Uncovering cortical activations of discourse comprehension and their overlaps with common large-scale neural networks

Neural mechanisms of the continued influence effect of misinformation: Analysis based on fMRI causal connectivity

The continued influence effect of misinformation (CIEM) can negatively affect individuals' reasoning and judgment processes. This research aims to enhance the correction of misinformation and foster rational judgement by investigating the internal brain mechanisms involved in the processing of the CIEM through the use of task-based functional magnetic resonance imaging combined with Granger causality analysis. Our findings demonstrate notable effective interactions in varying directions between the left inferior frontal gyrus and middle temporal gyrus during the encoding phase, and between the right anterior cingulate gyrus and left inferior occipital gyrus in the retrieval phase. These insights elucidate the roles of mental model updating and retrieval failure in the processing of CIEM, offering more granular evidence to support the differentiation in processing phases.

Predicting the next sentence (not word) in large language models: What model-brain alignment tells us about discourse comprehension

Current large language models (LLMs) rely on word prediction as their backbone pretraining task. Although word prediction is an important mechanism underlying language processing, human language comprehension occurs at multiple levels, involving the integration of words and sentences to achieve a full understanding of discourse. This study models language comprehension by using the next sentence prediction (NSP) task to investigate mechanisms of discourse-level comprehension. We show that NSP pretraining enhanced a model's alignment with brain data especially in the right hemisphere and in the multiple demand network, highlighting the contributions of nonclassical language regions to high-level language understanding. Our results also suggest that NSP can enable the model to better capture human comprehension performance and to better encode contextual information. Our study demonstrates that the inclusion of diverse learning objectives in a model leads to more human-like representations, and investigating the neurocognitive plausibility of pretraining tasks in LLMs can shed light on outstanding questions in language neuroscience.

How does the "default mode" network contribute to semantic cognition?

This review examines whether and how the "default mode" network (DMN) contributes to semantic processing. We review evidence implicating the DMN in the processing of individual word meanings and in sentence- and discourse-level semantics. Next, we argue that the areas comprising the DMN contribute to semantic processing by coordinating and integrating the simultaneous activity of local neuronal ensembles across multiple unimodal and multimodal cortical regions, creating a transient, global neuronal ensemble. The resulting ensemble implements an integrated simulation of phenomenological experience - that is, an embodied situation model - constructed from various modalities of experiential memory traces. These situation models, we argue, are necessary not only for semantic processing but also for aspects of cognition that are not traditionally considered semantic. Although many aspects of this proposal remain provisional, we believe it provides new insights into the relationships between semantic and non-semantic cognition and into the functions of the DMN.

Alignment of brain embeddings and artificial contextual embeddings in natural language points to common geometric patterns

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. We hypothesize that language areas in the human brain, similar to DLMs, rely on a continuous embedding space to represent language. To test this hypothesis, we densely record the neural activity patterns in the inferior frontal gyrus (IFG) of three participants using dense intracranial arrays while they listened to a 30-minute podcast. From these fine-grained spatiotemporal neural recordings, we derive a continuous vectorial representation for each word (i.e., a brain embedding) in each patient. Using stringent zero-shot mapping we demonstrate that brain embeddings in the IFG and the DLM contextual embedding space have common geometric patterns. The common geometric patterns allow us to predict the brain embedding in IFG of a given left-out word based solely on its geometrical relationship to other non-overlapping words in the podcast. Furthermore, we show that contextual embeddings capture the geometry of IFG embeddings better than static word embeddings. The continuous brain embedding space exposes a vector-based neural code for natural language processing in the human brain.

The Use of Narrative in Science and Health Communication: A Scoping Review

BACKGROUND

Many people deny science and reject health recommendations despite widely distributed facts and statistics. Didactic science and health communication is often dry, and relies on the false assumption that people make purely evidence-based decisions. Stories can be a powerful teaching tool by capturing attention and evoking emotion.

OBJECTIVE

We explore the impact and appeal of, and describe best practices for, using narrative (storytelling) versus didactic methods in science and health communication.

PATIENT INVOLVEMENT

No patients were involved in the review process.

METHODS

We searched PubMed and Web of Science for articles either: assessing effectiveness of narrative science/health communication; assessing acceptability of (or preference for) narrative science/health communication; giving advice on how best to use narrative; and/or providing science-based explanations for how/why narrative succeeds.

RESULTS

Narrative science/health communication is effective and appealing for audiences across a variety of topics and mediums, with supporting evidence across fields such as epidemiology, neuroscience, and psychology. Whether narrative or didactic messaging is most effective depends on the topic, audience, and objective, as well as message quality. However, combining narrative with didactic methods is likely to be more effective than using either strategy alone.

DISCUSSION

Narrative science/health communication merits wider implementation and further research. Narrative communication creates openness to information by delaying the formulation of counterarguments.

PRACTICAL VALUE

Science and health communicators should collaborate with cultural and storytelling experts, work directly with their target audiences throughout the message development and testing processes, and rely on popular story elements (e.g., first-person point of view, relatable protagonists) to improve the comprehension, engagement, and thoughtful consideration of their intended audience.

FUNDING

This work was funded by Thirty Meter Telescope, with which two authors (GKS and SD) were affiliated. Otherwise, the funding organization had no role in the study and/or submission.

Divergent interpersonal neural synchronization patterns in the first, second language and interlingual communication

An accumulating number of studies have highlighted the importance of interpersonal neural synchronization (INS) between interlocutors in successful verbal communications. The opportunities for communication across different language contexts are rapidly expanding, thanks to the frequent interactions among people all over the world. However, whether the INS changes in different language contexts and how language choice affects the INS remain scarcely explored. The study recruited twenty pairs of participants to communicate in the first language (L1), second language (L2) and interlingual contexts. Using functional near-infrared spectroscopy (fNIRS), we examined the neural activities of interlocutors and analyzed their wavelet transform coherence to assess the INS of dyads. Results showed that as compared to the resting state, stronger INS was observed at the left inferior temporal gyrus, middle temporal gyrus, pre-motor and supplementary motor cortex, dorsolateral prefrontal cortex, and inferior frontal gyrus in L1; at the left middle temporal gyrus, superior temporal gyrus, and inferior frontal gyrus in L2; at the left inferior temporal gyrus and inferior frontal gyrus in interlingual context. Additionally, INS at the left inferior frontal gyrus was significantly stronger in L2 than in L1. These findings reveal the differences of the INS in different language contexts and confirm the importance of language choice for the INS changes.

Situation updating during discourse comprehension recruits right posterior portion of the multiple-demand network

Discourse comprehension involves the construction of a mental representation of the situation model as well as a continuous update of this representation. This mental update is cognitively demanding and likely engages the multiple-demand network. However, there is little evidence for the involvement of the multiple-demand network during situation updating. In this study, we used fMRI to test whether situation updating based on the change of spatial location activated the multiple-demand network. In a discourse comprehension task, readers read two-sentence discourses in which the second sentence either continues or introduces a shift of the spatial location information presented in the first sentence. Compared to situation continuation, situation updating reliably activated the right superior parietal lobule. This area is a part of the multiple-demand network as defined by a digit N-back localizer task and locates within the dorsal attention network as defined in the previous study by Yeo et al. in 2011. Our results provide evidence for the reliable involvement of a specific area of the multiple-demand network in situation updating during high-level discourse processing.

Word-producing brain: Contribution of the left anterior middle temporal gyrus to word production patterns in spoken language

Vocabulary is based on semantic knowledge. The anterior temporal lobe (ATL) has been considered an essential region for processing semantic knowledge; nonetheless, the association between word production patterns and the structural and functional characteristics of the ATL remains unclear. To examine this, we analyzed over one million words from group conversations among community-dwelling older adults and their multimodal magnetic resonance imaging data. A quantitative index for the word production patterns, namely the exponent β of Heaps' law, positively correlated with the left anterior middle temporal gyrus volume. Moreover, β negatively correlated with its resting-state functional connectivity with the precuneus. There was no significant correlation with the diffusion tensor imaging metrics in any fiber. These findings suggest that the vocabulary richness in spoken language depends on the brain status characterized by the semantic knowledge-related brain structure and its activation dissimilarity with the precuneus, a core region of the default mode network.

Understanding an implicated causality: The brain network for processing concessive relations

Concessive relations, often indicated by conjunction words such as although, are semantically and pragmatically more complex than causal relations (expressed using because), as they involve more semantic features such as implicated meaning and negation. However, it remains unclear how linguistic-level complexity is manifested through different brain activities and functional connectivities. This fMRI study investigated how the neural underpinnings of concessive relations differ from those of causal relations. Pragmatically congruent and incongruent words were embedded in causal as well as concessive sentences. The whole-brain analysis revealed that relative to because-congruent sentences, although-congruent sentences evoked increased activations in a left network including IFG, bilateral MFG, mPFC, pMTG, and TPJ. DCM analysis showed that while the functional connectivity from IFG to MFG was commonly involved in processing concessive and causal relations, functional connectivities from pMTG to IFG and from pMTG to TPJ were involved in processing causal and concessive relations, respectively.

Neural basis of impaired narrative discourse comprehension in prodromal and mild dementia with lewy bodies

Narrative discourse (ND) comprehension is a complex task that implies not only linguistic abilities but also other cognitive abilities, including efficient executive functioning. An executive dysfunction has been described in dementia with Lewy bodies (DLB) from the early stage. Here, we question the link between executive dysfunction in DLB and narrative comprehension. The aim of our study was to evaluate ND comprehension and to investigate the neuroanatomical basis for its impairment in the early stage of DLB. DLB patients (N = 26) and controls (N = 19) underwent the ND comprehension test of the Montreal Protocol for Evaluation of Communication (MEC). An additional, qualitative analysis was conducted on their verbal productions. Cognitive tests assessing verbal episodic memory, executive functions, naming and oral syntactic comprehension were also performed. Brain gray matter correlates of the ND comprehension test were examined using voxel-based morphometry (VBM). An ND comprehension impairment was found for prodromal and mild DLB patients as compared to controls. These difficulties were correlated with the Frontal Assessment Battery (FAB) score. ND comprehension impairment in DLB was further characterized by a deficit in the organization and the logic of the discourse. Moreover, VBM analysis revealed a correlation between striatal gray matter volumes and DLB patients’ ability to extract and organize relevant information (p < 0.05, FDR correction, cluster level). The ND comprehension impairment in DLB patients could be related to their executive dysfunction through a deficit of information selection and organization that correlates with the volumetric reduction of striatal gray matter.

The longitudinal neural dynamics changes of whole brain connectome during natural recovery from poststroke aphasia

Poststroke aphasia is one of the most dramatic functional deficits that results from direct damage of focal brain regions and dysfunction of large-scale brain networks. The reconstruction of language function depends on the hierarchical whole-brain dynamic reorganization. However, investigations into the longitudinal neural changes of large-scale brain networks for poststroke aphasia remain scarce. Here we characterize large-scale brain dynamics in left-frontal-stroke aphasia through energy landscape analysis. Using fMRI during an auditory comprehension task, we find that aphasia patients suffer serious whole-brain dynamics perturbation in the acute and subacute stages after stroke, in which the brains were restricted into two major activity patterns. Following spontaneous recovery process, the brain flexibility improved in the chronic stage. Critically, we demonstrated that the abnormal neural dynamics are correlated with the aberrant brain network coordination. Taken together, the energy landscape analysis exhibited that the acute poststroke aphasia has a constrained, low dimensional brain dynamics, which were replaced by less constrained and high dimensional dynamics at chronic aphasia. Our study provides a new perspective to profoundly understand the pathological mechanisms of poststroke aphasia.

Deep neural networks reveal topic-level representations of sentences in medial prefrontal cortex, lateral anterior temporal lobe, precuneus, and angular gyrus

When reading a sentence, individual words can be combined to create more complex meaning. In this study, we sought to uncover brain regions that reflect the representation of the meaning of sentences at the topic level, as opposed to the meaning of their individual constituent words when considered irrespective of their context. Using fMRI, we recorded the neural activity of participants while reading sentences. We constructed a topic-level sentence representations using the final layer of a convolutional neural network (CNN) trained to classify Wikipedia sentences into broad semantic categories. This model was contrasted with word-level sentence representations constructed using the average of the word embeddings constituting the sentence. Using representational similarity analysis, we found that the medial prefrontal cortex, lateral anterior temporal lobe, precuneus, and angular gyrus more strongly represent sentence topic-level, compared to word-level, meaning, uncovering the important role of these semantic system regions in the representation of topic-level meaning. Results were comparable when sentence meaning was modelled with a multilayer perceptron that was not sensitive to word order within a sentence, suggesting that the learning objective, in the terms of the topic being modelled, is the critical factor in capturing these neural representational spaces.

Perceptual coupling and decoupling of the default mode network during mind-wandering and reading

While reading, our mind can wander to unrelated autobiographical information, creating a perceptually decoupled state detrimental to narrative comprehension. To understand how this mind-wandering state emerges, we asked whether retrieving autobiographical content necessitates functional disengagement from visual input. In Experiment 1, brain activity was recorded using functional magnetic resonance imaging (fMRI) in an experimental situation mimicking naturally occurring mind-wandering, allowing us to precisely delineate neural regions involved in memory and reading. Individuals read expository texts and ignored personally relevant autobiographical memories, as well as the opposite situation. Medial regions of the default mode network (DMN) were recruited during memory retrieval. In contrast, left temporal and lateral prefrontal regions of the DMN, as well as ventral visual cortex, were recruited when reading for comprehension. Experiment two used functional connectivity both at rest and during tasks to establish that (i) DMN regions linked to memory are more functionally decoupled from regions of ventral visual cortex than regions in the same network engaged when reading; and (ii) individuals with more self-generated mental contents and poorer comprehension, while reading in the lab, showed more decoupling between visually connected DMN sites important for reading and primary visual cortex. A similar pattern of connectivity was found in Experiment 1, with greater coupling between this DMN site and visual cortex when participants reported greater focus on reading in the face of conflict from autobiographical memory cues; moreover, the retrieval of personally relevant memories increased the decoupling of these sites. These converging data suggest we lose track of the narrative when our minds wander because generating autobiographical mental content relies on cortical regions within the DMN which are functionally decoupled from ventral visual regions engaged during reading.

As your eyes scan these words, you may be thinking about what to make for dinner, how to address an unexpected hurdle at work, or how many emails are sitting, unread, in your inbox. This type of mind-wandering disrupts our focus and limits how much information we comprehend, whilst also being conducive to creative thinking and problem-solving.

Despite being an everyday occurrence, exactly how our mind wanders remains elusive. One possible explanation is that the brain disengages from visual information from the external world and turns its attention inwards. A greater understanding of which neural circuits are involved in this process could reveal insights about focus, attention, and reading comprehension.

Here, Zhang et al. investigated whether the brain becomes disengaged from visual input when our mind wanders while reading. Recalling personal events was used as a proxy for mind-wandering. Brain activity was recorded as participants were shown written statements; sometimes these were preceded by cues to personal memories. People were asked to focus on reading the statements or they were instructed to concentrate on their memories while ignoring the text. The analyses showed that recalling memories and reading stimulated distinct parts of the brain, which were in direct competition during mind-wandering.

Further work examined how these regions were functionally connected. In individuals who remained focused on reading despite memory cues, the areas activated by reading showed strong links to the visual cortex. Conversely, these reading-related areas became ‘decoupled’ from visual processing centres in people who were focusing more on their internal thoughts.

These results shed light on why we lose track of what we are reading when our mind wanders: recalling personal memories activates certain brain areas which are functionally decoupled from the regions involved in processing external information – such as the words on a page.

In summary, the work by Zhang et al. builds a mechanistic understanding of mind-wandering, a natural feature of our daily brain activity. These insights may help to inform future interventions in education to improve reading, comprehension and focus.

Factors inducing complexities in musical embedded structure processing

The organizational structure of music is similar to that found in language, involving a large number of complicated hierarchical and embedded structures. The factors inducing complexities and difficulties in embedded structure processing are important subjects of inquiry in areas of cognitive neuroscience, such as music and language domains. Enlightened by relevant linguistic theories, this study investigated the influence of dependency lengthening and structural shift on musical embedded sequences processing. Results showed that final chords in sequences with long dependence elicited larger ERAN and N5 under near-key shift conditions, while elicited larger ERAN and LPC under far-key shift conditions, when compared to the sequences with short dependence; Further, the final chords in sequences with far-key shift elicited larger N5 under short dependence conditions, while elicited larger LPC under long dependence conditions when compared to the sequences with near-key shift. These results indicate that both dependency lengthening and structure shift could be the factors inducing complexities and difficulties in the processing of musical embedded structures, and there might be some common mechanisms underlying the processing of center-embedded structure across music and language domains.

The brain network in support of social semantic accumulation

Some studies have indicated that a specific 'social semantic network' represents the social meanings of words. However, studies of the comprehension of complex materials, such as sentences and narratives, have indicated that the same network supports the online accumulation of connected semantic information. In this study, we examined the hypothesis that this network does not simply represent the social meanings of words but also accumulates connected social meanings from texts. We defined the social semantic network by conducting a meta-analysis of previous studies on social semantic processing and then examined the effects of social semantic accumulation using a functional Magnetic Resonance Imaging (fMRI) experiment. Two important findings were obtained. First, the social semantic network showed a stronger social semantic effect in sentence and narrative reading than in word list reading, indicating the amplitude of social semantic activation can be accumulated in the network. Second, the activation of the social semantic network in sentence and narrative reading can be better explained by the holistic social-semantic-richness rating scores of the stimuli than by those of the constitutive words, indicating the social semantic contents can be integrated in the network. These two findings convergently indicate that the social semantic network supports the accumulation of connected social meanings.

Neural Correlates of Causal Inferences in Discourse Understanding and Logical Problem-Solving: A Meta-Analysis Study

In discourse comprehension, we need to draw inferences to make sense of discourse. Previous neuroimaging studies have investigated the neural correlates of causal inferences in discourse understanding. However, these findings have been divergent, and how these types of inferences are related to causal inferences in logical problem-solving remains unclear. Using the activation likelihood estimation (ALE) approach, the current meta-analysis analyzed 19 experiments on causal inferences in discourse understanding and 20 experiments on those in logical problem-solving to identify the neural correlates of these two cognitive processes and their shared and distinct neural correlates. We found that causal inferences in discourse comprehension recruited a left-lateralized frontotemporal brain system, including the left inferior frontal gyrus, the left middle temporal gyrus (MTG), and the bilateral medial prefrontal cortex (MPFC), while causal inferences in logical problem-solving engaged a nonoverlapping brain system in the frontal and parietal cortex, including the left inferior frontal gyrus, the bilateral middle frontal gyri, the dorsal MPFC, and the left inferior parietal lobule (IPL). Furthermore, the pattern similarity analyses showed that causal inferences in discourse understanding were primarily related to the terms about language processing and theory-of-mind processing. Both types of inferences were found to be related to the terms about memory and executive function. These findings suggest that causal inferences in discourse understanding recruit distinct neural bases from those in logical problem-solving and rely more on semantic knowledge and social interaction experiences.

The Role of the Dorsolateral Prefrontal Cortex for Speech and Language Processing

This review article summarizes various functions of the dorsolateral prefrontal cortex (DLPFC) that are related to language processing. To this end, its connectivity with the left-dominant perisylvian language network was considered, as well as its interaction with other functional networks that, directly or indirectly, contribute to language processing. Language-related functions of the DLPFC comprise various aspects of pragmatic processing such as discourse management, integration of prosody, interpretation of nonliteral meanings, inference making, ambiguity resolution, and error repair. Neurophysiologically, the DLPFC seems to be a key region for implementing functional connectivity between the language network and other functional networks, including cortico-cortical as well as subcortical circuits. Considering clinical aspects, damage to the DLPFC causes psychiatric communication deficits rather than typical aphasic language syndromes. Although the number of well-controlled studies on DLPFC language functions is still limited, the DLPFC might be an important target region for the treatment of pragmatic language disorders.

Constructing and evaluating a cortical surface atlas and analyzing cortical sex differences in young Chinese adults

Cortical surface templates are an important standardized coordinate frame for cortical structure and function analysis in magnetic resonance (MR) imaging studies. The widely used adult cortical surface templates (e.g., fsaverage, Conte69, and the HCP-MMP atlas) are based on the Caucasian population. Neuroanatomical differences related to environmental and genetic factors between Chinese and Caucasian populations make these templates unideal for analysis of the cortex in the Chinese population. We used a multimodal surface matching algorithm in an iterative procedure to create Chinese (sCN200) and Caucasian (sUS200) cortical surface atlases based on 200 demographically matched high-quality T1- and T2-weighted (T1w and T2w, respectively) MR images from the Chinese Human Connectome Project (CHCP) and the Human Connectome Project (HCP), respectively. Templates for anatomical cortical surfaces (white matter, pial, midthickness) and cortical feature maps of sulcal depth, curvature, thickness, T1w/T2w myelin, and cortical labels were generated. Using independent subsets from the CHCP and the HCP, we quantified the accuracy of cortical registration when using population-matched and mismatched atlases. The performance of the cortical registration and accuracy of curvature alignment when using population-matched atlases was significantly improved, thereby demonstrating the importance of using the sCN200 cortical surface atlas for Chinese adult population studies. Finally, we analyzed female and male cortical differences within the Chinese and Caucasian populations. We identified significant between-sex differences in cortical curvature, sulcal depth, thickness, and T1w/T2w myelin maps in the frontal, temporal, parietal, occipital, and insular lobes as well as the cingulate cortices.