot Lerousseau, Agnès Trébuchon, Benjamin Morillon and Daniele Schön
Topic areas: brain processing of speech and language correlates of auditory behavior/perception neural coding
Human Cortex Neural oscillations MEG iEEG Temporal coding Harmonic oscillatorThursday, 10/22 | 1:45PM - 2:45PM EDT | Virtual poster
Abstract
Rhythmic stimulation, either sensory or electrical, aiming at entraining oscillatory activity to reveal or optimize brain functions, relies on an untested hypothesis: it should produce a persistent effect, outlasting the stimulus duration. We tested this assumption by studying cortical neural oscillations during and after presentation of rhythmic auditory stimuli. Using intracranial and surface recordings in humans, we reveal consistent neural response properties throughout the cortex, with entrainment present only in high-gamma oscillations. Critically, during passive perception, neural oscillations do not outlast low-frequency acoustic dynamics, revealing that sensory driven exposure to a periodic stream does not automatically lead to auditory entrainment. We further show that our data are well-captured by a model of damped harmonic oscillator and can be classified into three classes of neural dynamics, with distinct damping properties and eigenfrequencies. This model thus provides a mechanistic and quantitative explanation of the frequency selectivity of auditory neural entrainment in the human cortex.
Lars Hausfeld, Federico De Martino and Elia Formisano
Topic areas: brain processing of speech and language correlates of auditory behavior/perception
high-field fMRI selective attention cocktail party speech tracking MVPAThursday, 10/22 | 11:15AM - 12:15PM EDT | Virtual poster + podium teaser
Abstract
Previous studies used ECoG, MEG and EEG measurements to examine brain responses to “cocktail-party”-like listening situations. These showed that neural activity “tracked” features of the speech signal like the amplitude of the speech envelope, especially of relevant speech. Furthermore, findings suggested that primary and non-primary auditory cortex contributed to the tracking of speech and its modulation by task. However, due to the measurements’ limited coverage and/or spatial resolution, the specific role of these regions and areas outside auditory cortex remains unclear. We measured brain responses of human participants with high-field fMRI at 7T who paid attention to one speaker presented without and with a concurrent distractor speaker (‘single speaker’ and ‘auditory scene’, respectively). Using voxel-wise models, we observed - for BOLD activation sampled at 1Hz - tracking of the speech envelope of single speakers in bilateral Heschl’s gyrus (HG) and middle superior temporal sulcus (mSTS). We found that the BOLD activity was either positively (HG) or negatively (mSTS) correlated to the speech signal. For these regions, the multi-voxel analysis of spatial patterns during auditory scenes suggested that tracking in HG reflected both relevant and (to a lesser extent) non-relevant speech whereas right mSTS selectively represented the relevant speech signal. Exploratory analyses of BOLD activity in the auditory scene supported a role of posterior superior temporal gyrus for processing non-relevant speech and revealed a wide network characterized by negative tracking of relevant speech in temporal, parietal and frontal cortex most likely reflecting the adaptation to and execution of task demands.
Natsumi Homma, Craig Atencio and Christoph Schreiner
Topic areas: neural coding
information theory maximally informative dimension primary auditory cortex rat spectrotemporal receptive field ventral auditory fieldFriday, 10/23 | 1:30PM - 2:30PM EDT | Virtual poster
Abstract
The nonlinear processing by auditory cortical neurons can be accounted for by the joint activity of two or more spectrotemporal receptive fields. The static or dynamic functional relationship between these filters is, however, not well understood. Here we examined temporal and spectral modulation properties and the cooperativity of multidimensional filters in two core rat auditory cortical fields, primary auditory cortex (A1) and ventral auditory field (VAF). We estimated two spectrotemporal receptive fields (STRFs) per single-unit neuron using maximally informative dimension (MID) analysis. While the spike information conveyed by the first STRF was higher in VAF than in A1, the additional second STRF contribution and nonlinear cooperativity of these two filters were larger for A1. We next tested how the acoustic environment affects the structure and relationship of these two filters. Rats were exposed to spectrotemporally-modulated noise during their developmental period and early adulthood. This moderate noise exposure (~60 dB SPL) altered the spectrotemporal preference of both filters and changed the interaction between the filters in A1. The results suggest that (i) noise exposure diminishes modulation parameter representation contained in the noise, (ii) A1 has more synergistic filter interactions than VAF, and (iii) multidimensional receptive fields shifted towards greater independence between filters following exposure to noise. It indicates that core fields not only differ in what they process but also how they process stimulus information.
Ira Kraemer and Catherine Carr
Topic areas: cross-species comparisons subcortical auditory processing
superior olivary nucleus barn owl auditory brainstem in vivo electrophysiology inhibitionThursday, 10/22 | 11:15AM - 12:15PM EDT | Virtual poster
Abstract
We recorded in the barn owl superior olivary nucleus (SON) to understand how inhibition regulates excitatory brainstem nuclei, since the barn owl has relatively simpler auditory brainstem circuitry compared to mammals (review in Nothwang 2016). Although important for regulating the sound localization circuit in chicks in vitro (Yang et al. 1999; Monsivais et al. 2000; Yang et al. 2012; review in Burger et al. 2011) and in vivo (Nishino et al. 2008), there is little research on the barn owl SON. Moiseff & Konishi (1983) recorded SON single units in barn owl and suggested that their search criteria were biased to the ipsilateral ear (Moiseff and Konishi 1983). We therefore recorded single units in barn owl SON, using lesions to confirm the recording location. We recorded a total of 128 auditory single units, 43 which were in SON. The majority of units responded preferentially to binaural noise and weakly to tones, with broad frequency tuning and low spontaneous rates. Response types were heterogeneous, and included off-responses, sustained, suppressed, primary-like and onset units. Because the SON projects back to the three brainstem excitatory nuclei and the inferior colliculus, and has a distinct homotypic projection to the contralateral SON, these heterogeneous response types may represent separate SON neuronal populations (Burger et al. 2005). Further experiments are needed to understand how SON regulates the activity in the auditory brainstem of the barn owl. Grant# R01 DC000436/NIDCD NIH HHS/United States Grant# T32 DC-00046/NIDCD NIH HHS/United States
Matthew McGill, Ariel Hight, Yurika Watanabe, Dongqin Cai, Aravindakshan Parthasarathy and Daniel Polley
Topic areas: auditory disorders correlates of auditory behavior/perception neural coding
Hearing loss Hyperacusis Hyperactivity Loudness Perception Central gainFriday, 10/23 | 11:15AM - 12:15PM EDT | Virtual poster
Abstract
Across sensory systems, a deprivation of peripheral input is associated with a compensatory increase in neural excitability in primary sensory cortex. Commonly, this compensatory process overshoots the mark, resulting in chronic hyperactivity along with perceptual hypersensitivity to sensory stimuli. Such observations have a long history in the auditory system: noise-induced damage to the cochlea is associated with paradoxical cortical hyperactivity and perceptual abnormalities including hyperacusis, an auditory disorder characterized by an increase in perceived loudness to moderately intense sounds. Although it is believed that cortical hyperactivity is a driving force behind perceptual hypersensitivity, the definitive experiments linking hyperactivity and hypersensitivity have yet to be performed. Here, we induced sensorineural hearing loss in mice by exposure to intense noise, and developed chronic two-photon imaging approaches as well as new Go/NoGo (GNG) and two-alternative forced choice (2AFC) operant behavioral tasks that demonstrate cortical involvement in perceptual hypersensitivity. Behavioral markers of hyperacusis included steepened tone detection functions in the GNG task and increased categorization of moderate intensity tones as ‘loud’ after acoustic trauma in the 2AFC task for sound frequencies bordering the cochlear lesion. In a modified GNG task, we interleaved tone detection with optogenetic activation of thalamocortical projection neurons to isolate a cortical contribution to perceptual hypersensitivity. Two-photon calcium imaging demonstrated daily changes in neural gain and map reorganization that match the time course of perceptual hypersensitivity. Further understanding of the neural circuit pathology underlying perceptual hypersensitivity will prove valuable for neurological disorders with core hypersensitivity phenotypes such as autism and migraine.
Yunan Wu, Yongyi Wu and Lori Holt
Topic areas: auditory memory and cognition brain processing of speech and language neural coding
Speech perception Acoustic dimensions Short-term regularities Adaptive plasticity MMN N100 P300Friday, 10/23 | 1:30PM - 2:30PM EDT | Virtual poster
Abstract
Dialects, accents, or even a stuffy nose can shift short-term speech acoustics away from language community norms and impact comprehension. Yet, perception exhibits rapid adaptive plasticity. When disambiguating information shifts perception of ambiguous speech input, its influence can persist even when the disambiguating information is no longer available. There is, as yet, no detailed neurobiological model of this adaptive plasticity. Here, we used electroencephalography to examine human auditory cortical response to speech exemplars that varied only in fundamental frequency (F0) as we manipulated short-term speech input regularities. In short blocks, participants overtly categorized speech sampled across voice onset time (VOT) and F0 space sampled to mirror the VOTxF0 relationship in English, or reversed to create an ‘accent.’ In intermixed blocks, participants passively listened to the two F0-disinguished stimuli presented in an 85:15 ratio to elicit a mismatch negativity (MMN) response. Behavioral reliance on F0 in speech categorization was robust when short-term regularities aligned with English. In the context of the accent, F0 exerted little influence on speech categorization. The impact of the short-term regularities was reflected in the magnitude of the MMN in both early temporal windows associated with encoding acoustic dimensions and later windows associated with category-level processing. Event-related potentials in the overt categorization task also were impacted by short-term regularities across both N100 and P300 temporal windows and, for P300, exhibited a strong correlation with behavioral measures. Short-term regularities impact even early auditory cortical evoked potentials, in passive listening and overt categorization.
Nihaad Paraouty and Dan H. Sanes
Topic areas: auditory memory and cognition correlates of auditory behavior/perception neural coding
social learning auditory discrimination primary auditory cortexThursday, 10/22 | 12:15 PM - 12:30PM EDT | Short talk
Abstract
The acquisition of new skills can be facilitated by social experience, typically studied by exposing a naïve animal to a conspecific that is performing a well-defined behavior. In a previous study, we showed that naïve gerbils can learn a Go-Nogo sound discrimination task following 5 days of exposure to a demonstrator gerbil. Specifically, facilitation of learning occurred even when observer and demonstrator were separated by an opaque divider, suggesting that observers were learning auditory cues. Here, we asked whether such social learning is associated with auditory cortex (AC) function. We first showed that auditory cortex is necessary for social learning. Naïve gerbils received bilateral infusions of an agent that blocks activity (muscimol) in AC to during the exposure sessions. These animals required a significantly greater number of days to learn the task during subsequent practice sessions, as compared to saline-infused controls. We then asked whether AC neuron coding properties were modified during social learning. Using 64-channel silicon probes, we recorded from freely moving naïve gerbils during the exposure sessions. AC neurons displayed significant changes in firing rate responses to the Go and Nogo stimuli across exposure days. Furthermore, using a spike pattern classifier, we found that, for the majority of AC neurons, the neural dprime values also improved across exposure sessions. Together, these results suggest that social exposure to a demonstrator performing a sound discrimination task leads to neural plasticity changes in the AC of a naïve gerbil, and these changes may be instrumental to the subsequent facilitation of learning. Crowdcast Session
Kelsey Mankel and Gavin Bidelman
Topic areas: brain processing of speech and language correlates of auditory behavior/perception
Auditory learning Categorical perception (CP) Event-related potentials (ERPs) EEGThursday, 10/22 | 11:15AM - 12:15PM EDT | Virtual poster
Abstract
Categorizing sounds into meaningful groups, as opposed to processing them individually, helps listeners more efficiently process the auditory scene. Yet, how auditory categories develop in the brain through learning is not well understood. In the current study, musically naïve listeners completed a brief (~20 min) training session where they learned to identify sounds from a minor and major 3rd musical interval continuum. Identification performance was assessed before and after training while multichannel EEG tracked neuroplastic changes in the auditory ERPs from pre- to post-test. Behavioral performance revealed steeper (i.e., more categorical) identification function slopes in the posttest that correlated with better training accuracy. At the neural level, successful learning was associated with a decrease in P2 amplitudes after training at frontocentral electrode sites, and the correspondence between neural response (i.e., N1 and P2) amplitudes and behavioral identification performance changed from pre- to post-test. These learning-related changes in the ERPs were not observed in control listeners who did not partake in training. Learning was also associated with stronger categorical neural responses over temporal sites in right compared to left hemisphere. Collectively, our results suggest successful auditory categorical learning of nonspeech sounds is associated with changes in early neural activity beginning no later than ~100-180 ms. The right hemisphere dominance for learning musical sound labels complements the left hemisphere bias reported for speech categorization.
Nathan Tardiff, Lalitta Suriya-Arunroj, Yale E. Cohen and Joshua I. Gold
Topic areas: auditory memory and cognition
perceptual decision-making evidence accumulation biases habituation adaptationFriday, 10/23 | 1:30PM - 2:30PM EDT | Virtual poster
Abstract
Auditory decisions can be affected by different sources of prior information in potentially different ways. We compared how top-down and bottom-up information, provided separately or together, induced choice and response-time (RT) biases of human subjects (N=50) performing an auditory decision-making task. Subjects categorized a test tone embedded in noise as low frequency or high frequency. Task difficulty was controlled by varying the signal-to-noise ratio (SNR) of the test tone, relative to the noise. In “top-down” conditions, a symbol presented just before the test tone indicated the most probable frequency. In “bottom-up” conditions, the test tone was preceded by a stream of tones that varied in the number and pattern of low and high frequencies. As expected, the top-down cue generated consistent choice and RT biases toward the more-probable tone, particularly when the SNR was low. In contrast, the bottom-up stimuli had two primary effects: 1) a similar choice and RT bias toward the most recently heard tones in the sequence when the SNR was low; but 2) a choice and RT bias in the opposite direction, away from the most recently heard tones in the sequence, when the SNR was high. We modeled these effects in the context of a drift-diffusion decision model as: 1) an additive, expectation-dependent bias that can come from top-down or bottom-up information; plus 2) a multiplicative, adaptation-like effect on sensitivity driven by the bottom-up sequence. These results provide a new, unified view of how top-down and bottom-up information combine to affect auditory decisions.
Nina Suess, Anne Hauswald, Sebastian Roesch and Nathan Weisz
Topic areas: auditory disorders brain processing of speech and language multisensory processes
MEG visual speech perception lip reading age-related hearing loss cross-modal integrationThursday, 10/22 | 1:45PM - 2:45PM EDT | Virtual poster
Abstract
Successful lip-reading requires the transformation from seen, but unheard speech to the corresponding phonological information. In a previous study we have shown that this process occurs already at the visual cortical level. Since lip-movements are an important cue for understanding speech under noisy conditions and hearing loss increases with age, we investigated visuo-phonological transformation processes as a function of age. Concretely, in an MEG study we investigated how the brain tracks the absent acoustic speech envelope when watching silent intelligible (forward speech) and unintelligible (backward speech) lip movements using coherence. We calculated an “intelligibility index” that contrasts coherence during forward and backward speech and calculated a voxel-wise correlation with age. Corroborating our previous report, overall visual cortex entrains stronger to the acoustic envelope of unheard intelligible than unintelligible speech. Interestingly this effect was also observed in the left auditory cortex of younger individuals. The intelligibility index correlated negatively with age especially in auditory areas, and to some extent in motor areas and visual areas. The absence of effects when using the visual envelope as reference signal (and calculating an analogous contrast) strengthens the previous claim that our results reflect genuine crossmodal audiovisual transformation processes. Our study suggests however that these processes are sensitive to age-related decline, in particular in cross-modally “mapped“ information engaging auditory cortical regions. This observation may have important real-life implications, as a reduced visuo-phonological transformation ability could exacerbate listening difficulties of aging individuals beyond the direct effects of hearing loss.
Eugenia Gonzalez Palomares, Luciana López Jury, Francisco García Rosales and Julio C. Hechavarria
Topic areas: brain processing of speech and language hierarchical sensory organization neural coding subcortical auditory processing
inferior colliculus auditory midbrain mutual information natural sounds brain-stimulus synchrony batsFriday, 10/23 | 11:15AM - 12:15PM EDT | Virtual poster
Abstract
The auditory midbrain (inferior colliculus, IC) plays an important role in sound processing, acting as hub for acoustic information extraction and for the implementation of fast audio-motor behaviors. IC neurons are topographically organized according to their sound frequency preference: dorsal IC regions encode low frequencies while ventral areas respond best to high frequencies, a type of sensory map defined as tonotopy. Tonotopic maps have been studied extensively using artificial stimuli (pure tones) but our knowledge of how these maps represent information about sequences of natural, spectro-temporally rich sounds is sparse. We studied this question by conducting simultaneous extracellular recordings across IC depths in awake bats (Carollia perspicillata) that listened to sequences of natural communication and echolocation sounds. The hypothesis was that information about these two types of sound streams is represented at different IC depths since they exhibit large differences in spectral composition, i.e. echolocation covers the high frequency portion of the bat soundscape (> 45 kHz), while communication sounds are broadband and carry most power at low frequencies (20-25 kHz). Our results showed that mutual information between neuronal responses and acoustic stimuli, as well as response redundancy in pairs of neurons recorded simultaneously, increase exponentially with IC depth. The latter occurs regardless of the sound type presented to the bats (echolocation or communication). Taken together, our results indicate the existence of mutual information and redundancy maps at the midbrain level whose response cannot be predicted based on the frequency composition of natural sounds and classic neuronal tuning curves.
Karol Przewrocki, Gloria Gutierrez Parras, Javier Nieto-Diego, Artoghrul Alishbayli, Manuel Malmierca and Bernhard Englitz
Topic areas: