Perceptual constraints and the learnability of simple grammars

Rules generalization in children with dyslexia

BACKGROUND

Rule learning (RL) is the ability to extract and generalize higher-order repetition-based structures. Children with Developmental Dyslexia (DD) often report difficulties in learning complex regularities in sequential stimuli, which might be due to the complexity of the rule to be learned. Learning high-order repetition-based rules represents a building block for the development of language skills.

AIMS

This study investigates the ability to extract and generalize simple, repetition-based visual rules (e.g., ABA) in 8-11-year-old children without (TD) and with a diagnosis of Development Dyslexia (DD) and its relationship with language and reading skills.

METHOD

Using a forced-choice paradigm, children were first exposed to a visual sequence containing a repetition-based rule (e.g., ABA) and were then asked to recognize familiar and novel rules generated by new visual elements. Standardized language and reading tests were also administered to both groups.

RESULTS

The accuracy in recognizing rules was above chance for both groups, even though DD children were less accurate than TD children, suggesting a less efficient RL mechanism in the DD group. Moreover, visual RL was positively correlated with both language and reading skills.

CONCLUSION

These results further confirm the crucial role of RL in the acquisition of linguistic skills and mastering reading abilities.

Learnability Advantage of Segmental Repetitions in Word Learning

To date, research on wordform learning biases has mostly focused on language-dependent factors, such as the phonotactics and neighborhood density of the language(s) known by the learner. Domain-general biases, by contrast, have received little attention. In this study, we focus on one such bias-an advantage for string-internal repetitions-and examine its effects on wordform learning. Importantly, we consider whether any type of segmental repetition is equally beneficial for word recall, or whether learning is favored more or only by repeated consonants, in line with previous research indicating that consonants play a larger role than vowels in lexical processing. In Experiment 1, adult English speakers learned artificial consonant-vowel-consonant-vowel words containing either a repeated consonant (e.g., /sesu/, "c-rep"), a repeated vowel (e.g., /sepe/, "v-rep"), or dissimilar consonants and vowels (e.g., /sepu/, "no-rep"). Recall results showed no advantage for v-reps but higher accuracy for c-reps compared with no-reps. In Experiment 2, participants performed a label preference task with the same stimuli. The results showed dispreference for both c-reps and v-reps relative to no-reps, indicating that the results of Experiment 1 are independent of wordlikeness effects. These outcomes reveal that there is a form-learning bias for words with identical consonants but not for words with identical vowels, suggesting that a domain-general advantage for repetitions within strings is modulated by a language-specific processing bias for consonants.

Exploring Abstract Pattern Representation in The Brain and Non-symbolic Neural Networks

Human cognitive and linguistic generativity depends on the ability to identify abstract relationships between perceptually dissimilar items. Marcus et al. (1999) found that human infants can rapidly discover and generalize patterns of syllable repetition (reduplication) that depend on the abstract property of identity, but simple recurrent neural networks (SRNs) could not. They interpreted these results as evidence that purely associative neural network models provide an inadequate framework for characterizing the fundamental generativity of human cognition. Here, we present a series of deep long short-term memory (LSTM) models that identify abstract syllable repetition patterns and words based on training with cochleagrams that represent auditory stimuli. We demonstrate that models trained to identify individual syllable trigram words and models trained to identify reduplication patterns discover representations that support classification of abstract repetition patterns. Simulations examined the effects of training categories (words vs. patterns) and pretraining to identify syllables, on the development of hidden node representations that support repetition pattern discrimination. Representational similarity analyses (RSA) comparing patterns of regional brain activity based on MRI-constrained MEG/EEG data to patterns of hidden node activation elicited by the same stimuli showed a significant correlation between brain activity localized in primarily posterior temporal regions and representations discovered by the models. These results suggest that associative mechanisms operating over discoverable representations that capture abstract stimulus properties account for a critical example of human cognitive generativity.

Abstract representations in temporal cortex support generative linguistic processing

Generativity, the ability to create and evaluate novel constructions, is a fundamental property of human language and cognition. The productivity of generative processes is determined by the scope of the representations they engage. Here we examine the neural representation of reduplication, a productive phonological process that can create novel forms through patterned syllable copying (e.g. ba-mih → ba-ba-mih, ba-mih-mih, or ba-mih-ba). Using MRI-constrained source estimates of combined MEG/EEG data collected during an auditory artificial grammar task, we identified localized cortical activity associated with syllable reduplication pattern contrasts in novel trisyllabic nonwords. Neural decoding analyses identified a set of predominantly right hemisphere temporal lobe regions whose activity reliably discriminated reduplication patterns evoked by untrained, novel stimuli. Effective connectivity analyses suggested that sensitivity to abstracted reduplication patterns was propagated between these temporal regions. These results suggest that localized temporal lobe activity patterns function as abstract representations that support linguistic generativity.

Statistical Learning in Vision

Vision and learning have long been considered to be two areas of research linked only distantly. However, recent developments in vision research have changed the conceptual definition of vision from a signal-evaluating process to a goal-oriented interpreting process, and this shift binds learning, together with the resulting internal representations, intimately to vision. In this review, we consider various types of learning (perceptual, statistical, and rule/abstract) associated with vision in the past decades and argue that they represent differently specialized versions of the fundamental learning process, which must be captured in its entirety when applied to complex visual processes. We show why the generalized version of statistical learning can provide the appropriate setup for such a unified treatment of learning in vision, what computational framework best accommodates this kind of statistical learning, and what plausible neural scheme could feasibly implement this framework. Finally, we list the challenges that the field of statistical learning faces in fulfilling the promise of being the right vehicle for advancing our understanding of vision in its entirety. Expected final online publication date for the Annual Review of Vision Science, Volume 8 is September 2022. Please see http://www.annualreviews.org/page/journal/pubdates for revised estimates.

Musical instrument familiarity affects statistical learning of tone sequences

Most listeners have an implicit understanding of the rules that govern how music unfolds over time. This knowledge is acquired in part through statistical learning, a robust learning mechanism that allows individuals to extract regularities from the environment. However, it is presently unclear how this prior musical knowledge might facilitate or interfere with the learning of novel tone sequences that do not conform to familiar musical rules. In the present experiment, participants listened to novel, statistically structured tone sequences composed of pitch intervals not typically found in Western music. Between participants, the tone sequences either had the timbre of artificial, computerized instruments or familiar instruments (piano or violin). Knowledge of the statistical regularities was measured as by a two-alternative forced choice recognition task, requiring discrimination between novel sequences that followed versus violated the statistical structure, assessed at three time points (immediately post-training, as well as one day and one week post-training). Compared to artificial instruments, training on familiar instruments resulted in reduced accuracy. Moreover, sequences from familiar instruments - but not artificial instruments - were more likely to be judged as grammatical when they contained intervals that approximated those commonly used in Western music, even though this cue was non-informative. Overall, these results demonstrate that instrument familiarity can interfere with the learning of novel statistical regularities, presumably through biasing memory representations to be aligned with Western musical structures. These results demonstrate that real-world experience influences statistical learning in a non-linguistic domain, supporting the view that statistical learning involves the continuous updating of existing representations, rather than the establishment of entirely novel ones.

Prosody facilitates learning the word order in a new language

One of the prominent ideas developed by Jacques Mehler and his colleagues was that perceptual tuning, present from birth on, enables infants, and language learners in general, to extract regularities from speech input. Here we discuss language learners'' ability to extract basic word order (VO or OV) structure from prosodic regularities in a language. The two are closely related: in phonological phrases of VO languages, the most prominent word is the rightmost one, and in OV languages, it is the leftmost one. In speech, this prominence is realized as extended duration, or as elevated pitch, sometimes combined with changes in intensity. When learning the first (L1) or the second language (L2), exposure to relevant rhythmic structure elicits implicit learning about syntactic structure, including the basic word order. However, it remains unclear whether triggering the learning process requires a certain level of familiarity with the relevant rhythm. It is moreover unknown whether prosodic information can help L2 learners to extract and learn the vocabulary of a new language. We tested Spanish- and Italian-speaking adults' ability to learn words from an artificial language with either non-native OV or native VO word order. The results show that learners used prosodic information to identify the most prominent words in short utterances when the artificial language was similar to the native language, with duration-based prominence in prosody and a VO word order. In contrast, when the artificial language had a non-native prominence marked by pitch alternations and an OV word order, prominent words were learned only after a three-day exposure to the relevant rhythmic structure. Thus, for adult L2 learners, only repeated exposure to the relevant prosody elicited learning new words from an unknown language with non-native prosodic marking, indicating that, with familiarity, prosodic cues can facilitate learning in L2.

A theory of memory for binary sequences: Evidence for a mental compression algorithm in humans

Working memory capacity can be improved by recoding the memorized information in a condensed form. Here, we tested the theory that human adults encode binary sequences of stimuli in memory using an abstract internal language and a recursive compression algorithm. The theory predicts that the psychological complexity of a given sequence should be proportional to the length of its shortest description in the proposed language, which can capture any nested pattern of repetitions and alternations using a limited number of instructions. Five experiments examine the capacity of the theory to predict human adults' memory for a variety of auditory and visual sequences. We probed memory using a sequence violation paradigm in which participants attempted to detect occasional violations in an otherwise fixed sequence. Both subjective complexity ratings and objective violation detection performance were well predicted by our theoretical measure of complexity, which simply reflects a weighted sum of the number of elementary instructions and digits in the shortest formula that captures the sequence in our language. While a simpler transition probability model, when tested as a single predictor in the statistical analyses, accounted for significant variance in the data, the goodness-of-fit with the data significantly improved when the language-based complexity measure was included in the statistical model, while the variance explained by the transition probability model largely decreased. Model comparison also showed that shortest description length in a recursive language provides a better fit than six alternative previously proposed models of sequence encoding. The data support the hypothesis that, beyond the extraction of statistical knowledge, human sequence coding relies on an internal compression using language-like nested structures.

A Simple, Biologically Plausible Feature Detector for Language Acquisition

Language has a complex grammatical system we still have to understand computationally and biologically. However, some evolutionarily ancient mechanisms have been repurposed for grammar so that we can use insight from other taxa into possible circuit-level mechanisms of grammar. Drawing upon recent evidence for the importance of disinhibitory circuits across taxa and brain regions, I suggest a simple circuit that explains the acquisition of core grammatical rules used in 85% of the world's languages: grammatical rules based on sameness/difference relations. This circuit acts as a sameness detector. “Different” items are suppressed through inhibition, but presenting two “identical” items leads to inhibition of inhibition. The items are thus propagated for further processing. This sameness detector thus acts as a feature detector for a grammatical rule. I suggest that having a set of feature detectors for elementary grammatical rules might make language acquisition feasible based on relatively simple computational mechanisms.

A review of computational models of basic rule learning: The neural-symbolic debate and beyond

We present a critical review of computational models of generalization of simple grammar-like rules, such as ABA and ABB. In particular, we focus on models attempting to account for the empirical results of Marcus et al. (Science, 283(5398), 77-80 1999). In that study, evidence is reported of generalization behavior by 7-month-old infants, using an Artificial Language Learning paradigm. The authors fail to replicate this behavior in neural network simulations, and claim that this failure reveals inherent limitations of a whole class of neural networks: those that do not incorporate symbolic operations. A great number of computational models were proposed in follow-up studies, fuelling a heated debate about what is required for a model to generalize. Twenty years later, this debate is still not settled. In this paper, we review a large number of the proposed models. We present a critical analysis of those models, in terms of how they contribute to answer the most relevant questions raised by the experiment. After identifying which aspects require further research, we propose a list of desiderata for advancing our understanding on generalization.

Preverbal infants utilize cross-modal semantic congruency in artificial grammar acquisition

Learning in a multisensory world is challenging as the information from different sensory dimensions may be inconsistent and confusing. By adulthood, learners optimally integrate bimodal (e.g. audio-visual, AV) stimulation by both low-level (e.g. temporal synchrony) and high-level (e.g. semantic congruency) properties of the stimuli to boost learning outcomes. However, it is unclear how this capacity emerges and develops. To approach this question, we examined whether preverbal infants were capable of utilizing high-level properties with grammar-like rule acquisition. In three experiments, we habituated pre-linguistic infants with an audio-visual (AV) temporal sequence that resembled a grammar-like rule (A-A-B). We varied the cross-modal semantic congruence of the AV stimuli (Exp 1: congruent syllables/faces; Exp 2: incongruent syllables/shapes; Exp 3: incongruent beeps/faces) while all the other low-level properties (e.g. temporal synchrony, sensory energy) were constant. Eight- to ten-month-old infants only learned the grammar-like rule from AV congruent stimuli pairs (Exp 1), not from incongruent AV pairs (Exp 2, 3). Our results show that similar to adults, preverbal infants' learning is influenced by a high-level multisensory integration gating system, pointing to a perceptual origin of bimodal learning advantage that was not previously acknowledged.

Are Words Easier to Learn From Infant- Than Adult-Directed Speech? A Quantitative Corpus-Based Investigation

We investigate whether infant-directed speech (IDS) could facilitate word form learning when compared to adult-directed speech (ADS). To study this, we examine the distribution of word forms at two levels, acoustic and phonological, using a large database of spontaneous speech in Japanese. At the acoustic level we show that, as has been documented before for phonemes, the realizations of words are more variable and less discriminable in IDS than in ADS. At the phonological level, we find an effect in the opposite direction: The IDS lexicon contains more distinctive words (such as onomatopoeias) than the ADS counterpart. Combining the acoustic and phonological metrics together in a global discriminability score reveals that the bigger separation of lexical categories in the phonological space does not compensate for the opposite effect observed at the acoustic level. As a result, IDS word forms are still globally less discriminable than ADS word forms, even though the effect is numerically small. We discuss the implication of these findings for the view that the functional role of IDS is to improve language learnability.

Reduplication facilitates early word segmentation

This study explores the possibility that early word segmentation is aided by infants' tendency to segment words with repeated syllables ('reduplication'). Twenty-four nine-month-olds were familiarized with passages containing one novel reduplicated word and one novel non-reduplicated word. Their central fixation times in response to these as well as new reduplicated and non-reduplicated words introduced at test showed that familiarized reduplicated words were segmented better than familiarized non-reduplicated words. These results demonstrate that infants are predisposed to segment words with repeated phonological elements, and suggest that register-specific words in infant-directed speech may have evolved in response to this learning bias.

Perceptual and Positional Saliencies Influence Children's Sequence Learning Differently with Age and Instructions at Test

There is growing evidence that, faced with a complex environment, participants subdivide the incoming information into small perceptual units, called chunks. Although statistical properties have been identified as playing a key role in chunking, we wanted to determine whether perceptual (repetitions) and positional (initial units) features might provide immediate guidance for the parsing of information into chunks. Children aged 5 and 8 years were exposed to sequences of 3, 4, or 5 colours. Sequence learning was assessed either through an explicit generation test (Experiment 1) or through a recognition test (Experiment 2). Experiment 1 showed that perceptual and positional saliencies benefited learning and that sensitivity to repetitions was age dependent and permitted the formation of longer chunks (trigrams) in the oldest children. Experiment 2 suggested that children became sensitive to perceptual and positional saliencies regardless of age and that the both types of saliencies supported the formation of longer chunks in the oldest children. The discussion focuses on the multiple factors intervening in sequence learning and their differential effects as a function of the instructions used at test to assess sequence learning.

Learning multiple rules simultaneously: Affixes are more salient than reduplications

Language learners encounter numerous opportunities to learn regularities, but need to decide which of these regularities to learn, because some are not productive in their native language. Here, we present an account of rule learning based on perceptual and memory primitives (Endress, Dehaene-Lambertz, & Mehler, Cognition, 105(3), 577–614, 2007; Endress, Nespor, & Mehler, Trends in Cognitive Sciences, 13(8), 348–353, 2009), suggesting that learners preferentially learn regularities that are more salient to them, and that the pattern of salience reflects the frequency of language features across languages. We contrast this view with previous artificial grammar learning research, which suggests that infants “choose” the regularities they learn based on rational, Bayesian criteria (Frank & Tenenbaum, Cognition, 120(3), 360–371, 2013; Gerken, Cognition, 98(3)B67–B74, 2006, Cognition, 115(2), 362–366, 2010). In our experiments, adult participants listened to syllable strings starting with a syllable reduplication and always ending with the same “affix” syllable, or to syllable strings starting with this “affix” syllable and ending with the “reduplication”. Both affixation and reduplication are frequently used for morphological marking across languages. We find three crucial results. First, participants learned both regularities simultaneously. Second, affixation regularities seemed easier to learn than reduplication regularities. Third, regularities in sequence offsets were easier to learn than regularities at sequence onsets. We show that these results are inconsistent with previous Bayesian rule learning models, but mesh well with the perceptual or memory primitives view. Further, we show that the pattern of salience revealed in our experiments reflects the distribution of regularities across languages. Ease of acquisition might thus be one determinant of the frequency of regularities across languages.

Artificial grammar learning in zebra finches and human adults: XYX versus XXY

Abstracting syntactic rules is critical to human language learning. It is debated whether this ability, already present in young infants, is human- and language specific or can also be found in non-human animals, indicating it may arise from more general cognitive mechanisms. Current studies are often ambiguous and few have directly compared rule learning by humans and non-human animals. In a series of discrimination experiments, we presented zebra finches and human adults with comparable training and tests with the same artificial stimuli consisting of XYX and XXY structures, in which X and Y were zebra finch song elements. Zebra finches readily discriminated the training stimuli. Some birds also discriminated novel stimuli when these were composed of familiar element types, but none of the birds generalized the discrimination to novel element types. We conclude that zebra finches show evidence of simple rule abstraction related to positional learning, suggesting stimulus-bound generalization, but found no evidence for a more abstract rule generalization. This differed from the human adults, who categorized novel stimuli consisting of novel element types into different groups according to their structure. The limited abilities for rule abstraction in zebra finches may indicate what the precursors of more complex abstraction as found in humans may have been like.

Generalization of visual regularities in newly hatched chicks (Gallus gallus)

Evidence of learning and generalization of visual regularities in a newborn organism is provided in the present research. Domestic chicks have been trained to discriminate visual triplets of simultaneously presented shapes, implementing AAB versus ABA (Experiment 1), AAB versus ABB and AAB versus BAA (Experiment 2). Chicks distinguished pattern-following and pattern-violating novel test triplets in all comparisons, showing no preference for repetition-based patterns. The animals generalized to novel instances even when the patterns compared were not discriminable by the presence or absence of reduplicated elements or by symmetry (e.g., AAB vs. ABB). These findings represent the first evidence of learning and generalization of regularities at the onset of life in an animal model, revealing intriguing differences with respect to human newborns and infants. Extensive prior experience seems to be unnecessary to drive the process, suggesting that chicks are predisposed to detect patterns characterizing the visual world.

Words, rules, and mechanisms of language acquisition

We review recent artificial language learning studies, especially those following Endress and Bonatti (Endress AD, Bonatti LL. Rapid learning of syllable classes from a perceptually continuous speech stream. Cognition 2007, 105:247-299), suggesting that humans can deploy a variety of learning mechanisms to acquire artificial languages. Several experiments provide evidence for multiple learning mechanisms that can be deployed in fluent speech: one mechanism encodes the positions of syllables within words and can be used to extract generalization, while the other registers co-occurrence statistics of syllables and can be used to break a continuum into its components. We review dissociations between these mechanisms and their potential role in language acquisition. We then turn to recent criticisms of the multiple mechanisms hypothesis and show that they are inconsistent with the available data. Our results suggest that artificial and natural language learning is best understood by dissecting the underlying specialized learning abilities, and that these data provide a rare opportunity to link important language phenomena to basic psychological mechanisms. For further resources related to this article, please visit the WIREs website.

On the phonetic and syntactic processing abilities of birds: from songs to speech and artificial grammars

Like speech and language, the songs of many songbirds consist of learned, rapidly produced, structured sequences of distinct vocal units, originating from an interplay between experience and learning biases. Songs are species specific, but also show considerable within species variation in elements or element sequencing. This variation implies that birds possess mechanisms to identify, categorize and combine sounds. I review the abilities for speech sound perception and categorization, as well as for grammatical rule learning by birds. Speech sound perception in birds is in many ways comparable to human speech perception. Birds can also detect and generalize patterns underlying artificially arranged strings of vocal elements. However, there is a need for more comparative studies to examine the limits of their rule learning abilities and how they relate to those of humans.

A hierarchy of cortical responses to sequence violations in three-month-old infants

The adult human brain quickly adapts to regular temporal sequences, and emits a sequence of novelty responses when these regularities are violated. These novelty responses have been interpreted as error signals that reflect the difference between the incoming signal and predictions generated at multiple cortical levels. Do infants already possess such a hierarchy of violation-detection mechanisms? Using high-density recordings of event-related potentials during an auditory local-global violation paradigm, we show that three-month-old infants process novelty in temporal sequences at two distinct levels. Violations of local expectancies, such as perceiving a deviant vowel "a" after repeated presentation of another vowel i-i-i, elicited an early auditory mismatch response. Conversely, violations of global expectancies, such as hearing the rare sequence a-a-a-a instead of the frequent sequence a-a-a-i, modulated this early mismatch response and led to a late frontal negative slow wave, whose cortical sources included the left inferior frontal region. These results suggest that the infant brain already possesses two dissociable systems for temporal sequence learning.

A perceptual account of symbolic reasoning

People can be taught to manipulate symbols according to formal mathematical and logical rules. Cognitive scientists have traditionally viewed this capacity-the capacity for symbolic reasoning-as grounded in the ability to internally represent numbers, logical relationships, and mathematical rules in an abstract, amodal fashion. We present an alternative view, portraying symbolic reasoning as a special kind of embodied reasoning in which arithmetic and logical formulae, externally represented as notations, serve as targets for powerful perceptual and sensorimotor systems. Although symbolic reasoning often conforms to abstract mathematical principles, it is typically implemented by perceptual and sensorimotor engagement with concrete environmental structures.

Learning non-adjacent regularities at age 0 ; 7

One important mechanism suggested to underlie the acquisition of grammar is rule learning. Indeed, infants aged 0 ; 7 are able to learn rules based on simple identity relations (adjacent repetitions, ABB: "wo fe fe" and non-adjacent repetitions, ABA: "wo fe wo", respectively; Marcus et al., 1999). One unexplored issue is whether young infants are able to process both adjacent and non-adjacent repetitions. As the previous studies always compared the two types of repetition structures directly, the ability to learn only one of them was sufficient for successful discrimination in these tasks. The present study reports two experiments, in which we test the ability of infants aged 0 ; 7 to discriminate adjacent and non-adjacent repetition structures against random controls (ABB vs. ABC and ABA vs. ABC). We show that, contrary to some previous proposals, infants aged 0 ; 7 successfully discriminate both repetition types from random controls, but show no spontaneous preference for either of them.

The nature of the memory buffer in implicit learning: learning Chinese tonal symmetries

Previous research has established that people can implicitly learn chunks, which (in terms of formal language theory) do not require a memory buffer to process. The present study explores the implicit learning of nonlocal dependencies generated by higher than finite-state grammars, specifically, Chinese tonal retrogrades (i.e. centre embeddings generated from a context-free grammar) and inversions (i.e. cross-serial dependencies generated from a mildly context-sensitive grammar), which do require buffers (for example, last in-first out and first in-first out, respectively). People were asked to listen to and memorize artificial poetry instantiating one of the two grammars; after this training phase, people were informed of the existence of rules and asked to classify new poems, while providing attributions of the basis of their judgments. People acquired unconscious structural knowledge of both tonal retrogrades and inversions. Moreover, inversions were implicitly learnt more easily than retrogrades constraining the nature of the memory buffer in computational models of implicit learning.

Bayesian learning and the psychology of rule induction

In recent years, Bayesian learning models have been applied to an increasing variety of domains. While such models have been criticized on theoretical grounds, the underlying assumptions and predictions are rarely made concrete and tested experimentally. Here, I use Frank and Tenenbaum's (2011) Bayesian model of rule-learning as a case study to spell out the underlying assumptions, and to confront them with the empirical results Frank and Tenenbaum (2011) propose to simulate, as well as with novel experiments. While rule-learning is arguably well suited to rational Bayesian approaches, I show that their models are neither psychologically plausible nor ideal observer models. Further, I show that their central assumption is unfounded: humans do not always preferentially learn more specific rules, but, at least in some situations, those rules that happen to be more salient. Even when granting the unsupported assumptions, I show that all of the experiments modeled by Frank and Tenenbaum (2011) either contradict their models, or have a large number of more plausible interpretations. I provide an alternative account of the experimental data based on simple psychological mechanisms, and show that this account both describes the data better, and is easier to falsify. I conclude that, despite the recent surge in Bayesian models of cognitive phenomena, psychological phenomena are best understood by developing and testing psychological theories rather than models that can be fit to virtually any data.

Uncertainty in visual and auditory series is coded by modality-general and modality-specific neural systems

Coding for the degree of disorder in a temporally unfolding sensory input allows for optimized encoding of these inputs via information compression and predictive processing. Prior neuroimaging work has examined sensitivity to statistical regularities within single sensory modalities and has associated this function with the hippocampus, anterior cingulate, and lateral temporal cortex. Here we investigated to what extent sensitivity to input disorder, quantified by Markov entropy, is subserved by modality-general or modality-specific neural systems when participants are not required to monitor the input. Participants were presented with rapid (3.3 Hz) auditory and visual series varying over four levels of entropy, while monitoring an infrequently changing fixation cross. For visual series, sensitivity to the magnitude of disorder was found in early visual cortex, the anterior cingulate, and the intraparietal sulcus. For auditory series, sensitivity was found in inferior frontal, lateral temporal, and supplementary motor regions implicated in speech perception and sequencing. Ventral premotor and central cingulate cortices were identified as possible candidates for modality-general uncertainty processing, exhibiting marginal sensitivity to disorder in both modalities. The right temporal pole differentiated the highest and lowest levels of disorder in both modalities, but did not show general sensitivity to the parametric manipulation of disorder. Our results indicate that neural sensitivity to input disorder relies largely on modality-specific systems embedded in extended sensory cortices, though uncertainty-related processing in frontal regions may be driven by both input modalities.

Rule learning by zebra finches in an artificial grammar learning task: which rule?

A hallmark of the human language faculty is the use of syntactic rules. The natural vocalizations of animals are syntactically simple, but several studies indicate that animals can detect and discriminate more complex structures in acoustic stimuli. However, how they discriminate such structures is often not clear. Using an artificial grammar learning paradigm, zebra finches were tested in a Go/No-go experiment for their ability to distinguish structurally different three-element sound sequences. In Experiment 1, zebra finches learned to discriminate ABA and BAB from ABB, AAB, BBA, and ABB sequences. Tests with probe sounds consisting of four elements suggested that the discrimination was based on attending to the presence or absence of repeated A- and B-elements. One bird generalized the discrimination to a new element type. In Experiment 2, we continued the training by adding four-element songs following a 'first and last identical versus different' rule that could not be solved by attending to repetitions. Only two out of five birds learned the overall discrimination. Testing with novel probes demonstrated that discrimination was not based on using the 'first and last identical' rule, but on attending to the presence or absence of the individual training stimuli. The two birds differed in the strategies used. Our results thus demonstrate only a limited degree of abstract rule learning but highlight the need for extensive and critical probe testing to examine the rules that animals (and humans) use to solve artificial grammar learning tasks. They also underline that rule learning strategies may differ between individuals.

Infant rule learning: advantage language, or advantage speech?

Infants appear to learn abstract rule-like regularities (e.g., la la da follows an AAB pattern) more easily from speech than from a variety of other auditory and visual stimuli (Marcus et al., 2007). We test if that facilitation reflects a specialization to learn from speech alone, or from modality-independent communicative stimuli more generally, by measuring 7.5-month-old infants’ ability to learn abstract rules from sign language-like gestures. Whereas infants appear to easily learn many different rules from speech, we found that with sign-like stimuli, and under circumstances comparable to those of Marcus et al. (1999), hearing infants were able to learn an ABB rule, but not an AAB rule. This is consistent with results of studies that demonstrate lower levels of infant rule learning from a variety of other non-speech stimuli, and we discuss implications for accounts of speech-facilitation.

A neuronal model of predictive coding accounting for the mismatch negativity

The mismatch negativity (MMN) is thought to index the activation of specialized neural networks for active prediction and deviance detection. However, a detailed neuronal model of the neurobiological mechanisms underlying the MMN is still lacking, and its computational foundations remain debated. We propose here a detailed neuronal model of auditory cortex, based on predictive coding, that accounts for the critical features of MMN. The model is entirely composed of spiking excitatory and inhibitory neurons interconnected in a layered cortical architecture with distinct input, predictive, and prediction error units. A spike-timing dependent learning rule, relying upon NMDA receptor synaptic transmission, allows the network to adjust its internal predictions and use a memory of the recent past inputs to anticipate on future stimuli based on transition statistics. We demonstrate that this simple architecture can account for the major empirical properties of the MMN. These include a frequency-dependent response to rare deviants, a response to unexpected repeats in alternating sequences (ABABAA…), a lack of consideration of the global sequence context, a response to sound omission, and a sensitivity of the MMN to NMDA receptor antagonists. Novel predictions are presented, and a new magnetoencephalography experiment in healthy human subjects is presented that validates our key hypothesis: the MMN results from active cortical prediction rather than passive synaptic habituation.

Consonants and vowels: different roles in early language acquisition

Language acquisition involves both acquiring a set of words (i.e. the lexicon) and learning the rules that combine them to form sentences (i.e. syntax). Here, we show that consonants are mainly involved in word processing, whereas vowels are favored for extracting and generalizing structural relations. We demonstrate that such a division of labor between consonants and vowels plays a role in language acquisition. In two very similar experimental paradigms, we show that 12-month-old infants rely more on the consonantal tier when identifying words (Experiment 1), but are better at extracting and generalizing repetition-based srtuctures over the vocalic tier (Experiment 2). These results indicate that infants are able to exploit the functional differences between consonants and vowels at an age when they start acquiring the lexicon, and suggest that basic speech categories are assigned to different learning mechanisms that sustain early language acquisition.

Learning implicitly to produce avoided behaviours

The literature on repetition processing reveals an intriguing paradox between the particular salience of repetitions, which makes them easy to learn, and a tendency to avoid them when generating sequences. The aim of this experiment was to study the extent to which children can learn to produce these avoided behaviours by means of an artificial grammar paradigm using generation tests with implicit or explicit instructions. The analysis of the control group's performance confirmed the presence of a spontaneous tendency to avoid generating repetitions. A comparison with chance revealed that the children learned to produce repetitions in the explicit test but not in the implicit test. However, a comparison with the control group showed that learning nonetheless occurred in the experimental group with the implicit test. The discussion focused on this antirepetition behavioural bias and how it interacted with the type of information processes elicited by the tests selected for assessing implicit learning effects.

Domain general constraints on statistical learning

All theories of language development suggest that learning is constrained. However, theories differ on whether these constraints arise from language-specific processes or have domain-general origins such as the characteristics of human perception and information processing. The current experiments explored constraints on statistical learning of patterns, such as the phonotactic patterns of an infants' native language. Infants in these experiments were presented with a visual analog of a phonotactic learning task used by J. R. Saffran and E. D. Thiessen (2003). Saffran and Thiessen found that infants' phonotactic learning was constrained such that some patterns were learned more easily than other patterns. The current results indicate that infants' learning of visual patterns shows the same constraints as infants' learning of phonotactic patterns. This is consistent with theories suggesting that constraints arise from domain-general sources and, as such, should operate over many kinds of stimuli in addition to linguistic stimuli.

Acoustic markers of prominence influence infants' and adults' segmentation of speech sequences

Two experiments investigated the way acoustic markers of prominence influence the grouping of speech sequences by adults and 7-month-old infants. In the first experiment, adults were familiarized with and asked to memorize sequences of adjacent syllables that alternated in either pitch or duration. During the test phase, participants heard pairs of syllables with constant pitch and duration and were asked whether the syllables had appeared adjacently during familiarization. Adults were better at remembering pairs of syllables that during familiarization had short syllables preceding long syllables, or high-pitched syllables preceding low-pitched syllables. In the second experiment, infants were familiarized and tested with similar stimuli as in the first experiment, and their preference for pairs of syllables was accessed using the head-turn preference paradigm.When familiarized with syllables alternating in pitch, infants showed a preference to listen to pairs of syllables that had high pitch in the first syllable. However, no preference was found when the familiarization stream alternated in duration. It is proposed that these perceptual biases help infants and adults find linguistic units in the continuous speech stream.While the bias for grouping based on pitch appears early in development, biases for durational grouping might rely on more extensive linguistic experience.

Perceptual and memory constraints on language acquisition

A wide variety of organisms employ specialized mechanisms to cope with the demands of their environment. We suggest that the same is true for humans when acquiring artificial grammars, and at least some basic properties of natural grammars. We show that two basic mechanisms can explain many results in artificial grammar learning experiments, and different linguistic regularities ranging from stress assignment to interfaces between different components of grammar. One mechanism is sensitive to identity relations, whereas the other uses sequence edges as anchor points for extracting positional regularities. This piecemeal approach to mental computations helps to explain otherwise perplexing data, and offers a working hypothesis on how statistical and symbolic accounts of cognitive processes could be bridged.

The effect of repetition and similarity on sequence learning

Repetition is a pervasive feature of children's environments, and may be an important contributor to learning such complex sequential structures as language. Endress, Dehaene-Lambertz, and Mehler (2007) found that repeated tone sequences were learned more easily than sequences containing ordinal relations, but there have been no direct comparisons of repeating sequences versus sequences that contain similar, but not identical, stimuli. In Experiment 1, we compared learning from repeating tone sequences to learning from tones that varied in similarity, and confirmed that repetition is a special case for learning. In Experiment 2 we showed that the learning distinction between repeated and similar elements is not affected by whether similarity is variable. We conclude by indicating that repetition provides an important constraint on learning, and we discuss the extent to which such constraints are consistent with general-purpose statistical learning mechanisms.

The neonate brain detects speech structure

What are the origins of the efficient language learning abilities that allow humans to acquire their mother tongue in just a few years very early in life? Although previous studies have identified different mechanisms underlying the acquisition of auditory and speech patterns in older infants and adults, the earliest sensitivities remain unexplored. To address this issue, we investigated the ability of newborns to learn simple repetition-based structures in two optical brain-imaging experiments. In the first experiment, 22 neonates listened to syllable sequences containing immediate repetitions (ABB; e.g., "mubaba," "penana"), intermixed with random control sequences (ABC; e.g., "mubage," "penaku"). We found increased responses to the repetition sequences in the temporal and left frontal areas, indicating that the newborn brain differentiated the two patterns. The repetition sequences evoked greater activation than the random sequences during the first few trials, suggesting the presence of an automatic perceptual mechanism to detect repetitions. In addition, over the subsequent trials, activation increased further in response to the repetition sequences but not in response to the random sequences, indicating that recognition of the ABB pattern was enhanced by repeated exposure. In the second experiment, in which nonadjacent repetitions (ABA; e.g., "bamuba," "napena") were contrasted with the same random controls, no discrimination was observed. These findings suggest that newborns are sensitive to certain input configurations in the auditory domain, a perceptual ability that might facilitate later language development.