How semantic biases in simple adjacencies affect learning a complex structure with non-adjacencies in AGL: a statistical account

A major theoretical debate in language acquisition research regards the learnability of hierarchical structures. The artificial grammar learning methodology is increasingly influential in approaching this question. Studies using an artificial centre-embedded A(n)B(n) grammar without semantics draw conflicting conclusions. This study investigates the facilitating effect of distributional biases in simple AB adjacencies in the input sample--caused in natural languages, among others, by semantic biases-on learning a centre-embedded structure. A mathematical simulation of the linguistic input and the learning, comparing various distributional biases in AB pairs, suggests that strong distributional biases might help us to grasp the complex A(n)B(n) hierarchical structure in a later stage. This theoretical investigation might contribute to our understanding of how distributional features of the input--including those caused by semantic variation--help learning complex structures in natural languages.     

Artificial Grammar Learning of Melody Is Constrained by Melodic Inconsistency: Narmour's Principles Affect Melodic Learning

Considerable evidence suggests that people acquire artificial grammars incidentally and implicitly, an indispensable capacity for the acquisition of music or language. However, less research has been devoted to exploring constraints affecting incidental learning. Within the domain of music, the extent to which Narmour''s (1990) melodic principles affect implicit learning of melodic structure was experimentally explored. Extending previous research (Rohrmeier, Rebuschat & Cross, 2011), the identical finite-state grammar is employed having terminals (the alphabet) manipulated so that melodies generated systematically violated Narmour''s principles. Results indicate that Narmour-inconsistent melodic materials impede implicit learning. This further constitutes a case in which artificial grammar learning is affected by prior knowledge or processing constraints.     

Explore the Functional Connectivity between Brain Regions during a Chemistry Working Memory Task

Previous studies have rarely examined how temporal dynamic patterns, event-related coherence, and phase-locking are related to each other. This study assessed reaction-time-sorted spectral perturbation and event-related spectral perturbation in order to examine the temporal dynamic patterns in the frontal midline (F), central parietal (CP), and occipital (O) regions during a chemistry working memory task at theta, alpha, and beta frequencies. Furthermore, the functional connectivity between F-CP, CP-O, and F-O were assessed by component event-related coherence (ERCoh) and component phase-locking (PL) at different frequency bands. In addition, this study examined whether the temporal dynamic patterns are consistent with the functional connectivity patterns across different frequencies and time courses. Component ERCoh/PL measured the interactions between different independent components decomposed from the scalp EEG, mixtures of time courses of activities arising from different brain, and artifactual sources. The results indicate that the O and CP regions’ temporal dynamic patterns are similar to each other. Furthermore, pronounced component ERCoh/PL patterns were found to exist between the O and CP regions across each stimulus and probe presentation, in both theta and alpha frequencies. The consistent theta component ERCoh/PL between the F and O regions was found at the first stimulus and after probe presentation. These findings demonstrate that temporal dynamic patterns at different regions are in accordance with the functional connectivity patterns. Such coordinated and robust EEG temporal dynamics and component ERCoh/PL patterns suggest that these brain regions’ neurons work together both to induce similar event-related spectral perturbation and to synchronize or desynchronize simultaneously in order to swiftly accomplish a particular goal. The possible mechanisms for such distinct component phase-locking and coherence patterns were also further discussed.     

Temporal difference models and reward-related learning in the human brain

Temporal difference learning has been proposed as a model for Pavlovian conditioning, in which an animal learns to predict delivery of reward following presentation of a conditioned stimulus (CS). A key component of this model is a prediction error signal, which, before learning, responds at the time of presentation of reward but, after learning, shifts its response to the time of onset of the CS. In order to test for regions manifesting this signal profile, subjects were scanned using event-related fMRI while undergoing appetitive conditioning with a pleasant taste reward. Regression analyses revealed that responses in ventral striatum and orbitofrontal cortex were significantly correlated with this error signal, suggesting that, during appetitive conditioning, computations described by temporal difference learning are expressed in the human brain.     

Sleep Promotes the Extraction of Grammatical Rules

Grammar acquisition is a high level cognitive function that requires the extraction of complex rules. While it has been proposed that offline time might benefit this type of rule extraction, this remains to be tested. Here, we addressed this question using an artificial grammar learning paradigm. During a short-term memory cover task, eighty-one human participants were exposed to letter sequences generated according to an unknown artificial grammar. Following a time delay of 15 min, 12 h (wake or sleep) or 24 h, participants classified novel test sequences as Grammatical or Non-Grammatical. Previous behavioral and functional neuroimaging work has shown that classification can be guided by two distinct underlying processes: (1) the holistic abstraction of the underlying grammar rules and (2) the detection of sequence chunks that appear at varying frequencies during exposure. Here, we show that classification performance improved after sleep. Moreover, this improvement was due to an enhancement of rule abstraction, while the effect of chunk frequency was unaltered by sleep. These findings suggest that sleep plays a critical role in extracting complex structure from separate but related items during integrative memory processing. Our findings stress the importance of alternating periods of learning with sleep in settings in which complex information must be acquired.     

The evolutionary dynamics of grammar acquisition

Grammar is the computational system of language. It is a set of rules that specifies how to construct sentences out of words. Grammar is the basis of the unlimited expressibility of human language. Children acquire the grammar of their native language without formal education simply by hearing a number of sample sentences. Children could not solve this learning task if they did not have some pre-formed expectations. In other words, children have to evaluate the sample sentences and choose one grammar out of a limited set of candidate grammars. The restricted search space and the mechanism which allows to evaluate the sample sentences is called universal grammar. Universal grammar cannot be learned; it must be in place when the learning process starts. In this paper, we design a mathematical theory that places the problem of language acquisition into an evolutionary context. We formulate equations for the population dynamics of communication and grammar learning. We ask how accurate children have to learn the grammar of their parents' language for a population of individuals to evolve and maintain a coherent grammatical system. It turns out that there is a maximum error tolerance for which a predominant grammar is stable. We calculate the maximum size of the search space that is compatible with coherent communication in a population. Thus, we specify the conditions for the evolution of universal grammar.     

Timing Rhythms: Perceived Duration Increases with a Predictable Temporal Structure of Short Interval Fillers

Variations in the temporal structure of an interval can lead to remarkable differences in perceived duration. For example, it has previously been shown that isochronous intervals, that is, intervals filled with temporally regular stimuli, are perceived to last longer than intervals left empty or filled with randomly timed stimuli. Characterizing the extent of such distortions is crucial to understanding how duration perception works. One account to explain effects of temporal structure is a non-linear accumulator-counter mechanism reset at the beginning of every subinterval. An alternative explanation based on entrainment to regular stimulation posits that the neural response to each filler stimulus in an isochronous sequence is amplified and a higher neural response may lead to an overestimation of duration. If entrainment is the key that generates response amplification and the distortions in perceived duration, then any form of predictability in the temporal structure of interval fillers should lead to the perception of an interval that lasts longer than a randomly filled one. The present experiments confirm that intervals filled with fully predictable rhythmically grouped stimuli lead to longer perceived duration than anisochronous intervals. No general over- or underestimation is registered for rhythmically grouped compared to isochronous intervals. However, we find that the number of stimuli in each group composing the rhythm also influences perceived duration. Implications of these findings for a non-linear clock model as well as a neural response magnitude account of perceived duration are discussed.     

When It Hurts (and Helps) to Try: The Role of Effort in Language Learning

Compared to children, adults are bad at learning language. This is counterintuitive; adults outperform children on most measures of cognition, especially those that involve effort (which continue to mature into early adulthood). The present study asks whether these mature effortful abilities interfere with language learning in adults and further, whether interference occurs equally for aspects of language that adults are good (word-segmentation) versus bad (grammar) at learning. Learners were exposed to an artificial language comprised of statistically defined words that belong to phonologically defined categories (grammar). Exposure occurred under passive or effortful conditions. Passive learners were told to listen while effortful learners were instructed to try to 1) learn the words, 2) learn the categories, or 3) learn the category-order. Effortful learners showed an advantage for learning words while passive learners showed an advantage for learning the categories. Effort can therefore hurt the learning of categories.     

The minimalist grammar of action

Language and action have been found to share a common neural basis and in particular a common 'syntax', an analogous hierarchical and compositional organization. While language structure analysis has led to the formulation of different grammatical formalisms and associated discriminative or generative computational models, the structure of action is still elusive and so are the related computational models. However, structuring action has important implications on action learning and generalization, in both human cognition research and computation. In this study, we present a biologically inspired generative grammar of action, which employs the structure-building operations and principles of Chomsky's Minimalist Programme as a reference model. In this grammar, action terminals combine hierarchically into temporal sequences of actions of increasing complexity; the actions are bound with the involved tools and affected objects and are governed by certain goals. We show, how the tool role and the affected-object role of an entity within an action drives the derivation of the action syntax in this grammar and controls recursion, merge and move, the latter being mechanisms that manifest themselves not only in human language, but in human action too.     

统计学习的认知机制及其神经基础

本文主要探讨了人类统计学习研究的起源和发展、影响统计学习的因素及统计学习的脑神经基础.人类统计学习的研究起源于20世纪90年代关于婴儿听觉语音流切分的实验,之后,大量研究又考察了对听觉非语言信息及视觉图形的统计学习.这些研究表明,统计学习是一种具有一般性的、涉及发现和提取规则的学习类型,但是不同范畴的统计学习又受到与特定范畴有关因素的制约.例如,语言统计学习受到语言因素的制约,而非语言统计学习与呈现刺激的特征以及呈现的通道有关.近些年,研究者开始采用事件相关电位(event-related potentials,ERPs)技术考察统计学习的时间进程,以及采用功能核磁共振成像(functional magnetic resonance imaging,fMRI)技术考察统计学习的脑激活模式.ERPs研究较一致地发现400 ms左右产生的负波与统计信息的提取有关,而脑成像研究结果表明,与统计学习相关的脑区有大脑左侧颞上回(superiortemporal gyrus)、右侧纹状体(right striatum)和右侧颞叶内侧记忆系统(right medial temporal memory system).     

Second language processing shows increased native-like neural responses after months of no exposure

Although learning a second language (L2) as an adult is notoriously difficult, research has shown that adults can indeed attain native language-like brain processing and high proficiency levels. However, it is important to then retain what has been attained, even in the absence of continued exposure to the L2--particularly since periods of minimal or no L2 exposure are common. This event-related potential (ERP) study of an artificial language tested performance and neural processing following a substantial period of no exposure. Adults learned to speak and comprehend the artificial language to high proficiency with either explicit, classroom-like, or implicit, immersion-like training, and then underwent several months of no exposure to the language. Surprisingly, proficiency did not decrease during this delay. Instead, it remained unchanged, and there was an increase in native-like neural processing of syntax, as evidenced by several ERP changes--including earlier, more reliable, and more left-lateralized anterior negativities, and more robust P600s, in response to word-order violations. Moreover, both the explicitly and implicitly trained groups showed increased native-like ERP patterns over the delay, indicating that such changes can hold independently of L2 training type. The results demonstrate that substantial periods with no L2 exposure are not necessarily detrimental. Rather, benefits may ensue from such periods of time even when there is no L2 exposure. Interestingly, both before and after the delay the implicitly trained group showed more native-like processing than the explicitly trained group, indicating that type of training also affects the attainment of native-like processing in the brain. Overall, the findings may be largely explained by a combination of forgetting and consolidation in declarative and procedural memory, on which L2 grammar learning appears to depend. The study has a range of implications, and suggests a research program with potentially important consequences for second language acquisition and related fields.     

Event-related brain potential of human newborns to pitch change of an acoustic stimulus

We report here event-related potentials (ERPs) of human newborns to occasional pitch changes in a repetitive sequence of tone pips. These pitch changes elicited a large slow negative ERP component which resembles the mismatch negativity (MMN) generated by the adult brain under similar conditions. This MMN-type of negativity in newborns suggests that already at this early ontogenetic stage the brain monitors the acoustic environment for a possible change in any of its repetitive aspects. Apart from its theoretical interest, this finding might provide a new way to test the development of the central nervous system and to diagnose cerebral dysfunction at a very early stage.     

Brain Responses before and after Intensive Second Language Learning: Proficiency Based Changes and First Language Background Effects in Adult Learners

This longitudinal study tracked the neuro-cognitive changes associated with second language (L2) grammar learning in adults in order to investigate how L2 processing is shaped by a learner’s first language (L1) background and L2 proficiency. Previous studies using event-related potentials (ERPs) have argued that late L2 learners cannot elicit a P600 in response to L2 grammatical structures that do not exist in the L1 or that are different in the L1 and L2. We tested whether the neuro-cognitive processes underlying this component become available after intensive L2 instruction. Korean- and Chinese late-L2-learners of English were tested at the beginning and end of a 9-week intensive English-L2 course. ERPs were recorded while participants read English sentences containing violations of regular past tense (a grammatical structure that operates differently in Korean and does not exist in Chinese). Whereas no P600 effects were present at the start of instruction, by the end of instruction, significant P600s were observed for both L1 groups. Latency differences in the P600 exhibited by Chinese and Korean speakers may be attributed to differences in L1–L2 reading strategies. Across all participants, larger P600 effects at session 2 were associated with: 1) higher levels of behavioural performance on an online grammaticality judgment task; and 2) with correct, rather than incorrect, behavioural responses. These findings suggest that the neuro-cognitive processes underlying the P600 (e.g., “grammaticalization”) are modulated by individual levels of L2 behavioural performance and learning.     

Language learning: How much evidence does a child need in order to learn to speak grammatically?

In order to learn grammar from a finite amount of evidence, children must begin with in-built expectations of what is grammatical. They clearly are not born, however, with fully developed grammars. Thus early language development involves refinement of the grammar hypothesis until a target grammar is learnt. Here we address the question of how much evidence is required for this refinement process, by considering two standard learning algorithms and a third algorithm which is presumably as efficient as a child for some value of its memory capacity. We reformulate this algorithm in the context of Chomsky’s ‘principles and parameters’ and show that it is possible to bound the amount of evidence required to almost certainly speak almost grammatically.     

Reliable short-term memory in the trion model: toward a cortical language and grammar

 It has previously been shown that Hebb learning in a single column in the trion model of cortical organization occurs by selection. Motivated by von Neumann's solution for obtaining reliability and by models of circulating cortical activity, we introduce Hebb intercolumnar couplings to achieve dramatic enhancements in reliability in the firing of connected columns. In order for these enhancements to occur, specific temporal phase differences must exist between the same inherent spatial-temporal memory patterns in connected columns. We then generalize the criteria of large enhancements in the global firing of the entire connected columnar network to investigate the case when different inherent memory patterns are in the columns. The spatial rotations as well as the temporal phases now are crucial. Only certain combinations of inherent memory patterns meet these criteria with the symmetry properties playing a major role. The columnar order of these memory patterns not in the same symmetry family can be extremely important. This yields the first higher-level architecture of a cortical language and grammar within the trion model. The implications of this result with regard to an innate human language and grammar are discussed. Received: 14 June 2000 / Accepted in revised form: 25 July 2000     

Comparative analysis of changes in short EEG segments during music perception based on event-related synchronization/desynchronization and wavelet synchronicity

The goal of the pilot study was to analyze the characteristics of changes in short EEG segments recorded from 32 sites during perception of musical melodies by healthy subjects depending on logical (recognition) and emotional (pleasant/unpleasant) estimation of the melody. For this purpose, the changes in event-related synchronization/desynchronization and the indices of wavelet synchrony of EEG responses were compared in 31 healthy subjects (18 to 60 years old). It has been shown that melody recognition during logical estimation of music is accompanied by event-related desynchronization in the left frontal-parietal-temporal area. Emotional estimation of a melody is characterized by event-related synchronization in the left frontal-temporal area for pleasant melodies, desynchronization in the temporal area for unpleasant melodies, and desynchronization in the occipital area for melodies inducing no emotional response. The analysis of EEG wavelet synchronization characterizing reactive changes in the interaction between cortical areas shows that the most distinct topographic differences are associated with the type of music processing: logical (familiar/unfamiliar) or emotional (pleasant/unpleasant). The changes in interhemispheric connections between the associative cortical areas (central, frontal, temporal) are greater during emotional estimation, while the changes in inter- and intrahemispheric connections between the projection areas of the acoustic analyzer (temporal area) are greater during logical estimation. It is assumed that the revealed event-related synchronization/desynchronization is most likely to reflect the activation component of musical fragment estimation, whereas wavelet analysis provides insight into the character of musical stimulus processing.     

Brain networks of explicit and implicit learning

Are explicit versus implicit learning mechanisms reflected in the brain as distinct neural structures, as previous research indicates, or are they distinguished by brain networks that involve overlapping systems with differential connectivity? In this functional MRI study we examined the neural correlates of explicit and implicit learning of artificial grammar sequences. Using effective connectivity analyses we found that brain networks of different connectivity underlie the two types of learning: while both processes involve activation in a set of cortical and subcortical structures, explicit learners engage a network that uses the insula as a key mediator whereas implicit learners evoke a direct frontal-striatal network. Individual differences in working memory also differentially impact the two types of sequence learning.     

Why Movement Is Captured by Music,but Less by Speech: Role of Temporal Regularity

Music has a pervasive tendency to rhythmically engage our body. In contrast, synchronization with speech is rare. Music’s superiority over speech in driving movement probably results from isochrony of musical beats, as opposed to irregular speech stresses. Moreover, the presence of regular patterns of embedded periodicities (i.e., meter) may be critical in making music particularly conducive to movement. We investigated these possibilities by asking participants to synchronize with isochronous auditory stimuli (target), while music and speech distractors were presented at one of various phase relationships with respect to the target. In Exp. 1, familiar musical excerpts and fragments of children poetry were used as distractors. The stimuli were manipulated in terms of beat/stress isochrony and average pitch to achieve maximum comparability. In Exp. 2, the distractors were well-known songs performed with lyrics, on a reiterated syllable, and spoken lyrics, all having the same meter. Music perturbed synchronization with the target stimuli more than speech fragments. However, music superiority over speech disappeared when distractors shared isochrony and the same meter. Music’s peculiar and regular temporal structure is likely to be the main factor fostering tight coupling between sound and movement.     

Cognitive components of regularity processing in the auditory domain

Background

Music-syntactic irregularities often co-occur with the processing of physical irregularities. In this study we constructed chord-sequences such that perceived differences in the cognitive processing between regular and irregular chords could not be due to the sensory processing of acoustic factors like pitch repetition or pitch commonality (the major component of ‘sensory dissonance’).

Methodology/Principal Findings

Two groups of subjects (musicians and nonmusicians) were investigated with electroencephalography (EEG). Irregular chords elicited an early right anterior negativity (ERAN) in the event-related brain potentials (ERPs). The ERAN had a latency of around 180 ms after the onset of the music-syntactically irregular chords, and had maximum amplitude values over right anterior electrode sites.

Conclusions/Significance

Because irregular chords were hardly detectable based on acoustical factors (such as pitch repetition and sensory dissonance), this ERAN effect reflects for the most part cognitive (not sensory) components of regularity-based, music-syntactic processing. Our study represents a methodological advance compared to previous ERP-studies investigating the neural processing of music-syntactically irregular chords.