Learning at a distance II. Statistical learning of non-adjacent dependencies in a non-human primate

The dynamics of statistical learning in visual search and its interaction with salience processing: An EEG study

Visual attention can be guided by statistical regularities in the environment, that people implicitly learn from past experiences (statistical learning, SL). Moreover, a perceptually salient element can automatically capture attention, gaining processing priority through a bottom-up attentional control mechanism. The aim of our study was to investigate the dynamics of SL and if it shapes attentional target selection additively with salience processing, or whether these mechanisms interact, e.g. one gates the other. In a visual search task, we therefore manipulated target frequency (high vs. low) across locations while, in some trials, the target was salient in terms of colour. Additionally, halfway through the experiment, the high-frequency location changed to the opposite hemifield. EEG activity was simultaneously recorded, with a specific interest in two markers related to target selection and post-selection processing, respectively: N2pc and SPCN. Our results revealed that both SL and saliency significantly enhanced behavioural performance, but also interacted with each other, with an attenuated saliency effect at the high-frequency target location, and a smaller SL effect for salient targets. Concerning processing dynamics, the benefit of salience processing was more evident during the early stage of target selection and processing, as indexed by a larger N2pc and early-SPCN, whereas SL modulated the underlying neural activity particularly later on, as revealed by larger late-SPCN. Furthermore, we showed that SL was rapidly acquired and adjusted when the spatial imbalance changed. Overall, our findings suggest that SL is flexible to changes and, combined with salience processing, jointly contributes to establishing attentional priority.

Integrated effects of top-down attention and statistical learning during visual search: An EEG study

The present study aims to investigate how the competition between visual elements is solved by top-down and/or statistical learning (SL) attentional control (AC) mechanisms when active together. We hypothesized that the "winner" element that will undergo further processing is selected either by one AC mechanism that prevails over the other, or by the joint activity of both mechanisms. To test these hypotheses, we conducted a visual search experiment that combined an endogenous cueing protocol (valid vs. neutral cue) and an imbalance of target frequency distribution across locations (high- vs. low-frequency location). The unique and combined effects of top-down control and SL mechanisms were measured on behaviour and amplitudes of three evoked-response potential (ERP) components (i.e., N2pc, P1, CNV) related to attentional processing. Our behavioural results showed better performance for validly cued targets and for targets in the high-frequency location. The two factors were found to interact, so that SL effects emerged only in the absence of top-down guidance. Whereas the CNV and P1 only displayed a main effect of cueing, for the N2pc we observed an interaction between cueing and SL, revealing a cueing effect for targets in the low-frequency condition, but not in the high-frequency condition. Thus, our data support the view that top-down control and SL work in a conjoint, integrated manner during target selection. In particular, SL mechanisms are reduced or even absent when a fully reliable top-down guidance of attention is at play.

No need to forget, just keep the balance: Hebbian neural networks for statistical learning

Language processing in humans has long been proposed to rely on sophisticated learning abilities including statistical learning. Endress and Johnson (E&J, 2021) recently presented a neural network model for statistical learning based on Hebbian learning principles. This model accounts for word segmentation tasks, one primary paradigm in statistical learning. In this discussion paper we review this model and compare it with the Hebbian model previously presented by Tovar and Westermann (T&W, 2017a; 2017b; 2018) that has accounted for serial reaction time tasks, cross-situational learning, and categorization paradigms, all relevant in the study of statistical learning. We discuss the similarities and differences between both models, and their key findings. From our analysis, we question the concept of "forgetting" in the model of E&J and their suggestion of considering forgetting as the critical ingredient for successful statistical learning. We instead suggest that a set of simple but well-balanced mechanisms including spreading activation, activation persistence, and synaptic weight decay, all based on biologically grounded principles, allow modeling statistical learning in Hebbian neural networks, as demonstrated in the T&W model which successfully covers learning of nonadjacent dependencies and accounts for differences between typical and atypical populations, both aspects that have not been fully demonstrated in the E&J model. We outline the main computational and theoretical differences between the E&J and T&W approaches, present new simulation results, and discuss implications for the development of a computational cognitive theory of statistical learning.

Memory for nonadjacent dependencies in the first year of life and its relation to sleep

Grammar learning requires memory for dependencies between nonadjacent elements in speech. Immediate learning of nonadjacent dependencies has been observed in very young infants, but their memory of such dependencies has remained unexplored. Here we used event-related potentials to investigate whether 6- to 8-month-olds retain nonadjacent dependencies and if sleep after learning affects this memory. Infants were familiarised with two rule-based morphosyntactic dependencies, presented in sentences of an unknown language. Brain responses after a retention period reveal memory of the nonadjacent dependencies, independent of whether infants napped or stayed awake. Napping, however, altered a specific processing stage, suggesting that memory evolves during sleep. Infants with high left frontal spindle activity show an additional brain response indicating memory of individual speech phrases. Results imply that infants as young as 6 months are equipped with memory mechanisms relevant to grammar learning. They also suggest that during sleep, consolidation of highly specific information can co-occur with changes in the nature of generalised memory.

Statistical Learning of Language: A Meta-Analysis Into 25 Years of Research

Statistical learning is a key concept in our understanding of language acquisition. Ample work has highlighted its role in numerous linguistic functions-yet statistical learning is not a unitary construct, and its consistency across different language properties remains unclear. In a meta-analysis of auditory-linguistic statistical learning research spanning the last 25 years, we evaluated how learning varies across different language properties in infants, children, and adults and surveyed the methodological trends in the literature. We found robust learning across stimuli (syllables, words, etc.) in infants, and across stimuli and structures (adjacent dependencies, non-adjacent dependencies, etc.) in adults, with larger effect sizes when multiple cues were present. However, the analysis also showed significant publication bias and revealed a tendency toward using a narrow range of simplified language properties, including in the strength of the transitional probabilities used during training. Bayes factor analyses revealed prevalent data insensitivity of moderators commonly hypothesized to impact learning, such as the amount of exposure and transitional probability strength, which contradict core theoretical assumptions in the field. Methodological factors, such as the tasks used at test, also significantly impacted effect sizes in adults and children, suggesting that choice of task may critically constrain current theories of how statistical learning operates. Collectively, our results suggest that auditory-linguistic statistical learning has the kind of robustness needed to play a foundational role in language acquisition, but that more research is warranted to reveal its full potential.

Chimpanzees produce diverse vocal sequences with ordered and recombinatorial properties

The origins of human language remains a major question in evolutionary science. Unique to human language is the capacity to flexibly recombine a limited sound set into words and hierarchical sequences, generating endlessly new sentences. In contrast, sequence production of other animals appears limited, stunting meaning generation potential. However, studies have rarely quantified flexibility and structure of vocal sequence production across the whole repertoire. Here, we used such an approach to examine the structure of vocal sequences in chimpanzees, known to combine calls used singly into longer sequences. Focusing on the structure of vocal sequences, we analysed 4826 recordings of 46 wild adult chimpanzees from Taï National Park. Chimpanzees produced 390 unique vocal sequences. Most vocal units emitted singly were also emitted in two-unit sequences (bigrams), which in turn were embedded into three-unit sequences (trigrams). Bigrams showed positional and transitional regularities within trigrams with certain bigrams predictably occurring in either head or tail positions in trigrams, and predictably co-occurring with specific other units. From a purely structural perspective, the capacity to organize single units into structured sequences offers a versatile system potentially suitable for expansive meaning generation. Further research must show to what extent these structural sequences signal predictable meanings.

Gray Matter Variation in the Posterior Superior Temporal Gyrus Is Associated with Polymorphisms in the KIAA0319 Gene in Chimpanzees (Pan troglodytes)

Determining the impact that the KIAA0319 gene has on primate brain morphology can provide insight into the evolution of human cognition and language systems. Here, we tested whether polymorphisms in KIAA0319 in chimpanzees account for gray matter volumetric variation in brain regions implicated in language and communication (particularly within the posterior superior temporal gyrus and inferior frontal gyrus). First, we identified the nature and frequencies of single nucleotide variants (SNVs) in KIAA0319 in a sample of unrelated chimpanzees (Pan troglodytes spp.). Next, we genotyped a subset of SNVs (those important for gene regulation or likely to alter protein structure/function) in a sample of chimpanzees for which in vivo T1-structural magnetic resonance imaging scans had been obtained. We then used source-based morphometry (SBM) to test for whole-brain gray matter covariation differences between chimpanzees with different KIAA0319 alleles. Finally, using histologic sections of 15 postmortem chimpanzee brains, we analyzed microstructural variation related to KIAA0319 polymorphisms in the posterior superior temporal cortex. We found that the SNVs were associated with variation in gray matter within several brain regions, including the posterior superior temporal gyrus (a region associated with language comprehension and production in humans). The microstructure analysis further revealed hemispheric differences in neuropil fraction, indicating that KIAA0319 expression may be involved in regulation of processes related to the formation and maintenance of synapses, dendrites, or axons within regions associated with communication.

Nonhuman primates learn adjacent dependencies but fail to learn nonadjacent dependencies in a statistical learning task with a salient cue

There is ample evidence that humans and nonhuman animals can learn complex statistical regularities presented within various types of input. However, humans outperform their nonhuman primate counterparts when it comes to recognizing relationships that exist across one or several intervening stimuli (nonadjacent dependencies). This is especially true when the two elements in the dependency do not share any perceptual similarity (arbitrary associations). In the present study, we investigated whether manipulating the saliency of the predictive stimulus would enhance nonadjacent dependency learning in nonhuman primates. Rhesus macaques and tufted capuchins engaged in a computerized signal detection task that included sequences that were random in nature, included an adjacent dependency, or included a nonadjacent dependency. We manipulated the saliency of the predictive stimulus, such that the predictor jittered in place on the screen in some grammar blocks, as well as the transitional probability (the likelihood of the stimulus preceding the target to accurately predict the target's appearance) from block to block. Some monkeys evidenced learning of adjacent dependencies by faster response times to targets that followed a predictive stimulus compared to targets that were not preceded by a predictor. However, consistent with the body of evidence that indicates that nonhuman animals' statistical learning mechanisms are not at the same level of sophistication as humans', there was no evidence that monkeys learned nonadjacent dependencies of arbitrary associations, even when the salient cue was present.

Arc-shaped pitch contours facilitate item recognition in non-human animals

Acoustic changes linked to natural prosody are a key source of information about the organization of language. Both human infants and adults readily take advantage of such changes to discover and memorize linguistic patterns. Do they so because our brain is efficiently wired to specifically process linguistic stimuli? Or are we co-opting for language acquisition purposes more general principles that might be inherited from our animal ancestors? Here, we address this question by exploring if other species profit from prosody to better process acoustic sequences. More specifically, we test whether arc-shaped pitch contours defining natural prosody might facilitate item recognition and memorization in rats. In two experiments, we presented to the rats nonsense words with flat, natural, inverted and random prosodic contours. We observed that the animals correctly recognized the familiarization words only when arc-shaped pitch contours were implemented over them. Our results suggest that other species might also benefit from prosody for the memorization of items in a sequence. Such capacity seems to be rooted in general principles of how biological sounds are produced and processed.

Nonadjacent dependency processing in monkeys, apes, and humans

The ability to track syntactic relationships between words, particularly over distances ("nonadjacent dependencies"), is a critical faculty underpinning human language, although its evolutionary origins remain poorly understood. While some monkey species are reported to process auditory nonadjacent dependencies, comparative data from apes are missing, complicating inferences regarding shared ancestry. Here, we examined nonadjacent dependency processing in common marmosets, chimpanzees, and humans using "artificial grammars": strings of arbitrary acoustic stimuli composed of adjacent (nonhumans) or nonadjacent (all species) dependencies. Individuals from each species (i) generalized the grammars to novel stimuli and (ii) detected grammatical violations, indicating that they processed the dependencies between constituent elements. Furthermore, there was no difference between marmosets and chimpanzees in their sensitivity to nonadjacent dependencies. These notable similarities between monkeys, apes, and humans indicate that nonadjacent dependency processing, a crucial cognitive facilitator of language, is an ancestral trait that evolved at least ~40 million years before language itself.

Emotional Voice Intonation: A Communication Code at the Origins of Speech Processing and Word-Meaning Associations?

The aim of the present work is to investigate the facilitating effect of vocal emotional intonation on the evolution of the following processes involved in language: (a) identifying and producing phonemes, (b) processing compositional rules underlying vocal utterances, and (c) associating vocal utterances with meanings. To this end, firstly, I examine research on the presence of these abilities in animals, and the biologically ancient nature of emotional vocalizations. Secondly, I review research attesting to the facilitating effect of emotional voice intonation on these abilities in humans. Thirdly, building on these studies in animals and humans, and through taking an evolutionary perspective, I provide insights for future empirical work on the facilitating effect of emotional intonation on these three processes in animals and preverbal humans. In this work, I highlight the importance of a comparative approach to investigate language evolution empirically. This review supports Darwin’s hypothesis, according to which the ability to express emotions through voice modulation was a key step in the evolution of spoken language.

Non-adjacent Dependency Learning in Humans and Other Animals

Learning and processing natural language requires the ability to track syntactic relationships between words and phrases in a sentence, which are often separated by intervening material. These nonadjacent dependencies can be studied using artificial grammar learning paradigms and structured sequence processing tasks. These approaches have been used to demonstrate that human adults, infants and some nonhuman animals are able to detect and learn dependencies between nonadjacent elements within a sequence. However, learning nonadjacent dependencies appears to be more cognitively demanding than detecting dependencies between adjacent elements, and only occurs in certain circumstances. In this review, we discuss different types of nonadjacent dependencies in language and in artificial grammar learning experiments, and how these differences might impact learning. We summarize different types of perceptual cues that facilitate learning, by highlighting the relationship between dependent elements bringing them closer together either physically, attentionally, or perceptually. Finally, we review artificial grammar learning experiments in human adults, infants, and nonhuman animals, and discuss how similarities and differences observed across these groups can provide insights into how language is learned across development and how these language-related abilities might have evolved.

Structured Sequence Learning: Animal Abilities, Cognitive Operations, and Language Evolution

Human language is a salient example of a neurocognitive system that is specialized to process complex dependencies between sensory events distributed in time, yet how this system evolved and specialized remains unclear. Artificial Grammar Learning (AGL) studies have generated a wealth of insights into how human adults and infants process different types of sequencing dependencies of varying complexity. The AGL paradigm has also been adopted to examine the sequence processing abilities of nonhuman animals. We critically evaluate this growing literature in species ranging from mammals (primates and rats) to birds (pigeons, songbirds, and parrots) considering also cross-species comparisons. The findings are contrasted with seminal studies in human infants that motivated the work in nonhuman animals. This synopsis identifies advances in knowledge and where uncertainty remains regarding the various strategies that nonhuman animals can adopt for processing sequencing dependencies. The paucity of evidence in the few species studied to date and the need for follow-up experiments indicate that we do not yet understand the limits of animal sequence processing capacities and thereby the evolutionary pattern. This vibrant, yet still budding, field of research carries substantial promise for advancing knowledge on animal abilities, cognitive substrates, and language evolution.

Common marmosets (Callithrix jacchus) cannot recognize global configurations of sound patterns but can recognize adjacent relations of sounds

Processing the temporal configuration of discrete sounds to extract a regular pattern is fundamental to humans' faculties of perceiving words and musical phrases. To investigate such auditory pattern perception in monkeys, I trained two common marmosets to discriminate between AB-AB and AA-BB patterns under two training paradigms. One was an absolute discrimination task, in which the discrimination between these stimuli without reference cues was required. The other was a relative discrimination task, in which the detection of a change from one stimulus to the other was required. The marmosets failed in the absolute discrimination task but achieved the relative discrimination task. Failure in the absolute task indicated that the marmosets were unable to form a representation of the global sound patterns in their long-term memory stores. In contrast, success in the relative task indicated that the marmosets had short-term memory of ongoing sounds that enabled an online monitoring to detect deviations between incoming sounds and the anticipated upcoming sounds. Thus, the current findings imply that marmosets can at least perceive adjacent tone relations in an auditory stream regardless of the temporal configuration of the global sound patterns.

Network architectures supporting learnability

Human learners acquire complex interconnected networks of relational knowledge. The capacity for such learning naturally depends on two factors: the architecture (or informational structure) of the knowledge network itself and the architecture of the computational unit-the brain-that encodes and processes the information. That is, learning is reliant on integrated network architectures at two levels: the epistemic and the computational, or the conceptual and the neural. Motivated by a wish to understand conventional human knowledge, here, we discuss emerging work assessing network constraints on the learnability of relational knowledge, and theories from statistical physics that instantiate the principles of thermodynamics and information theory to offer an explanatory model for such constraints. We then highlight similarities between those constraints on the learnability of relational networks, at one level, and the physical constraints on the development of interconnected patterns in neural systems, at another level, both leading to hierarchically modular networks. To support our discussion of these similarities, we employ an operational distinction between the modeller (e.g. the human brain), the model (e.g. a single human's knowledge) and the modelled (e.g. the information present in our experiences). We then turn to a philosophical discussion of whether and how we can extend our observations to a claim regarding explanation and mechanism for knowledge acquisition. What relation between hierarchical networks, at the conceptual and neural levels, best facilitate learning? Are the architectures of optimally learnable networks a topological reflection of the architectures of comparably developed neural networks? Finally, we contribute to a unified approach to hierarchies and levels in biological networks by proposing several epistemological norms for analysing the computational brain and social epistemes, and for developing pedagogical principles conducive to curious thought. This article is part of the theme issue 'Unifying the essential concepts of biological networks: biological insights and philosophical foundations'.

Adaptive Efficient Coding of Correlated Acoustic Properties

Natural sounds such as vocalizations often have covarying acoustic attributes, resulting in redundancy in neural coding. The efficient coding hypothesis proposes that sensory systems are able to detect such covariation and adapt to reduce redundancy, leading to more efficient neural coding. Recent psychoacoustic studies have shown the auditory system can rapidly adapt to efficiently encode two covarying dimensions as a single dimension, following passive exposure to sounds in which temporal and spectral attributes covaried in a correlated fashion. However, these studies observed a cost to this adaptation, which was a loss of sensitivity to the orthogonal dimension. Here we explore the neural basis of this psychophysical phenomenon by recording single-unit responses from the primary auditory cortex in awake ferrets exposed passively to stimuli with two correlated attributes, similar in stimulus design to the psychoacoustic experiments in humans. We found: (1) the signal-to-noise ratio of spike-rate coding of cortical responses driven by sounds with correlated attributes remained unchanged along the exposure dimension, but was reduced along the orthogonal dimension; (2) performance of a decoder trained with spike data to discriminate stimuli along the orthogonal dimension was equally reduced; (3) correlations between neurons tuned to the two covarying attributes decreased after exposure; and (4) these exposure effects still occurred if sounds were correlated along two acoustic dimensions, but varied randomly along a third dimension. These neurophysiological results are consistent with the efficient coding hypothesis and may help deepen our understanding of how the auditory system encodes and represents acoustic regularities and covariance.SIGNIFICANCE STATEMENT The efficient coding (EC) hypothesis (Attneave, 1954; Barlow, 1961) proposes that the neural code in sensory systems efficiently encodes natural stimuli by minimizing the number of spikes to transmit a sensory signal. Results of recent psychoacoustic studies in humans are consistent with the EC hypothesis in that, following passive exposure to stimuli with correlated attributes, the auditory system rapidly adapts so as to more efficiently encode the two covarying dimensions as a single dimension. In the current neurophysiological experiments, using a similar stimulus design and the experimental paradigm to the psychoacoustic studies of Stilp et al. (2010) and Stilp and Kluender (2011, 2012, 2016), we recorded responses from single neurons in the auditory cortex of the awake ferret, showing adaptive efficient neural coding of two correlated acoustic attributes.

A Mathematical Approach to Correlating Objective Spectro-Temporal Features of Non-linguistic Sounds With Their Subjective Perceptions in Humans

Non-linguistic sounds (NLSs) are a core feature of our everyday life and many evoke powerful cognitive and emotional outcomes. The subjective perception of NLSs by humans has occasionally been defined for single percepts, e.g., their pleasantness, whereas many NLSs evoke multiple perceptions. There has also been very limited attempt to determine if NLS perceptions are predicted from objective spectro-temporal features. We therefore examined three human perceptions well-established in previous NLS studies ("Complexity," "Pleasantness," and "Familiarity"), and the accuracy of identification, for a large NLS database and related these four measures to objective spectro-temporal NLS features, defined using rigorous mathematical descriptors including stimulus entropic and algorithmic complexity measures, peaks-related measures, fractal dimension estimates, and various spectral measures (mean spectral centroid, power in discrete frequency ranges, harmonicity, spectral flatness, and spectral structure). We mapped the perceptions to the spectro-temporal measures individually and in combinations, using complex multivariate analyses including principal component analyses and agglomerative hierarchical clustering.

Is the consonant bias specifically human? Long-Evans rats encode vowels better than consonants in words

In natural languages, vowels tend to convey structures (syntax, prosody) while consonants are more important lexically. The consonant bias, which is the tendency to rely more on consonants than on vowels to process words, is well attested in human adults and infants after the first year of life. Is the consonant bias based on evolutionarily ancient mechanisms, potentially present in other species? The current study investigated this issue in a species phylogenetically distant from humans: Long-Evans rats. During training, the animals were presented with four natural word-forms (e.g., mano, "hand"). We then compared their responses to novel words carrying either a consonant (pano) or a vowel change (meno). Results show that the animals were less disrupted by consonantal alterations than by vocalic alterations of words. That is, word recognition was more affected by the alteration of a vowel than a consonant. Together with previous findings in very young human infants, this reliance on vocalic information we observe in rats suggests that the emergence of the consonant bias may require a combination of vocal, cognitive and auditory skills that rodents do not seem to possess.

Common marmosets are sensitive to simple dependencies at variable distances in an artificial grammar

Recognizing that two elements within a sequence of variable length depend on each other is a key ability in understanding the structure of language and music. Perception of such interdependencies has previously been documented in chimpanzees in the visual domain and in human infants and common squirrel monkeys with auditory playback experiments, but it remains unclear whether it typifies primates in general. Here, we investigated the ability of common marmosets (Callithrix jacchus) to recognize and respond to such dependencies. We tested subjects in a familiarization-discrimination playback experiment using stimuli composed of pure tones that either conformed or did not conform to a grammatical rule. After familiarization to sequences with dependencies, marmosets spontaneously discriminated between sequences containing and lacking dependencies ('consistent' and 'inconsistent', respectively), independent of stimulus length. Marmosets looked more often to the sound source when hearing sequences consistent with the familiarization stimuli, as previously found in human infants. Crucially, looks were coded automatically by computer software, avoiding human bias. Our results support the hypothesis that the ability to perceive dependencies at variable distances was already present in the common ancestor of all anthropoid primates (Simiiformes).

Capacities and neural mechanisms for auditory statistical learning across species

Statistical learning has been proposed as a possible mechanism by which individuals can become sensitive to the structures of language fundamental for speech perception. Since its description in human infants, statistical learning has been described in human adults and several non-human species as a general process by which animals learn about stimulus-relevant statistics. The neurobiology of statistical learning is beginning to be understood, but many questions remain about the underlying mechanisms. Why is the developing brain particularly sensitive to stimulus and environmental statistics, and what neural processes are engaged in the adult brain to enable learning from statistical regularities in the absence of external reward or instruction? This review will survey the statistical learning abilities of humans and non-human animals with a particular focus on communicative vocalizations. We discuss the neurobiological basis of statistical learning, and specifically what can be learned by exploring this process in both humans and laboratory animals. Finally, we describe advantages of studying vocal communication in rodents as a means to further our understanding of the cortical plasticity mechanisms engaged during statistical learning. We examine the use of rodents in the context of pup retrieval, which is an auditory-based and experience-dependent form of maternal behavior.

Auditory sequence perception in common marmosets (Callithrix jacchus)

One of the essential linguistic and musical faculties of humans is the ability to recognize the structure of sound configurations and to extract words and melodies from continuous sound sequences. However, monkeys' ability to process the temporal structure of sounds is controversial. Here, to investigate whether monkeys can analyze the temporal structure of auditory patterns, two common marmosets were trained to discriminate auditory patterns in three experiments. In Experiment 1, the marmosets were able to discriminate trains of either 0.5- or 2-kHz tones repeated in either 50- or 200-ms intervals. However, the marmosets were not able to discriminate ABAB from AABB patterns consisting of A (0.5-kHz/50-ms pulse) and B (2-kHz/200-ms pulse) elements in Experiment 2, and A (0.5-kHz/50-ms pulse) and B (0.5-kHz/200-ms pulse) [or A (0.5-kHz/200-ms pulse) and B (2-kHz/200-ms pulse)] in Experiment 3. Consequently, the results indicated that the marmosets could not perceive tonal structures in terms of the temporal configuration of discrete sounds, whereas they could recognize the acoustic features of the stimuli. The present findings were supported by cognitive and brain studies that indicated a limited ability to process sound sequences. However, more studies are needed to confirm the ability of auditory sequence perception in common marmosets.

Auditory and Visual Sequence Learning in Humans and Monkeys using an Artificial Grammar Learning Paradigm

Language flexibly supports the human ability to communicate using different sensory modalities, such as writing and reading in the visual modality and speaking and listening in the auditory domain. Although it has been argued that nonhuman primate communication abilities are inherently multisensory, direct behavioural comparisons between human and nonhuman primates are scant. Artificial grammar learning (AGL) tasks and statistical learning experiments can be used to emulate ordering relationships between words in a sentence. However, previous comparative work using such paradigms has primarily investigated sequence learning within a single sensory modality. We used an AGL paradigm to evaluate how humans and macaque monkeys learn and respond to identically structured sequences of either auditory or visual stimuli. In the auditory and visual experiments, we found that both species were sensitive to the ordering relationships between elements in the sequences. Moreover, the humans and monkeys produced largely similar response patterns to the visual and auditory sequences, indicating that the sequences are processed in comparable ways across the sensory modalities. These results provide evidence that human sequence processing abilities stem from an evolutionarily conserved capacity that appears to operate comparably across the sensory modalities in both human and nonhuman primates. The findings set the stage for future neurobiological studies to investigate the multisensory nature of these sequencing operations in nonhuman primates and how they compare to related processes in humans.

Implicit Memory for Complex Sounds in Higher Auditory Cortex of the Ferret

Responses of auditory cortical neurons encode sound features of incoming acoustic stimuli and also are shaped by stimulus context and history. Previous studies of mammalian auditory cortex have reported a variable time course for such contextual effects ranging from milliseconds to minutes. However, in secondary auditory forebrain areas of songbirds, long-term stimulus-specific neuronal habituation to acoustic stimuli can persist for much longer periods of time, ranging from hours to days. Such long-term habituation in the songbird is a form of long-term auditory memory that requires gene expression. Although such long-term habituation has been demonstrated in avian auditory forebrain, this phenomenon has not previously been described in the mammalian auditory system. Utilizing a similar version of the avian habituation paradigm, we explored whether such long-term effects of stimulus history also occur in auditory cortex of a mammalian auditory generalist, the ferret. Following repetitive presentation of novel complex sounds, we observed significant response habituation in secondary auditory cortex, but not in primary auditory cortex. This long-term habituation appeared to be independent for each novel stimulus and often lasted for at least 20 min. These effects could not be explained by simple neuronal fatigue in the auditory pathway, because time-reversed sounds induced undiminished responses similar to those elicited by completely novel sounds. A parallel set of pupillometric response measurements in the ferret revealed long-term habituation effects similar to observed long-term neural habituation, supporting the hypothesis that habituation to passively presented stimuli is correlated with implicit learning and long-term recognition of familiar sounds.SIGNIFICANCE STATEMENT Long-term habituation in higher areas of songbird auditory forebrain is associated with gene expression and is correlated with recognition memory. Similar long-term auditory habituation in mammals has not been previously described. We studied such habituation in single neurons in the auditory cortex of awake ferrets that were passively listening to repeated presentations of various complex sounds. Responses exhibited long-lasting habituation (at least 20 min) in the secondary, but not primary auditory cortex. Habituation ceased when stimuli were played backward, despite having identical spectral content to the original sound. This long-term neural habituation correlated with similar habituation of ferret pupillary responses to repeated presentations of the same stimuli, suggesting that stimulus habituation is retained as a long-term behavioral memory.

A Comparative Perspective on the Role of Acoustic Cues in Detecting Language Structure

Most human language learners acquire language primarily via the auditory modality. This is one reason why auditory artificial grammars play a prominent role in the investigation of the development and evolutionary roots of human syntax. The present position paper brings together findings from human and non-human research on the impact of auditory cues on learning about linguistic structures with a special focus on how different types of cues and biases in auditory cognition may contribute to success and failure in artificial grammar learning (AGL). The basis of our argument is the link between auditory cues and syntactic structure across languages and development. Cross-species comparison suggests that many aspects of auditory cognition that are relevant for language are not human specific and are present even in rather distantly related species. Furthermore, auditory cues and biases impact on learning, which we will discuss in the example of auditory perception and AGL studies. This observation, together with the significant role of auditory cues in language processing, supports the idea that auditory cues served as a bootstrap to syntax during language evolution. Yet this also means that potentially human-specific syntactic abilities are not due to basic auditory differences between humans and non-human animals but are based upon more advanced cognitive processes.

Non-adjacent Dependencies Processing in Human and Non-human Primates

Human and non-human primates share the ability to extract adjacent dependencies and, under certain conditions, non-adjacent dependencies (i.e., predictive relationships between elements that are separated by one or several intervening elements in a sequence). In this study, we explore the online extraction dynamics of non-adjacent dependencies in humans and baboons using a serial reaction time task. Participants had to produce three-target sequences containing deterministic relationships between the first and last target locations. In Experiment 1, participants from the two species could extract these non-adjacent dependencies, but humans required less exposure than baboons. In Experiment 2, the data show for the first time in a non-human primate species the successful generalization of sequential non-adjacent dependencies over novel intervening items. These findings provide new evidence to further constrain current theories about the nature and the evolutionary origins of the learning mechanisms allowing the extraction of non-adjacent dependencies.

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.

Broca's region: A causal role in implicit processing of grammars with crossed non-adjacent dependencies

Non-adjacent dependencies are challenging for the language learning machinery and are acquired later than adjacent dependencies. In this transcranial magnetic stimulation (TMS) study, we show that participants successfully discriminated between grammatical and non-grammatical sequences after having implicitly acquired an artificial language with crossed non-adjacent dependencies. Subsequent to transcranial magnetic stimulation of Broca's region, discrimination was impaired compared to when a language-irrelevant control region (vertex) was stimulated. These results support the view that Broca's region is engaged in structured sequence processing and extend previous functional neuroimaging results on artificial grammar learning (AGL) in two directions: first, the results establish that Broca's region is a causal component in the processing of non-adjacent dependencies, and second, they show that implicit processing of non-adjacent dependencies engages Broca's region. Since patients with lesions in Broca's region do not always show grammatical processing difficulties, the result that Broca's region is causally linked to processing of non-adjacent dependencies is a step towards clarification of the exact nature of syntactic deficits caused by lesions or perturbation to Broca's region. Our findings are consistent with previous results and support a role for Broca's region in general structured sequence processing, rather than a specific role for the processing of hierarchically organized sentence structure.

Conserved Sequence Processing in Primate Frontal Cortex

An important aspect of animal perception and cognition is learning to recognize relationships between environmental events that predict others in time, a form of relational knowledge that can be assessed using sequence-learning paradigms. Humans are exquisitely sensitive to sequencing relationships, and their combinatorial capacities, most saliently in the domain of language, are unparalleled. Recent comparative research in human and nonhuman primates has obtained behavioral and neuroimaging evidence for evolutionarily conserved substrates involved in sequence processing. The findings carry implications for the origins of domain-general capacities underlying core language functions in humans. Here, we synthesize this research into a 'ventrodorsal gradient' model, where frontal cortex engagement along this axis depends on sequencing complexity, mapping onto the sequencing capacities of different species.

Evolutionary origins of non-adjacent sequence processing in primate brain potentials

There is considerable interest in understanding the ontogeny and phylogeny of the human language system, yet, neurobiological work at the interface of both fields is absent. Syntactic processes in language build on sensory processing and sequencing capabilities on the side of the receiver. While we better understand language-related ontogenetic changes in the human brain, it remains a mystery how neurobiological processes at specific human development stages compare with those in phylogenetically closely related species. To address this knowledge gap, we measured EEG event-related potentials (ERPs) in two macaque monkeys using a paradigm developed to evaluate human infant and adult brain potentials associated with the processing of non-adjacent ordering relationships in sequences of syllable triplets. Frequent standard triplet sequences were interspersed with infrequent voice pitch or non-adjacent rule deviants. Monkey ERPs show early pitch and rule deviant mismatch responses that are strikingly similar to those previously reported in human infants. This stands in contrast to adults' later ERP responses for rule deviants. The results reveal how non-adjacent sequence ordering relationships are processed in the primate brain and provide evidence for evolutionarily conserved neurophysiological effects, some of which are remarkably like those seen at an early human developmental stage.

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.

Something Old, Something New

Infants face the paramount task of learning a language. Here, I review recent literature on two separate topics that suggest they use a combination of both evolutionarily old and new cognitive tools to face this task. Research on the principles that guide how humans and nonhuman animals group sequences of sounds has shown that we share with other species perceptual biases that we apply to linguistic stimuli. On the contrary, research on processing differences between consonants and vowels suggests humans, but not other animals, benefit from a “division of labor” across phonological representations. This division would help to extract regularities from the speech signal and facilitate language learning. The studies reviewed here provide support for the idea that perceptual biases together with language-specific representations guide the discovery of linguistic structures.

Can Chunk Size Differences Explain Developmental Changes in Lexical Learning?

In three experiments, we investigated Hebb repetition learning (HRL) differences between children and adults, as a function of the type of item (lexical vs. sub-lexical) and the level of item-overlap between sequences. In a first experiment, it was shown that when non-repeating and repeating (Hebb) sequences of words were all permutations of the same words, HRL was slower than when the sequences shared no words. This item-overlap effect was observed in both children and adults. In a second experiment, we used syllable sequences and we observed reduced HRL due to item-overlap only in children. The findings are explained within a chunking account of the HRL effect on the basis of which we hypothesize that children, compared with adults, chunk syllable sequences in smaller units. By hypothesis, small chunks are more prone to interference from anagram representations included in the filler sequences, potentially explaining the item-overlap effect in children. This hypothesis was tested in a third experiment with adults where we experimentally manipulated the chunk size by embedding pauses in the syllable sequences. Interestingly, we showed that imposing a small chunk size caused adults to show the same behavioral effects as those observed in children. Departing from the analogy between verbal HRL and lexical development, the results are discussed in light of the less-is-more hypothesis of age-related differences in language acquisition.

Auditory sequence processing reveals evolutionarily conserved regions of frontal cortex in macaques and humans

An evolutionary account of human language as a neurobiological system must distinguish between human-unique neurocognitive processes supporting language and evolutionarily conserved, domain-general processes that can be traced back to our primate ancestors. Neuroimaging studies across species may determine whether candidate neural processes are supported by homologous, functionally conserved brain areas or by different neurobiological substrates. Here we use functional magnetic resonance imaging in Rhesus macaques and humans to examine the brain regions involved in processing the ordering relationships between auditory nonsense words in rule-based sequences. We find that key regions in the human ventral frontal and opercular cortex have functional counterparts in the monkey brain. These regions are also known to be associated with initial stages of human syntactic processing. This study raises the possibility that certain ventral frontal neural systems, which play a significant role in language function in modern humans, originally evolved to support domain-general abilities involved in sequence processing.

The Temporal Dynamics of Regularity Extraction in Non-Human Primates

Extracting the regularities of our environment is one of our core cognitive abilities. To study the fine-grained dynamics of the extraction of embedded regularities, a method combining the advantages of the artificial language paradigm (Saffran, Aslin, & Newport, ) and the serial response time task (Nissen & Bullemer, ) was used with a group of Guinea baboons (Papio papio) in a new automatic experimental device (Fagot & Bonté, ). After a series of random trials, monkeys were exposed to language-like patterns. We found that the extraction of embedded patterns positioned at the end of larger patterns was faster than the extraction of initial embedded patterns. This result suggests that there is a learning advantage for the final element of a sequence that benefits from the contextual information provided by previous elements.

The Relationship Between Artificial and Second Language Learning

Artificial language learning (ALL) experiments have become an important tool in exploring principles of language and language learning. A persistent question in all of this work, however, is whether ALL engages the linguistic system and whether ALL studies are ecologically valid assessments of natural language ability. In the present study, we considered these questions by examining the relationship between performance in an ALL task and second language learning ability. Participants enrolled in a Spanish language class were evaluated using a number of different measures of Spanish ability and classroom performance, which was compared to IQ and a number of different measures of ALL performance. The results show that success in ALL experiments, particularly more complex artificial languages, correlates positively with indices of L2 learning even after controlling for IQ. These findings provide a key link between studies involving ALL and our understanding of second language learning in the classroom.

Learned states of preparatory attentional control

Individuals regularly experience fluctuations in the ability to perform cognitive operations. Although previous research has focused on predicting cognitive flexibility from persistent individual traits, as well as from spontaneous fluctuations in neural activity, the role of learning in shaping preparatory attentional control remains poorly understood. Across 3 experiments, we manipulated the statistical regularities of an attentional orienting paradigm to examine whether individuals modulated attentional flexibility, the readiness to perform a spatial shift of attention, across learned contexts. We found evidence of learning-based modulations in preparatory attentional control settings when the probability of shifting the focus of attention differed based on temporally or color-defined contexts. Furthermore, in the case of color-defined contexts, these modulations in preparatory control persisted even after a change in the underlying statistical properties. Our results indicate that dynamic adjustments in preparatory attentional control are sensitive to the underlying statistical regularities of an environment. This finding has implications for understanding disordered patterns of attentional control and how these patterns might be modified with training.

Mixed-complexity artificial grammar learning in humans and macaque monkeys: evaluating learning strategies

Artificial grammars (AG) can be used to generate rule-based sequences of stimuli. Some of these can be used to investigate sequence-processing computations in non-human animals that might be related to, but not unique to, human language. Previous AG learning studies in non-human animals have used different AGs to separately test for specific sequence-processing abilities. However, given that natural language and certain animal communication systems (in particular, song) have multiple levels of complexity, mixed-complexity AGs are needed to simultaneously evaluate sensitivity to the different features of the AG. Here, we tested humans and Rhesus macaques using a mixed-complexity auditory AG, containing both adjacent (local) and non-adjacent (longer-distance) relationships. Following exposure to exemplary sequences generated by the AG, humans and macaques were individually tested with sequences that were either consistent with the AG or violated specific adjacent or non-adjacent relationships. We observed a considerable level of cross-species correspondence in the sensitivity of both humans and macaques to the adjacent AG relationships and to the statistical properties of the sequences. We found no significant sensitivity to the non-adjacent AG relationships in the macaques. A subset of humans was sensitive to this non-adjacent relationship, revealing interesting between- and within-species differences in AG learning strategies. The results suggest that humans and macaques are largely comparably sensitive to the adjacent AG relationships and their statistical properties. However, in the presence of multiple cues to grammaticality, the non-adjacent relationships are less salient to the macaques and many of the humans.

Non-adjacent visual dependency learning in chimpanzees

Humans have a strong proclivity for structuring and patterning stimuli: Whether in space or time, we tend to mentally order stimuli in our environment and organize them into units with specific types of relationships. A crucial prerequisite for such organization is the cognitive ability to discern and process regularities among multiple stimuli. To investigate the evolutionary roots of this cognitive capacity, we tested chimpanzees—which, along with bonobos, are our closest living relatives—for simple, variable distance dependency processing in visual patterns. We trained chimpanzees to identify pairs of shapes either linked by an arbitrary learned association (arbitrary associative dependency) or a shared feature (same shape, feature-based dependency), and to recognize strings where items related to either of these ways occupied the first (leftmost) and the last (rightmost) item of the stimulus. We then probed the degree to which subjects generalized this pattern to new colors, shapes, and numbers of interspersed items. We found that chimpanzees can learn and generalize both types of dependency rules, indicating that the ability to encode both feature-based and arbitrary associative regularities over variable distances in the visual domain is not a human prerogative. Our results strongly suggest that these core components of human structural processing were already present in our last common ancestor with chimpanzees.

Statistical learning of recurring sound patterns encodes auditory objects in songbird forebrain

Auditory neurophysiology has demonstrated how basic acoustic features are mapped in the brain, but it is still not clear how multiple sound components are integrated over time and recognized as an object. We investigated the role of statistical learning in encoding the sequential features of complex sounds by recording neuronal responses bilaterally in the auditory forebrain of awake songbirds that were passively exposed to long sound streams. These streams contained sequential regularities, and were similar to streams used in human infants to demonstrate statistical learning for speech sounds. For stimulus patterns with contiguous transitions and with nonadjacent elements, single and multiunit responses reflected neuronal discrimination of the familiar patterns from novel patterns. In addition, discrimination of nonadjacent patterns was stronger in the right hemisphere than in the left, and may reflect an effect of top-down modulation that is lateralized. Responses to recurring patterns showed stimulus-specific adaptation, a sparsening of neural activity that may contribute to encoding invariants in the sound stream and that appears to increase coding efficiency for the familiar stimuli across the population of neurons recorded. As auditory information about the world must be received serially over time, recognition of complex auditory objects may depend on this type of mnemonic process to create and differentiate representations of recently heard sounds.

Auditory artificial grammar learning in macaque and marmoset monkeys

Artificial grammars (AG) are designed to emulate aspects of the structure of language, and AG learning (AGL) paradigms can be used to study the extent of nonhuman animals' structure-learning capabilities. However, different AG structures have been used with nonhuman animals and are difficult to compare across studies and species. We developed a simple quantitative parameter space, which we used to summarize previous nonhuman animal AGL results. This was used to highlight an under-studied AG with a forward-branching structure, designed to model certain aspects of the nondeterministic nature of word transitions in natural language and animal song. We tested whether two monkey species could learn aspects of this auditory AG. After habituating the monkeys to the AG, analysis of video recordings showed that common marmosets (New World monkeys) differentiated between well formed, correct testing sequences and those violating the AG structure based primarily on simple learning strategies. By comparison, Rhesus macaques (Old World monkeys) showed evidence for deeper levels of AGL. A novel eye-tracking approach confirmed this result in the macaques and demonstrated evidence for more complex AGL. This study provides evidence for a previously unknown level of AGL complexity in Old World monkeys that seems less evident in New World monkeys, which are more distant evolutionary relatives to humans. The findings allow for the development of both marmosets and macaques as neurobiological model systems to study different aspects of AGL at the neuronal level.

Action at a distance: dependency sensitivity in a New World primate

Sensitivity to dependencies (correspondences between distant items) in sensory stimuli plays a crucial role in human music and language. Here, we show that squirrel monkeys (Saimiri sciureus) can detect abstract, non-adjacent dependencies in auditory stimuli. Monkeys discriminated between tone sequences containing a dependency and those lacking it, and generalized to previously unheard pitch classes and novel dependency distances. This constitutes the first pattern learning study where artificial stimuli were designed with the species' communication system in mind. These results suggest that the ability to recognize dependencies represents a capability that had already evolved in humans’ last common ancestor with squirrel monkeys, and perhaps before.

Rule learning over consonants and vowels in a non-human animal

Perception studies have shown similarities between humans and other animals in a wide array of language-related processes. However, the components of language that make it uniquely human have not been fully identified. Here we show that nonhuman animals extract rules over speech sequences that are difficult for humans. Specifically, animals easily learn rules over both consonants and vowels, while humans do it only over vowels. In Experiment 1, rats learned a rule implemented over vowels in CVCVCV nonsense words. In Experiment 2, rats learned the rule when it was implemented over the consonants. In both experiments, rats generalized such knowledge to novel words they had not heard before. Using the same stimuli, human adults learned the rules over the vowels but not over the consonants. These results suggest differences between humans and animals on speech processing might lie on the constraints they face while extracting information from the signal.

Revisiting the syntactic abilities of non-human animals: natural vocalizations and artificial grammar learning

The domain of syntax is seen as the core of the language faculty and as the most critical difference between animal vocalizations and language. We review evidence from spontaneously produced vocalizations as well as from perceptual experiments using artificial grammars to analyse animal syntactic abilities, i.e. abilities to produce and perceive patterns following abstract rules. Animal vocalizations consist of vocal units (elements) that are combined in a species-specific way to create higher order strings that in turn can be produced in different patterns. While these patterns differ between species, they have in common that they are no more complex than a probabilistic finite-state grammar. Experiments on the perception of artificial grammars confirm that animals can generalize and categorize vocal strings based on phonetic features. They also demonstrate that animals can learn about the co-occurrence of elements or learn simple 'rules' like attending to reduplications of units. However, these experiments do not provide strong evidence for an ability to detect abstract rules or rules beyond finite-state grammars. Nevertheless, considering the rather limited number of experiments and the difficulty to design experiments that unequivocally demonstrate more complex rule learning, the question of what animals are able to do remains open.