Biology subjects

Morillon, B.

Publications and source records attributed to Morillon, B..

At least 19 recordsLinked to original sources

bioRxiv2026

Cortical encoding of probabilistic temporal predictions during speech perception

The temporal structure of speech has traditionally been characterized by the rhythmicity of its canonical linguistic units (phonemes, syllables, words), each summarized by a mean occurrence rate. While valid, this view overlooks whether speech carries a finer, context-dependent and probabilistic temporal structure that could support temporal predictive coding during listening. Using large French and English speech corpora, we trained models of increasing complexity to predict the onsets of linguistic units. Recurrent neural networks (RNNs) outperform mean-rate and hazard-rate models, showing that the variability around these rates is not noise but a temporal structure shaped by local context, statistically predictable across phonemes, syllables and words. Recording from 7,698 intracerebral electrodes in 53 neurosurgical patients listening to natural speech, we next show that the models' output), the continuous probability of an upcoming onset (when), explains neural activity beyond acoustic and linguistic content (what) features, with markedly stronger effects for RNNs than for mean- or hazard-rate models. This dynamic neural prediction of when an onset will occur is dissociable from the encoding of linguistic content, relying on largely distinct channel populations. Temporal predictions engage a distributed cortical network extending from bilateral temporal cortex into left frontal and sensorimotor regions. Together, these results establish temporal prediction in speech as a dynamic, context-dependent and probabilistic process in its own right.

neuroscience↗

bioRxiv2026

Surprise gates two distinct mechanisms to support memorability in music

Music is a uniquely memorable human creation that, when skillfully composed, can persist in individual memory (as in e.g., earworms) and cultural transmission (e.g. global hits or anthems). While both acoustic and statistical properties are known to influence a songs memorability, the neural mechanisms that facilitate the engramming of certain musical sequences remain unclear. Current theories suggest that memory systems function as predictive internal models, enhancing learning when expectations are violated. Yet expected stimuli, by aligning with and reinforcing prior knowledge, also enhance memorability. How these two opposing processes arise from the brains sensitivity to statistical regularities, especially in naturalistic sequences, is not well understood. Here, we leveraged musics intrinsic balance between expectancy and surprise, and examined the neural correlates of minutes-scale memorability using intracranial EEG recordings from nine patients with epilepsy performing a musical memory task. Quantifying the statistical surprise of each musical excerpt with PolyRNN, a polyphonic model of musical expectations, we uncovered a U-shaped relationship between musical surprise and memory performance: both highly expected and highly surprising melodies led to greater memorability. While neural pattern similarity between song repetitions was enhanced for low-surprise stimuli, high-surprise stimuli enhanced neural pattern separability in medial temporal regions, each mediating memorability in distinguishable ways. These findings reveal complementary neural mechanisms through which statistical structure shapes musical sequence memory, clarifying how the brain encodes complex, ecologically valid stimuli.

neuroscience↗

bioRxiv2025

Darwin's entangled bank through deep time: Structural stability of mutualistic networks over evolutionary timescales

Mutualistic interactions form species-rich, complex networks that play essential roles for ecosystem function. Over macroevolutionary time scales, global- and continental-level networks change as species emerge and go extinct, yet the stability of their structural organization remains poorly understood. Here, we develop ELEFANT, a novel method for reconstructing ancestral interaction networks based on the hypothesis that species interactions are shaped by unobserved traits that evolve over evolutionary time. We show that ancestral interaction networks can be reliably reconstructed from present-day phylogenetic and interaction data. We infer the ancestral networks of plant mutualisms involving arbuscular mycorrhizal fungi, bat pollinators, and bird seed dispersers at large biogeographic scales. We find that these mutualistic networks exhibit a modular structure that seems to have persisted for millions of years, in part maintained by the evolutionary conservatism of species interactions. As species diversify, they tend to show limited shifts in mutualistic partners, which results in a remarkable long-term stability of mutualistic network structure at large spatial scales.

evolutionary biology↗

bioRxiv2025

Two forms of uncertainty hierarchically shape lexical predictions during natural speech processing

Prediction under uncertainty is a hallmark of human cognition(Friston, 2010), yet how the brain represents uncertainty remains unresolved(Bach and Dolan, 2012). Speech provides a powerful test case: lexical predictions involve thou-sands of alternatives, making classical Shannon entropy an ambiguous measure of uncertainty. Here we explored a generalization of Shannon entropy, the Renyi entropy family (Renyi, 1961), combined with information-theoretic tools and whole-brain intracranial EEG during naturalistic listening. Word-related neural activity was best explained by two complementary forms of uncertainty: dis-persion, reflecting the number of plausible alternatives, and strength, reflecting the probability of the most likely word. These uncertainties were consistently encoded in each participant by distinct neural populations, with opposing com-putational goals. Dispersion shapes anticipatory sensory activity and dampens responses to unexpected inputs, consistent with Bayesian optimization(Pouget et al., 2013). Strength instead emerges only for surprising words in lexical circuits, amplifying prediction errors and promoting information seeking. Despite bidirec-tional exchange between these two populations, their dynamics are dominated by local computations(Heilbron et al., 2022), challenging purely hierarchical models of predictive processing. Our findings reveal that the brain approximates large lexical probability distributions using multiple summary statistics, providing a flexible, low-dimensional representation of uncertainty in complex cognitive tasks and advancing principles of predictive processing across cognition.

neuroscience↗

bioRxiv2025

Neural coding of spectrotemporal modulations in the auditory cortex supports speech and music categorization

Auditory processing is typically described as hierarchical, culminating in neural representation of abstract categories. However, it remains unclear whether category-selective responses in auditory cortex require representational mechanisms beyond the coding of acoustic features, or whether the acoustic representations already available in the auditory cortex are sufficient to account for categorization. Here, we test whether cortical coding of spectrotemporal modulation (STM) features is sufficient to support speech-music categorization by combining human intracranial recordings with continuous behavioral judgments of a naturalistic soundtrack in which speech and music occur both separately and simultaneously. We show that temporal and spectral modulation patterns largely characterize speech and music, respectively, and that cortical auditory regions robustly track these features over time, with distinct oscillatory frequency bands preferentially encoding temporal and spectral modulations. Critically, cortical representations of STMs predicted perceptual categorical judgments gathered in an independent sample. Finally, speech- and music-related STM representations showed stronger tracking of category-specific acoustical features in left versus right cortical auditory regions, respectively. These findings indicate that the efficient neural coding of acoustical features provides a sufficient basis for the categorical distinction between speech and music, and that the temporal and spectral components of this representation are implemented through distinct oscillatory mechanisms.

neuroscience↗

bioRxiv2025

Deep bidirectional interplay between sensory integration and latent rule discovery

Deciphering the structure of variable sensory input is key to building an accurate model of ones environment. Humans can integrate stimulus sequences to estimate their sensory statistics, predict the timing of upcoming stimuli, but also discover rules governing sequence generation. However, whether these three forms of inference operate independently or synergistically remains untested. Here we report selective interactions between sensory integration, temporal prediction, and rule discovery in humans. Participants were exposed to rhythmic sequences of 10 stimuli governed or not by a latent rule - a predictable change in stimulus statistics after 5 stimuli - and then asked to predict the 10th stimulus from incomplete sequences. Individual differences in sensory integration timescale for rule-free sequences predicted efficient rule discovery. Conversely, discovering the latent rule shaped the timescale and format of sensory integration for rule-based sequences. Tampering with the rhythmicity of stimulus presentation impaired rule discovery without affecting sensory integration accuracy. Selective perturbations of recurrent neural networks (RNNs) trained in the same conditions confirmed these specific interactions. Together, these findings provide novel insights into the flexibility of human inferences based on variable yet predictable sensory input.

neuroscience↗

bioRxiv2024

Multi-stream predictions in human auditory cortex during natural music listening

Musical expectations shape how we perceive and process music, yet current computational models are limited to monophonic or simplified stimuli. The study of the neural processes underlying musical expectations in real-world music therefore requires significant advances in our statistical modeling of these stimuli. We present PolyRNN, a recurrent neural network designed to model expectations in naturalistic, polyphonic music. We recorded neurophysiological activity non invasively (MEG) and within the human brain (intracranial EEG) while participants listened to naturally expressive piano recordings. The musical expectations estimated by the model are encoded in evoked P2- and P3-like components in auditory regions. Comparing PolyRNN to a state-of-the-art generative music model, we show that piano roll representations are best suited to represent expectations in polyphonic contexts. Overall, our approach provides a new way to capture the musical expectations emerging from natural music listening, and enables the study of predictive processes in more ecologically valid settings.

neuroscience↗

bioRxiv2024

Behavioral Evidence for Two Modes of Attention

Attention modulates sensory gain to select and optimize the processing of behaviorally relevant events. It has been hypothesized that attention can operate in either a rhythmic or continuous mode, depending on the nature of sensory stimulation. Despite this conceptual framework, direct behavioral evidence has been scarce. Our study explores when attention operates in a rhythmic mode through a series of nine interrelated behavioral experiments with varying stream lengths, stimulus types, attended features, and tasks. The rhythmic mode optimally operates at approximately 1.5 Hz and is prevalent in perceptual tasks involving long (> 7 s) auditory streams. Our results are supported by a model of coupled oscillators, illustrating that variations in the systems noise level can induce shifts between continuous and rhythmic modes. Finally, the rhythmic mode is absent in syllable categorization tasks. Overall, this study provides empirical evidence for two modes of attention and defines their conditions of operation.

neuroscience↗

bioRxiv2024

Bridging verbal coordination and neural dynamics

AbstractOur use of language, which is profoundly social in nature, essentially takes place in interactive contexts and is shaped by precise coordination dynamics that interlocutors must observe. Thus, language interaction is highly demanding on fast adjustment of speech production. Here, we developed a real-time coupled-oscillators virtual partner that allows - by changing the coupling strength parameters - to modulate the ability to synchronise speech with a virtual speaker. Then, we recorded the intracranial brain activity of 16 patients with drug-resistant epilepsy while they performed a verbal coordination task with the virtual partner (VP). More precisely, patients had to repeat short sentences synchronously with the VP. This synchronous speech task is efficient to highlight both the dorsal and ventral language pathways. Importantly, combining time-resolved verbal coordination and neural activity shows more spatially differentiated patterns and different types of neural sensitivity along the dorsal pathway. More precisely, high-frequency activity in left secondary auditory regions is highly sensitive to verbal coordinative dynamics, while primary regions are not. Finally, while bilateral engagement was observed in the high-frequency activity of the IFG BA44--which seems to index online coordinative adjustments that are continuously required to compensate deviation from synchronisation-- interpretation of right hemisphere involvement should be approached cautiously due to relatively sparse electrode coverage. These findings illustrate the possibility and value of using a fully dynamic, adaptive and interactive language task to gather deeper understanding of the subtending neural dynamics involved in speech perception, production as well as their interaction.

neuroscience↗

bioRxiv2023

Neuronal Avalanches in Naturalistic Speech and Music Listening

Neuronal avalanches are cascade-like events ubiquitously observed across imaging modalities and scales. Aperiodic timing and topographic distribution of these events have been related to the systemic physiology of brain states. However, it is still unknown whether neuronal avalanches are correlates of cognition, or purely reflect physiological properties. In this work, we investigate this question by analyzing intracranial recordings of epileptic participants during rest and passive listening of naturalistic speech and music stimuli. During speech or music listening, but not rest, participants brains "tick" together, as the timing of neuronal avalanches is stimulus-driven and hence correlated across participants. Auditory regions are strongly participating in coordinated neuronal avalanches, but also associative regions, indicating both the specificity and distributivity of cognitive processing. The subnetworks where such processing takes place during speech and music largely overlap, especially in auditory regions, but also diverge in associative cortical sites. Finally, differential pathways of avalanche propagation across auditory and non-auditory regions differentiate brain network dynamics during speech, music and rest. Overall, these results highlight the potential of neuronal avalanches as a neural index of cognition. Authors summaryNeuronal avalanches consist of collective network events propagating across the brain in short-lived and aperiodic instances. These salient events have garnered a great interest for studying the physics of cortical dynamics, and bear potential for studying brain data also in purely neuroscientific contexts. In this work we investigated neuronal avalanches to index cognition, analyzing an intracranial stereo electroencephalography (iEEG) dataset during speech, music listening and resting state in epileptic patients. We show that neuronal avalanches are consistently driven by music and speech stimuli: avalanches co-occur in participants listening to the same auditory stimulus; avalanche topography differs from resting state, presenting partial similarities during speech and music; avalanche propagation changes during speech, music, and rest conditions, especially along the pathways between auditory and non auditory regions. Our work underlines the distributed nature of auditory stimulus processing, supporting neuronal avalanches as a valuable and computationally advantageous framework for the study of cognition in humans.

neuroscience↗

bioRxiv2023

Neural dynamics of predictive timing and motor engagement in music listening

Why do humans spontaneously dance to music? To test the hypothesis that motor dynamics reflect predictive timing during music listening, we built melodies with varying degrees of rhythmic predictability. Magnetoencephalography data showed that while auditory regions track the rhythm of melodies, intrinsic neural dynamics at delta (1.4 Hz) and beta (20-30 Hz) rates in the dorsal auditory pathway embody the experience of groove. Critically, neural dynamics are organized along this pathway in a spectral gradient, with the left sensorimotor cortex acting as a hub coordinating groove-related delta and beta activity. Combined with predictions of a neurodynamic model, this indicate that spontaneous motor engagement during music listening is a manifestation of predictive timing effected by interaction of neural dynamics along the dorsal auditory pathway. One-Sentence SummaryInteracting neural dynamics along the dorsal auditory pathway effect the experience of groove during music listening.

neuroscience↗

bioRxiv2023

Auditory hemispheric asymmetry as a specialization for actions and objects

What is the function of auditory hemispheric asymmetry? We propose that the identification of sound sources relies on two complementary and perceptually relevant acoustic invariants -- actions and objects -- that are processed asymmetrically and sufficient to model and categorize any sound. We observed that environmental sounds are an independent combination of purely temporal and spectral acoustic modulations. Behaviorally, discrimination of actions relied on temporal modulations, while discrimination of objects relied on spectral modulations. Functional magnetic resonance imaging data showed that actions and objects are respectively decoded in the left and right hemispheres, in bilateral superior temporal and left inferior frontal regions. This asymmetry reffects a generic differential processing -- through differential neural sensitivity to temporal and spectral modulations present in all environmental sounds -- that supports the efficient categorization of actions and objects. These results provide an ecologically valid framework of the functional role of auditory brain asymmetry.

neuroscience↗

bioRxiv2022

Speech and music recruit frequency-specific distributed and overlapping cortical networks

To what extent does speech and music processing rely on domain-specific and domain-general neural networks? Using whole-brain intracranial EEG recordings in 18 epilepsy patients listening to natural, continuous speech or music, we investigated the presence of frequency-specific and network-level brain activity. We combined it with a statistical approach in which a clear operational distinction is made between shared, preferred, and domain-selective neural responses. We show that the majority of focal and network-level neural activity is shared between speech and music processing. Our data also reveal an absence of anatomical regional selectivity. Instead, domain-selective neural responses are restricted to distributed and frequency-specific coherent oscillations, typical of spectral fingerprints. Our work highlights the importance of considering natural stimuli and brain dynamics in their full complexity to map cognitive and brain functions.

neuroscience↗

bioRxiv2021

The channel capacity of multilevel linguistic features constrains speech comprehension

Humans are expert at processing speech but how this feat is accomplished remains a major question in cognitive neuroscience. Capitalizing on the concept of channel capacity, we developed a unified measurement framework to investigate the respective influence of seven acoustic and linguistic features on speech comprehension, encompassing acoustic, sub-lexical, lexical and supra-lexical levels of description. We show that comprehension is independently impacted by all these features, but at varying degrees and with a clear dominance of the syllabic rate. Comparing comprehension of French words and sentences further reveals that when supra-lexical contextual information is present, the impact of all other features is dramatically reduced. Finally, we estimated the channel capacity associated with each linguistic feature and compared them with their generic distribution in natural speech. Our data point towards supra-lexical contextual information as the feature limiting the flow of natural speech. Overall, this study reveals how multilevel linguistic features constrain speech comprehension.

neuroscience↗

bioRxiv2021

Stimulating music supports attention in listeners with attentional difficulties

Background music is widely used to sustain attention, but little is known about what musical properties aid attention. This may be due to inter-individual variability in neural responses to music. We test the hypothesis that music can sustain attention by affecting oscillations via acoustic amplitude modulation, differentially for those with varying levels of attentional difficulty. We first show that heavily-modulated music improves sustained attention for participants with more ADHD symptoms. FMRI showed this music elicited greater activity in attentional networks in this group only, and EEG showed greater stimulus-brain coupling for this group in response to the heavily-modulated music. Finally, we parametrically manipulated the depth and rate of amplitude modulations inserted in otherwise-identical music, and found that beta-range modulations helped more than other frequency ranges for participants with more ADHD symptoms. Results suggest the possibility of an oscillation-based neural mechanism for targeted music to support improved cognitive performance.

neuroscience↗

bioRxiv2019

Persistent neural entrainment in the human cortex is frequency selective

Cortical oscillations have been proposed to play a functional role in speech and music perception, attentional selection and working memory, via the mechanism of neural entrainment. One of the most compelling arguments for neural entrainment is that its modulatory effect on ongoing oscillations outlasts rhythmic stimulation. We tested the existence of this phenomenon by studying cortical neural oscillations during and after presentation of melodic stimuli in a passive perception paradigm. Melodies were composed of [~]60 and [~]80 Hz tones embedded in a 2.5 Hz stream. Using intracranial and surface recordings in humans, we reveal consistent neural response properties throughout the cortex, well beyond the auditory regions. Persistent oscillatory activity in the high-gamma band was observed in response to the tones. By contrast, in response to the 2.5 Hz stream, no persistent activity in any frequency band was observed. 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 provides a mechanistic and quantitative explanation of the frequency selectivity of auditory neural entrainment in the human cortex. Significance statementIt has been proposed that the functional role of cortical oscillations is subtended by a mechanism of entrainment, the synchronisation in phase or amplitude of neural oscillations to a periodic stimulation. We tested whether the modulatory effect on ongoing oscillations outlasts the rhythmic stimulation, a phenomenon considered to be one of the most compelling arguments for entrainment. Using intracranial and surface recordings of human listening to rhythmic auditory stimuli, we reveal consistent oscillatory responses throughout the cortex, with persistent activity of high-gamma oscillations. On the contrary, neural oscillations do not outlast low-frequency acoustic dynamics. We interpret our results as reflecting harmonic oscillator properties - a model ubiquitous in physics but rarely used in neuroscience.

neuroscience↗

bioRxiv2019

Attention to rhythms: sensory-specific constrained sampling of temporal regularities

That attention is a fundamentally rhythmic process has recently received abundant empirical evidence. The essence of temporal attention, however, is to flexibly focus in time. Whether this function is hampered by an underlying rhythmic mechanism is unknown. In six interrelated experiments, we behaviourally quantify the sampling capacities of periodic temporal attention during auditory or visual perception. We reveal the presence of limited attentional capacities, with an optimal sampling rate of ~1.4 Hz in audition and ~0.7 Hz in vision. Investigating the motor contribution to temporal attention, we show that it scales with motor rhythmic precision, maximal at ~1.7 Hz. Critically, the motor modulation is beneficial to auditory but detrimental to visual temporal attention. These results are captured by a computational model of coupled oscillators, that reveals the underlying structural constraints governing the temporal alignment between motor and attention fluctuations.

neuroscience↗

bioRxiv2019

Converging intracortical signatures of two separated processing timescales in human early auditory cortex.

Neural oscillations in auditory cortex are argued to support parsing and representing speech constituents at their corresponding temporal scales. Yet, how incoming sensory information interacts with ongoing spontaneous brain activity, what features of the neuronal microcircuitry underlie spontaneous and stimulus-evoked spectral fingerprints, and what these fingerprints entail for stimulus encoding, remain largely open questions. We used a combination of human invasive electrophysiology, computational modeling and decoding techniques to assess the information encoding properties of brain activity and to relate them to a plausible underlying neuronal microarchitecture. We analyzed intracortical auditory EEG activity from 10 patients while they were listening to short sentences. Pre-stimulus neural activity in early auditory cortical regions often exhibited power spectra with a shoulder in the delta range and a small bump in the beta range. Speech decreased power in the beta range, and increased power in the delta-theta and gamma ranges. Using multivariate machine learning techniques, we assessed the spectral profile of information content for two aspects of speech processing: detection and discrimination. We obtained better phase than power information decoding, and a bimodal spectral profile of information content with better decoding at low (delta-theta) and high (gamma) frequencies than at intermediate (beta) frequencies. These experimental data were reproduced by a simple rate model made of two subnetworks with different timescales, each composed of coupled excitatory and inhibitory units, and connected via a negative feedback loop. Modeling and experimental results were similar in terms of pre-stimulus spectral profile (except for the iEEG beta bump), spectral modulations with speech, and spectral profile of information content. Altogether, we provide converging evidence from both univariate spectral analysis and decoding approaches for a dual timescale processing infrastructure in human auditory cortex, and show that it is consistent with the dynamics of a simple rate model.\n\nAuthor summaryLike most animal vocalizations, speech results from a pseudo-rhythmic process that reflects the convergence of motor and auditory neural substrates and the natural resonance properties of the vocal apparatus towards efficient communication. Here, we leverage the excellent temporal and spatial resolution of intracranial EEG to demonstrate that neural activity in human early auditory cortical areas during speech perception exhibits a dual-scale spectral profile of power changes, with speech increasing power in low (delta-theta) and high (gamma - high-gamma) frequency ranges, while decreasing power in intermediate (alpha-beta) frequencies. Single-trial multivariate decoding also resulted in a bimodal spectral profile of information content, with better decoding at low and high frequencies than at intermediate ones. From both spectral and informational perspectives, these patterns are consistent with the activity of a relatively simple computational model comprising two reciprocally connected excitatory/inhibitory sub-networks operating at different (low and high) timescales. By combining experimental, decoding and modeling approaches, we provide consistent evidence for the existence, information coding value and underlying neuronal architecture of dual timescale processing in human auditory cortex.

neuroscience↗