The design of ballantines

The shared variance between two variables can be expressed graphically by overlapping circles. A procedure is presented for locating the circles so that the graphical and statistical relations correspond exactly. The procedure does not generalize to the representation of multiple, part, and partial correlations between three or more variables. The reasons for the nonextension are explained, and some alternative graphical representations are suggested.     

Neural Representations of Nondeclarative Memories

ABSTRACT— Nondeclarative learning refers to abilities characterized by a lack of awareness of what has been learned and an independence from medial temporal lobe structures that support conscious memories of facts and events. Neuroimaging approaches have been used extensively in two domains of nondeclarative learning—priming and skill learning—to investigate the neural substrates supporting performance. Recent neuroimaging studies have attempted to understand what is being learned in different tasks in order to inform psychological theories of nondeclarative memory. For example, priming may be considered a form of perceptual learning or a form of stimulus–response learning, and correlations between performance and activation patterns in different regions may suggest the nature of the brain changes that support behavior. The attainment of expertise in a skill has been characterized as greater efficiency of processing in the same neural structures that support novice performance or, alternatively, as the recruitment of additional regions. Current research suggests that, within the domains of priming and skill learning, there is much heterogeneity in the underlying brain representations and psychological theories will need to account for these variations.     

Neural network processing of natural language: II. Towards a unified model of corticostriatal function in learning sentence comprehension and non-linguistic sequencing

A central issue in cognitive neuroscience today concerns how distributed neural networks in the brain that are used in language learning and processing can be involved in non-linguistic cognitive sequence learning. This issue is informed by a wealth of functional neurophysiology studies of sentence comprehension, along with a number of recent studies that examined the brain processes involved in learning non-linguistic sequences, or artificial grammar learning (AGL). The current research attempts to reconcile these data with several current neurophysiologically based models of sentence processing, through the specification of a neural network model whose architecture is constrained by the known cortico-striato-thalamo-cortical (CSTC) neuroanatomy of the human language system. The challenge is to develop simulation models that take into account constraints both from neuranatomical connectivity, and from functional imaging data, and that can actually learn and perform the same kind of language and artificial syntax tasks. In our proposed model, structural cues encoded in a recurrent cortical network in BA47 activate a CSTC circuit to modulate the flow of lexical semantic information from BA45 to an integrated representation of meaning at the sentence level in BA44/6. During language acquisition, corticostriatal plasticity is employed to allow closed class structure to drive thematic role assignment. From the AGL perspective, repetitive internal structure in the AGL strings is encoded in BA47, and activates the CSTC circuit to predict the next element in the sequence. Simulation results from Caplan's [Caplan, D., Baker, C., & Dehaut, F. (1985). Syntactic determinants of sentence comprehension in aphasia. Cognition, 21, 117-175] test of syntactic comprehension, and from Gomez and Schvaneveldts' [Gomez, R. L., & Schvaneveldt, R. W. (1994). What is learned from artificial grammars?. Transfer tests of simple association. Journal of Experimental Psychology: Learning, Memory and Cognition, 20, 396-410] artificial grammar learning experiments are presented. These results are discussed in the context of a brain architecture for learning grammatical structure for multiple natural languages, and non-linguistic sequences.     

Electrifying diagrams for learning: principles for complex representational systems

Six characteristics of effective representational systems for conceptual learning in complex domains have been identified. Such representations should: (1) integrate levels of abstraction; (2) combine globally homogeneous with locally heterogeneous representation of concepts; (3) integrate alternative perspectives of the domain; (4) support malleable manipulation of expressions; (5) possess compact procedures; and (6) have uniform procedures. The characteristics were discovered by analysing and evaluating a novel diagrammatic representation that has been invented to support students' comprehension of electricity—AVOW diagrams (Amps, Volts, Ohms, Watts). A task analysis is presented that demonstrates that problem solving using a conventional algebraic approach demands more effort than AVOW diagrams. In an experiment comparing two groups of learners using the alternative approaches, the group using AVOW diagrams learned more than the group using equations and were better able to solve complex transfer problems and questions involving multiple constraints. Analysis of verbal protocols and work scratchings showed that the AVOW diagram group, in contrast to the equations group, acquired a coherently organised network of concepts, learnt effective problem solving procedures, and experienced more positive learning events. The six principles of effective representations were proposed on the basis of these findings. AVOW diagrams are Law Encoding Diagrams, a general class of representations that have been shown to support learning in other scientific domains.     

Basing Categorization on Individuals and Events

Exemplar, prototype, and connectionist models typically assume that events constitute the basic unit of learning and representation in categorization. In these models, each learning event updates a statistical representation of a category independently of other learning events. An implication is that events involving the same individual affect learning independently and are not integrated into a single structure that represents the individual in an internal model of the world. A series of experiments demonstrates that human subjects track individuals across events, establish representations of them, and use these representations in categorization. These findings are consistent with “representationalism,” the view that an internal model of the world constitutes a physical level of representation in the brain, and that the brain does not simply capture the statistical properties of events in an undifferentiated dynamical system. Although categorization is an inherently statistical process that produces generalization, pattern completion, frequency effects, and adaptive learning, it is also an inherently representational process that establishes an internal model of the world. As a result, representational structures evolve in memory to track the histories of individuals, accumulate information about them, and simulate them in events.     

View-invariant object category learning, recognition, and search: how spatial and object attention are coordinated using surface-based attentional shrouds

How does the brain learn to recognize an object from multiple viewpoints while scanning a scene with eye movements? How does the brain avoid the problem of erroneously classifying parts of different objects together? How are attention and eye movements intelligently coordinated to facilitate object learning? A neural model provides a unified mechanistic explanation of how spatial and object attention work together to search a scene and learn what is in it. The ARTSCAN model predicts how an object's surface representation generates a form-fitting distribution of spatial attention, or "attentional shroud". All surface representations dynamically compete for spatial attention to form a shroud. The winning shroud persists during active scanning of the object. The shroud maintains sustained activity of an emerging view-invariant category representation while multiple view-specific category representations are learned and are linked through associative learning to the view-invariant object category. The shroud also helps to restrict scanning eye movements to salient features on the attended object. Object attention plays a role in controlling and stabilizing the learning of view-specific object categories. Spatial attention hereby coordinates the deployment of object attention during object category learning. Shroud collapse releases a reset signal that inhibits the active view-invariant category in the What cortical processing stream. Then a new shroud, corresponding to a different object, forms in the Where cortical processing stream, and search using attention shifts and eye movements continues to learn new objects throughout a scene. The model mechanistically clarifies basic properties of attention shifts (engage, move, disengage) and inhibition of return. It simulates human reaction time data about object-based spatial attention shifts, and learns with 98.1% accuracy and a compression of 430 on a letter database whose letters vary in size, position, and orientation. The model provides a powerful framework for unifying many data about spatial and object attention, and their interactions during perception, cognition, and action.     

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.     

Learning of control in a neural architecture of grounded language processing

Human cognition is characterized by three important features: productivity, dynamics and grounding. These features can be integrated in a neural architecture of language processing. The representations in this architecture always remain “in situ”, because they are grounded in perception, action, emotion, associations and (semantic) relations. The neural architecture shows how these representations can be combined in a productive manner, and how dynamics influences this process. The constraints that each of these features impose on each other result in an architecture in which local and global aspects interact in processing and learning. The architecture consists of neural “binding” mechanisms that produce (novel) sentence structures on the fly. Here, we discuss how the control of this binding process can be learned. We trained a feedforward network (FFN) for this task. The results show that information from the architecture is needed as input to learn control of binding. Thus, the control system is recurrent. We show that this recurrent system can learn control of binding for basic (but recursive) sentence structures. After learning, the binding process behaves well on a series of test sentences, including sentences with (unlimited) embeddings. However, for some of these sentences, difficulties arise due to dynamical binding conflicts in the architecture. We also discuss and illustrate the potential influence that the dynamics in the architecture could have on the binding process.     

Automatically deriving readers’ knowledge structures from texts

Latent semantic analysis (LSA) serves as both a theory and a method for representing the meaning of words based on a statistical analysis of their contextual usage (Foltz, 1996; Landauer & Dumais, 1997). In experiments in the domains of psychology and history, we compared the representation of readers’ knowledge structures of information learned from texts with the representation generated by LSA. Results indicated that LSA’s representation is similar to readers’ representations. In addition, the degree to which the reader’s representation is similar to LSA’s representation is indicative of the amount of knowledge the reader has acquired and of the reader’s reading ability. This approach has implications both as a model of learning from text and as a practical tool for performing knowledge assessment.     

Integration of spatial information across vision and language

Three experiments investigated whether spatial information acquired from vision and language is maintained in distinct spatial representations on the basis of the input modality. Participants studied a visual and a verbal layout of objects at different times from either the same (Experiments 1 and 2) or different learning perspectives (Experiment 3) and then carried out a series of pointing judgments involving objects from the same or different layouts. Results from Experiments 1 and 2 indicated that participants pointed equally fast on within- and between-layout trials; coupled with verbal reports from participants, this result suggests that they integrated all locations in a single spatial representation during encoding. However, when learning took place from different perspectives in Experiment 3, participants were faster to respond to within- than between-layout trials and indicated that they kept separate representations during learning. Results are compared to those from similar studies that involved layouts learned from perception only.     

大脑腹侧视觉通路知觉表征神经机制研究

大脑腹侧视觉通路知觉表征的神经机制是认知神经科学研究面临的基本问题。本文系统介绍了该问题研究中比较有影响力的理论模型,归纳分析了模型之间的分歧与各自的局限。文章分析指出大脑自上而下的调控机制是腹侧视觉通路神经表征机制问题研究的另一重要维度,如何有效整合知觉表征模型与大脑调控机制的相关研究是进一步深化知觉表征神经机制问题研究的关键。     

Acquiring experiential traces in word-referent learning

In two experiments, we investigated the activation of perceptual representations of referent objects during word processing. In both experiments, participants learned to associate pictures of novel three-dimensional objects with pseudowords. They subsequently performed a recognition task (Experiment 1) or a naming task (Experiment 2) on the object names while being primed with different types of visual stimuli. Only the stimuli that the participants had encountered as referent objects during the training phase facilitated recognition or naming responses. New stimuli did not facilitate the processing of object names, even if they matched a schematic or prototypical representation of the referent object that the participants might have abstracted during word-referent learning. These results suggest that words learned by way of examples of referent objects are associated with experiential traces of encounters with these objects.     

Aligning Logical and Psychological Contributions to the Understanding of Human Reasoning

Modern logic is an abstract study of consequence rather than a mechanistic study of reasoning. This abstract view has much to offer psychological studies of the representations which implement human reasoning. This paper reviews recent results showing the abstract equivalence of all the main psychological competence models of syllogistic reasoning. Their apparently contrasting representations are computationally equivalent for the kind of data presented. The true empirical contrasts between theories lie in other details of implementation — in whether integrated representations are constructed and in the sequence of cases considered. These contrasts are independent of sentential, graphical or mental model representation.Logic is also a guide to representation requirements. Linear syllogism models are representable by sequences of tokens, whereas categorial syllogism models generally require attribute binding. However, one consequence of the present semantic analysis is that all categorial syllogism competence models converge on an algorithm which avoids representation of more than a single sequence.The paper goes on to consider what evidence could distinguish the representations implementing reasoning in working memory. The example of dual task paradigms is considered and some conceptual questions are raised about the identification of the modalities of working memory subsystems with linguistic and spatial representations. Carefully distinguishing memory medium from memory interpretation leads to the conclusion that it is the interpretation of the contents of memory which determines whether it holds spatial or linguistic information. It is argued that the critical dimension for understanding reasoning processes is the expressiveness of interpretation of representations.     

Three-Dimensional Information in Face Representations Revealed by Identity Aftereffects

ABSTRACT— Representations of individual faces evolve with experience to support progressively more robust recognition. Knowledge of three-dimensional face structure is required to predict an image of a face as illumination and viewpoint change. Robust recognition across such transformations can be achieved with representations based on multiple two-dimensional views, three-dimensional structure, or both. We used face-identity adaptation in a familiarization paradigm to address a long-standing controversy concerning the role of two-dimensional versus three-dimensional information in face representations. We reasoned that if three-dimensional information is coded in the representations of familiar faces, then learning a new face using images generated by one three-dimensional transformation should enhance the robustness of the representation to another type of three-dimensional transformation. Familiarization with multiple views of faces enhanced the transfer of face-identity adaptation effects across changes in illumination by compensating for a generalization cost at a novel test viewpoint. This finding demonstrates a role for three-dimensional information in representations of familiar faces.     

The processing of verbs and nouns in neural networks: insights from synthetic brain imaging

The paper presents a computational model of language in which linguistic abilities evolve in organisms that interact with an environment. Each individual's behavior is controlled by a neural network and we study the consequences in the network's internal functional organization of learning to process different classes of words. Agents are selected for reproduction according to their ability to manipulate objects and to understand nouns (objects' names) and verbs (manipulation tasks). The weights of the agents' neural networks are evolved using a genetic algorithm. Synthetic brain imaging techniques are then used to examine the functional organization of the neural networks. Results show that nouns produce more integrated neural activity in the sensory-processing hidden layer, while verbs produce more integrated synaptic activity in the layer where sensory information is integrated with proprioceptive input. Such findings are qualitatively compared with human brain imaging data that indicate that nouns activate more the posterior areas of the brain related to sensory and associative processing, while verbs activate more the anterior motor areas.     

Neuroconstructivism

Neuroconstructivism is a theoretical framework focusing on the construction of representations in the developing brain. Cognitive development is explained as emerging from the experience-dependent development of neural structures supporting mental representations. Neural development occurs in the context of multiple interacting constraints acting on different levels, from the individual cell to the external environment of the developing child. Cognitive development can thus be understood as a trajectory originating from the constraints on the underlying neural structures. This perspective offers an integrated view of normal and abnormal development as well as of development and adult processing, and it stands apart from traditional cognitive approaches in taking seriously the constraints on cognition inherent to the substrate that delivers it.     

Structural Priming as Implicit Learning: A Comparison of Models of Sentence Production

Structural priming reflects a tendency to generalize recently spoken or heard syntactic structures to different utterances. We propose that it is a form of implicit learning. To explore this hypothesis, we developed and tested a connectionist model of language production that incorporated mechanisms previously used to simulate implicit learning. In the model, the mechanism that learned to produce structured sequences of phrases from messages also exhibited structural priming. The ability of the model to account for structural priming depended on representational assumptions about the nature of messages and the relationship between comprehension and production. Modeling experiments showed that comprehension-based representations were important for the model's generalizations in production and that nonatomic message representations allowed a better fit to existing data on structural priming than traditional thematic-role representations.     

The bilingual brain: cerebral representation of languages.

The present article deals with theoretical and experimental aspects of language representation in the multilingual brain. Two general approaches were adopted in the study of the bilingual brain. The study of bilingual aphasics allows us to describe dissociations and double dissociations between the different subcomponents of the various languages. Furthermore, symptoms peculiar to bilingual aphasia were reported (pathological mixing and switching and translations disorders) which allowed the correlation of some abilities specific to bilinguals with particular neurofunctional systems. Another approach to the study of the bilingual brain is of the experimental type, such as electrophysiological investigations (electrocorticostimulation during brain surgery and event-related potentials) and functional neuroanatomy studies (positron emission tomography and functional magnetic resonance imaging). Functional neuroanatomy studies investigated the brain representation of languages when processing lexical and syntactic stimuli and short stories. Neurophysiologic and neuroimaging studies evidenced a similar cerebral representation of L1 and L2 lexicons both in early and late bilinguals. The representation of grammatical aspects of languages seems to be different between the two languages if L2 is acquired after the age of 7, with automatic processes and correctness being lower than those of the native language. These results are in line with a greater representation of the two lexicons in the declarative memory systems, whereas morphosyntactic aspects may be organized in different systems according to the acquisition vs learning modality.     

Mental representation and learning: The influence of practice on the development of mental representation structure in complex action

ObjectivesRecent research has elicited distinct differences in mental representations between athletes of different skill levels. Such differences suggest that the structure of mental representations changes as a function of skill level. However, research examining how such mental representation structures develop over the course of learning is lacking. In the present study, we examine the effects of practice on the development of one's mental representation of a complex action during early skill acquisition.DesignFor this purpose, we created a controllable learning situation, using a repeated-measures design with a control group. More specifically, novice golfers were randomly assigned to either a practice group (n = 12) or a control group (n = 12). Both groups were tested before and after an acquisition phase of three days as well as after a three day retention interval.MethodsMental representation structures of the putt were recorded, employing the structural dimensional analysis of mental representation (SDA-M), which provides psychometric data on the structure and grouping of action concepts in long-term memory. In addition, outcome performance of the practice group was measured, using two-dimensional error scores of the putt.ResultsFindings revealed a significant improvement in task performance, as well as functional changes in the structure of the practice group's mental representation. In contrast, no functional adaptations were evident in the mental representation of the control group.ConclusionOur findings suggest that motor skill acquisition is associated with functional adaptations of action-related knowledge in long-term memory.     

Roles of egocentric and allocentric spatial representations in locomotion and reorientation

Four experiments investigated the nature of spatial representations used in locomotion. Participants learned the layout of several objects and then pointed to the objects while blindfolded in 3 conditions: before turning (baseline), after turning to a new heading (updating), and after disorientation (disorientation). The internal consistency of pointing in the disorientation condition was relatively high and equivalent to that in the baseline and updating conditions, when the layout had salient intrinsic axes and the participants learned the locations of the objects on the periphery of the layout. The internal consistency of pointing was disrupted by disorientation when participants learned the locations of objects while standing amid them and the layout did not have salient intrinsic axes. It was also observed that many participants retrieved spatial relations after disorientation from the original learning heading. These results indicate that people form an allocentric representation of object-to-object spatial relations when they learn the layout of a novel environment and use that representation to locate objects around them. Egocentric representations may be used to locate objects when allocentric representations are not of high fidelity.