Biology subjects

Hasson, U.

Publications and source records attributed to Hasson, U..

At least 19 recordsLinked to original sources

bioRxiv2026

Content-Sensitive Linguistic Representations in the Human Multiple-Demand Network

Neuroimaging dissociates specialized language regions from the domain-general multiple-demand (MD) network, yet the functional contribution of MD regions to language processing remains unresolved. Because MD recruitment during linguistic tasks is conventionally attributed to domain-general cognitive load, prior research has largely prioritized activation magnitude over representational content. Consequently, it remains unclear whether MD cortices function merely as nonspecific amplifiers of effort or actively encode fine-grained linguistic structures. Here we show that the MD network reliably encodes content-sensitive linguistic representations during naturalistic comprehension independently of peak cognitive demand. Using a voxelwise encoding framework with architecturally identical language models trained on distinct semantic domains (BERT and SciBERT) across three naturalistic fMRI datasets, we find that multiple-demand regions track linguistic structure during both high-demand scientific lectures and low-demand narratives. Furthermore, comparative analyses reveal that these representations are sensitive to domain-specific semantic regularities rather than surface-level form alone. Finally, by evaluating relative model alignment alongside the language network, we demonstrate that the representational balance between these two systems shifts dynamically as a function of task context and domain relevance, revealing a complementary division of labor. Together, these findings reframe the multiple-demand network from a purely extrinsic control mechanism into an active, content-sensitive component of a distributed semantic architecture.

neuroscience↗

bioRxiv2026

The Richness of Experience: Agent-Based Learning Reveals theMechanisms of Early Language Acquisition

A longstanding challenge in developmental science is to understand how children learn language from naturalistic everyday input. To study this process, we leveraged the First 1,000 Days (1kD) dataset, which provides longitudinal, ultra-dense daily audiovisual recordings for individual children in their home environments. This unusually detailed, child-specific record of early experience enabled us to pair each childs rich language input with a cognitively grounded learning agent, linking naturalistic experience ("nurture") to internal learning mechanisms ("nature"). Trained incrementally on each childs input without prior linguistic knowledge, the learning agent discovered speech units corresponding to the English phoneme inventory and acquired thousands of words, closely mirroring individual developmental trajectories. Learning generalized across children while preserving individual differences in rate and timing. Interestingly, learning relied not only on linguistic input but also on the rehearsal of past experiences at the end of each training day. These findings demonstrate that everyday environments provide sufficient structure for language acquisition and establish a unified mechanistic framework for studying development in real-world contexts.

neuroscience↗

bioRxiv2026

The First 1,000 Days (1kD) Project - Collecting and Analyzing an Ultra-Dense Naturalistic Dataset of Human Baby Development

Human development unfolds in continuous, multimodal environments across seconds, days, and years, yet most developmental datasets capture sparse, context-limited samples of everyday life. We introduce the First 1,000 Days (1kD) Project, an initiative designed to collect ultra-dense, longitudinal, child-centered data that capture developmental trajectories within their full ecological context. Fifteen U.S. homes with 17 infants were recorded 12-14 hours per day over a median of 944 days, yielding [~]1.18 million hours of raw audiovisual data. We present an end-to-end framework for large-scale longitudinal naturalistic measurement and a scalable analysis pipeline of the collected data. In a case study, we describe how we utilized our pipeline to isolate child-centered speech, resulting in the collection of 2,000 to 6,000 hours of transcribed speech for each infant. We demonstrate that dense sampling within the home environment reveals a stable, household-specific lexical structure, which sparse sampling methods consistently fail to capture. The 1kD project offers a blueprint for teams aiming to collect and analyze natural behavior at scale in real-world settings.

neuroscience↗

bioRxiv2025

Music-evoked reactivation during continuous perception is associated with enhanced subsequent recall of naturalistic events

Music is a potent cue for recalling personal experiences, yet the neural basis of music-evoked memory remains elusive. We address this question by using the full-length film Eternal Sunshine of the Spotless Mind to examine how repeated musical themes reactivate previously encoded events in cortex and shape next-day recall. Participants in an fMRI study viewed either the original film (with repeated musical themes) or a no-music version. By comparing neural activity patterns between these groups, we found that music-evoked reactivation of neural patterns linked to earlier scenes in the default mode network was associated with improved subsequent recall. This relationship was specific to the music condition and persisted when we controlled for a proxy measure of initial encoding strength (spatial intersubject correlation), suggesting that music-evoked reactivation may play a role in making event memories stick that is distinct from what happens at initial encoding.

neuroscience↗

bioRxiv2025

Cortical language areas are coupled via a soft hierarchy of model-based linguistic features

Natural language comprehension is a complex task that relies on coordinated activity across a network of cortical regions. In this study, we propose that regions of the language network are coupled to one another through subspaces of shared linguistic features. To test this idea, we developed a model-based connectivity framework to quantify stimulus-driven, feature-specific functional connectivity between language areas during natural language comprehension. Using fMRI data acquired while subjects listened to spoken narratives, we tested three types of features extracted from a unified neural network model for speech and language: low-level acoustic embeddings, mid-level speech embeddings, and high-level language embeddings. Our modeling framework enabled us to quantify the proportions of stimulus features driving connectivity between regions: early auditory areas were coupled to intermediate language areas via lower-level acoustic and speech features; in contrast, higher-order language and default-mode regions were predominantly coupled through more abstract language features. We observed a clear progression of feature-specific connectivity from early auditory to lateral temporal areas, advancing from acoustic connectivity to speech- and finally to language-driven connectivity. Our findings suggest that higher-order language areas are coupled along increasingly higher level, more contextualized language features.

neuroscience↗

bioRxiv2025

Linguistic coupling between neural systems for speech production and comprehension during real-time dyadic conversations

The core use of human language is communicating complex ideas from one mind to another in everyday conversations. In conversations, comprehension and production processes are intertwined, as speakers soon become listeners, and listeners become speakers. Nonetheless, the neural systems underlying these faculties are typically studied in isolation, using paradigms that cannot fully engage our capacity for interactive communication, and with indirect measures of similarity. Here, we used an fMRI hyperscanning paradigm to measure neural activity simultaneously in pairs of subjects engaged in real-time, interactive conversations. We used contextual word embeddings from a large language model to quantify the linguistic coupling between production and comprehension systems within and across individual brains. We found a highly overlapping network of regions involved in both production and comprehension spanning much of the cortical language network. Our findings reveal that shared representations for both processes extend beyond the language network into areas associated with social cognition. Together, these results suggest that the specialized neural systems for speech perception and production align on a common set of linguistic features encoded in a broad cortical network for language and communication.

neuroscience↗

bioRxiv2025

The "Podcast" ECoG dataset for modeling neural activity during natural language comprehension

Naturalistic electrocorticography (ECoG) data are a rare but essential resource for studying the brains linguistic capabilities. ECoG offers a high temporal resolution suitable for investigating processes at multiple temporal timescales and frequency bands. It also provides broad spatial coverage, often along critical language areas. Here, we share a dataset of nine ECoG participants with 1,330 electrodes listening to a 30-minute audio podcast. The richness of this naturalistic stimulus can be used for various research endeavors, from auditory perception to semantic integration. In addition to the neural data, we extract linguistic features of the stimulus ranging from phonetic information to large language model word embeddings. We use these linguistic features in encoding models that relate stimulus properties to neural activity. Finally, we provide detailed tutorials for preprocessing raw data, extracting stimulus features, and running encoding analyses that can serve as a pedagogical resource or a springboard for new research.

neuroscience↗

bioRxiv2025

Fast-timescale hippocampal processes bridge between slowly unfurling neocortical states during memory search

Prior behavioral work showed that event structure supports our ability to search through continuous naturalistic memories. We hypothesized that, neurally, this memory search process involves a division of labor between slowly unfurling neocortical states representing event knowledge and fast hippocampal-neocortical communication supporting retrieval of new information at transitions between events. In a sample of ten intracranial EEG patients we tracked slow neural state-patterns as patients viewed a movie and then searched their memories in a structured naturalistic interview. As patients answered questions ("after A, when does B happen next?"), state-patterns from movie-viewing were reinstated in neocortex. During the critical memory-search period, states unfurled in a forward direction, realizing a trajectory from A to B in memory. Punctate moments of state-transition during memory-search were marked by low-frequency power decreases in cortex and preceded by power decreases in hippocampus that correlated with reinstatement. Connectivity-analysis revealed information-flow from hippocampus to cortex underpinning state-transitions.

neuroscience↗

bioRxiv2024

Neuronal correlates of sleep in honey bees

Honey bees Apis mellifera follow the day-night cycle for their foraging activity, entering rest periods during darkness. Despite considerable research on sleep behaviour in bees, its underlying neurophysiological mechanisms are not well understood, partly due to the lack of brain imaging data that allow for analysis from a network- or system-level perspective. This study aims to fill this gap by investigating whether neuronal activity during rest periods exhibits stereotypic patterns comparable to sleep signatures observed in vertebrates. Using two-photon calcium imaging of the antennal lobes (AL) in head-fixed bees, we analysed brain dynamics across motion and rest epochs during the nocturnal period. The recorded activity was computationally characterised, and machine learning was applied to determine whether a classifier could distinguish the two states after motion correction. Out-of-sample classification accuracy reached 93%, and a feature importance analysis suggested network features to be decisive. Accordingly, the glomerular connectivity was found to be significantly increased in the rest-state patterns. A full simulation of the AL using a leaky spiking neural network revealed that such a transition in network connectivity could be achieved by weakly correlated input noise and a reduction of synaptic conductance of the inhibitive local neurons (LNs) which couple the AL network nodes. The difference in the AL response maps between awake- and sleep-like states generated by the simulation showed a decreased specificity of the odour code in the sleep state, suggesting reduced information processing during sleep. Since LNs in the bee brain are GABAergic, this suggests that the GABAergic system plays a central role in sleep regulation in bees as in many higher species including humans. Our findings provide the first evidence that sleep-related network modulation mechanisms may be conserved throughout evolution, highlighting the bees potential as an invertebrate model for studying sleep at the level of single neurons.

neuroscience↗

bioRxiv2024

Information-making processes in the speaker's brain drive human conversations forward

Natural human conversation is driven by the exchange of information-rich messages that surprise the listener and deviate from predictable context. While extensive research has characterized how the brain processes unexpected linguistic input during comprehension, the neural mechanisms underlying the generation of such information by the speaker remain poorly understood. In this study, we hypothesize that the speakers brain actively creates information through extra neural computation, transforming internal thoughts into novel, meaningful linguistic outputs. Utilizing a unique 24/7 dataset of continuous electrocorticography (ECoG) comprising approximately 100 hours of spontaneous natural conversations, we contrast the neural basis of speech production and comprehension within the same participants. Behaviorally, we find that speakers take longer pausing for an additional 100-150 milliseconds before producing information-rich, improbable words, even when controlling for word identity and frequency. Neurally, we provide converging evidence for a previously unreported process: generating information-rich words elicits significantly stronger neural activity (ERP) and enhanced neural encoding and decoding in language-related areas starting 100-500 ms before articulation. This pattern contrasts sharply with speech comprehension, where enhanced neural activity for predictable words occurs before onset, while responses to improbable words emerge only after onset as prediction errors. Furthermore, we demonstrate that large language models (LLMs) mirror this biological process, requiring deeper internal computation across layers to generate improbable versus probable words. Together, these results reveal that the speakers brain is not merely a transmission channel but an active generator of information, recruiting extra neural resources to create novel content that diverges from listener expectations.

neuroscience↗

bioRxiv2024

Scale matters: Large language models with billions (rather than millions) of parameters better match neural representations of natural language

Recent research has used large language models (LLMs) to study the neural basis of naturalistic language processing in the human brain. LLMs have rapidly grown in complexity, leading to improved language processing capabilities. However, neuroscience researchers havent kept up with the quick progress in LLM development. Here, we utilized several families of transformer-based LLMs to investigate the relationship between model size and their ability to capture linguistic information in the human brain. Crucially, a subset of LLMs were trained on a fixed training set, enabling us to dissociate model size from architecture and training set size. We used electrocorticography (ECoG) to measure neural activity in epilepsy patients while they listened to a 30-minute naturalistic audio story. We fit electrode-wise encoding models using contextual embeddings extracted from each hidden layer of the LLMs to predict word-level neural signals. In line with prior work, we found that larger LLMs better capture the structure of natural language and better predict neural activity. We also found a logarithmic relationship where the encoding performance peaks in relatively earlier layers as model size increases. We also observed variations in the best-performing layer across different brain regions, corresponding to an organized language processing hierarchy.

neuroscience↗

bioRxiv2024

ALIGNING BRAINS INTO A SHARED SPACE IMPROVES THEIR ALIGNMENT TO LARGE LANGUAGE MODELS

AO_SCPLOWBSTRACTC_SCPLOWRecent studies have shown that large language models (LLMs) can accurately predict neural activity measured using electrocorticography (ECoG) during natural language processing. To predict word-by-word neural activity, most prior work has estimated and evaluated encoding models within each electrode and subject--without evaluating how these models generalize across individual brains. In this paper, we analyze neural responses in 8 subjects while they listened to the same 30-minute podcast episode. We use a shared response model (SRM) to estimate a shared information space across subjects. We show that SRM significantly improves LLM-based encoding model performance. We also show that we can use this shared space to denoise the individual brain responses by projecting back into the individualized electrode space, and this process achieves a mean 38% improvement in encoding performance. The strongest improvement was observed for brain areas specialized for language comprehension, specifically in the superior temporal gyrus (STG) and inferior frontal gyrus (IFG). Critically, estimating a shared space allows us to construct encoding models that better generalize across individuals.

neuroscience↗

bioRxiv2024

An Incremental Large Language Model for long text processing in the Brain

Accumulated evidence suggests that Large Language Models (LLMs) are beneficial in predicting neural signals related to narrative processing. The way LLMs integrate context over large timescales, however, is fundamentally different from the way the brain does it. In this study, we show that unlike LLMs that apply parallel processing of large contextual windows, the incoming context to the brain is limited to short windows of a few tens of words. We hypothesize that whereas lower-level brain areas process short contextual windows, higher-order areas in the default-mode network (DMN) engage in an online incremental mechanism where the incoming short context is summarized and integrated with information accumulated across long timescales. Consequently, we introduce a novel LLM that instead of processing the entire context at once, it incrementally generates a concise summary of previous information. As predicted, we found that neural activities at the DMN were better predicted by the incremental model, and conversely, lower-level areas were better predicted with short-context-window LLM.

neuroscience↗

bioRxiv2023

How a speaker herds the audience: Multi-brain neural convergence over time during naturalistic storytelling

Storytelling--an ancient way for humans to share individual experiences with others--has been found to induce neural synchronization among listeners. In our exploration of the dynamic fluctuations in listener-listener (LL) coupling throughout stories, we uncover a significant correlation between LL and lag-speaker-listener (lag-SL) couplings over time. Using the analogy of neural pattern (dis)similarity as distances between participants, we term this phenomenon the "herding effect": like a shepherd guiding a group of sheep, the more closely listeners follow the speakers prior brain activity patterns (higher lag-SL similarity), the more tightly they cluster together (higher LL similarity). This herding effect is particularly pronounced in brain regions where neural synchronization among listeners tracks with behavioral ratings of narrative engagement, highlighting the mediating role of narrative content in the observed multi-brain neural coupling dynamics. By integrating LL and SL neural couplings, this study illustrates how unfolding stories shape a dynamic multi-brain functional network and how the configuration of this network may be associated with moment-by-moment efficacy of communication. Significance StatementDifferent stories have been found to evoke distinct brain activation patterns in the audience. This study delves into how the storyteller guides the audience through the multi-dimensional space of brain states, reflected in a series of shared activation patterns. We reveal that the listeners follow along the trajectory outlined by the speakers brain activity moments before, forming a tighter cluster at the more engaging moments of the story. This phenomenon is localized to high-level cortical areas supporting event representation. Our investigation illustrates how storytelling dynamically sculpts multi-brain neural dynamics in both the listeners and the speaker, shedding light on the potential association between the configuration of this network and communication efficacy.

neuroscience↗

bioRxiv2023

A shared linguistic space for transmitting our thoughts from brain to brain in natural conversations

Effective communication hinges on a mutual understanding of word meaning in different contexts. The embedding space learned by large language models can serve as an explicit model of the shared, context-rich meaning space humans use to communicate their thoughts. We recorded brain activity using electrocorticography during spontaneous, face-to-face conversations in five pairs of epilepsy patients. We demonstrate that the linguistic embedding space can capture the linguistic content of word-by-word neural alignment between speaker and listener. Linguistic content emerged in the speakers brain before word articulation, and the same linguistic content rapidly reemerged in the listeners brain after word articulation. These findings establish a computational framework to study how human brains transmit their thoughts to one another in real-world contexts.

neuroscience↗

bioRxiv2023

Deep speech-to-text models capture the neural basis of spontaneous speech in everyday conversations

Humans effortlessly use the continuous acoustics of speech to communicate rich linguistic meaning during everyday conversations. In this study, we leverage 100 hours (half a million words) of spontaneous open-ended conversations and concurrent high-quality neural activity recorded using electrocorticography (ECoG) to decipher the neural basis of real-world speech production and comprehension. Employing a deep multimodal speech-to-text model named Whisper, we develop encoding models capable of accurately predicting neural responses to both acoustic and semantic aspects of speech. Our encoding models achieved high accuracy in predicting neural responses in hundreds of thousands of words across many hours of left-out recordings. We uncover a distributed cortical hierarchy for speech and language processing, with sensory and motor regions encoding acoustic features of speech and higher-level language areas encoding syntactic and semantic information. Many electrodes--including those in both perceptual and motor areas--display mixed selectivity for both speech and linguistic features. Notably, our encoding model reveals a temporal progression from language-to-speech encoding before word onset during speech production and from speech-to-language encoding following word articulation during speech comprehension. This study offers a comprehensive account of the unfolding neural responses during fully natural, unbounded daily conversations. By leveraging a multimodal deep speech recognition model, we highlight the power of deep learning for unraveling the neural mechanisms of language processing in real-world contexts.

neuroscience↗

bioRxiv2023

An intact oculomotor neural circuit in congenital blindness

In the past three decades, multiple studies revealed that congenital blindness is associated with functional and structural reorganization in early visual areas and its interaction with other neural systems. Among the most reproducible findings is the weaker connectivity between the visual and sensorimotor cortices, which in sighted individuals plays a role in eye-motor coordination. Here we demonstrate an important exception to this reorganization phenomena: we find that in congenitally blind individuals (as for normally sighted ones), spontaneous, non-controlled eye movements correlate with connectivity between visual and sensorimotor cortices. Furthermore, using time-lagged regression, we show that eye movements drive activity in the visual cortex, which subsequently drives sensorimotor activity. Remarkably, this phenomenon persists even though blind participants often exhibit unsynchronized eye movements and cannot sense or regulate their eye positions. These findings provide evidence of a preserved, "non-functional" connectivity in congenitally blind individuals and reveal a firm, hard-wired constraint on connectivity that remains immune to any reorganization phenomena produced by lack of visual experience and oculomotor control.

neuroscience↗

bioRxiv2022

Test-Retest Reliability of the Human Connectome: An OPM-MEG study

Magnetoencephalography with optically pumped magnetometers (OPM-MEG) offers a new way to record electrophysiological brain function, with significant advantages over conventional MEG including adaptability to head shape/size, free movement during scanning, better spatial resolution, increased signal, and no reliance on cryogenics. However, OPM-MEG remains in its infancy, with significant questions to be answered regarding optimal system design and robustness. Here, we present an open-source dataset acquired using a newly constructed OPM-MEG system with a triaxial sensor design averaging 168 channels. Using OPM-optimised magnetic shielding and active background-field control, we measure the test-retest reliability of the human connectome. We employ amplitude envelope correlation to measure whole-brain functional connectivity in 10 individuals whilst they watch a 600 s move clip. Our results show high repeatability between experimental runs at the group level, with a correlation coefficient of 0.81 in the theta, 0.93 in alpha and 0.94 in beta frequency ranges. At the individual subject level, we found marked differences between individuals, but high within-subject robustness (correlations of 0.56 {+/-} 0.25, 0.72 {+/-} 0.15 and 0.78 {+/-} 0.13 in theta, alpha and beta respectively). These results compare well to previously reported findings using conventional MEG; they show that OPM-MEG is a viable way to characterise whole brain connectivity and add significant weight to a growing argument that OPMs can overtake cryogenic sensors as the fundamental building block of MEG systems.

neuroscience↗