Mark my words: High frequency marker words impact early stages of language learning

High frequency words have been suggested to benefit both speech segmentation and grammatical categorization of the words around them. Despite utilizing similar information, these tasks are usually investigated separately in studies examining learning. We determined whether including high frequency words in continuous speech could support categorization when words are being segmented for the first time. We familiarized learners with continuous artificial speech comprising repetitions of target words, which were preceded by high-frequency marker words. Crucially, marker words distinguished targets into 2 distributionally defined categories. We measured learning with segmentation and categorization tests and compared performance against a control group that heard the artificial speech without these marker words (i.e., just the targets, with no cues for categorization). Participants segmented the target words from speech in both conditions, but critically when the marker words were present, they influenced acquisition of word-referent mappings in a subsequent transfer task, with participants demonstrating better early learning for mappings that were consistent (rather than inconsistent) with the distributional categories. We propose that high-frequency words may assist early grammatical categorization, while speech segmentation is still being learned. (PsycINFO Database Record (c) 2019 APA, all rights reserved).

Statistical regularities bias overt attention

A previous study employing the additional singleton paradigm showed that a singleton distractor that appeared more often in one specific location interfered less with target search than when it appeared at any other location. These findings suggested that through statistical learning the location that was likely to contain a distractor was suppressed relative to all other locations. Even though feasible, it is also possible that this effect is due to faster disengagement of attention from the high-probability distractor location. The present study tested this hypothesis using a variant of the additional singleton task adapted for eye tracking in which observers made a speeded saccade to a shape singleton and gave a manual response. The singleton distractor was presented more often at one location than all other locations. Consistent with the suppression hypothesis, we found that fewer saccades landed at the high-probability distractor location than any other location. Also, when a target appeared at the high-probability location, saccade latencies towards the target were higher than latencies towards the target when it was presented at other locations. Furthermore, in addition to suppression, we also found evidence for faster disengagement from the high-probability distractor location than the low-probability distractor location; however, this effect was relatively small. The current findings support the notion that through statistical learning plasticity is induced in the spatial priority map of attentional selection so that the high-probability distractor location is suppressed compared to any other location.

A Single Paradigm for Implicit and Statistical Learning

Implicit learning generally refers to the acquisition of structures that, like knowledge of natural language grammar, are not available to awareness. In contrast, statistical learning has frequently been related to learning language structures that are explicitly available, such as vocabulary. In this paper, we report an experimental paradigm that enables testing of both classic implicit and statistical learning in language. The paradigm employs an artificial language comprising sentences that accompany visual scenes that they represent, thus combining artificial grammar learning with cross-situational statistical learning of vocabulary. We show that this methodology enables a comparison between acquisition of grammar and vocabulary, and the influences on their learning. We show that both grammar and vocabulary are promoted by explicit information about the language structure, that awareness of structure affects acquisition during learning, and awareness precedes learning, but is not distinctive at the endpoint of learning. The two traditions of learning-implicit and statistical-can be conjoined in a single paradigm to explore both the phenomenological and learning consequences of statistical structural knowledge.

Attentional instructions modulate differential context-switch effects after short and long training in human predictive learning

Three experiments in human predictive learning assessed the modulating role of instructions on context-switch effects on performance after different levels of training. Cue X (a food name) was paired with an outcome (gastric malaise) in Context A (a specific restaurant), whereas another cue, Y, was presented in the absence of outcome in Context B. The series manipulated the testing context (same or different from the acquisition context), the length of training (short vs. long), and the instructions participants received before testing (attentional or neutral). Attentional instructions intended to either focusing participants' attention on the context (Experiments 1 and 2) or to take attention away from the context (Experiment 3). In agreement with the predictions of the attentional theory of context processing, instructions that focused participants' attention on the context made retrieval of information after long training context specific, something that did not occur in the absence of attentional instructions. Conversely, instructions that took participants' attention away from the context (by focusing their attention on the cue) attenuated context-switch effects that otherwise appear after short training. (PsycINFO Database Record (c) 2019 APA, all rights reserved).

Statistics learned are statistics forgotten: Children's retention and retrieval of cross-situational word learning

Children are able to resolve the referential ambiguity of learning new words by tracking co-occurrence probabilities across moments in time, a behavior termed cross-situational word learning (CSWL). Although we know that children can use co-occurrence data to map words to objects, the literature has a striking limitation: research has focused on encoding of language and, consequently, children's CSWL has only been assessed at an immediate test. The current research addressed this gap in the literature by examining whether children can retain and retrieve learned words after a retention interval, and whether children's age and individual cognitive abilities contribute to their CSWL performance. The results revealed that children were able to retain and retrieve co-occurrence statistics, but only reliably so at the end of early childhood. Moreover, children's visual recognition memory abilities and the timing of learning events were the two key factors that contributed to children's performance. These findings have implications for theories and computational models of CSWL, and suggest that more research is needed to understand the processes that support CSWL after encoding. (PsycINFO Database Record (c) 2019 APA, all rights reserved).

Human Foragers: Searchers by Nature and Experience

Diverse studies of human foraging have revealed behavioral strategies that may have evolved as adaptations for foraging. Here, we used an outdoor experimental search task to explore the effect of three sources of information on participants' performance: (i) information obtained directly from performing a search, (ii) information obtained prior to testing in the form of a distilled snippet of knowledge intended to experimentally simulate information acquired culturally about the environment, and (iii) information obtained from experience of foraging for natural resources for economic gain. We found that (i) immediate searching experience improved performance from the beginning to the end of the short, 2-min task, (ii) information priming improved performance notably from the very beginning of the task, and (iii) natural resource foraging experience improved performance to a lesser extent. Our results highlight the role of culturally transmitted information as well as the presence of mechanisms to rapidly integrate and implement new information into searching choices, which ultimately influence performance in a foraging task.

Daily-life stress differentially impacts ventral striatal dopaminergic modulation of reward processing in first-degree relatives of individuals with psychosis

Emerging evidence shows that stress can impair the ability to learn from and pursue rewards, which in turn has been linked to motivational impairments characteristic of the psychotic disorder. Ventral striatal dopaminergic neurotransmission has been found to modulate reward processing, and appears to be disrupted by exposure to stress. We investigated the hypothesis that stress experienced in the everyday life has a blunting effect on reward-induced dopamine release in the ventral striatum of 16 individuals at a familial risk for psychosis compared to 16 matched control subjects. Six days of ecological momentary assessments quantified the amount of daily-life stress prior to [¹⁸F]fallypride PET imaging while performing a probabilistic reinforcement learning task. Relative to the controls, individuals at a familial risk for psychosis who encountered more daily-life stress showed significantly diminished extent of reward-induced dopamine release in the right ventral striatum, as well as poorer performance on the reward task. These findings provide the first neuromolecular evidence for stress-related deregulation of reward processing in familial predisposition to psychosis. The implication of daily-life stress in compromised modulation of reward function may facilitate the design of targeted neuropharmacological and ecological interventions.

Attractor-like Dynamics in Belief Updating in Schizophrenia

Subjects with a diagnosis of schizophrenia (Scz) overweight unexpected evidence in probabilistic inference: such evidence becomes "aberrantly salient." A neurobiological explanation for this effect is that diminished synaptic gain (e.g., hypofunction of cortical NMDARs) in Scz destabilizes quasi-stable neuronal network states (or "attractors"). This attractor instability account predicts that (1) Scz would overweight unexpected evidence but underweight consistent evidence, (2) belief updating would be more vulnerable to stochastic fluctuations in neural activity, and (3) these effects would correlate. Hierarchical Bayesian belief updating models were tested in two independent datasets (n = 80 male and n = 167 female) comprising human subjects with Scz, and both clinical and nonclinical controls (some tested when unwell and on recovery) performing the "probability estimates" version of the beads task (a probabilistic inference task). Models with a standard learning rate, or including a parameter increasing updating to "disconfirmatory evidence," or a parameter encoding belief instability were formally compared. The "belief instability" model (based on the principles of attractor dynamics) had most evidence in all groups in both datasets. Two of four parameters differed between Scz and nonclinical controls in each dataset: belief instability and response stochasticity. These parameters correlated in both datasets. Furthermore, the clinical controls showed similar parameter distributions to Scz when unwell, but were no different from controls once recovered. These findings are consistent with the hypothesis that attractor network instability contributes to belief updating abnormalities in Scz, and suggest that similar changes may exist during acute illness in other psychiatric conditions.SIGNIFICANCE STATEMENT Subjects with a diagnosis of schizophrenia (Scz) make large adjustments to their beliefs following unexpected evidence, but also smaller adjustments than controls following consistent evidence. This has previously been construed as a bias toward "disconfirmatory" information, but a more mechanistic explanation may be that in Scz, neural firing patterns ("attractor states") are less stable and hence easily altered in response to both new evidence and stochastic neural firing. We model belief updating in Scz and controls in two independent datasets using a hierarchical Bayesian model, and show that all subjects are best fit by a model containing a belief instability parameter. Both this and a response stochasticity parameter are consistently altered in Scz, as the unstable attractor hypothesis predicts.

Successful structure learning from observational data

Previous work suggests that humans find it difficult to learn the structure of causal systems given observational data alone. We identify two conditions that enable successful structure learning from observational data: people succeed if the underlying causal system is deterministic, and if each pattern of observations has a single root cause. In four experiments, we show that either condition alone is sufficient to enable high levels of performance, but that performance is poor if neither condition applies. A fifth experiment suggests that neither determinism nor root sparsity takes priority over the other. Our data are broadly consistent with a Bayesian model that embodies a preference for structures that make the observed data not only possible but probable.

Delay Discounting and Probability Discounting, But Not Response Inhibition, Are Associated With Sexual Risk Taking in Adults

Sexual risk behavior (SRB) is associated with impulsivity, but it is not clear whether different aspects of impulsivity are uniquely related to SRBs in adults. Unmarried adults ages 18 to 30 (N = 296; 41.0% male) recruited from the community reported their frequency of SRBs over the past six months using the Sexual Risk Survey (SRS; Turchik & Garske, 2009) and completed three behavioral measures of impulsivity: insensitivity to delayed consequences (delay discounting [DD]), insensitivity to probabilistic consequences (probability discounting [PD]) and response inhibition (stop signal task [SST]). SRS total scores were significantly related to both DD and PD but not SST. These findings suggest that insensitivity to delayed and probabilistic consequences, but not response inhibition, are meaningfully associated with SRBs in young adults. This may be important for the development of strategies to prevent or reduce sexual risk behaviors.

Learning unfamiliar pitch intervals: A novel paradigm for demonstrating the learning of statistical associations between musical pitches

While mastering a musical instrument takes years, becoming familiar with a new music system requires less time and skills. In this study, we examine whether musically untrained Western listeners can incidentally learn an unfamiliar, microtonal musical scale from simply engaging in a timbre discrimination task. The experiment is comprised of an Exposure and a Test phase. During Exposure, 21 non-musicians were instructed to detect a timbre shift (TS) within short microtonal melodies, and we hypothesised that they would incidentally learn about the pitch interval structure of the microtonal scale from attending to the melodies during the task. In a follow-up Test phase, the tone before the TS was either a member (congruent) or a non-member (incongruent) of the scale. Based on our statistical manipulation of the stimuli, incongruent tones would be a better predictor of an incoming TS than the congruent tones. We therefore expect a faster response time to the shift after the participants have heard an incongruent tone. Specifically, a faster response time observed after an incongruent tone would imply participants’ ability to differentiate tones from the microtonal and the diatonic scale, and reflect their learning of the microtonal pitch intervals. Results are consistent with our predictions. In investigating the learning of a microtonal scale, our study can offer directions for future research on the perception of computer music and new musical genres.

Human Pavlovian fear conditioning conforms to probabilistic learning

Learning to predict threat from environmental cues is a fundamental skill in changing environments. This aversive learning process is exemplified by Pavlovian threat conditioning. Despite a plethora of studies on the neural mechanisms supporting the formation of associations between neutral and aversive events, our computational understanding of this process is fragmented. Importantly, different computational models give rise to different and partly opposing predictions for the trial-by-trial dynamics of learning, for example expressed in the activity of the autonomic nervous system (ANS). Here, we investigate human ANS responses to conditioned stimuli during Pavlovian fear conditioning. To obtain precise, trial-by-trial, single-subject estimates of ANS responses, we build on a statistical framework for psychophysiological modelling. We then consider previously proposed non-probabilistic models, a simple probabilistic model, and non-learning models, as well as different observation functions to link learning models with ANS activity. Across three experiments, and both for skin conductance (SCR) and pupil size responses (PSR), a probabilistic learning model best explains ANS responses. Notably, SCR and PSR reflect different quantities of the same model: SCR track a mixture of expected outcome and uncertainty, while PSR track expected outcome alone. In summary, by combining psychophysiological modelling with computational learning theory, we provide systematic evidence that the formation and maintenance of Pavlovian threat predictions in humans may rely on probabilistic inference and includes estimation of uncertainty. This could inform theories of neural implementation of aversive learning.

Bayesian validation of grammar productions for the language of thought

Probabilistic proposals of Language of Thoughts (LoTs) can explain learning across different domains as statistical inference over a compositionally structured hypothesis space. While frameworks may differ on how a LoT may be implemented computationally, they all share the property that they are built from a set of atomic symbols and rules by which these symbols can be combined. In this work we propose an extra validation step for the set of atomic productions defined by the experimenter. It starts by expanding the defined LoT grammar for the cognitive domain with a broader set of arbitrary productions and then uses Bayesian inference to prune the productions from the experimental data. The result allows the researcher to validate that the resulting grammar still matches the intuitive grammar chosen for the domain. We then test this method in the language of geometry, a specific LoT model for geometrical sequence learning. Finally, despite the fact of the geometrical LoT not being a universal (i.e. Turing-complete) language, we show an empirical relation between a sequence’s probability and its complexity consistent with the theoretical relationship for universal languages described by Levin’s Coding Theorem.

Habitual versus goal-driven attention

Recent research has expanded the list of factors that control spatial attention. Beside current goals and perceptual salience, statistical learning, reward, motivation and emotion also affect attention. But do these various factors influence spatial attention in the same manner, as suggested by the integrated framework of attention, or do they target different aspects of spatial attention? Here I present evidence that the control of attention may be implemented in two ways. Whereas current goals typically modulate where in space attention is prioritized, search habits affect how one moves attention in space. Using the location probability learning paradigm, I show that a search habit forms when people frequently find a visual search target in one region of space. Attentional cuing by probability learning differs from that by current goals. Probability cuing is implicit and persists long after the probability cue is no longer valid. Whereas explicit goal-driven attention codes space in an environment-centered reference frame, probability cuing is viewer-centered and is insensitive to secondary working memory load and aging. I propose a multi-level framework that separates the source of attentional control from its implementation. Similar to the integrated framework, the multi-level framework considers current goals, perceptual salience, and selection history as major sources of attentional control. However, these factors are implemented in two ways, controlling where spatial attention is allocated and how one shifts attention in space.

A universal cue for grammatical categories in the input to children: Frequent frames

•

Data from typologically diverse languages shows common distributional patterns.

•

Discontinuous repetitive patterns in the input provide cues for category assignment.

•

Morphological frames accurately predict nouns and verbs in the input to children.

How does a child map words to grammatical categories when words are not overtly marked either lexically or prosodically? Recent language acquisition theories have proposed that distributional information encoded in sequences of words or morphemes might play a central role in forming grammatical classes. To test this proposal, we analyze child-directed speech from seven typologically diverse languages to simulate maximum variation in the structures of the world’s languages. We ask whether the input to children contains cues for assigning syntactic categories in frequent frames, which are frequently occurring nonadjacent sequences of words or morphemes. In accord with aggregated results from previous studies on individual languages, we find that frequent word frames do not provide a robust distributional pattern for accurately predicting grammatical categories. However, our results show that frames are extremely accurate cues cross-linguistically at the morpheme level. We theorize that the nonadjacent dependency pattern captured by frequent frames is a universal anchor point for learners on the morphological level to detect and categorize grammatical categories. Whether frames also play a role on higher linguistic levels such as words is determined by grammatical features of the individual language.

Explicit goal-driven attention, unlike implicitly learned attention, spreads to secondary tasks

To what degree does spatial attention for one task spread to all stimuli in the attended region, regardless of task relevance? Most models imply that spatial attention acts through a unitary priority map in a task-general manner. We show that implicit learning, unlike endogenous spatial cuing, can bias spatial attention within one task without biasing attention to a spatially overlapping secondary task. Participants completed a visual search task superimposed on a background containing scenes, which they were told to encode for a later memory task. Experiments 1 and 2 used explicit instructions to bias spatial attention to one region for visual search; Experiment 3 used location probability cuing to implicitly bias spatial attention. In location probability cuing, a target appeared in one region more than others despite participants not being told of this. In all experiments, search performance was better in the cued region than in uncued regions. However, scene memory was better in the cued region only following endogenous guidance, not after implicit biasing of attention. These data support a dual-system view of top-down attention that dissociates goal-driven and implicitly learned attention. Goal-driven attention is task general, amplifying processing of a cued region across tasks, whereas implicit statistical learning is task-specific. (PsycINFO Database Record

Infant statistical-learning ability is related to real-time language processing

Infants are adept at learning statistical regularities in artificial language materials, suggesting that the ability to learn statistical structure may support language development. Indeed, infants who perform better on statistical learning tasks tend to be more advanced in parental reports of infants' language skills. Work with adults suggests that one way statistical learning ability affects language proficiency is by facilitating real-time language processing. Here we tested whether 15-month-olds' ability to learn sequential statistical structure in artificial language materials is related to their ability to encode and interpret native-language speech. Specifically, we tested their ability to learn sequential structure among syllables (Experiment 1) and words (Experiment 2), as well as their ability to encode familiar English words in sentences. The results suggest that infants' ability to learn sequential structure among syllables is related to their lexical-processing efficiency, providing continuity with findings from children and adults, though effects were modest.

Relationship between effortful motivation and neurocognition in schizophrenia

Effortful motivation and reward valuation learning deficits are associated with negative symptoms and impaired cognition in schizophrenia (SZ) patients. Whereas clinical assessments of motivation and reward value typically rely upon clinician ratings or self-report scales, behavioral measures often confound these constructs. Simple reverse-translated behavioral tasks that independently quantify motivation and reward valuation-which could then be linked to cognition-may facilitate the development of pro-cognitive therapeutics by bridging the "preclinical-to-clinical" gap. This study determined whether novel behavioral measures of effortful motivation and reward valuation are associated with impaired cognition in SZ patients (n=36). Patients completed the Progressive Ratio Breakpoint task (PRBT; physical effort motivation) and the Probabilistic Learning Task (PLT; reward learning/valuation) in conjunction with the MATRICS Consensus Cognitive Battery (MCCB). SZ patients exhibited statistically significant deficits in global cognition and all individual MCCB subdomains. Significant correlations were observed between PRBT and MCCB global cognition (r=0.52), speed of processing (r=0.56) and attention vigilance (r=0.48) subdomains, but not with PLT or clinical symptoms. Results indicate that effort and reward learning deficits are dissociable targets that can improve our understanding of cognitive impairments associated among patients with SZ. More importantly, the results support the long-standing notion that the measurement of cognitive impairments in SZ is highly linked to a willingness to expend effort. The availability of a PRBT designed for use in both rodents and humans could improve our understanding of the nature of cognitive impairments in neuropsychiatric disorders and accelerate the development of novel pro-cognitive therapeutics.

The longevity of statistical learning: When infant memory decays, isolated words come to the rescue

Research over the past 2 decades has demonstrated that infants are equipped with remarkable computational abilities that allow them to find words in continuous speech. Infants can encode information about the transitional probability (TP) between syllables to segment words from artificial and natural languages. As previous research has tested infants immediately after familiarization, infants' ability to retain sequential statistics beyond the immediate familiarization context remains unknown. Here, we examine infants' memory for statistically defined words 10 min after familiarization with an Italian corpus. Eight-month-old English-learning infants were familiarized with Italian sentences that contained 4 embedded target words-2 words had high internal TP (HTP, TP = 1.0) and 2 had low TP (LTP, TP = .33)-and were tested on their ability to discriminate HTP from LTP words using the Headturn Preference Procedure. When tested after a 10-min delay, infants failed to discriminate HTP from LTP words, suggesting that memory for statistical information likely decays over even short delays (Experiment 1). Experiments 2-4 were designed to test whether experience with isolated words selectively reinforces memory for statistically defined (i.e., HTP) words. When 8-month-olds were given additional experience with isolated tokens of both HTP and LTP words immediately after familiarization, they looked significantly longer on HTP than LTP test trials 10 min later. Although initial representations of statistically defined words may be fragile, our results suggest that experience with isolated words may reinforce the output of statistical learning by helping infants create more robust memories for words with strong versus weak co-occurrence statistics. (PsycINFO Database Record

Altered Medial Frontal Feedback Learning Signals in Anorexia Nervosa

BACKGROUND

In their relentless pursuit of thinness, individuals with anorexia nervosa (AN) engage in maladaptive behaviors (restrictive food choices and overexercising) that may originate in altered decision making and learning.

METHODS

In this functional magnetic resonance imaging study, we employed computational modeling to elucidate the neural correlates of feedback learning and value-based decision making in 36 female patients with AN and 36 age-matched healthy volunteers (12-24 years). Participants performed a decision task that required adaptation to changing reward contingencies. Data were analyzed within a hierarchical Gaussian filter model that captures interindividual variability in learning under uncertainty.

RESULTS

Behaviorally, patients displayed an increased learning rate specifically after punishments. At the neural level, hemodynamic correlates for the learning rate, expected value, and prediction error did not differ between the groups. However, activity in the posterior medial frontal cortex was elevated in AN following punishment.

CONCLUSIONS

Our findings suggest that the neural underpinning of feedback learning is selectively altered for punishment in AN.

Predictions and causal estimations are not supported by the same associative structure

Studies performed by different researchers have shown that judgements about cue–outcome relationships are systematically influenced by the type of question used to request those judgements. It is now recognized that judgements about the strength of the causal link between a cue and an outcome are mostly determined by the cue–outcome contingency, whereas predictions of the outcome are more influenced by the probability of the outcome given the cue. Although these results make clear that those different types of judgement are mediated by some knowledge of the normative differences between causal estimations and outcome predictions, they do not speak to the underlying processes of these effects. The experiment presented here reveals an interaction between the type of question and the order of trials that challenges standard models of causal and predictive learning that are framed exclusively in associative terms or exclusively in higher order reasoning terms. However, this evidence could be easily explained by assuming the combined intervention of both types of process.

Inferences about unobserved causes in human contingency learning

Estimates of the causal efficacy of an event need to take into account the possible presence and influence of other unobserved causes that might have contributed to the occurrence of the effect. Current theoretical approaches deal differently with this problem. Associative theories assume that at least one unobserved cause is always present. In contrast, causal Bayes net theories (including Power PC theory) hypothesize that unobserved causes may be present or absent. These theories generally assume independence of different causes of the same event, which greatly simplifies modelling learning and inference. In two experiments participants were requested to learn about the causal relation between a single cause and an effect by observing their co-occurrence (Experiment 1) or by actively intervening in the cause (Experiment 2). Participants’ assumptions about the presence of an unobserved cause were assessed either after each learning trial or at the end of the learning phase. The results show an interesting dissociation. Whereas there was a tendency to assume interdependence of the causes in the online judgements during learning, the final judgements tended to be more in the direction of an independence assumption. Possible explanations and implications of these findings are discussed.

Mechanisms of the associated nontargets effect: Processes influenced by statistical learning in a simple visual environment

In three visual search experiments participants were asked to make a target response if either of two targets was present and to make a nontarget response if neither target was present. Some target-absent displays included only nontarget features that never occurred in the same displays as target features, whereas other target-absent displays included nontarget features that did sometimes occur with target features. Nontarget responses were reliably faster in the former case than in the latter. This “associated nontargets effect” appears to arise from participants’ ability to learn and to use contingencies between the presence of certain nontargets and the absence of any target.

Prior belief and polarity in multicue learning

We report two experiments in which participants are trained using a multicue probability learning (MCPL) task, which attempts to simulate the acquisition of expert judgement by experience in the real world. Participants were asked to predict performance in certain occupations given a profile of personality test results with trial-by-trial outcome feedback. Only some cues were relevant, and the polarity of the cues (positive or negative predictors) was unspecified. In addition, 25% of random noise was added to the feedback to simulate real world uncertainty. The main factor of interest was that the role of prior belief (determined in a separate study of stereotypes) interfered with the learning process. Experiment 1 failed to find any influence of prior belief in the cues that were irrelevant to the criterion being trained. However, in Experiment 2 people learned to use the relevant cues better when their effect conformed with rather than conflicted with prior belief. Both experiments showed strong effects of cue polarity, with positive predictors much more easily learned. The results are discussed with reference to the cognitive processes involved in MCPL and closely related tasks.