Biology subjects

Hasson, U.

Publications and source records attributed to Hasson, U..

At least 37 records · Page 2Linked to original sources

bioRxiv2022

Correspondence between the layered structure of deep language models and temporal structure of natural language processing in the human brain

Deep language models (DLMs) provide a novel computational paradigm for how the brain processes natural language. Unlike symbolic, rule-based models described in psycholinguistics, DLMs encode words and their context as continuous numerical vectors. These "embeddings" are constructed by a sequence of computations organized in "layers" to ultimately capture surprisingly sophisticated representations of linguistic structures. How does this layered hierarchy map onto the human brain during natural language comprehension? In this study, we used electrocorticography (ECoG) to record neural activity in language areas along the superior temporal gyrus and inferior frontal gyrus while human participants listened to a 30-minute spoken narrative. We supplied this same narrative to a high-performing DLM (GPT2-XL) and extracted the contextual embeddings for each word in the story across all 48 layers of the model. We next trained a set of linear encoding models to predict the temporally-evolving neural activity from the embeddings at each layer. We found a striking correspondence between the layer-by-layer sequence of embeddings from GPT2-XL and the temporal sequence of neural activity in language areas. In addition, we found evidence for the gradual accumulation of recurrent information along the linguistic processing hierarchy. However, we also noticed additional neural processes in the brain, but not in DLMs, during the processing of surprising (unpredictable) words. These findings point to a connection between human language processing and DLMs where the layer-by-layer accumulation of contextual information in DLM embeddings matches the temporal dynamics of neural activity in high-order language areas.

neuroscience↗

bioRxiv2022

Reconstructing the cascade of language processing in the brain using the internal computations of a transformer-based language model

Humans use complex linguistic structures to transmit ideas to one another. The brain is thought to deploy specialized computations to process these structures. Recently, a new class of artificial neural networks based on the Transformer architecture has revolutionized the field of language modeling, attracting attention from neuroscientists seeking to understand the neurobiology of language in silico. Transformers integrate information across words via multiple layers of structured circuit computations, forming increasingly contextualized representations of linguistic content. Prior work has focused on the internal representations (the "embeddings") generated by these circuits. In this paper, we instead analyze the circuit computations directly: we deconstruct these computations into functionally-specialized "transformations" to provide a complementary window onto linguistic computations in the human brain. Using functional MRI data acquired while participants listened to naturalistic spoken stories, we first verify that the transformations account for considerable variance in brain activity across the cortical language network. We then demonstrate that the emergent syntactic computations performed by individual, functionally-specialized "attention heads" differentially predict brain activity in specific cortical regions. These heads fall along gradients corresponding to different layers, contextual distances, and syntactic dependencies in a low-dimensional cortical space. Our findings indicate that large language models and the cortical language network may converge on similar trends of functional specialization for processing natural language.

neuroscience↗

bioRxiv2022

Brain embeddings with shared geometry to artificial contextual embeddings, as a code for representing language in the human brain

Contextual embeddings, derived from deep language models (DLMs), provide a continuous vectorial representation of language. This embedding space differs fundamentally from the symbolic representations posited by traditional psycholinguistics. Do language areas in the human brain, similar to DLMs, rely on a continuous embedding space to represent language? To test this hypothesis, we densely recorded the neural activity in the Inferior Frontal Gyrus (IFG, also known as Brocas area) of three participants using dense intracranial arrays while they listened to a 30-minute podcast. From these fine-grained spatiotemporal neural recordings, we derived for each patient a continuous vectorial representation for each word (i.e., a brain embedding). Using stringent, zero-shot mapping, we demonstrated that brain embeddings in the IFG and the DLM contextual embedding space have strikingly similar geometry. This shared geometry allows us to precisely triangulate the position of unseen words in both the brain embedding space (zero-shot encoding) and the DLM contextual embedding space (zero-shot decoding). The continuous brain embedding space provides an alternative computational framework for how natural language is represented in cortical language areas.

neuroscience↗

bioRxiv2022

Real-time neurofeedback to alter interpretations of a naturalistic narrative

We explored the potential of using real-time fMRI (rt-fMRI) neurofeedback training to bias interpretations of naturalistic narrative stimuli. Participants were randomly assigned to one of two possible conditions, each corresponding to a different interpretation of an ambiguous spoken story. While participants listened to the story in the scanner, neurofeedback was used to reward neural activity corresponding to the assigned interpretation. After scanning, final interpretations were assessed. While neurofeedback did not change story interpretations on average, participants with higher levels of decoding accuracy during the neurofeedback procedure were more likely to adopt the assigned interpretation. Thus, we believe that individualized neurofeedback shaped interpretations successfully when the signal was accurate, although more work is needed to improve this method and validate the result. While naturalistic stimuli introduce a unique set of challenges in providing effective and individualized neurofeedback, we believe that this technique holds promise for individualized cognitive therapy.

neuroscience↗

bioRxiv2021

Information flow across the cortical timescales hierarchy during narrative comprehension

When listening to spoken narratives, we must integrate information over multiple, concurrent timescales, building up from words to sentences to paragraphs to a coherent narrative. Recent evidence suggests that the brain relies on a chain of hierarchically organized areas with increasing temporal receptive windows to process naturalistic narratives. We hypothesized that the structure of this cortical processing hierarchy should result in an observable sequence of response lags between networks comprising the hierarchy during narrative comprehension. This study uses functional MRI to estimate the response lags between functional networks during narrative comprehension. We use inter-subject cross-correlation analysis to capture network connectivity driven by the shared stimulus. We found a fixed temporal sequence of response lags--on the scale of several seconds--starting in early auditory areas, followed by language areas, the attention network, and lastly the default mode network. This gradient is consistent across eight distinct stories but absent in data acquired during rest or using a scrambled story stimulus, supporting our hypothesis that narrative construction gives rise to inter-network lags. Finally, we build a simple computational model for the neural dynamics underlying the construction of nested narrative features. Our simulations illustrate how the gradual accumulation of information within the boundaries of nested linguistic events, accompanied by increased activity at each level of the processing hierarchy, can give rise to the observed lag gradient. Significance StatementOur findings reveal a consistent, stimulus-driven gradient of lags in connectivity along the cortical processing hierarchy--from early auditory cortex to the language network, then to the default mode network--during the comprehension of naturalistic, spoken narratives. We provide a simple computational model for the neural dynamics underlying the construction of nested narrative features, allowing us to systematically explore the conditions under which the lag gradient emerges and synthesize our results with previous findings based on simple well-controlled language stimuli. Our results illustrate the isomorphism between hierarchically structured neural dynamics and hierarchically structured, real-world narrative inputs.

neuroscience↗

bioRxiv2021

Fine-tuning of deep language models as a computational framework of modeling listeners' perspective during language comprehension

Computational Deep Language Models (DLMs) have been shown to be effective in predicting neural responses during natural language processing. This study introduces a novel computational framework, based on the concept of fine-tuning (Hinton, 2007), for modeling differences in interpretation of narratives based on the listeners perspective (i.e. their prior knowledge, thoughts, and beliefs). We draw on an fMRI experiment conducted by Yeshurun et al. (2017), in which two groups of listeners were listening to the same narrative but with two different perspectives (cheating versus paranoia). We collected a dedicated dataset of ~3000 stories, and used it to create two modified (fine-tuned) versions of a pre-trained DLM, each representing the perspective of a different group of listeners. Information extracted from each of the two fine-tuned models was better fitted with neural responses of the corresponding group of listeners. Furthermore, we show that the degree of difference between the listeners interpretation of the story - as measured both neurally and behaviorally - can be approximated using the distances between the representations of the story extracted from these two fine-tuned models. These models-brain associations were expressed in many language-related brain areas, as well as in several higher-order areas related to the default-mode and the mentalizing networks, therefore implying that computational fine-tuning reliably captures relevant aspects of human language comprehension across different levels of cognitive processing.

neuroscience↗

bioRxiv2021

Here's the twist: How the brain updates naturalistic event memories as our understanding of the past changes

The brain actively reshapes our understanding of past events in light of new incoming information. In the current study, we ask how the brain supports this updating process during the encoding and recall of naturalistic stimuli. One group of participants watched a movie ("The Sixth Sense") with a cinematic "twist" at the end that dramatically changed the interpretation of previous events. Next, participants were asked to verbally recall the movie events, taking into account the new "twist" information. Most participants updated their recall to incorporate the twist. Two additional groups recalled the movie without having to update their memories during recall: one group never saw the twist; another group was exposed to the twist prior to the beginning of the movie, and thus the twist information was incorporated both during encoding and recall. We found that providing participants with information about the twist beforehand altered neural response patterns during movie-viewing in the default mode network (DMN). Moreover, presenting participants with the twist at the end of the movie changed the neural representation of the previously-encoded information during recall in a subset of DMN regions. Further evidence for this transformation was obtained by comparing the neural activation patterns during encoding and recall and correlating them with behavioral signatures of memory updating. Our results demonstrate that neural representations of past events encoded in the DMN are dynamically integrated with new information that reshapes our understanding in natural contexts.

neuroscience↗

bioRxiv2021

Improved prediction of behavioral and neural similarity spaces using pruned DNNs

Deep Neural Networks (DNNs) have become an important tool for modeling brain and behaviour. One key area of interest has been to apply these networks to model human similarity judgements. Several previous works have used the embeddings from the penultimate layer of vision DNNs and showed that a reweighting of these features improves the fit between human similarity judgments and DNNs. These studies underline the idea that these embeddings form a good basis set but lack the correct level of salience. Here we re-examined the grounds for this idea and on the contrary, we hypothesized that these embeddings, beyond forming a good basis set, also have the correct level of salience to account for similarity judgments. It is just that the huge dimensional embedding needs to be pruned to select those features relevant for the considered domain for which a similarity space is modeled. In Study 1 we supervised DNN pruning based on a subset of human similarity judgments. We found that pruning: i) improved out-of-sample prediction of human similarity judgments from DNN embeddings, ii) produced better alignment with WordNet hierarchy, and iii) retained much higher classification accuracy than reweighting. Study 2 showed that pruning by neurobiological data is highly effective in improving out-of-sample prediction of brain-derived representational dissimilarity matrices from DNN embeddings, at times fleshing out isomorphisms not otherwise observable. Pruning supervised by human brain/behavior therefore effectively identifies alignable dimensions of semantic knowledge between DNNs and humans and constitutes an effective method for understanding the organization of knowledge in neural networks.

neuroscience↗

bioRxiv2021

High-order areas and auditory cortex both represent the high-level event structure of music

Recent fMRI studies of event segmentation have found that default mode regions represent high-level event structure during movie watching. In these regions, neural patterns are relatively stable during events and shift at event boundaries. Music, like narratives, contains hierarchical event structure (e.g., sections are composed of phrases). Here, we tested the hypothesis that brain activity patterns in default mode regions reflect the high-level event structure of music. We used fMRI to record brain activity from 25 participants (male and female) as they listened to a continuous playlist of 16 musical excerpts, and additionally collected annotations for these excerpts by asking a separate group of participants to mark when meaningful changes occurred in each one. We then identified temporal boundaries between stable patterns of brain activity using a hidden Markov model and compared the location of the model boundaries to the location of the human annotations. We identified multiple brain regions with significant matches to the observer-identified boundaries, including auditory cortex, medial prefrontal cortex, parietal cortex, and angular gyrus. From these results, we conclude that both higher-order and sensory areas contain information relating to the high-level event structure of music. Moreover, the higher-order areas in this study overlap with areas found in previous studies of event perception in movies and audio narratives, including regions in the default mode network. Significance StatementListening to music requires the brain to track dynamics at multiple hierarchical timescales. In our study, we had fMRI participants listen to real-world music (classical and jazz pieces) and then used an unsupervised learning algorithm (a hidden Markov model) to model the high-level event structure of music within participants brain data. This approach revealed that default mode brain regions involved in representing the high-level event structure of narratives are also involved in representing the high-level event structure of music. These findings provide converging support for the hypothesis that these regions play a domain-general role in processing events occurring over long timescales.

neuroscience↗

bioRxiv2020

Narratives: fMRI data for evaluating models of naturalistic language comprehension

The "Narratives" collection aggregates a variety of functional MRI datasets collected while human subjects listened to naturalistic spoken stories. The current release includes 345 subjects, 891 functional scans, and 27 diverse stories of varying duration totaling ~4.6 hours of unique stimuli (~43,000 words). This data collection is well-suited for naturalistic neuroimaging analysis, and is intended to serve as a benchmark for models of language and narrative comprehension. We provide standardized MRI data accompanied by rich metadata, preprocessed versions of the data ready for immediate use, and the spoken story stimuli with time-stamped phoneme- and word-level transcripts. All code and data are publicly available with full provenance in keeping with current best practices in transparent and reproducible neuroimaging.

neuroscience↗

bioRxiv2020

Learning to use episodic memory for event prediction

Recent human behavioral and neuroimaging results suggest that people are selective in when they encode and retrieve episodic memories. To explain these findings, we trained a memory-augmented neural network to use its episodic memory to support prediction of upcoming states in an environment where past situations sometimes reoccur. We found that the network learned to retrieve selectively as a function of several factors, including its uncertainty about the upcoming state. Additionally, we found that selectively encoding episodic memories at the end of an event (but not mid-event) led to better subsequent prediction performance. In all of these cases, the benefits of selective retrieval and encoding can be explained in terms of reducing the risk of retrieving irrelevant memories. Overall, these modeling results provide a resource-rational account of why episodic retrieval and encoding should be selective and lead to several testable predictions.

neuroscience↗

bioRxiv2020

Moment-by-moment tracking of naturalistic learning and its underlying hippocampo-cortical interactions

Every day our memory system achieves a remarkable feat: We form lasting memories of stimuli that were only encountered once. Here we investigate such learning as it naturally occurs during story listening, with the goal of uncovering when and how memories are stored and retrieved during processing of continuous, naturalistic stimuli. In behavioral experiments we confirm that, after a single exposure to a naturalistic story, participants can learn about its structure and are able to recall upcoming words in the story. In patients undergoing electrocorticographic recordings, we then track mnemonic information in high frequency activity (70 - 200Hz) as patients listen to a story twice. In auditory processing regions we demonstrate the rapid reinstatement of upcoming information after a single exposure; this neural measure of predictive recall correlates with behavioral measures of event segmentation and learning. Connectivity analyses on the neural data reveal information-flow from cortex to hippocampus at the end of events. On the second time of listening information-flow from hippocampus to cortex precedes moments of successful reinstatement.

neuroscience↗

bioRxiv2020

Thinking ahead: prediction in context as a keystone of language in humans and machines

Departing from traditional linguistic models, advances in deep learning have resulted in a new type of predictive (autoregressive) deep language models (DLMs). Using a self-supervised next-word prediction task, these models are trained to generate appropriate linguistic responses in a given context. We provide empirical evidence that the human brain and autoregressive DLMs share three fundamental computational principles as they process natural language: 1) both are engaged in continuous next-word prediction before word-onset; 2) both match their pre-onset predictions to the incoming word to calculate post-onset surprise (i.e., prediction error signals); 3) both represent words as a function of the previous context. In support of these three principles, our findings indicate that: a) the neural activity before word-onset contains context-dependent predictive information about forthcoming words, even hundreds of milliseconds before the words are perceived; b) the neural activity after word-onset reflects the surprise level and prediction error; and c) autoregressive DLM contextual embeddings capture the neural representation of context-specific word meaning better than arbitrary or static semantic embeddings. Together, our findings suggest that autoregressive DLMs provide a novel and biologically feasible computational framework for studying the neural basis of language.

neuroscience↗

bioRxiv2020

Spontaneous eye-movements during eyes-open rest reduce resting-state-network modularity by increasing visual-sensorimotor connectivity

During wakeful rest, individuals make small eye movements when asked to fixate. We examined how these endogenously-driven oculomotor patterns impact topography and topology of functional brain networks. We used a dataset consisting of eyes-open resting-state (RS) fMRI data with simultaneous eye-tracking (Nilsonne et al., 2016). The eye-tracking data indicated minor movements during rest, on the order of 1.0 degree on average when analyzed over 2sec epochs, which correlated modestly with RS BOLD data. However, the eye-tracking data correlated well with echo-planar imaging (EPI) time series sampled from the area of the Eye-Orbit (EO-EPI), which is a signal previously used to identify eye movements during exogenous saccades and movie viewing. We found that EO-EPI data correlated with activity in an extensive motor and sensory-motor network, but also some components of the dorsal attention network including the frontal and supplementary eye fields. Partialling out variance related to EO-EPI from RS data reduced connectivity, primarily between sensory-motor and visual areas. For three different network sparsity levels, the resulting RS connectivity networks showed higher modularity, lower mean connectivity strength, and lower mean clustering coefficient. Our results highlight new aspects of endogenous eye movement control during wakeful rest. They show that oculomotor-related contributions form an important component of RS network topology, and that those should be considered in interpreting differences in network structure between populations, or as a function of different experimental conditions.

neuroscience↗

bioRxiv2020

Teacher-student neural coupling during teaching and learning

Human communication is remarkably versatile, enabling teachers to share highly abstracted and novel information with their students. What neural processes enable such transfer of information across brains during naturalistic teaching and learning? Here, we show that during lectures, wherein information transmission is unidirectional and flows from the teacher to the student, the students brain mirrors the teachers brain and that this neural coupling is correlated with learning outcomes. A teacher was scanned in fMRI giving an oral lecture with slides on a scientific topic followed by a review lecture. Students were then scanned watching either the intact lecture and review (N = 20) or a temporally scrambled version of the lecture (N = 20). Using intersubject correlation (ISC), we observed widespread teacher-student neural coupling spanning sensory cortex and language regions along the superior temporal sulcus as well as higher-level regions including posterior medial cortex (PMC), superior parietal lobule (SPL), and dorsolateral and dorsomedial prefrontal cortex. Teacher-student alignment in higher-level areas was not observed when learning was disrupted by temporally scrambling the lecture. Moreover, teacher-student coupling in PMC was significantly correlated with learning outcomes: the more closely the students brain mirrored the teachers brain, the more the student improved between behavioral pre-learning and post-learning assessments. Together, these results suggest that the alignment of neural responses between teacher and students may underlie effective communication of complex information across brains in classroom settings. Significance statementHow is technical, non-narrative information communicated from one brain to another during teaching and learning? In this fMRI study, we show that the DMN activity of teachers and students are coupled during naturalistic teaching. This teacher-student neural coupling emerges only during intact learning and is correlated with learning outcomes. Together, these findings suggest that teacher-student neural alignment underlies effective communication during teaching.

neuroscience↗

bioRxiv2020

Think Like an Expert: Neural Alignment Predicts Understanding in Students Taking an Introduction to Computer Science Course

How do students understand and remember new information? Despite major advances in measuring human brain activity during and after educational experiences, it is unclear how learners internalize new content, especially in real-life and online settings. In this work, we introduce a neural measure for predicting and assessing learning outcomes. Our approach hinges on the idea that successful learning involves forming the "right" set of neural representations, which are captured in "canonical" activity patterns shared across individuals. Specifically, we hypothesized that understanding is mirrored in "neural alignment": the degree to which an individual learners neural representations match those of experts, as well as those of other learners. We tested this hypothesis in a longitudinal functional MRI study that regularly scanned college students enrolled in an introduction to computer science course. We additionally scanned graduate student "experts" in computer science. We found that alignment among students successfully predicted overall performance in a final exam. Furthermore, within individual students, concepts that evoked better alignment with the experts and with their fellow students were better understood, revealing neural patterns associated with understanding specific concepts. These results provide support for a novel neural measure of concept understanding that can be used to assess and predict learning outcomes in real-life contexts.

neuroscience↗

bioRxiv2020

Behavioral, physiological, and neural signatures of surprise during naturalistic sports viewing

Surprise signals a discrepancy between past and current beliefs. It is theorized to be linked to affective experiences, the creation of particularly resilient memories, and segmentation of the flow of experience into discrete perceived events. However, the ability to precisely measure naturalistic surprise has remained elusive. We used advanced basketball analytics to derive a quantitative measure of surprise and characterized its behavioral, physiological, and neural correlates in human subjects observing basketball games. We found that surprise was associated with segmentation of ongoing experiences, as reflected by subjectively perceived event boundaries and shifts in neocortical patterns underlying belief states. Interestingly, these effects differed by whether surprising moments contradicted or bolstered current predominant beliefs. Surprise also positively correlated with pupil dilation, activation in subcortical regions associated with dopamine, game enjoyment, and long-term memory. These investigations support key predictions from event segmentation theory and extend theoretical conceptualizations of surprise to real-world contexts.

neuroscience↗

bioRxiv2020

Relating the past with the present: Information integration and segregation during ongoing narrative processing

This study examined how the brain dynamically updates event representations by integrating new information over multiple minutes while segregating irrelevant input. A professional writer custom-designed a narrative with two independent storylines, interleaving across minute-long segments (ABAB). In the last (C) part, characters from the two storylines meet and their shared history is revealed. Part C is designed to induce the spontaneous recall of past events, upon the recurrence of narrative motifs from A/B, and to shed new light on them. Our fMRI results showed storyline-specific neural patterns, which were reinstated (i.e. became more active) during storyline transitions. This effect increased along the processing timescale hierarchy, peaking in the default mode network. Similarly, the neural reinstatement of motifs was found during part C. Furthermore, participants showing stronger motif reinstatement performed better in integrating A/B and C events, demonstrating the role of memory reactivation in information integration over intervening irrelevant events.

neuroscience↗