Interference between sentence processing and probabilistic implicit sequence learning

Background

During sentence processing we decode the sequential combination of words, phrases or sentences according to previously learned rules. The computational mechanisms and neural correlates of these rules are still much debated. Other key issue is whether sentence processing solely relies on language-specific mechanisms or is it also governed by domain-general principles.

Methodology/Principal Findings

In the present study, we investigated the relationship between sentence processing and implicit sequence learning in a dual-task paradigm in which the primary task was a non-linguistic task (Alternating Serial Reaction Time Task for measuring probabilistic implicit sequence learning), while the secondary task were a sentence comprehension task relying on syntactic processing. We used two control conditions: a non-linguistic one (math condition) and a linguistic task (word processing task). Here we show that the sentence processing interfered with the probabilistic implicit sequence learning task, while the other two tasks did not produce a similar effect.

Conclusions/Significance

Our findings suggest that operations during sentence processing utilize resources underlying non-domain-specific probabilistic procedural learning. Furthermore, it provides a bridge between two competitive frameworks of language processing. It appears that procedural and statistical models of language are not mutually exclusive, particularly for sentence processing. These results show that the implicit procedural system is engaged in sentence processing, but on a mechanism level, language might still be based on statistical computations.     

Neural bases of syntax–semantics interface processing

The binding problem—question of how information between the modules of the linguistic system is integrated during language processing—is as yet unresolved. The remarkable speed of language processing and comprehension (Pulvermüller et al. 2009) suggests that at least coarse semantic information (e.g. noun animacy) and syntactically-relevant information (e.g. verbal template) are integrated rapidly to allow for coarse comprehension. This EEG study investigated syntax–semantics interface processing during word-by-word sentence reading. As alpha-band neural activity serves as an inhibition mechanism for local networks, we used topographical distribution of alpha power to help identify the timecourse of the binding process. We manipulated the syntactic parameter of verbal event structure, and semantic parameter of noun animacy in reduced relative clauses (RRCs, e.g. “The witness/mansion seized/protected by the agent was in danger”), to investigate the neural bases of interaction between syntactic and semantic networks during sentence processing. The word-by-word stimulus presentation method in the present experiment required manipulation of both syntactic structure and semantic features in the working memory. The results demonstrated a gradient distribution of early components (biphasic posterior P1–N2 and anterior N1–P2) over function words “by” and “the”, and the verb, corresponding to facilitation or conflict resulting from the syntactic (telicity) and semantic (animacy) cues in the preceding portion of the sentence. This was followed by assimilation of power distribution in the α band at the second noun. The flattened distribution of α power during the mental manipulation with high demand on working memory—thematic role re-assignment—demonstrates a state of α equilibrium with strong functional coupling between posterior and anterior regions. These results demonstrate that the processing of semantic and syntactic features during sentence comprehension proceeds in highly integrated fashion using gating of attentional resources to facilitate rapid comprehension, with attentional suppression of global alpha power to facilitate interaction of local networks.     

Dynamical correlation patterns and corresponding community structure in neural spontaneous activity at criticality

It has been considered that the state in the vicinity of a critical point, which is the point between ordered and disordered states, can underlie and facilitate information processing of the brain in various aspects. In this research, we numerically study the influence of criticality on one aspect of brain information processing, i.e., the community structure, which is an important characteristic of complex networks. We examine community structure of the functional connectivity in simulated brain spontaneous activity, which is based on dynamical correlations between neural activity patterns at different positions. The brain spontaneous activity is simulated by a neural field model whose parameter covers subcritical, critical, and supercritical regions. Then, the corresponding dynamical correlation patterns and community structure are compared. In the critical region, we found some distinctive properties, namely high correlation and correlation switching, high modularity and a low number of modules, high stability of the dynamical functional connectivity, and moderate flexibility of the community structure across temporal scales. We also discuss how these characteristics might improve information processing of the brain.     

The utility of behavioral models and modules in molecular analyses of social behavior

It is extremely difficult to trace the causal pathway relating gene products or molecular pathways to the expression of behavior. This is especially true for social behavior, which being dependent on interactions and communication between individuals is even further removed from molecular-level events. In this review, we discuss how behavioral models can aid molecular analyses of social behavior. Various models of behavior exist, each of which suggest strategies to dissect complex behavior into simpler behavioral 'modules.' The resulting modules are easier to relate to neural processes and thus suggest hypotheses for neural and molecular function. Here we discuss how three different models of behavior have facilitated understanding the molecular bases of aspects of social behavior. We discuss the response threshold model and two different approaches to modeling motivation, the state space model and models of reinforcement and reward processing. The examples we have chosen illustrate how models can generate testable hypotheses for neural and molecular function and also how molecular analyses probe the validity of a model of behavior. We do not champion one model over another; rather, our examples illustrate how modeling and molecular analyses can be synergistic in exploring the molecular bases of social behavior.     

A corpus of full-text journal articles is a robust evaluation tool for revealing differences in performance of biomedical natural language processing tools

ABSTRACT: BACKGROUND: We introduce the linguistic annotation of a corpus of 97 full-text biomedical publications, known as the Colorado Richly Annotated Full Text (CRAFT) corpus. We further assess the performance of existing tools for performing sentence splitting, tokenization, syntactic parsing, and named entity recognition on this corpus. RESULTS: Many biomedical natural language processing systems demonstrated large differences between their previously published results and their performance on the CRAFT corpus when tested with the publicly available models or rule sets. Trainable systems differed widely with respect to their ability to build high-performing models based on this data. CONCLUSIONS: The finding that some systems were able to train high-performing models based on this corpus is additional evidence, beyond high inter-annotator agreement, that the quality of the CRAFT corpus is high. The overall poor performance of various systems indicates that considerable work needs to be done to enable natural language processing systems to work well when the input is full-text journal articles. The CRAFT corpus provides avaluable resource to the biomedical natural language processing community for evaluation and training of new models for biomedical full text publications.     

Left inferior frontal activations depending on the canonicity determined by the argument structures of ditransitive sentences: an MEG study

To elucidate the relationships between syntactic and semantic processes, one interesting question is how syntactic structures are constructed by the argument structure of a verb, where each argument corresponds to a semantic role of each noun phrase (NP). Here we examined the effects of possessivity [sentences with or without a possessor] and canonicity [canonical or noncanonical word orders] using Japanese ditransitive sentences. During a syntactic decision task, the syntactic structure of each sentence would be constructed in an incremental manner based on the predicted argument structure of the ditransitive verb in a verb-final construction. Using magnetoencephalography, we found a significant canonicity effect on the current density in the left inferior frontal gyrus (IFG) at 530-550 ms after the verb onset. This effect was selective to canonical sentences, and significant even when the precedent NP was physically identical. We suggest that the predictive effects associated with syntactic processing became larger for canonical sentences, where the NPs and verb were merged with a minimum structural distance, leading to the left IFG activations. For monotransitive and intransitive verbs, in which structural computation of the sentences was simpler than that of ditransitive sentences, we observed a significant effect selective to noncanonical sentences in the temporoparietal regions during 480-670 ms. This effect probably reflects difficulty in semantic processing of noncanonical sentences. These results demonstrate that the left IFG plays a predictive role in syntactic processing, which depends on the canonicity determined by argument structures, whereas other temporoparietal regions would subserve more semantic aspects of sentence processing.     

How hierarchical is language use?

It is generally assumed that hierarchical phrase structure plays a central role in human language. However, considerations of simplicity and evolutionary continuity suggest that hierarchical structure should not be invoked too hastily. Indeed, recent neurophysiological, behavioural and computational studies show that sequential sentence structure has considerable explanatory power and that hierarchical processing is often not involved. In this paper, we review evidence from the recent literature supporting the hypothesis that sequential structure may be fundamental to the comprehension, production and acquisition of human language. Moreover, we provide a preliminary sketch outlining a non-hierarchical model of language use and discuss its implications and testable predictions. If linguistic phenomena can be explained by sequential rather than hierarchical structure, this will have considerable impact in a wide range of fields, such as linguistics, ethology, cognitive neuroscience, psychology and computer science.     

Neuro-fuzzy prediction of uricase production

Recent biotechnology requires implementation of new modelling methods based on knowledge principles and learning structures, comprised in fuzzy knowledge-based systems (FKBS), neural networks (NN) and different hybrid methods. The intelligent modelling approaches solve sufficiently a very important problem - processing of scarce, uncertainty and incomplete numerical and linguistic information about multivariate non-linear and non-stationary systems as well as biotechnological processes. The paper deals with prediction of an enzyme oxidizing uric acid to alantoin - the uricase, produced by Candida utilis 90-12 employing neuro-fuzzy knowledge-based approach. The implemented predictive technique exploits the fact that the fuzzy model can be seen as a network structure, similar to artificial NN, which on computational level assure a high model accuracy. The predictors implemented are four different by nature variables. The developed predictive model shows that best predictors of uricase production are biomass and limiting substrate concentrations.     

Aedes-AI: Neural network models of mosquito abundance

We present artificial neural networks as a feasible replacement for a mechanistic model of mosquito abundance. We develop a feed-forward neural network, a long short-term memory recurrent neural network, and a gated recurrent unit network. We evaluate the networks in their ability to replicate the spatiotemporal features of mosquito populations predicted by the mechanistic model, and discuss how augmenting the training data with time series that emphasize specific dynamical behaviors affects model performance. We conclude with an outlook on how such equation-free models may facilitate vector control or the estimation of disease risk at arbitrary spatial scales.     

Modulation of Arm Reaching Movements during Processing of Arm/Hand-Related Action Verbs with and without Emotional Connotation

The theory of embodied language states that language comprehension relies on an internal reenactment of the sensorimotor experience associated with the processed word or sentence. Most evidence in support of this hypothesis had been collected using linguistic material without any emotional connotation. For instance, it had been shown that processing of arm-related verbs, but not of those leg-related verbs, affects the planning and execution of reaching movements; however, at present it is unknown whether this effect is further modulated by verbs evoking an emotional experience. Showing such a modulation might shed light on a very debated issue, i.e. the way in which the emotional meaning of a word is processed. To this end, we assessed whether processing arm/hand-related verbs describing actions with negative connotations (e.g. to stab) affects reaching movements differently from arm/hand-related verbs describing actions with neutral connotation (e.g. to comb). We exploited a go/no-go paradigm in which healthy participants were required to perform arm-reaching movements toward a target when verbs expressing emotional hand actions, neutral hand actions or foot actions were shown, and to refrain from moving when no-effector-related verbs were presented. Reaction times and percentages of errors increased when the verb involved the same effector as used to give the response. However, we also found that the size of this interference decreased when the arm/hand-related verbs had a negative emotional connotation. Crucially, we show that such modulation only occurred when the verb semantics had to be retrieved. These results suggest that the comprehension of negatively valenced verbs might require the simultaneous reenactment of the neural circuitry associated with the processing of the emotion evoked by their meaning and of the neural circuitry associated with their motor features.     

Matching consumer feeding behaviours and resource traits: a fourth‐corner problem in food‐web theory

For decades, food web theory has proposed phenomenological models for the underlying structure of ecological networks. Generally, these models rely on latent niche variables that match the feeding behaviour of consumers with their resource traits. In this paper, we used a comprehensive database to evaluate different hypotheses on the best dependency structure of trait‐matching patterns between consumers and resource traits. We found that consumer feeding behaviours had complex interactions with resource traits; however, few dimensions (i.e. latent variables) could reproduce the trait‐matching patterns. We discuss our findings in the light of three food web models designed to reproduce the multidimensionality of food web data; additionally, we discuss how using species traits clarify food webs beyond species pairwise interactions and enable studies to infer ecological generality at larger scales, despite potential taxonomic differences, variations in ecological conditions and differences in species abundance between communities.     

A multi-differential neuromorphic approach to motion detection

This paper presents a multi-differential neuromorphic approach to motion detection. The model is based evidence for a differential operators interpretation of the properties of the cortical motion pathway. We discuss how this strategy, which provides a robust measure of speed for a range of types of image motion using a single computational mechanism, forms a useful framework in which to develop future neuromorphic motion systems. We also discuss both our approaches to developing computational motion models, and constraints in the design strategy for transferring motion models to other domains of early visual processing.     

Processing coordinate structures in Chinese: evidence from eye movements

This article reports the results of an eye-tracking experiment that investigated the processing of coordinate structures in Chinese sentence comprehension. The study tracked the eye movements of native Chinese readers as they read sentences consisting of two independent clauses connected by the word huo zhe. The data strongly confirmed readers' preference for an initial noun phrase (NP)-coordination parsing in Chinese coordination structure. When huo zhe was absent from the beginning of a sentence, we identified a cost associated with abandoning the NP-coordination analysis, which was evident with regard to the second NP when the coordination was unambiguous. Otherwise, this cost was evident with regard to the verb, the syntactically disambiguating region, when the coordination was ambiguous. However, the presence of a sentence-initial huo zhe reduced reading times and regressions in the huo zhe NP and the verb regions. We believe that the word huo zhe at the beginning of a sentence helps the reader predict that the sentence contains a parallel structure. Before the corresponding phrases appear, the readers can use the word huo zhe and the language structure thereafter to predicatively construct the syntactic structure. Such predictive capability can eliminate the reader's preference for NP-coordination analysis. Implications for top-down parsing theory and models of initial syntactic analysis and reanalysis are discussed.     

A single microcircuit with multiple functions: state dependent information processing in the hippocampus

Many neural circuits process information in multiple distinct modes. For example, the hippocampus is involved in memory encoding, retrieval, and consolidation processes. These different mnemonic computations require processing of differing balances of current sensory input and previously stored associations. Here we explore patterns of activity in hippocampal output area CA1 associated with different information processing states. We discuss the evidence linking these patterns to specific inputs to CA1 and describe behavioral factors that are related to the balance of synaptic drive. We suggest that understanding the factors that influence information flow in the hippocampal circuit could provide important new insights into how neural circuits are reconfigured on the fly to perform different functions at different times.     

Hallucinations on demand: the utility of experimentally induced phenomena in hallucination research

Despite the desire to delve deeper into hallucinations of all types, methodological obstacles have frustrated development of more rigorous quantitative experimental techniques, thereby hampering research progress. Here, we discuss these obstacles and, with reference to visual phenomena, argue that experimentally induced phenomena (e.g. hallucinations induced by flickering light and classical conditioning) can bring hallucinations within reach of more objective behavioural and neural measurement. Expanding the scope of hallucination research raises questions about which phenomena qualify as hallucinations, and how to identify phenomena suitable for use as laboratory models of hallucination. Due to the ambiguity inherent in current hallucination definitions, we suggest that the utility of phenomena for use as laboratory hallucination models should be represented on a continuous spectrum, where suitability varies with the degree to which external sensory information constrains conscious experience. We suggest that existing strategies that group pathological hallucinations into meaningful subtypes based on hallucination characteristics (including phenomenology, disorder and neural activity) can guide extrapolation from hallucination models to other hallucinatory phenomena. Using a spectrum of phenomena to guide scientific hallucination research should help unite the historically separate fields of psychophysics, cognitive neuroscience and clinical research to better understand and treat hallucinations, and inform models of consciousness.This article is part of the theme issue ‘Offline perception: voluntary and spontaneous perceptual experiences without matching external stimulation’.     

Extensive feature detection of N-terminal protein sorting signals

MOTIVATION: The prediction of localization sites of various proteins is an important and challenging problem in the field of molecular biology. TargetP, by Emanuelsson et al. (J. Mol. Biol., 300, 1005-1016, 2000) is a neural network based system which is currently the best predictor in the literature for N-terminal sorting signals. One drawback of neural networks, however, is that it is generally difficult to understand and interpret how and why they make such predictions. In this paper, we aim to generate simple and interpretable rules as predictors, and still achieve a practical prediction accuracy. We adopt an approach which consists of an extensive search for simple rules and various attributes which is partially guided by human intuition. RESULTS: We have succeeded in finding rules whose prediction accuracies come close to that of TargetP, while still retaining a very simple and interpretable form. We also discuss and interpret the discovered rules.     

Using deep neural networks to evaluate object vision tasks in rats

In the last two decades rodents have been on the rise as a dominant model for visual neuroscience. This is particularly true for earlier levels of information processing, but a number of studies have suggested that also higher levels of processing such as invariant object recognition occur in rodents. Here we provide a quantitative and comprehensive assessment of this claim by comparing a wide range of rodent behavioral and neural data with convolutional deep neural networks. These networks have been shown to capture hallmark properties of information processing in primates through a succession of convolutional and fully connected layers. We find that performance on rodent object vision tasks can be captured using low to mid-level convolutional layers only, without any convincing evidence for the need of higher layers known to simulate complex object recognition in primates. Our approach also reveals surprising insights on assumptions made before, for example, that the best performing animals would be the ones using the most abstract representations–which we show to likely be incorrect. Our findings suggest a road ahead for further studies aiming at quantifying and establishing the richness of representations underlying information processing in animal models at large.     

How the emotional content of discourse affects language comprehension

Emotion effects on cognition have often been reported. However, only few studies investigated emotional effects on subsequent language processing, and in most cases these effects were induced by non-linguistic stimuli such as films, faces, or pictures. Here, we investigated how a paragraph of positive, negative, or neutral emotional valence affects the processing of a subsequent emotionally neutral sentence, which contained either semantic, syntactic, or no violation, respectively, by means of event-related brain potentials (ERPs). Behavioral data revealed strong effects of emotion; error rates and reaction times increased significantly in sentences preceded by a positive paragraph relative to negative and neutral ones. In ERPs, the N400 to semantic violations was not affected by emotion. In the syntactic experiment, however, clear emotion effects were observed on ERPs. The left anterior negativity (LAN) to syntactic violations, which was not visible in the neutral condition, was present in the negative and positive conditions. This is interpreted as reflecting modulatory effects of prior emotions on syntactic processing, which is discussed in the light of three alternative or complementary explanations based on emotion-induced cognitive styles, working memory, and arousal models. The present effects of emotion on the LAN are especially remarkable considering that syntactic processing has often been regarded as encapsulated and autonomous.