A predictive coding model of the N400.

The N400 event-related component has been widely used to investigate the neural mechanisms underlying real-time language comprehension. However, despite decades of research, there is still no unifying theory that can explain both its temporal dynamics and functional properties. In this work, we show that predictive coding - a biologically plausible algorithm for approximating Bayesian inference - offers a promising framework for characterizing the N400. Using an implemented predictive coding computational model, we demonstrate how the N400 can be formalized as the lexico-semantic prediction error produced as the brain infers meaning from the linguistic form of incoming words. We show that the magnitude of lexico-semantic prediction error mirrors the functional sensitivity of the N400 to various lexical variables, priming, contextual effects, as well as their higher-order interactions. We further show that the dynamics of the predictive coding algorithm provides a natural explanation for the temporal dynamics of the N400, and a biologically plausible link to neural activity. Together, these findings directly situate the N400 within the broader context of predictive coding research. More generally, they raise the possibility that the brain may use the same computational mechanism for inference across linguistic and non-linguistic domains.

Dissociating the pre-activation of word meaning and form during sentence comprehension: Evidence from EEG representational similarity analysis.

During language comprehension, the processing of each incoming word is facilitated in proportion to its predictability. Here, we asked whether anticipated upcoming linguistic information is actually pre-activated before new bottom-up input becomes available, and if so, whether this pre-activation is limited to the level of semantic features, or whether extends to representations of individual word-forms (orthography/phonology). We carried out Representational Similarity Analysis on EEG data while participants read highly constraining sentences. Prior to the onset of the expected target words, sentence pairs predicting semantically related words (financial "bank" - "loan") and form-related words (financial "bank" - river "bank") produced more similar neural patterns than pairs predicting unrelated words ("bank" - "lesson"). This provides direct neural evidence for item-specific semantic and form predictive pre-activation. Moreover, the semantic pre-activation effect preceded the form pre-activation effect, suggesting that top-down pre-activation is propagated from higher to lower levels of the linguistic hierarchy over time.

Natural Language Processing Markers for Psychosis and Other Psychiatric Disorders: Emerging Themes and Research Agenda From a Cross-Linguistic Workshop.

This workshop summary on natural language processing (NLP) markers for psychosis and other psychiatric disorders presents some of the clinical and research issues that NLP markers might address and some of the activities needed to move in that direction. We propose that the optimal development of NLP markers would occur in the context of research efforts to map out the underlying mechanisms of psychosis and other disorders. In this workshop, we identified some of the challenges to be addressed in developing and implementing NLP markers-based Clinical Decision Support Systems (CDSSs) in psychiatric practice, especially with respect to psychosis. Of note, a CDSS is meant to enhance decision-making by clinicians by providing additional relevant information primarily through software (although CDSSs are not without risks). In psychiatry, a field that relies on subjective clinical ratings that condense rich temporal behavioral information, the inclusion of computational quantitative NLP markers can plausibly lead to operationalized decision models in place of idiosyncratic ones, although ethical issues must always be paramount.

Predictive coding across the left fronto-temporal hierarchy during language comprehension.

We used magnetoencephalography (MEG) and event-related potentials (ERPs) to track the time-course and localization of evoked activity produced by expected, unexpected plausible, and implausible words during incremental language comprehension. We suggest that the full pattern of results can be explained within a hierarchical predictive coding framework in which increased evoked activity reflects the activation of residual information that was not already represented at a given level of the fronto-temporal hierarchy ("error" activity). Between 300 and 500 ms, the three conditions produced progressively larger responses within left temporal cortex (lexico-semantic prediction error), whereas implausible inputs produced a selectively enhanced response within inferior frontal cortex (prediction error at the level of the event model). Between 600 and 1,000 ms, unexpected plausible words activated left inferior frontal and middle temporal cortices (feedback activity that produced top-down error), whereas highly implausible inputs activated left inferior frontal cortex, posterior fusiform (unsuppressed orthographic prediction error/reprocessing), and medial temporal cortex (possibly supporting new learning). Therefore, predictive coding may provide a unifying theory that links language comprehension to other domains of cognition.

We both say tomato: Intact lexical alignment in schizophrenia and bipolar disorder.

In people with schizophrenia and related disorders, impairments in communication and social functioning can negatively impact social interactions and quality of life. In the present study, we investigated the cognitive basis of a specific aspect of linguistic communication-lexical alignment-in people with schizophrenia and bipolar disorder. We probed lexical alignment as participants played a collaborative picture-naming game with the experimenter, in which the two players alternated between naming a dual-name picture (e.g., rabbit/bunny) and listening to their partner name a picture. We found evidence of lexical alignment in all three groups, with no differences between the patient groups and the controls. We argue that these typical patterns of lexical alignment in patients were supported by preserved-and in some cases increased-bottom-up mechanisms, which balanced out impairments in top-down perspective-taking.

Word predictability effects are linear, not logarithmic: Implications for probabilistic models of sentence comprehension.

During language comprehension, we routinely use information from the prior context to help identify the meaning of individual words. While measures of online processing difficulty, such as reading times, are strongly influenced by contextual predictability, there is disagreement about the mechanisms underlying this lexical predictability effect, with different models predicting different linking functions - linear (Reichle, Rayner & Pollatsek, 2003) or logarithmic (Levy, 2008). To help resolve this debate, we conducted two highly-powered experiments (self-paced reading, N = 216; cross-modal picture naming, N = 36), and a meta-analysis of prior eye-tracking while reading studies (total N = 218). We observed a robust linear relationship between lexical predictability and word processing times across all three studies. Beyond their methodological implications, these findings also place important constraints on predictive processing models of language comprehension. In particular, these results directly contradict the empirical predictions of surprisal theory, while supporting a proportional pre-activation account of lexical prediction effects in comprehension.

Tea With Milk? A Hierarchical Generative Framework of Sequential Event Comprehension.

To make sense of the world around us, we must be able to segment a continual stream of sensory inputs into discrete events. In this review, I propose that in order to comprehend events, we engage hierarchical generative models that "reverse engineer" the intentions of other agents as they produce sequential action in real time. By generating probabilistic predictions for upcoming events, generative models ensure that we are able to keep up with the rapid pace at which perceptual inputs unfold. By tracking our certainty about other agents' goals and the magnitude of prediction errors at multiple temporal scales, generative models enable us to detect event boundaries by inferring when a goal has changed. Moreover, by adapting flexibly to the broader dynamics of the environment and our own comprehension goals, generative models allow us to optimally allocate limited resources. Finally, I argue that we use generative models not only to comprehend events but also to produce events (carry out goal-relevant sequential action) and to continually learn about new events from our surroundings. Taken together, this hierarchical generative framework provides new insights into how the human brain processes events so effortlessly while highlighting the fundamental links between event comprehension, production, and learning.

Going the Extra Mile: Effects of Discourse Context on Two Late Positivities During Language Comprehension.

During language comprehension, online neural processing is strongly influenced by the constraints of the prior context. While the N400 ERP response (300-500ms) is known to be sensitive to a word's semantic predictability, less is known about a set of late positive-going ERP responses (600-1000ms) that can be elicited when an incoming word violates strong predictions about upcoming content (late frontal positivity) or about what is possible given the prior context (late posterior positivity/P600). Across three experiments, we systematically manipulated the length of the prior context and the source of lexical constraint to determine their influence on comprehenders' online neural responses to these two types of prediction violations. In Experiment 1, within minimal contexts, both lexical prediction violations and semantically anomalous words produced a larger N400 than expected continuations (James unlocked the door/laptop/gardener), but no late positive effects were observed. Critically, the late posterior positivity/P600 to semantic anomalies appeared when these same sentences were embedded within longer discourse contexts (Experiment 2a), and the late frontal positivity appeared to lexical prediction violations when the preceding context was rich and globally constraining (Experiment 2b). We interpret these findings within a hierarchical generative framework of language comprehension. This framework highlights the role of comprehension goals and broader linguistic context, and how these factors influence both top-down prediction and the decision to update or reanalyze the prior context when these predictions are violated.

Impairments in Probabilistic Prediction and Bayesian Learning Can Explain Reduced Neural Semantic Priming in Schizophrenia.

It has been proposed that abnormalities in probabilistic prediction and dynamic belief updating explain the multiple features of schizophrenia. Here, we used electroencephalography (EEG) to ask whether these abnormalities can account for the well-established reduction in semantic priming observed in schizophrenia under nonautomatic conditions. We isolated predictive contributions to the neural semantic priming effect by manipulating the prime's predictive validity and minimizing retroactive semantic matching mechanisms. We additionally examined the link between prediction and learning using a Bayesian model that probed dynamic belief updating as participants adapted to the increase in predictive validity. We found that patients were less likely than healthy controls to use the prime to predictively facilitate semantic processing on the target, resulting in a reduced N400 effect. Moreover, the trial-by-trial output of our Bayesian computational model explained between-group differences in trial-by-trial N400 amplitudes as participants transitioned from conditions of lower to higher predictive validity. These findings suggest that, compared with healthy controls, people with schizophrenia are less able to mobilize predictive mechanisms to facilitate processing at the earliest stages of accessing the meanings of incoming words. This deficit may be linked to a failure to adapt to changes in the broader environment. This reciprocal relationship between impairments in probabilistic prediction and Bayesian learning/adaptation may drive a vicious cycle that maintains cognitive disturbances in schizophrenia.

Neural Evidence for the Prediction of Animacy Features during Language Comprehension: Evidence from MEG and EEG Representational Similarity Analysis.

It has been proposed that people can generate probabilistic predictions at multiple levels of representation during language comprehension. We used magnetoencephalography (MEG) and electroencephalography (EEG), in combination with representational similarity analysis, to seek neural evidence for the prediction of animacy features. In two studies, MEG and EEG activity was measured as human participants (both sexes) read three-sentence scenarios. Verbs in the final sentences constrained for either animate or inanimate semantic features of upcoming nouns, and the broader discourse context constrained for either a specific noun or for multiple nouns belonging to the same animacy category. We quantified the similarity between spatial patterns of brain activity following the verbs until just before the presentation of the nouns. The MEG and EEG datasets revealed converging evidence that the similarity between spatial patterns of neural activity following animate-constraining verbs was greater than following inanimate-constraining verbs. This effect could not be explained by lexical-semantic processing of the verbs themselves. We therefore suggest that it reflected the inherent difference in the semantic similarity structure of the predicted animate and inanimate nouns. Moreover, the effect was present regardless of whether a specific word could be predicted, providing strong evidence for the prediction of coarse-grained semantic features that goes beyond the prediction of individual words.SIGNIFICANCE STATEMENT Language inputs unfold very quickly during real-time communication. By predicting ahead, we can give our brains a "head start," so that language comprehension is faster and more efficient. Although most contexts do not constrain strongly for a specific word, they do allow us to predict some upcoming information. For example, following the context of "they cautioned the…," we can predict that the next word will be animate rather than inanimate (we can caution a person, but not an object). Here, we used EEG and MEG techniques to show that the brain is able to use these contextual constraints to predict the animacy of upcoming words during sentence comprehension, and that these predictions are associated with specific spatial patterns of neural activity.

Having your cake and eating it too: Flexibility and power with mass univariate statistics for ERP data.

ERP studies produce large spatiotemporal data sets. These rich data sets are key to enabling us to understand cognitive and neural processes. However, they also present a massive multiple comparisons problem, potentially leading to a large number of studies with false positive effects (a high Type I error rate). Standard approaches to ERP statistical analysis, which average over time windows and regions of interest, do not always control for Type I error, and their inflexibility can lead to low power to detect true effects. Mass univariate approaches offer an alternative analytic method. However, they have thus far been viewed as appropriate primarily for exploratory statistical analysis and only applicable to simple designs. Here, we present new simulation studies showing that permutation-based mass univariate tests can be employed with complex factorial designs. Most importantly, we show that mass univariate approaches provide slightly greater power than traditional spatiotemporal averaging approaches when strong a priori time windows and spatial regions are used. Moreover, their power decreases only modestly when more exploratory spatiotemporal parameters are used. We argue that mass univariate approaches are preferable to traditional spatiotemporal averaging analysis approaches for many ERP studies.

A Tale of Two Positivities and the N400: Distinct Neural Signatures Are Evoked by Confirmed and Violated Predictions at Different Levels of Representation.

It has been proposed that hierarchical prediction is a fundamental computational principle underlying neurocognitive processing. Here, we ask whether the brain engages distinct neurocognitive mechanisms in response to inputs that fulfill versus violate strong predictions at different levels of representation during language comprehension. Participants read three-sentence scenarios in which the third sentence constrained for a broad event structure, for example, {Agent caution animate-Patient}. High constraint contexts additionally constrained for a specific event/lexical item, for example, a two-sentence context about a beach, lifeguards, and sharks constrained for the event, {Lifeguards cautioned Swimmers}, and the specific lexical item swimmers. Low constraint contexts did not constrain for any specific event/lexical item. We measured ERPs on critical nouns that fulfilled and/or violated each of these constraints. We found clear, dissociable effects to fulfilled semantic predictions (a reduced N400), to event/lexical prediction violations (an increased late frontal positivity), and to event structure/animacy prediction violations (an increased late posterior positivity/P600). We argue that the late frontal positivity reflects a large change in activity associated with successfully updating the comprehender's current situation model with new unpredicted information. We suggest that the late posterior positivity/P600 is triggered when the comprehender detects a conflict between the input and her model of the communicator and communicative environment. This leads to an initial failure to incorporate the unpredicted input into the situation model, which may be followed by second-pass attempts to make sense of the discourse through reanalysis, repair, or reinterpretation. Together, these findings provide strong evidence that confirmed and violated predictions at different levels of representation manifest as distinct spatiotemporal neural signatures.

Functional MRI reveals evidence of a self-positivity bias in the medial prefrontal cortex during the comprehension of social vignettes.

A large literature in social neuroscience has associated the medial prefrontal cortex (mPFC) with the processing of self-related information. However, only recently have social neuroscience studies begun to consider the large behavioral literature showing a strong self-positivity bias, and these studies have mostly focused on its correlates during self-related judgments and decision-making. We carried out a functional MRI (fMRI) study to ask whether the mPFC would show effects of the self-positivity bias in a paradigm that probed participants' self-concept without any requirement of explicit self-judgment. We presented social vignettes that were either self-relevant or non-self-relevant with a neutral, positive or negative outcome described in the second sentence. In previous work using event-related potentials, this paradigm has shown evidence of a self-positivity bias that influences early stages of semantically processing incoming stimuli. In the present fMRI study, we found evidence for this bias within the mPFC: an interaction between self-relevance and valence, with only positive scenarios showing a self vs other effect within the mPFC. We suggest that the mPFC may play a role in maintaining a positively biased self-concept and discuss the implications of these findings for the social neuroscience of the self and the role of the mPFC.

Neural evidence for Bayesian trial-by-trial adaptation on the N400 during semantic priming.

When semantic information is activated by a context prior to new bottom-up input (i.e. when a word is predicted), semantic processing of that incoming word is typically facilitated, attenuating the amplitude of the N400 event related potential (ERP) - a direct neural measure of semantic processing. N400 modulation is observed even when the context is a single semantically related "prime" word. This so-called "N400 semantic priming effect" is sensitive to the probability of encountering a related prime-target pair within an experimental block, suggesting that participants may be adapting the strength of their predictions to the predictive validity of their broader experimental environment. We formalize this adaptation using a Bayesian learning model that estimates and updates the probability of encountering a related versus an unrelated prime-target pair on each successive trial. We found that our model's trial-by-trial estimates of target word probability accounted for significant variance in trial-by-trial N400 amplitude. These findings suggest that Bayesian principles contribute to how comprehenders adapt their semantic predictions to the statistical structure of their broader environment, with implications for the functional significance of the N400 component and the predictive nature of language processing.

Multimodal neuroimaging evidence for looser lexico-semantic networks in schizophrenia:Evidence from masked indirect semantic priming.

It has been hypothesized that schizophrenia is characterized by overly broad automatic activity within lexico-semantic networks. We used two complementary neuroimaging techniques, Magnetoencephalography (MEG) and functional Magnetic Resonance Imaging (fMRI), in combination with a highly automatic indirect semantic priming paradigm, to spatiotemporally localize this abnormality in the brain. Eighteen people with schizophrenia and 20 demographically-matched control participants viewed target words ("bell") preceded by directly related ("church"), indirectly related ("priest"), or unrelated ("hammer") prime words in MEG and fMRI sessions. To minimize top-down processing, the prime was masked, the target appeared only 140 ms after prime onset, and participants simply monitored for words within a particular semantic category that appeared in filler trials. Both techniques revealed a significantly larger automatic indirect priming effect in people with schizophrenia than in control participants. MEG temporally localized this enhanced effect to the N400 time window (300-500 ms) - the critical stage of accessing meaning from words. fMRI spatially localized the effect to the left temporal fusiform cortex, which plays a role in mapping of orthographic word-form on to meaning. There was no evidence of an enhanced automatic direct semantic priming effect in the schizophrenia group. These findings provide converging neural evidence for abnormally broad highly automatic lexico-semantic activity in schizophrenia. We argue that, rather than arising from an unconstrained spread of automatic activation across semantic memory, this broader automatic lexico-semantic activity stems from looser mappings between the form and meaning of words.

Priming production: Neural evidence for enhanced automatic semantic activity preceding language production in schizophrenia.

Introduction: Lexico-semantic disturbances are considered central to schizophrenia. Clinically, their clearest manifestation is in language production. However, most studies probing their underlying mechanisms have used comprehension or categorization tasks. Here, we probed automatic semantic activity prior to language production in schizophrenia using event-related potentials (ERPs). Methods: 19 people with schizophrenia and 16 demographically-matched healthy controls named target pictures that were very quickly preceded by masked prime words. To probe automatic semantic activity prior to production, we measured the N400 ERP component evoked by these targets. To determine the origin of any automatic semantic abnormalities, we manipulated the type of relationship between prime and target such that they overlapped in (a) their semantic features (semantically related, e.g. "cake" preceding a < picture of a pie >, (b) their initial phonemes (phonemically related, e.g. "stomach" preceding a < picture of a starfish >), or (c) both their semantic features and their orthographic/phonological word form (identity related, e.g. "socks" preceding a < picture of socks >). For each of these three types of relationship, the same targets were paired with unrelated prime words (counterbalanced across lists). We contrasted ERPs and naming times to each type of related target with its corresponding unrelated target. Results: People with schizophrenia showed abnormal N400 modulation prior to naming identity related (versus unrelated) targets: whereas healthy control participants produced a smaller amplitude N400 to identity related than unrelated targets, patients showed the opposite pattern, producing a larger N400 to identity related than unrelated targets. This abnormality was specific to the identity related targets. Just like healthy control participants, people with schizophrenia produced a smaller N400 to semantically related than to unrelated targets, and showed no difference in the N400 evoked by phonemically related and unrelated targets. There were no differences between the two groups in the pattern of naming times across conditions. Conclusion: People with schizophrenia can show abnormal neural activity associated with automatic semantic processing prior to language production. The specificity of this abnormality to the identity related targets suggests that that, rather than arising from abnormalities of either semantic features or lexical form alone, it may stem from disruptions of mappings (connections) between the meaning of words and their form.

When Proactivity Fails: An Electrophysiological Study of Establishing Reference in Schizophrenia.

BACKGROUND: Schizophrenia is characterized by abnormalities in referential communication, which may be linked to more general deficits in proactive cognitive control. We used event-related potentials to probe the timing and nature of the neural mechanisms engaged as people with schizophrenia linked pronouns to their preceding referents during word-by-word sentence comprehension. METHODS: We measured event-related potentials to pronouns in two-clause sentences in 16 people with schizophrenia and 20 demographically matched control participants. Our design crossed the number of potential referents (1-referent, 2-referent) with whether the pronoun matched the gender of its preceding referent(s) (matching, mismatching). This gave rise to four conditions: 1) 1-referent matching ("Edward took courses in accounting but he . . ."); 2) 2-referent matching ("Edward and Phillip took courses but he . . . "); 3) 1-referent mismatching ("Edward took courses in accounting but she . . ."); and 4) 2-referent mismatching ("Edward and Phillip took courses but she . . ."). RESULTS: Consistent with previous findings, healthy control participants produced a larger left anteriorly distributed negativity between 400 and 600 ms to 2-referent matching than to 1-referent matching pronouns (the "Nref effect"). In contrast, people with schizophrenia produced a larger centroposterior positivity effect between 600 and 800 ms. Both patient and control groups produced a larger positivity between 400 and 800 ms to mismatching than to matching pronouns. CONCLUSIONS: These findings suggest that proactive mechanisms of referential processing, reflected by the Nref effect, are impaired in schizophrenia, while reactive mechanisms, reflected by the positivity effects, are relatively spared. Indeed, patients may compensate for proactive deficits by retroactively engaging with context to influence the processing of inputs at a later stage of analysis.

Separate streams or probabilistic inference? What the N400 can tell us about the comprehension of events.

Since the early 2000s, several ERP studies have challenged the assumption that we always use syntactic contextual information to influence semantic processing of incoming words, as reflected by the N400 component. One approach for explaining these findings is to posit distinct semantic and syntactic processing mechanisms, each with distinct time courses. While this approach can explain specific datasets, it cannot account for the wider body of findings. I propose an alternative explanation: a dynamic generative framework in which our goal is to infer the underlying event that best explains the set of inputs encountered at any given time. Within this framework, combinations of semantic and syntactic cues with varying reliabilities are used as evidence to weight probabilistic hypotheses about this event. I further argue that the computational principles of this framework can be extended to understand how we infer situation models during discourse comprehension, and intended messages during spoken communication.

What do we mean by prediction in language comprehension?

We consider several key aspects of prediction in language comprehension: its computational nature, the representational level(s) at which we predict, whether we use higher level representations to predictively pre-activate lower level representations, and whether we 'commit' in any way to our predictions, beyond pre-activation. We argue that the bulk of behavioral and neural evidence suggests that we predict probabilistically and at multiple levels and grains of representation. We also argue that we can, in principle, use higher level inferences to predictively pre-activate information at multiple lower representational levels. We also suggest that the degree and level of predictive pre-activation might be a function of the expected utility of prediction, which, in turn, may depend on comprehenders' goals and their estimates of the relative reliability of their prior knowledge and the bottom-up input. Finally, we argue that all these properties of language understanding can be naturally explained and productively explored within a multi-representational hierarchical actively generative architecture whose goal is to infer the message intended by the producer, and in which predictions play a crucial role in explaining the bottom-up input.

The Yin and the Yang of Prediction: An fMRI Study of Semantic Predictive Processing.

Probabilistic prediction plays a crucial role in language comprehension. When predictions are fulfilled, the resulting facilitation allows for fast, efficient processing of ambiguous, rapidly-unfolding input; when predictions are not fulfilled, the resulting error signal allows us to adapt to broader statistical changes in this input. We used functional Magnetic Resonance Imaging to examine the neuroanatomical networks engaged in semantic predictive processing and adaptation. We used a relatedness proportion semantic priming paradigm, in which we manipulated the probability of predictions while holding local semantic context constant. Under conditions of higher (versus lower) predictive validity, we replicate previous observations of reduced activity to semantically predictable words in the left anterior superior/middle temporal cortex, reflecting facilitated processing of targets that are consistent with prior semantic predictions. In addition, under conditions of higher (versus lower) predictive validity we observed significant differences in the effects of semantic relatedness within the left inferior frontal gyrus and the posterior portion of the left superior/middle temporal gyrus. We suggest that together these two regions mediated the suppression of unfulfilled semantic predictions and lexico-semantic processing of unrelated targets that were inconsistent with these predictions. Moreover, under conditions of higher (versus lower) predictive validity, a functional connectivity analysis showed that the left inferior frontal and left posterior superior/middle temporal gyrus were more tightly interconnected with one another, as well as with the left anterior cingulate cortex. The left anterior cingulate cortex was, in turn, more tightly connected to superior lateral frontal cortices and subcortical regions-a network that mediates rapid learning and adaptation and that may have played a role in switching to a more predictive mode of processing in response to the statistical structure of the wider environmental context. Together, these findings highlight close links between the networks mediating semantic prediction, executive function and learning, giving new insights into how our brains are able to flexibly adapt to our environment.