How is spatial context learning integrated over signal versus noise? A primacy effect in contextual cueing

Over repeated exposure to particular visual search displays, subjects are able to implicitly extract regularities that then make search more efficient-a phenomenon known as contextual cueing. Here we explore how the learning involved in contextual cueing is formed, maintained, and updated over experience. During an initial training phase, a group of signal first subjects searched through a series of predictive displays (where distractor locations were perfectly correlated with the target location), followed with no overt break by a series of unpredictive displays (where repeated contexts were uncorrelated with target locations). A second noise first group of subjects encountered the unpredictive displays followed by the predictive displays. Despite the fact that both groups had the same overall exposure to signal and noise, only the signal first group demonstrated subsequent contextual cueing. This primacy effect indicates that initial experience can result in hypotheses about regularities in displays-or the lack thereof-which then become resistant to updating. The absence of regularities in early stages of training even blocked observers from learning predictive regularities later on.     

Channeling Power Across Ecological Systems: Social Regularities in Community Organizing

Relational and social network perspectives provide opportunities for more holistic conceptualizations of phenomena of interest in community psychology, including power and empowerment. In this article, we apply these tools to build on multilevel frameworks of empowerment by proposing that networks of relationships between individuals constitute the connective spaces between ecological systems. Drawing on an example of a model for grassroots community organizing practiced by WISDOM—a statewide federation supporting local community organizing initiatives in Wisconsin—we identify social regularities (i.e., relational and temporal patterns) that promote empowerment and the development and exercise of social power through building and altering relational ties. Through an emphasis on listening-focused one-to-one meetings, reflection, and social analysis, WISDOM organizing initiatives construct and reinforce social regularities that develop social power in the organizing initiatives and advance psychological empowerment among participant leaders in organizing. These patterns are established by organizationally driven brokerage and mobilization of interpersonal ties, some of which span ecological systems. Hence, elements of these power-focused social regularities can be conceptualized as cross-system channels through which micro-level empowerment processes feed into macro-level exercise of social recommendations for theory and designpower, and vice versa. We describe examples of these channels in action, and offer of future action research.     

远距离规则的内隐学习使用了何种记忆存储器：来自神经网络模拟的证据

关于远距离规则的知识是如何被内隐学习的,研究尚未得出结论。该研究通过采用和人类被试相同的实验材料和程序,考察了简单循环网络模型（SRN）对两种汉语声调远距离规则——倒映和逆行规则的内隐学习。结果发现：1.在广泛的参数范围上,SRN能够学会倒映和逆行规则,表明模型的记忆缓冲器可以模拟人类远距离规则的内隐学习;2.SRN对倒映规则的学习比对逆行规则的学习更好,表明在功能上远距离规则的内隐学习可能优先使用了先进先出的记忆存储器及信息加工模式。该研究为探究远距离规则内隐学习的机制提供了新的证据和视角。     

The large-scale structure of semantic networks: statistical analyses and a model of semantic growth

We present statistical analyses of the large-scale structure of 3 types of semantic networks: word associations, WordNet, and Roget's Thesaurus. We show that they have a small-world structure, characterized by sparse connectivity, short average path lengths between words, and strong local clustering. In addition, the distributions of the number of connections follow power laws that indicate a scale-free pattern of connectivity, with most nodes having relatively few connections joined together through a small number of hubs with many connections. These regularities have also been found in certain other complex natural networks, such as the World Wide Web, but they are not consistent with many conventional models of semantic organization, based on inheritance hierarchies, arbitrarily structured networks, or high-dimensional vector spaces. We propose that these structures reflect the mechanisms by which semantic networks grow. We describe a simple model for semantic growth, in which each new word or concept is connected to an existing network by differentiating the connectivity pattern of an existing node. This model generates appropriate small-world statistics and power-law connectivity distributions, and it also suggests one possible mechanistic basis for the effects of learning history variables (age of acquisition, usage frequency) on behavioral performance in semantic processing tasks.     

Reading skills of hearing and congenitally deaf children

The present study examined the nature of reading skills in congenitally deaf and hearing children 7–19 years of age. Deaf children were drawn from oralist and total communication programs. A visual detection task was designed to assess the extent of phonological coding and chunking used in reading a story of various degrees of syntactic, semantic, and orthographic complexity. The results provide evidence that (1) like hearing children, deaf children tend to use orthographic regularities in their reading: (2) there is no relation in the deaf child's performance between sensitivity to orthographic regularities and the type of communication method used in training; and (3) hearing and deaf readers use qualitatively similar psycholinguistic strategies in their processing of a story.     

Learning multiple rules simultaneously: Affixes are more salient than reduplications

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

Stimulus-dependent modulation of perceptual and motor learning in a serial reaction time task

In two experiments, we investigated the impact of spatial attributes on the representation acquired during a serial reaction time task. Two sequences were used, in which structural regularities occurred either in the horizontal or in the vertical locations of successive stimuli. After training with the dominant hand, participants were required to respond with the non-dominant hand to either the original sequence or to a mirror-ordered version of the original sequence that required finger movements homologous to those used during training. We observed that a difference in reaction times between the two transfer conditions was smaller in the vertical sequence than in the horizontal sequence. This pattern of results was independent of whether three fingers (Experiment 1) were used or only one finger (Experiment 2) was used for responding. This result suggests that perceptual and motor learning mechanisms may be weighted differently depending on the context in which the stimulus is presented.     

Ultrastructural plasticity associated with hippocampal-dependent learning: a meta-analysis

In order to develop a profile of how individual synapses in the hippocampal formation alter their structure following learning experience, a meta-analysis synthesized the available literature on morphological change following hippocampal-dependent learning. Analysis of the 132 calculated effect sizes suggest a consistent profile of morphological change in the hippocampus following learning experience. Across the hippocampal formation, dendritic complexity, spine density, and the size of perforated postsynaptic densities showed consistent increases following training. Both the density of synapses in general and perforated synapses in particular showed unique responses to training, depending on the duration of training and/or different cell layers of the hippocampal formation. Most importantly, it seems that this profile, while consistent, is small and specific--only a select few of the morphological parameters typically measured in anatomical studies of plasticity showed significant change following training. Collectively, these data suggest that the distinct electrophysiological properties of neocortical versus hippocampal synapses may be at least partially mediated by distinct morphological cascades. That is, on the basis of theory, and with the support of the current data, it seems that synaptogenesis correlates with enduring neocortical plasticity, while structural changes correlate with more transient hippocampal plasticity. To be able to state these conclusions with conviction, however, more data are needed in several key areas for continued pursuit of the morphological correlates of hippocampal-dependent learning.     

Explaining derivational morphology as the convergence of codes

Language users have a remarkable ability to create, produce and comprehend complex words. Words such as undercut and bakery appear to be composed of units, traditionally called morphemes, that recombine in rule-like ways to form other words, such as underline and cannery. However, morphological systems are quasiregular: they are systematic and productive but admit many seemingly irregular forms. Thus, bakery is related to bake and cannery to can but what is the groce in grocery? There is no bread in sweetbreads, liver in deliver, corn in corner or ginger in gingerly. Such words exhibit partial regularities concerning the correspondences between form and meaning, the treatment of which has important implications for linguistic and psycholinguistic theories. This article describes an approach to morphological phenomena called the convergence theory, in which morphology is a graded, inter-level representation that reflects correlations among orthography, phonology and semantics.     

Generation of hypotheses in Wason's 2-4-6 task: an information theory approach.

We explored the "context of discovery" in Wason's 2-4-6 task, focusing on how the first hypothesis is generated. According to Oaksford and Chater (1994a) people generate hypotheses extracting "common features", or regularities, from the available triples, but their model does not explain why some regularities contribute to the hypothesis more than do other regularities. Our conjecture is that some regularities contribute to the hypothesis more than do other regularities because people estimate the amount of information in the perceived regularities and try to preserve as much information as possible in their initial hypotheses. Experiment 1, which used two initial triples, showed that the presence of high-information relational regularities in the initial triples affected the information in the initial hypotheses more than did the presence of low-information object regularities. Experiment 2 extended the results to the classic situation in which only one initial triple is given. It also suggested that amount of information is the only aspect of the structure of the triple that affects hypotheses generation. Experiment 3 confirmed the latter finding: Although relations are commonly distinguished between first-order and higher order relations, the latter being most important for generating hypotheses (Gentner, 1983), higher order relations do have an effect on Wason's 2-4-6 task only if their presence incre ases information. In the conclusion we discuss the statistical soundness of human hypotheses generation processes, and we ask an unanswered question: Amount of information explains why some regularities are preferred to others, but only within a set of "nonarbitrary" regularities; there are object regularities that are rich in information content, but are considered "arbitrary", and are not used in generating hypotheses. Which formal property can distinguish between these two sets of regularities?     

On simulating one-trial learning using morphological neural networks

“Learning once, remembering forever”, this wonderful cognitive phenomenon sometimes occurs in the learning process of human beings. Psychologists call this psychological phenomenon “one-trial learning”. The traditional artificial neural networks can simulate the psychological phenomenon of “implicit learning”, but can’t simulate the cognitive phenomenon of “one-trial learning”. Therefore, cognitive psychology gives a challenge to the traditional artificial neural networks. From two aspects of theory and practice in this paper, the possibility of simulating this kind of psychological phenomenon was explored by using morphological neural networks. This paper takes advantage of morphological associative memory networks to realize the simulation of “one-trial learning” for the first time, and gives 5 simulating practical examples. Theoretical analysis and simulation experiments show that the morphological associative memory networks are a higher effective machine learning method, and can better simulate the cognitive phenomenon of “one-trial learning”, therefore provide a theoretical basis and technological support for the study of intelligent science and cognitive science.     

Long-term sensitization training primes Aplysia for further learning

Repetitive, unilateral stimulation of Aplysia induces long-term sensitization (LTS) of ipsilaterally elicited siphon-withdrawal responses. Whereas some morphological effects of training appear only on ipsilateral sensory neurons, others appear bilaterally. We tested the possibility that contralateral morphological modifications may have functional significance. Therefore, we examined whether LTS training primes subsequent sensitization. Twenty-four hours after LTS training the effects of brief shock treatment (BST) were examined. BST failed to sensitize animals that had previously received either 4-d control treatment or 4-d ipsilateral LTS training. In contrast, BST did sensitize animals that had previously received 4-d contralateral LTS training, suggesting the presence of a latent trace that primes the animal for further learning.     

Statistical regularities guide the spatial scale of attention

Visual scenes contain information on both a local scale (e.g., a tree) and a global scale (e.g., a forest). The question of whether the visual system prioritizes local or global elements has been debated for over a century. Given that visual scenes often contain distinct individual objects, here we examine how regularities between individual objects prioritize local or global processing. Participants viewed Navon-like figures consisting of small local objects making up a global object, and were asked to identify either the shape of the local objects or the shape of the global object, as fast and accurately as possible. Unbeknown to the participants, local regularities (i.e., triplets) or global regularities (i.e., quadruples) were embedded among the objects. We found that the identification of the local shape was faster when individual objects reliably co-occurred immediately next to each other as triplets (local regularities, Experiment 1). This result suggested that local regularities draw attention to the local scale. Moreover, the identification of the global shape was faster when objects co-occurred at the global scale as quadruples (global regularities, Experiment 2). This result suggested that global regularities draw attention to the global scale. No participant was explicitly aware of the regularities in the experiments. The results suggest that statistical regularities can determine whether attention is directed to the individual objects or to the entire scene. The findings provide evidence that regularities guide the spatial scale of attention in the absence of explicit awareness.     

Incidental learning of real-world regularities

The large literature on incidental learning relies almost exclusively on laboratory experiments. Whenever researchers have attempted to demonstrate incidental learning of real-world regularities, they have typically failed to show learning. For example, it is well established that people do not learn regularities in everyday objects, such as the left-right orientation of faces on coins, despite a very large exposure to them. In this report, we examine this apparent contradiction. We argue that most studies exploring real-life incidental learning use tests that are not as sensitive to low-confidence information as those traditionally used in laboratory tasks. Using more sensitive measures, we show that it is possible to learn regularities from British and Japanese cultural life as a direct result of exposure to these regularities. Further, confidence measures suggest that although the information may be acquired incidentally, it can be expressed with and without concomitant awareness of that knowledge.     

Electrophysiological evidence for modulation of lexical processing after repetitive exposure to foreign phonotactic rules

In two experiments we investigate how repeated exposure to native and non-native phonotactic regularities alters the N400, an event-related potential related to lexico-semantic access. Participants underwent a Passive Listening (Experiment 1) or a Categorization Training (Experiment 2) for monosyllabic pseudowords over 3 days. During Passive Listening participants solely listened to the stimuli while for Categorization Training they learned to assign items to two arbitrary categories by feedback. Notably, this task did not rely on phonotactic regularities. Before training, N400 was larger for legal compared to illegal items. Over the 3-day exposure Passive Listening yielded a significant decrease in N400-amplitude for illegal pseudowords, however, this effect was abolished and partially inverted by the Categorization Training. We suggest the decrease in N400-amplitude indicates more efficient discrimination between native and non-native pseudowords since only the former are potential lexical candidates. On the contrary, Categorization Training introduces a ‘protosemantic’ context overriding prelexical selection processes.     

Statistical regularities modulate attentional capture independent of search strategy

An earlier study using the additional singleton task showed that statistical regularities regarding the distractor location can cause an attentional bias that affects the amount of attentional capture by distractors and the efficiency of selection of targets. The distractor singleton was systematically present more often in one location than in all other locations. The present study investigated whether this bias also occurs when observers adopt a feature search mode, i.e., when they search for a specific feature (circle) between elements with different shapes, while ignoring a colored distractor singleton. It is assumed that in feature search, observers can ignore distractors in a top-down way and as such one expects that statistical regularities about the distractor location should not play a role. Contrary to this prediction, we found that even in feature search, both attentional capture by the distractors and the efficiency of selecting the target were impacted by these statistical regularities. Moreover, statistical regularities regarding the feature value of the distractor (its color) had no effect on the amount of capture or the efficiency of selection. We claim that statistical regularities cause passive lingering biases of attention such that on the priority map, the location containing a high probability distractor competes less for attention than locations that are less likely to contain distractors.     

The emergence of explicit knowledge from implicit learning

Substantial evidence has highlighted the ability of observers to incidentally extract statistical contingencies present in visual environments. This study examined whether the knowledge extracted regarding statistical contingencies is unconscious initially, even when it becomes fully accessible to conscious awareness after extensive training. Using a “typical” contextual cuing procedure adapted to real-world scenes, we first observed that, after extensive training in searching for a target within repeated scenes, knowledge about regularities was associated with conscious awareness (Experiment 1). However, both subjective and objective measures of consciousness revealed that in the early phase of training, learning of regular structures first takes place at an unconscious level (Experiments 2 and 3). These results are discussed in the light of the causal relationships between learning and consciousness.     

Learning multiple layers of representation

To achieve its impressive performance in tasks such as speech perception or object recognition, the brain extracts multiple levels of representation from the sensory input. Backpropagation was the first computationally efficient model of how neural networks could learn multiple layers of representation, but it required labeled training data and it did not work well in deep networks. The limitations of backpropagation learning can now be overcome by using multilayer neural networks that contain top-down connections and training them to generate sensory data rather than to classify it. Learning multilayer generative models might seem difficult, but a recent discovery makes it easy to learn nonlinear distributed representations one layer at a time.