Language structure in the brain: A fixation-related fMRI study of syntactic surprisal in reading

fMRI evidence that left posterior temporal cortex contributes to N400 effects of predictability independent of congruity

Previous electrophysiological work argues that predictability and semantic incongruity rapidly impact comprehension, as indicated by modulation of the N400 component between ~300 and 500 ms. An ongoing question is whether effects of predictability in fact reflect pre-activation in long-term memory as opposed to modulating the kind of integration processes triggered by incongruity. Using fMRI, we compared the impact of predictability and incongruity in adjective-noun phrases, in regions identified with lexical and phrasal localizer scans. We found that predictability impacted activity in left posterior middle temporal gyrus (pMTG), while incongruity impacted activity in left precentral gyrus. Together with parallel data from ERP, these data are consistent with the hypothesis that left pMTG activity is a key contributor to N400 effects of predictability and that the relevant mechanism is reduced activation of stored lexical representations. We tentatively suggest that the left precentral region may play a role in reanalysis when incongruity is encountered.

Shared neural resources of rhythm and syntax: An ALE meta-analysis

A growing body of evidence has highlighted behavioral connections between musical rhythm and linguistic syntax, suggesting that these abilities may be mediated by common neural resources. Here, we performed a quantitative meta-analysis of neuroimaging studies using activation likelihood estimate (ALE) to localize the shared neural structures engaged in a representative set of musical rhythm (rhythm, beat, and meter) and linguistic syntax (merge movement, and reanalysis) operations. Rhythm engaged a bilateral sensorimotor network throughout the brain consisting of the inferior frontal gyri, supplementary motor area, superior temporal gyri/temporoparietal junction, insula, intraparietal lobule, and putamen. By contrast, syntax mostly recruited the left sensorimotor network including the inferior frontal gyrus, posterior superior temporal gyrus, premotor cortex, and supplementary motor area. Intersections between rhythm and syntax maps yielded overlapping regions in the left inferior frontal gyrus, left supplementary motor area, and bilateral insula-neural substrates involved in temporal hierarchy processing and predictive coding. Together, this is the first neuroimaging meta-analysis providing detailed anatomical overlap of sensorimotor regions recruited for musical rhythm and linguistic syntax.

Cortical control of eye movements in natural reading: Evidence from MVPA

Language comprehension during reading requires fine-grained management of saccadic eye movements. A critical question, therefore, is how the brain controls eye movements in reading. Neural correlates of simple eye movements have been found in multiple cortical regions, but little is known about how this network operates in reading. To investigate this question in the present study, participants were presented with normal text, pseudo-word text, and consonant string text in a magnetic resonance imaging (MRI) scanner with eyetracking. Participants read naturally in the normal text condition and moved their eyes "as if they were reading" in the other conditions. Multi-voxel pattern analysis was used to analyze the fMRI signal in the oculomotor network. We found that activation patterns in a subset of network regions differentiated between stimulus types. These results suggest that the oculomotor network reflects more than simple saccade generation and are consistent with the hypothesis that specific network areas interface with cognitive systems.

Is there a prediction network? Meta-analytic evidence for a cortical-subcortical network likely subserving prediction

Predictive coding is an increasingly influential and ambitious concept in neuroscience viewing the brain as a 'hypothesis testing machine' that constantly strives to minimize prediction error, the gap between its predictions and the actual sensory input. Despite the invaluable contribution of this framework to the formulation of brain function, its neuroanatomical foundations have not been fully defined. To address this gap, we conducted activation likelihood estimation (ALE) meta-analysis of 39 neuroimaging studies of three functional domains (action perception, language and music) inherently involving prediction. The ALE analysis revealed a widely distributed brain network encompassing regions within the inferior and middle frontal gyri, anterior insula, premotor cortex, pre-supplementary motor area, temporoparietal junction, striatum, thalamus/subthalamus and the cerebellum. This network is proposed to subserve domain-general prediction and its relevance to motor control, attention, implicit learning and social cognition is discussed in light of the predictive coding scheme. Better understanding of the presented network may help advance treatments of neuropsychiatric conditions related to aberrant prediction processing and promote cognitive enhancement in healthy individuals.

Sentence processing is modulated by the current linguistic environment and a priori information: An fMRI study

INTRODUCTION

Words are not processed in isolation but in rich contexts that are used to modulate and facilitate language comprehension. Here, we investigate distinct neural networks underlying two types of contexts, the current linguistic environment and verb-based syntactic preferences.

METHODS

We had two main manipulations. The first was the current linguistic environment, where the relative frequencies of two syntactic structures (prepositional object [PO] and double-object [DO]) would either follow everyday linguistic experience or not. The second concerned the preference toward one or the other structure depending on the verb; learned in everyday language use and stored in memory. German participants were reading PO and DO sentences in German while brain activity was measured with functional magnetic resonance imaging.

RESULTS

First, the anterior cingulate cortex (ACC) showed a pattern of activation that integrated the current linguistic environment with everyday linguistic experience. When the input did not match everyday experience, the unexpected frequent structure showed higher activation in the ACC than the other conditions and more connectivity from the ACC to posterior parts of the language network. Second, verb-based surprisal of seeing a structure given a verb (PO verb preference but DO structure presentation) resulted, within the language network (left inferior frontal and left middle/superior temporal gyrus) and the precuneus, in increased activation compared to a predictable verb-structure pairing.

CONCLUSION

In conclusion, (1) beyond the canonical language network, brain areas engaged in prediction and error signaling, such as the ACC, might use the statistics of syntactic structures to modulate language processing, (2) the language network is directly engaged in processing verb preferences. These two networks show distinct influences on sentence processing.

Co-registration of eye movements and neuroimaging for studying contextual predictions in natural reading

Sixteen years ago, Sereno and Rayner (2003. Measuring word recognition in reading: eye movements and event-related potentials. Trends in Cognitive Sciences, 7(11), 489-493) illustrated how "by means of review and comparison" eye movement (EM) and event-related potential (ERP) studies may advance our understanding of visual word recognition. Attempts to simultaneously record EMs and ERPs soon followed. Recently, this co-registration approach has also been transferred to fMRI and oscillatory EEG. With experimental settings close to natural reading, co-registration enables us to directly integrate insights from EM and neuroimaging studies. This should extend current experimental paradigms by moving the field towards studying sentence-level processing including effects of context and parafoveal preview. This article will introduce the basic principles and applications of co-registration and selectively review how this approach may shed light on one of the most controversially discussed issues in reading research, contextual predictions in online language processing.

Hierarchical structure guides rapid linguistic predictions during naturalistic listening

The grammar, or syntax, of human language is typically understood in terms of abstract hierarchical structures. However, theories of language processing that emphasize sequential information, not hierarchy, successfully model diverse phenomena. Recent work probing brain signals has shown mixed evidence for hierarchical information in some tasks. We ask whether sequential or hierarchical information guides the expectations that a human listener forms about a word's part-of-speech when simply listening to every-day language. We compare the predictions of three computational models against electroencephalography signals recorded from human participants who listen passively to an audiobook story. We find that predictions based on hierarchical structure correlate with the human brain response above-and-beyond predictions based only on sequential information. This establishes a link between hierarchical linguistic structure and neural signals that generalizes across the range of syntactic structures found in every-day language.

Integration and Prediction in Language Processing: A Synthesis of Old and New

Current theories of language processing emphasize prediction as a mechanism to facilitate comprehension, which contrasts with the state of the field a few decades ago, when prediction was rarely mentioned. We argue that the field of psycholinguistics would benefit from revisiting these earlier theories of comprehension which attempted to explain integration and the processes underlying the formation of rich representations of linguistic input, and that emphasized informational newness over redundancy. We suggest further that integration and anticipation may be complementary mechanisms that operate on elaborated, coherent discourse representations, supporting enhanced comprehension and memory. In addition, the traditional emphasis on language as a tool for communication implies that much linguistic content will be nonredundant; moreover, the purpose of anticipation is probably not to permit the prediction of exact lexical or syntactic forms, but instead to induce a state of preparedness that allows the comprehender to be receptive to new information, thus facilitating its processing.

Cortical regions supporting reading comprehension skill for single words and discourse

A substantial amount of variation in reading comprehension skill is explained by listening comprehension skill, suggesting tight links between printed and spoken discourse processing. In addition, both word level (e.g., vocabulary) and discourse-level sub-skills (e.g., inference-making) support overall comprehension. However, while these contributions to variation in comprehension skill have been well-studied behaviorally, the underlying neurobiological basis of these relationships is less well understood. In order to examine the neural bases of individual differences in reading comprehension as a function of input modality and processing level, we examined functional neural activation to both spoken and printed single words and passages in adolescents with a range of comprehension skill. Data driven Partial Least Squares Correlation (PLSC) analyses revealed that comprehension skill was positively related to activation in a number of regions associated with discourse comprehension and negatively related to activation in regions associated with executive function and memory across processing levels and input modalities.

Rapid syntactic pre-activation in Broca's area: Concurrent electrophysiological and haemodynamic recordings

Listeners are constantly trying to predict what the speaker will say next. We concurrently measured the electrophysiological and haemodynamic correlates of syntactic pre-activation, investigating when and where the brain processes speech melody cues to upcoming word order structure. Pre-activation of syntactic structure was reflected in a left-lateralised pre-activation negativity (PrAN), which was subserved by Broca's area in the left inferior frontal gyrus, as well as the contiguous left anterior insula.

Word Frequency Effects in Naturalistic Reading

Word frequency is a central psycholinguistic variable that accounts for substantial variance in language processing. A number of neuroimaging studies have examined frequency at a single word level, typically demonstrating a strong negative, and sometimes positive correlation between frequency and hemodynamic response. Here, 40 subjects read passages of text in an MRI scanner while their eye movements were recorded. We used fixation-related analysis to identify neural activity tied to the frequency of each fixated word. We found that negative correlations with frequency were reduced, while strong positive correlations were found in the temporal and parietal areas associated with semantics. We propose that the processing cost of low frequency words is reduced due to contextual cues. Meanings of high frequency words are more readily accessed and integrated with context resulting in enhanced processing in the semantic system. The results demonstrate similarities and differences between single word and naturalistic text processing.

A robust dissociation among the language, multiple demand, and default mode networks: Evidence from inter-region correlations in effect size

Complex cognitive processes, including language, rely on multiple mental operations that are carried out by several large-scale functional networks in the frontal, temporal, and parietal association cortices of the human brain. The central division of cognitive labor is between two fronto-parietal bilateral networks: (a) the multiple demand (MD) network, which supports executive processes, such as working memory and cognitive control, and is engaged by diverse task domains, including language, especially when comprehension gets difficult; and (b) the default mode network (DMN), which supports introspective processes, such as mind wandering, and is active when we are not engaged in processing external stimuli. These two networks are strongly dissociated in both their functional profiles and their patterns of activity fluctuations during naturalistic cognition. Here, we focus on the functional relationship between these two networks and a third network: (c) the fronto-temporal left-lateralized "core" language network, which is selectively recruited by linguistic processing. Is the language network distinct and dissociated from both the MD network and the DMN, or is it synchronized and integrated with one or both of them? Recent work has provided evidence for a dissociation between the language network and the MD network. However, the relationship between the language network and the DMN is less clear, with some evidence for coordinated activity patterns and similar response profiles, perhaps due to the role of both in semantic processing. Here we use a novel fMRI approach to examine the relationship among the three networks: we measure the strength of activations in different language, MD, and DMN regions to functional contrasts typically used to identify each network, and then test which regions co-vary in their contrast effect sizes across 60 individuals. We find that effect sizes correlate strongly within each network (e.g., one language region and another language region, or one DMN region and another DMN region), but show little or no correlation for region pairs across networks (e.g., a language region and a DMN region). Thus, using our novel method, we replicate the language/MD network dissociation discovered previously with other approaches, and also show that the language network is robustly dissociated from the DMN, overall suggesting that these three networks contribute to high-level cognition in different ways and, perhaps, support distinct computations. Inter-individual differences in effect sizes therefore do not simply reflect general differences in vascularization or attention, but exhibit sensitivity to the functional architecture of the brain. The strength of activation in each network can thus be probed separately in studies that attempt to link neural variability to behavioral or genetic variability.

Neural Evidence of Cross-domain Structural Interaction between Language and Arithmetic

The presence of a shared neural system for the syntactic processing in language and arithmetic is controversial. Recent behavioral studies reported a cross-domain structural priming between language and arithmetic. Using functional magnetic resonance imaging, we examined whether the neural activation reflects the structural interaction between language and arithmetic. We prepared sentences and arithmetic expressions (A-expressions) with same and different syntactic structures and presented structurally congruent/incongruent pairs consecutively. By directly comparing activations in the congruent and incongruent conditions, we observed significant repetition suppression effect in the regions including the bilateral inferior frontal gyrus, i.e., neural activation with an A-expression decreased after a sentence with the same syntactic structure (and vice versa). The results strongly support the idea that arithmetic and language share the neural basis for processing syntactic structures.

Information and learning in processing adjective inflection

We investigated the processing of inflected Serbian adjective forms to bring together quantitative linguistic measures from two frameworks - information theory and discrimination learning. From each framework we derived several quantitative descriptions of an inflectional morphological system and fitted two separate regression models to the processing latencies that were elicited by inflected adjectival forms presented in a visual lexical decision task. The model, which was based on lexical distributional and information theory revealed a dynamic interplay of information. The information was sensitive to syntagmatic and paradigmatic dimensions of variation; the paradigmatic information (formalized as respective relative entropies) was also modulated by lemma frequency. The discrimination learning based model revealed an equally complex pattern, involving several learning-based variables. The two models revealed strikingly similar patterns of results, as confirmed by the very high proportion of shared variance in model predictions (85.83%). Our findings add to the body of research demonstrating that complex morphological phenomena can arise as a consequence of the basic principles of discrimination learning. Learning discriminatively about inflectional paradigms and classes, and about their contextual or syntagmatic embedding, sheds light on human language-processing efficiency and on the fascinating complexity of naturally emerged language systems.

Lexical Predictability During Natural Reading: Effects of Surprisal and Entropy Reduction

What are the effects of word-by-word predictability on sentence processing times during the natural reading of a text? Although information complexity metrics such as surprisal and entropy reduction have been useful in addressing this question, these metrics tend to be estimated using computational language models, which require some degree of commitment to a particular theory of language processing. Taking a different approach, this study implemented a large-scale cumulative cloze task to collect word-by-word predictability data for 40 passages and compute surprisal and entropy reduction values in a theory-neutral manner. A separate group of participants read the same texts while their eye movements were recorded. Results showed that increases in surprisal and entropy reduction were both associated with increases in reading times. Furthermore, these effects did not depend on the global difficulty of the text. The findings suggest that surprisal and entropy reduction independently contribute to variation in reading times, as these metrics seem to capture different aspects of lexical predictability.

Left-Lateralized Contributions of Saccades to Cortical Activity During a One-Back Word Recognition Task

Saccadic eye movements are an inherent component of natural reading, yet their contribution to information processing at subsequent fixation remains elusive. Here we use anatomically-constrained magnetoencephalography (MEG) to examine cortical activity following saccades as healthy human subjects engaged in a one-back word recognition task. This activity was compared with activity following external visual stimulation that mimicked saccades. A combination of procedures was employed to eliminate saccadic ocular artifacts from the MEG signal. Both saccades and saccade-like external visual stimulation produced early-latency responses beginning ~70 ms after onset in occipital cortex and spreading through the ventral and dorsal visual streams to temporal, parietal and frontal cortices. Robust differential activity following the onset of saccades vs. similar external visual stimulation emerged during 150-350 ms in a left-lateralized cortical network. This network included: (i) left lateral occipitotemporal (LOT) and nearby inferotemporal (IT) cortex; (ii) left posterior Sylvian fissure (PSF) and nearby multimodal cortex; and (iii) medial parietooccipital (PO), posterior cingulate and retrosplenial cortices. Moreover, this left-lateralized network colocalized with word repetition priming effects. Together, results suggest that central saccadic mechanisms influence a left-lateralized language network in occipitotemporal and temporal cortex above and beyond saccadic influences at preceding stages of information processing during visual word recognition.

Music models aberrant rule decoding and reward valuation in dementia

Aberrant rule- and reward-based processes underpin abnormalities of socio-emotional behaviour in major dementias. However, these processes remain poorly characterized. Here we used music to probe rule decoding and reward valuation in patients with frontotemporal dementia (FTD) syndromes and Alzheimer's disease (AD) relative to healthy age-matched individuals. We created short melodies that were either harmonically resolved ('finished') or unresolved ('unfinished'); the task was to classify each melody as finished or unfinished (rule processing) and rate its subjective pleasantness (reward valuation). Results were adjusted for elementary pitch and executive processing; neuroanatomical correlates were assessed using voxel-based morphometry. Relative to healthy older controls, patients with behavioural variant FTD showed impairments of both musical rule decoding and reward valuation, while patients with semantic dementia showed impaired reward valuation but intact rule decoding, patients with AD showed impaired rule decoding but intact reward valuation and patients with progressive non-fluent aphasia performed comparably to healthy controls. Grey matter associations with task performance were identified in anterior temporal, medial and lateral orbitofrontal cortices, previously implicated in computing diverse biological and non-biological rules and rewards. The processing of musical rules and reward distils cognitive and neuroanatomical mechanisms relevant to complex socio-emotional dysfunction in major dementias.

Probabilistic language models in cognitive neuroscience: Promises and pitfalls

Cognitive neuroscientists of language comprehension study how neural computations relate to cognitive computations during comprehension. On the cognitive part of the equation, it is important that the computations and processing complexity are explicitly defined. Probabilistic language models can be used to give a computationally explicit account of language complexity during comprehension. Whereas such models have so far predominantly been evaluated against behavioral data, only recently have the models been used to explain neurobiological signals. Measures obtained from these models emphasize the probabilistic, information-processing view of language understanding and provide a set of tools that can be used for testing neural hypotheses about language comprehension. Here, we provide a cursory review of the theoretical foundations and example neuroimaging studies employing probabilistic language models. We highlight the advantages and potential pitfalls of this approach and indicate avenues for future research.

Auditory conflict and congruence in frontotemporal dementia

Impaired analysis of signal conflict and congruence may contribute to diverse socio-emotional symptoms in frontotemporal dementias, however the underlying mechanisms have not been defined. Here we addressed this issue in patients with behavioural variant frontotemporal dementia (bvFTD; n = 19) and semantic dementia (SD; n = 10) relative to healthy older individuals (n = 20). We created auditory scenes in which semantic and emotional congruity of constituent sounds were independently probed; associated tasks controlled for auditory perceptual similarity, scene parsing and semantic competence. Neuroanatomical correlates of auditory congruity processing were assessed using voxel-based morphometry. Relative to healthy controls, both the bvFTD and SD groups had impaired semantic and emotional congruity processing (after taking auditory control task performance into account) and reduced affective integration of sounds into scenes. Grey matter correlates of auditory semantic congruity processing were identified in distributed regions encompassing prefrontal, parieto-temporal and insular areas and correlates of auditory emotional congruity in partly overlapping temporal, insular and striatal regions. Our findings suggest that decoding of auditory signal relatedness may probe a generic cognitive mechanism and neural architecture underpinning frontotemporal dementia syndromes.

Neural Pattern Similarity in the Left IFG and Fusiform Is Associated with Novel Word Learning

Previous studies have revealed that greater neural pattern similarity across repetitions is associated with better subsequent memory. In this study, we used an artificial language training paradigm and representational similarity analysis to examine whether neural pattern similarity across repetitions before training was associated with post-training behavioral performance. Twenty-four native Chinese speakers were trained to learn a logographic artificial language for 12 days and behavioral performance was recorded using the word naming and picture naming tasks. Participants were scanned while performing a passive viewing task before training, after 4-day training and after 12-day training. Results showed that pattern similarity in the left pars opercularis (PO) and fusiform gyrus (FG) before training was negatively associated with reaction time (RT) in both word naming and picture naming tasks after training. These results suggest that neural pattern similarity is an effective neurofunctional predictor of novel word learning in addition to word memory.

Neural correlates of individual differences in fixation duration during natural reading

Reading requires integration of language and cognitive processes with attention and eye movement control. Individuals differ in their reading ability, but little is known about the neurocognitive processes associated with these individual differences. To investigate this issue, we combined eyetracking and fMRI, simultaneously recording eye movements and BOLD activity while subjects read text passages. We found that the variability and skew of fixation duration distributions across individuals, as assessed by ex-Gaussian analyses, decreased with increasing neural activity in regions associated with the cortical eye movement control network (Left FEF, Left IPS, Left IFG, and Right IFG). The results suggest that individual differences in fixation duration during reading are related to underlying neurocognitive processes associated with the eye movement control system and its relationship to language processing. The results also show that eye movements and fMRI can be combined to investigate the neural correlates of individual differences in natural reading.

Neurophysiological dynamics of phrase-structure building during sentence processing

Although sentences unfold sequentially, one word at a time, most linguistic theories propose that their underlying syntactic structure involves a tree of nested phrases rather than a linear sequence of words. Whether and how the brain builds such structures, however, remains largely unknown. Here, we used human intracranial recordings and visual word-by-word presentation of sentences and word lists to investigate how left-hemispheric brain activity varies during the formation of phrase structures. In a broad set of language-related areas, comprising multiple superior temporal and inferior frontal sites, high-gamma power increased with each successive word in a sentence but decreased suddenly whenever words could be merged into a phrase. Regression analyses showed that each additional word or multiword phrase contributed a similar amount of additional brain activity, providing evidence for a merge operation that applies equally to linguistic objects of arbitrary complexity. More superficial models of language, based solely on sequential transition probability over lexical and syntactic categories, only captured activity in the posterior middle temporal gyrus. Formal model comparison indicated that the model of multiword phrase construction provided a better fit than probability-based models at most sites in superior temporal and inferior frontal cortices. Activity in those regions was consistent with a neural implementation of a bottom-up or left-corner parser of the incoming language stream. Our results provide initial intracranial evidence for the neurophysiological reality of the merge operation postulated by linguists and suggest that the brain compresses syntactically well-formed sequences of words into a hierarchy of nested phrases.

Teasing apart coercion and surprisal: Evidence from eye-movements and ERPs

Previous behavioral and electrophysiological studies have presented evidence suggesting that coercion expressions (e.g., began the book) are more difficult to process than control expressions like read the book. While this processing cost has been attributed to a specific coercion operation for recovering an event-sense of the complement (e.g., began reading the book), an alternative view based on the Surprisal Theory of language processing would attribute the cost to the relative unpredictability of the complement noun in the coercion compared to the control condition, with no need to postulate coercion-specific mechanisms. In two experiments, monitoring eye-tracking and event-related potentials (ERPs), respectively, we sought to determine whether there is any evidence for coercion-specific processing cost above-and-beyond the difficulty predicted by surprisal, by contrasting coercing and control expressions with a further control condition in which the predictability of the complement noun was similar to that in the coercion condition (e.g., bought the book). While the eye-tracking study showed significant effects of surprisal and a marginal effect of coercion on late reading measures, the ERP study clearly supported the surprisal account. Overall, our findings suggest that the coercion cost largely reflects the surprisal of the complement noun with coercion specific operations possibly influencing later processing stages.

MEG Evidence for Incremental Sentence Composition in the Anterior Temporal Lobe

Research investigating the brain basis of language comprehension has associated the left anterior temporal lobe (ATL) with sentence-level combinatorics. Using magnetoencephalography (MEG), we test the parsing strategy implemented in this brain region. The number of incremental parse steps from a predictive left-corner parsing strategy that is supported by psycholinguistic research is compared with those from a less-predictive strategy. We test for a correlation between parse steps and source-localized MEG activity recorded while participants read a story. Left-corner parse steps correlated with activity in the left ATL around 350-500 ms after word onset. No other correlations specific to sentence comprehension were observed. These data indicate that the left ATL engages in combinatoric processing that is well characterized by a predictive left-corner parsing strategy.