Visual statistical learning in the newborn infant

Infant fMRI: A Model System for Cognitive Neuroscience

Our understanding of the typical human brain has benefitted greatly from studying different kinds of brains and their associated behavioral repertoires, including animal models and neuropsychological patients. This same comparative perspective can be applied to early development - the environment, behavior, and brains of infants provide a model system for understanding how the mature brain works. This approach requires noninvasive methods for measuring brain function in awake, behaving infants. fMRI is becoming increasingly viable for this purpose, with the unique ability to precisely measure the entire brain, including both cortical and subcortical structures. Here we discuss potential lessons from infant fMRI for several domains of adult cognition and consider the challenges of conducting such research and how they might be mitigated.

Online incidental statistical learning of audiovisual word sequences in adults: a registered report

Statistical learning has been proposed as a key mechanism in language learning. Our main goal was to examine whether adults are capable of simultaneously extracting statistical dependencies in a task where stimuli include a range of structures amenable to statistical learning within a single paradigm. We devised an online statistical learning task using real word auditory-picture sequences that vary in two dimensions: (i) predictability and (ii) adjacency of dependent elements. This task was followed by an offline recall task to probe learning of each sequence type. We registered three hypotheses with specific predictions. First, adults would extract regular patterns from continuous stream (effect of grammaticality). Second, within grammatical conditions, they would show differential speeding up for each condition as a factor of statistical complexity of the condition and exposure. Third, our novel approach to measure online statistical learning would be reliable in showing individual differences in statistical learning ability. Further, we explored the relation between statistical learning and a measure of verbal short-term memory (STM). Forty-two participants were tested and retested after an interval of at least 3 days on our novel statistical learning task. We analysed the reaction time data using a novel regression discontinuity approach. Consistent with prediction, participants showed a grammaticality effect, agreeing with the predicted order of difficulty for learning different statistical structures. Furthermore, a learning index from the task showed acceptable test-retest reliability (r = 0.67). However, STM did not correlate with statistical learning. We discuss the findings noting the benefits of online measures in tracking the learning process.

Unsupervised statistical learning in newly hatched chicks

The ability to extract probabilistic information from visual inputs has been reported in human adults and infants (reviewed in [1,2]), and in adults of non-human species, though only under supervised (conditioning) procedures [3]. Here, we report spontaneous sensitivity to the probabilistic structure underlying sequences of visual stimuli in newly hatched domestic chicks using filial imprinting, suggesting that statistical learning may be fully operating at the onset of life in precocial avian species.

Splitting the variance of statistical learning performance: A parametric investigation of exposure duration and transitional probabilities

What determines individuals' efficacy in detecting regularities in visual statistical learning? Our theoretical starting point assumes that the variance in performance of statistical learning (SL) can be split into the variance related to efficiency in encoding representations within a modality and the variance related to the relative computational efficiency of detecting the distributional properties of the encoded representations. Using a novel methodology, we dissociated encoding from higher-order learning factors, by independently manipulating exposure duration and transitional probabilities in a stream of visual shapes. Our results show that the encoding of shapes and the retrieving of their transitional probabilities are not independent and additive processes, but interact to jointly determine SL performance. The theoretical implications of these findings for a mechanistic explanation of SL are discussed.

Constraints on Statistical Learning Across Species

Both human and nonhuman organisms are sensitive to statistical regularities in sensory inputs that support functions including communication, visual processing, and sequence learning. One of the issues faced by comparative research in this field is the lack of a comprehensive theory to explain the relevance of statistical learning across distinct ecological niches. In the current review we interpret cross-species research on statistical learning based on the perceptual and cognitive mechanisms that characterize the human and nonhuman models under investigation. Considering statistical learning as an essential part of the cognitive architecture of an animal will help to uncover the potential ecological functions of this powerful learning process.

Measuring individual differences in statistical learning: Current pitfalls and possible solutions

Most research in statistical learning (SL) has focused on the mean success rates of participants in detecting statistical contingencies at a group level. In recent years, however, researchers have shown increased interest in individual abilities in SL, either to predict other cognitive capacities or as a tool for understanding the mechanism underlying SL. Most if not all of this research enterprise has employed SL tasks that were originally designed for group-level studies. We argue that from an individual difference perspective, such tasks are psychometrically weak, and sometimes even flawed. In particular, the existing SL tasks have three major shortcomings: (1) the number of trials in the test phase is often too small (or, there is extensive repetition of the same targets throughout the test); (2) a large proportion of the sample performs at chance level, so that most of the data points reflect noise; and (3) the test items following familiarization are all of the same type and an identical level of difficulty. These factors lead to high measurement error, inevitably resulting in low reliability, and thereby doubtful validity. Here we present a novel method specifically designed for the measurement of individual differences in visual SL. The novel task we offer displays substantially superior psychometric properties. We report data regarding the reliability of the task and discuss the importance of the implementation of such tasks in future research.

Towards a theory of individual differences in statistical learning

In recent years, statistical learning (SL) research has seen a growing interest in tracking individual performance in SL tasks, mainly as a predictor of linguistic abilities. We review studies from this line of research and outline three presuppositions underlying the experimental approach they employ: (i) that SL is a unified theoretical construct; (ii) that current SL tasks are interchangeable, and equally valid for assessing SL ability; and (iii) that performance in the standard forced-choice test in the task is a good proxy of SL ability. We argue that these three critical presuppositions are subject to a number of theoretical and empirical issues. First, SL shows patterns of modality- and informational-specificity, suggesting that SL cannot be treated as a unified construct. Second, different SL tasks may tap into separate sub-components of SL that are not necessarily interchangeable. Third, the commonly used forced-choice tests in most SL tasks are subject to inherent limitations and confounds. As a first step, we offer a methodological approach that explicitly spells out a potential set of different SL dimensions, allowing for better transparency in choosing a specific SL task as a predictor of a given linguistic outcome. We then offer possible methodological solutions for better tracking and measuring SL ability. Taken together, these discussions provide a novel theoretical and methodological approach for assessing individual differences in SL, with clear testable predictions.This article is part of the themed issue 'New frontiers for statistical learning in the cognitive sciences'.

Learning predictive statistics from temporal sequences: Dynamics and strategies

Human behavior is guided by our expectations about the future. Often, we make predictions by monitoring how event sequences unfold, even though such sequences may appear incomprehensible. Event structures in the natural environment typically vary in complexity, from simple repetition to complex probabilistic combinations. How do we learn these structures? Here we investigate the dynamics of structure learning by tracking human responses to temporal sequences that change in structure unbeknownst to the participants. Participants were asked to predict the upcoming item following a probabilistic sequence of symbols. Using a Markov process, we created a family of sequences, from simple frequency statistics (e.g., some symbols are more probable than others) to context-based statistics (e.g., symbol probability is contingent on preceding symbols). We demonstrate the dynamics with which individuals adapt to changes in the environment's statistics-that is, they extract the behaviorally relevant structures to make predictions about upcoming events. Further, we show that this structure learning relates to individual decision strategy; faster learning of complex structures relates to selection of the most probable outcome in a given context (maximizing) rather than matching of the exact sequence statistics. Our findings provide evidence for alternate routes to learning of behaviorally relevant statistics that facilitate our ability to predict future events in variable environments.

Enhanced Neonatal Brain Responses To Sung Streams Predict Vocabulary Outcomes By Age 18 Months

Words and melodies are some of the basic elements infants are able to extract early in life from the auditory input. Whether melodic cues contained in songs can facilitate word-form extraction immediately after birth remained unexplored. Here, we provided converging neural and computational evidence of the early benefit of melodies for language acquisition. Twenty-eight neonates were tested on their ability to extract word-forms from continuous flows of sung and spoken syllabic sequences. We found different brain dynamics for sung and spoken streams and observed successful detection of word-form violations in the sung condition only. Furthermore, neonatal brain responses for sung streams predicted expressive vocabulary at 18 months as demonstrated by multiple regression and cross-validation analyses. These findings suggest that early neural individual differences in prosodic speech processing might be a good indicator of later language outcomes and could be considered as a relevant factor in the development of infants' language skills.

Infants are superior in implicit crossmodal learning and use other learning mechanisms than adults

During development internal models of the sensory world must be acquired which have to be continuously adapted later. We used event-related potentials (ERP) to test the hypothesis that infants extract crossmodal statistics implicitly while adults learn them when task relevant. Participants were passively exposed to frequent standard audio-visual combinations (A1V1, A2V2, p=0.35 each), rare recombinations of these standard stimuli (A1V2, A2V1, p=0.10 each), and a rare audio-visual deviant with infrequent auditory and visual elements (A3V3, p=0.10). While both six-month-old infants and adults differentiated between rare deviants and standards involving early neural processing stages only infants were sensitive to crossmodal statistics as indicated by a late ERP difference between standard and recombined stimuli. A second experiment revealed that adults differentiated recombined and standard combinations when crossmodal combinations were task relevant. These results demonstrate a heightened sensitivity for crossmodal statistics in infants and a change in learning mode from infancy to adulthood.

Infant Statistical Learning

Perception involves making sense of a dynamic, multimodal environment. In the absence of mechanisms capable of exploiting the statistical patterns in the natural world, infants would face an insurmountable computational problem. Infant statistical learning mechanisms facilitate the detection of structure. These abilities allow the infant to compute across elements in their environmental input, extracting patterns for further processing and subsequent learning. In this selective review, we summarize findings that show that statistical learning is both a broad and flexible mechanism (supporting learning from different modalities across many different content areas) and input specific (shifting computations depending on the type of input and goal of learning). We suggest that statistical learning not only provides a framework for studying language development and object knowledge in constrained laboratory settings, but also allows researchers to tackle real-world problems, such as multilingualism, the role of ever-changing learning environments, and differential developmental trajectories.

Problems with visual statistical learning in developmental dyslexia

Previous research shows that dyslexic readers are impaired in their recognition of faces and other complex objects, and show hypoactivation in ventral visual stream regions that support word and object recognition. Responses of these brain regions are shaped by visual statistical learning. If such learning is compromised, people should be less sensitive to statistically likely feature combinations in words and other objects, and impaired visual word and object recognition should be expected. We therefore tested whether people with dyslexia showed diminished capability for visual statistical learning. Matched dyslexic and typical readers participated in tests of visual statistical learning of pairs of novel shapes that frequently appeared together. Dyslexic readers on average recognized fewer pairs than typical readers, indicating some problems with visual statistical learning. These group differences were not accounted for by differences in intelligence, ability to remember individual shapes, or spatial attention paid to the stimuli, but other attentional problems could play a mediating role. Deficiencies in visual statistical learning may in some cases prevent appropriate experience-driven shaping of neuronal responses in the ventral visual stream, hampering visual word and object recognition.

Second Language Experience Facilitates Statistical Learning of Novel Linguistic Materials

Recent research has begun to explore individual differences in statistical learning, and how those differences may be related to other cognitive abilities, particularly their effects on language learning. In this research, we explored a different type of relationship between language learning and statistical learning: the possibility that learning a new language may also influence statistical learning by changing the regularities to which learners are sensitive. We tested two groups of participants, Mandarin Learners and Naïve Controls, at two time points, 6 months apart. At each time point, participants performed two different statistical learning tasks: an artificial tonal language statistical learning task and a visual statistical learning task. Only the Mandarin-learning group showed significant improvement on the linguistic task, whereas both groups improved equally on the visual task. These results support the view that there are multiple influences on statistical learning. Domain-relevant experiences may affect the regularities that learners can discover when presented with novel stimuli.

Sleep confers a benefit for retention of statistical language learning in 6.5month old infants

Infants show robust ability to track transitional probabilities within language and can use this information to extract words from continuous speech. The degree to which infants remember these words across a delay is unknown. Given well-established benefits of sleep on long-term memory retention in adults, we examine whether sleep similarly facilitates memory in 6.5month olds. Infants listened to an artificial language for 7minutes, followed by a period of sleep or wakefulness. After a time-matched delay for sleep and wakefulness dyads, we measured retention using the head-turn-preference procedure. Infants who slept retained memory for the extracted words that was prone to interference during the test. Infants who remained awake showed no retention. Within the nap group, retention correlated with three electrophysiological measures (1) absolute theta across the brain, (2) absolute alpha across the brain, and (3) greater fronto-central slow wave activity (SWA).

Online neural monitoring of statistical learning

The extraction of patterns in the environment plays a critical role in many types of human learning, from motor skills to language acquisition. This process is known as statistical learning. Here we propose that statistical learning has two dissociable components: (1) perceptual binding of individual stimulus units into integrated composites and (2) storing those integrated representations for later use. Statistical learning is typically assessed using post-learning tasks, such that the two components are conflated. Our goal was to characterize the online perceptual component of statistical learning. Participants were exposed to a structured stream of repeating trisyllabic nonsense words and a random syllable stream. Online learning was indexed by an EEG-based measure that quantified neural entrainment at the frequency of the repeating words relative to that of individual syllables. Statistical learning was subsequently assessed using conventional measures in an explicit rating task and a reaction-time task. In the structured stream, neural entrainment to trisyllabic words was higher than in the random stream, increased as a function of exposure to track the progression of learning, and predicted performance on the reaction time (RT) task. These results demonstrate that monitoring this critical component of learning via rhythmic EEG entrainment reveals a gradual acquisition of knowledge whereby novel stimulus sequences are transformed into familiar composites. This online perceptual transformation is a critical component of learning.

TRACX2: a connectionist autoencoder using graded chunks to model infant visual statistical learning

Even newborn infants are able to extract structure from a stream of sensory inputs; yet how this is achieved remains largely a mystery. We present a connectionist autoencoder model, TRACX2, that learns to extract sequence structure by gradually constructing chunks, storing these chunks in a distributed manner across its synaptic weights and recognizing these chunks when they re-occur in the input stream. Chunks are graded rather than all-or-nothing in nature. As chunks are learnt their component parts become more and more tightly bound together. TRACX2 successfully models the data from five experiments from the infant visual statistical learning literature, including tasks involving forward and backward transitional probabilities, low-salience embedded chunk items, part-sequences and illusory items. The model also captures performance differences across ages through the tuning of a single-learning rate parameter. These results suggest that infant statistical learning is underpinned by the same domain-general learning mechanism that operates in auditory statistical learning and, potentially, in adult artificial grammar learning.This article is part of the themed issue 'New frontiers for statistical learning in the cognitive sciences'.

Statistical learning: a powerful mechanism that operates by mere exposure

How do infants learn so rapidly and with little apparent effort? In 1996, Saffran, Aslin, and Newport reported that 8-month-old human infants could learn the underlying temporal structure of a stream of speech syllables after only 2 min of passive listening. This demonstration of what was called statistical learning, involving no instruction, reinforcement, or feedback, led to dozens of confirmations of this powerful mechanism of implicit learning in a variety of modalities, domains, and species. These findings reveal that infants are not nearly as dependent on explicit forms of instruction as we might have assumed from studies of learning in which children or adults are taught facts such as math or problem solving skills. Instead, at least in some domains, infants soak up the information around them by mere exposure. Learning and development in these domains thus appear to occur automatically and with little active involvement by an instructor (parent or teacher). The details of this statistical learning mechanism are discussed, including how exposure to specific types of information can, under some circumstances, generalize to never-before-observed information, thereby enabling transfer of learning. WIREs Cogn Sci 2017, 8:e1373. doi: 10.1002/wcs.1373 For further resources related to this article, please visit the WIREs website.

Central tendency effects in time interval reproduction in autism

Central tendency, the tendency of judgements of quantities (lengths, durations etc.) to gravitate towards their mean, is one of the most robust perceptual effects. A Bayesian account has recently suggested that central tendency reflects the integration of noisy sensory estimates with prior knowledge representations of a mean stimulus, serving to improve performance. The process is flexible, so prior knowledge is weighted more heavily when sensory estimates are imprecise, requiring more integration to reduce noise. In this study we measure central tendency in autism to evaluate a recent theoretical hypothesis suggesting that autistic perception relies less on prior knowledge representations than typical perception. If true, autistic children should show reduced central tendency than theoretically predicted from their temporal resolution. We tested autistic and age- and ability-matched typical children in two child-friendly tasks: (1) a time interval reproduction task, measuring central tendency in the temporal domain; and (2) a time discrimination task, assessing temporal resolution. Central tendency reduced with age in typical development, while temporal resolution improved. Autistic children performed far worse in temporal discrimination than the matched controls. Computational simulations suggested that central tendency was much less in autistic children than predicted by theoretical modelling, given their poor temporal resolution.

A smoothness constraint on the development of object recognition

Understanding how the brain learns to recognize objects is one of the ultimate goals in the cognitive sciences. To date, however, we have not yet characterized the environmental factors that cause object recognition to emerge in the newborn brain. Here, I present the results of a high-throughput controlled-rearing experiment that examined whether the development of object recognition requires experience with temporally smooth visual objects. When newborn chicks (Gallus gallus) were raised with virtual objects that moved smoothly over time, the chicks developed accurate color recognition, shape recognition, and color-shape binding abilities. In contrast, when newborn chicks were raised with virtual objects that moved non-smoothly over time, the chicks' object recognition abilities were severely impaired. These results provide evidence for a "smoothness constraint" on newborn object recognition. Experience with temporally smooth objects facilitates the development of object recognition.

Prediction suppression in monkey inferotemporal cortex depends on the conditional probability between images

When monkeys view two images in fixed sequence repeatedly over days and weeks, neurons in area TE of the inferotemporal cortex come to exhibit prediction suppression. The trailing image elicits only a weak response when presented following the leading image that preceded it during training. Induction of prediction suppression might depend either on the contiguity of the images, as determined by their co-occurrence and captured in the measure of joint probability P(A,B), or on their contingency, as determined by their correlation and as captured in the measures of conditional probability P(A|B) and P(B|A). To distinguish between these possibilities, we measured prediction suppression after imposing training regimens that held P(A,B) constant but varied P(A|B) and P(B|A). We found that reducing either P(A|B) or P(B|A) during training attenuated prediction suppression as measured during subsequent testing. We conclude that prediction suppression depends on contingency, as embodied in the predictive relations between the images, and not just on contiguity, as embodied in their co-occurrence.

Many faces, one rule: the role of perceptual expertise in infants' sequential rule learning

Rule learning is a mechanism that allows infants to recognize and generalize rule-like patterns, such as ABB or ABA. Although infants are better at learning rules from speech vs. non-speech, rule learning can be applied also to frequently experienced visual stimuli, suggesting that perceptual expertise with material to be learned is critical in enhancing rule learning abilities. Yet infants' rule learning has never been investigated using one of the most commonly experienced visual stimulus category available in infants' environment, i.e., faces. Here, we investigate 7-month-olds' ability to extract rule-like patterns from sequences composed of upright faces and compared their results to those of infants who viewed inverted faces, which presumably are encountered far less frequently than upright faces. Infants were habituated with face triads in either an ABA or ABB condition followed by a test phase with ABA and ABB triads composed of faces that differed from those showed during habituation. When upright faces were used, infants generalized the pattern presented during habituation to include the new face identities showed during testing, but when inverted faces were presented, infants failed to extract the rule. This finding supports the idea that perceptual expertise can modulate 7-month-olds' abilities to detect rule-like patterns.

Visual statistical learning is not reliably modulated by selective attention to isolated events

Recent studies of visual statistical learning (VSL) have indicated that the visual system can automatically extract the temporal and spatial relationships between objects. We report several attempts to replicate and extend earlier work (Journal of Experimental Psychology: General, 134, 552-564, 2005) in which observers performed a cover task on one of two interleaved stimulus sets, resulting in the learning of temporal relationships that occurred in the attended stream, but not those present in the unattended stream. Across four experiments, we exposed observers to similar or identical familiarization protocols, directing attention to one of two interleaved stimulus sets; afterward, we assessed the VSL efficacy for both sets using either implicit response time measures or explicit familiarity judgments. In line with prior work, we observed learning for the attended stimulus set. However, unlike in previous reports, we also observed learning for the unattended stimulus set. When instructed to selectively attend to only one of the stimulus sets and ignore the other set, observers could extract temporal regularities for both sets. Our efforts to experimentally decrease this effect by changing the cover task (Exp. 1) or the complexity of the statistical regularities (Exp. 3) were unsuccessful. A fourth experiment using a different assessment of learning likewise failed to show an attentional effect. Simulations drawing random samples from our first three experiments (n = 64) confirmed that the distribution of attentional effects in our sample closely approximated the null. We offer several potential explanations for our failure to replicate earlier findings and discuss how our results suggest limiting conditions on the relevance of attention to VSL.

Domain general learning: Infants use social and non-social cues when learning object statistics

Previous research has shown that infants can learn from social cues. But is a social cue more effective at directing learning than a non-social cue? This study investigated whether 9-month-old infants (N = 55) could learn a visual statistical regularity in the presence of a distracting visual sequence when attention was directed by either a social cue (a person) or a non-social cue (a rectangle). The results show that both social and non-social cues can guide infants' attention to a visual shape sequence (and away from a distracting sequence). The social cue more effectively directed attention than the non-social cue during the familiarization phase, but the social cue did not result in significantly stronger learning than the non-social cue. The findings suggest that domain general attention mechanisms allow for the comparable learning seen in both conditions.

Statistical learning as an individual ability: Theoretical perspectives and empirical evidence

Although the power of statistical learning (SL) in explaining a wide range of linguistic functions is gaining increasing support, relatively little research has focused on this theoretical construct from the perspective of individual differences. However, to be able to reliably link individual differences in a given ability such as language learning to individual differences in SL, three critical theoretical questions should be posed: Is SL a componential or unified ability? Is it nested within other general cognitive abilities? Is it a stable capacity of an individual? Following an initial mapping sentence outlining the possible dimensions of SL, we employed a battery of SL tasks in the visual and auditory modalities, using verbal and non-verbal stimuli, with adjacent and non-adjacent contingencies. SL tasks were administered along with general cognitive tasks in a within-subject design at two time points to explore our theoretical questions. We found that SL, as measured by some tasks, is a stable and reliable capacity of an individual. Moreover, we found SL to be independent of general cognitive abilities such as intelligence or working memory. However, SL is not a unified capacity, so that individual sensitivity to conditional probabilities is not uniform across modalities and stimuli.

Infants' statistical learning: 2- and 5-month-olds' segmentation of continuous visual sequences

Past research suggests that infants have powerful statistical learning abilities; however, studies of infants' visual statistical learning offer differing accounts of the developmental trajectory of and constraints on this learning. To elucidate this issue, the current study tested the hypothesis that young infants' segmentation of visual sequences depends on redundant statistical cues to segmentation. A sample of 20 2-month-olds and 20 5-month-olds observed a continuous sequence of looming shapes in which unit boundaries were defined by both transitional probability and co-occurrence frequency. Following habituation, only 5-month-olds showed evidence of statistically segmenting the sequence, looking longer to a statistically improbable shape pair than to a probable pair. These results reaffirm the power of statistical learning in infants as young as 5 months but also suggest considerable development of statistical segmentation ability between 2 and 5 months of age. Moreover, the results do not support the idea that infants' ability to segment visual sequences based on transitional probabilities and/or co-occurrence frequencies is functional at the onset of visual experience, as has been suggested previously. Rather, this type of statistical segmentation appears to be constrained by the developmental state of the learner. Factors contributing to the development of statistical segmentation ability during early infancy, including memory and attention, are discussed.

A chicken model for studying the emergence of invariant object recognition

“Invariant object recognition” refers to the ability to recognize objects across variation in their appearance on the retina. This ability is central to visual perception, yet its developmental origins are poorly understood. Traditionally, nonhuman primates, rats, and pigeons have been the most commonly used animal models for studying invariant object recognition. Although these animals have many advantages as model systems, they are not well suited for studying the emergence of invariant object recognition in the newborn brain. Here, we argue that newly hatched chicks (Gallus gallus) are an ideal model system for studying the emergence of invariant object recognition. Using an automated controlled-rearing approach, we show that chicks can build a viewpoint-invariant representation of the first object they see in their life. This invariant representation can be built from highly impoverished visual input (three images of an object separated by 15° azimuth rotations) and cannot be accounted for by low-level retina-like or V1-like neuronal representations. These results indicate that newborn neural circuits begin building invariant object representations at the onset of vision and argue for an increased focus on chicks as an animal model for studying invariant object recognition.

Domain generality versus modality specificity: the paradox of statistical learning

Statistical learning (SL) is typically considered to be a domain-general mechanism by which cognitive systems discover the underlying distributional properties of the input. However, recent studies examining whether there are commonalities in the learning of distributional information across different domains or modalities consistently reveal modality and stimulus specificity. Therefore, important questions are how and why a hypothesized domain-general learning mechanism systematically produces such effects. Here, we offer a theoretical framework according to which SL is not a unitary mechanism, but a set of domain-general computational principles that operate in different modalities and, therefore, are subject to the specific constraints characteristic of their respective brain regions. This framework offers testable predictions and we discuss its computational and neurobiological plausibility.

Visual statistical learning in children and young adults: how implicit?

Visual statistical learning (VSL) is the ability to extract the joint and conditional probabilities of shapes co-occurring during passive viewing of complex visual configurations. Evidence indicates that even infants are sensitive to these regularities (e.g., Kirkham et al., 2002). However, there is continuing debate as to whether VSL is accompanied by conscious awareness of the statistical regularities between sequence elements. Bertels et al. (2012) addressed this question in young adults. Here, we adapted their paradigm to investigate VSL and conscious awareness in children. Using the same version of the paradigm, we also tested young adults so as to directly compare results from both age groups. Fifth graders and undergraduates were exposed to a stream of visual shapes arranged in triplets. Learning of these sequences was then assessed using both direct and indirect measures. In order to assess the extent to which learning occurred explicitly, we also measured confidence through subjective measures in the direct task (i.e., binary confidence judgments). Results revealed that both children and young adults learned the statistical regularities between shapes. In both age groups, participants who performed above chance in the completion task had conscious access to their knowledge. Nevertheless, although adults performed above chance even when they claimed to guess, there was no evidence of implicit knowledge in children. These results suggest that the role of implicit and explicit influences in VSL may follow a developmental trajectory.

From local to global processing: the development of illusory contour perception

Global visual processing is important for segmenting scenes, extracting form from background, and recognizing objects. Local processing involves attention to the local elements, contrast, and boundaries of an image at the expense of extracting a global percept. Previous work is inconclusive regarding the relative development of local and global processing. Some studies suggest that global perception is already present by 8 months of age, whereas others suggest that the ability arises during childhood and continues to develop during adolescence. We used a novel method to assess the development of global processing in 3- to 10-year-old children and an adult comparison group. We used Kanizsa illusory contours as an assay of global perception and measured responses on a touch-sensitive screen while monitoring eye position with a head-mounted eye tracker. Participants were tested using a similarity match-to-sample paradigm. Using converging measures, we found a clear developmental progression with age such that the youngest children performed near chance on the illusory contour discrimination, whereas 7- and 8-year-olds performed nearly perfectly, as did adults. There was clear evidence of a gradual shift from a local processing strategy to a global one; young children looked predominantly at and touched the "pacman" inducers of the illusory form, whereas older children and adults looked predominantly at and touched the middle of the form. These data show a prolonged developmental trajectory in appreciation of global form, with a transition from local to global visual processing between 4 and 7 years of age.

Developmental insights into mature cognition

Three cases are described that illustrate new ways in which developmental research is informing the study of cognition in adults: statistical learning, neural substrates of cognition, and extended concepts. Developmental research has made clear the ubiquity of statistical learning while also revealing is limitations as a stand-alone way to acquire knowledge. With respect to neural substrates, development has uncovered links between executive processing and fronto-striatal circuits while also pointing to many aspects of high-level cognition that may not be neatly reducible to coherent neural descriptions. For extended concepts, children have made especially clear the weaknesses of intuitive theories in both children and adults while also illustrating other cognitive capacities that are used at all ages to navigate the socially distributed aspects of knowledge.

Infant perceptual development for faces and spoken words: an integrated approach

There are obvious differences between recognizing faces and recognizing spoken words or phonemes that might suggest development of each capability requires different skills. Recognizing faces and perceiving spoken language, however, are in key senses extremely similar endeavors. Both perceptual processes are based on richly variable, yet highly structured input from which the perceiver needs to extract categorically meaningful information. This similarity could be reflected in the perceptual narrowing that occurs within the first year of life in both domains. We take the position that the perceptual and neurocognitive processes by which face and speech recognition develop are based on a set of common principles. One common principle is the importance of systematic variability in the input as a source of information rather than noise. Experience of this variability leads to perceptual tuning to the critical properties that define individual faces or spoken words versus their membership in larger groupings of people and their language communities. We argue that parallels can be drawn directly between the principles responsible for the development of face and spoken language perception.

Statistical learning of serial visual transitions by neurons in monkey inferotemporal cortex

If monkeys repeatedly, over the course of weeks, view displays in which two images appear in fixed sequence, then neurons of inferotemporal cortex (ITC) come to exhibit prediction suppression. The response to the trailing image is weaker if it follows the leading image with which it was paired during training than if it follows some other leading image. Prediction suppression is a plausible neural mechanism for statistical learning of visual transitions such as has been demonstrated in behavioral studies of human infants and adults. However, in the human studies, subjects are exposed to continuous sequences in which the same image can be both predicted and predicting and statistical dependency can exist between nonadjacent items. The aim of the present study was to investigate whether prediction suppression in ITC develops under such circumstances. To resolve this issue, we exposed monkeys repeatedly to triplets of images presented in fixed order. The results indicate that prediction suppression can be induced by training not only with pairs of images but also with longer sequences.

Development of context dependency in human space perception

Perception is a complex process, where prior knowledge exerts a fundamental influence over what we see. The use of priors is at the basis of the well-known phenomenon of central tendency: Judgments of almost all quantities (such as length, duration, and number) tend to gravitate toward their mean magnitude. Although such context dependency is universal in adult perceptual judgments, how it develops with age remains unknown. We asked children from 7 to 14 years of age and adults to reproduce lengths of stimuli drawn from different distributions and evaluated whether judgments were influenced by stimulus context. All participants reproduced the presented length differently depending on the context: The same stimulus was reproduced as shorter, when on average stimuli were short, and as longer, when on average stimuli were long. Interestingly, the relative importance given to the current sensory signal and to priors was almost constant during childhood. This strategy, which in adults is optimal in Bayesian terms, is apparently successful in holding the sensory noise at bay even during development. Hence, the influence of previous knowledge on perception is present already in young children, suggesting that context dependency is established early in the developing brain.

Causal learning from probabilistic events in 24-month-olds: an action measure

How do young children learn about causal structure in an uncertain and variable world? We tested whether they can use observed probabilistic information to solve causal learning problems. In two experiments, 24-month-olds observed an adult produce a probabilistic pattern of causal evidence. The toddlers then were given an opportunity to design their own intervention. In Experiment 1, toddlers saw one object bring about an effect with a higher probability than a second object. In Experiment 2, the frequency of the effect was held constant, though its probability differed. After observing the probabilistic evidence, toddlers in both experiments chose to act on the object that was more likely to produce the effect. The results demonstrate that toddlers can learn about cause and effect without trial-and-error or linguistic instruction on the task, simply by observing the probabilistic patterns of evidence resulting from the imperfect actions of other social agents. Such observational causal learning from probabilistic displays supports human children's rapid cultural learning.

Electrophysiological evidence of heterogeneity in visual statistical learning in young children with ASD

Statistical learning is characterized by detection of regularities in one's environment without an awareness or intention to learn, and it may play a critical role in language and social behavior. Accordingly, in this study we investigated the electrophysiological correlates of visual statistical learning in young children with autism spectrum disorder (ASD) using an event-related potential shape learning paradigm, and we examined the relation between visual statistical learning and cognitive function. Compared to typically developing (TD) controls, the ASD group as a whole showed reduced evidence of learning as defined by N1 (early visual discrimination) and P300 (attention to novelty) components. Upon further analysis, in the ASD group there was a positive correlation between N1 amplitude difference and non-verbal IQ, and a positive correlation between P300 amplitude difference and adaptive social function. Children with ASD and a high non-verbal IQ and high adaptive social function demonstrated a distinctive pattern of learning. This is the first study to identify electrophysiological markers of visual statistical learning in children with ASD. Through this work we have demonstrated heterogeneity in statistical learning in ASD that maps onto non-verbal cognition and adaptive social function.

Human infants' preference for left-to-right oriented increasing numerical sequences

While associations between number and space, in the form of a spatially oriented numerical representation, have been extensively reported in human adults, the origins of this phenomenon are still poorly understood. The commonly accepted view is that this number-space association is a product of human invention, with accounts proposing that culture, symbolic knowledge, and mathematics education are at the roots of this phenomenon. Here we show that preverbal infants aged 7 months, who lack symbolic knowledge and mathematics education, show a preference for increasing magnitude displayed in a left-to-right spatial orientation. Infants habituated to left-to-right oriented increasing or decreasing numerical sequences showed an overall higher looking time to new left-to-right oriented increasing numerical sequences at test (Experiment 1). This pattern did not hold when infants were presented with the same ordinal numerical information displayed from right to left (Experiment 2). The different pattern of results was congruent with the presence of a malleable, context-dependent baseline preference for increasing, left-to-right oriented, numerosities (Experiment 3). These findings are suggestive of an early predisposition in humans to link numerical order with a left-to-right spatial orientation, which precedes the acquisition of symbolic abilities, mathematics education, and the acquisition of reading and writing skills.

Infants segment continuous events using transitional probabilities

Throughout their 1st year, infants adeptly detect statistical structure in their environment. However, little is known about whether statistical learning is a primary mechanism for event segmentation. This study directly tests whether statistical learning alone is sufficient to segment continuous events. Twenty-eight 7- to 9-month-old infants viewed a sequence of continuous actions performed by a novel agent in which there were no transitional movements that could have constrained the possible upcoming actions. At test, infants distinguished statistically intact units from less predictable ones. The ability to segment events using statistical structure may help infants discover other cues to event boundaries, such as intentions, and carve up the world of continuous motion in meaningful ways.

Developmental differences in effects of task pacing on implicit sequence learning

Although there is now substantial evidence that developmental change occurs in implicit learning abilities over the lifespan, disparate results exist regarding the specific developmental trajectory of implicit learning skills. One possible reason for discrepancies across implicit learning studies may be that younger children show an increased sensitivity to variations in implicit learning task procedures and demands relative to adults. Studies using serial-reaction time (SRT) tasks have suggested that in adults, measurements of implicit learning are robust across variations in task procedures. Most classic SRT tasks have used response-contingent pacing in which the participant's own reaction time determines the duration of each trial. However, recent paradigms with adults and children have used fixed trial pacing, which leads to alterations in both response and attention demands, accuracy feedback, perceived agency, and task motivation for participants. In the current study, we compared learning on fixed-paced and self-paced versions of a spatial sequence learning paradigm in 4-year-old children and adults. Results indicated that preschool-aged children showed reduced evidence of implicit sequence learning in comparison to adults, regardless of the SRT paradigm used. In addition, we found the preschoolers showed significantly greater learning when stimulus presentation was self-paced. These data provide evidence for developmental differences in implicit sequence learning that are dependent on specific task demands such as stimulus pacing, which may be related to developmental changes in the impact of broader constructs such as attention and task motivation on implicit learning.

Newborn chickens generate invariant object representations at the onset of visual object experience

To recognize objects quickly and accurately, mature visual systems build invariant object representations that generalize across a range of novel viewing conditions (e.g., changes in viewpoint). To date, however, the origins of this core cognitive ability have not yet been established. To examine how invariant object recognition develops in a newborn visual system, I raised chickens from birth for 2 weeks within controlled-rearing chambers. These chambers provided complete control over all visual object experiences. In the first week of life, subjects' visual object experience was limited to a single virtual object rotating through a 60° viewpoint range. In the second week of life, I examined whether subjects could recognize that virtual object from novel viewpoints. Newborn chickens were able to generate viewpoint-invariant representations that supported object recognition across large, novel, and complex changes in the object's appearance. Thus, newborn visual systems can begin building invariant object representations at the onset of visual object experience. These abstract representations can be generated from sparse data, in this case from a visual world containing a single virtual object seen from a limited range of viewpoints. This study shows that powerful, robust, and invariant object recognition machinery is an inherent feature of the newborn brain.