Hebbian learning of cognitive control: dealing with specific and nonspecific adaptation

The conflict monitoring model of M. M. Botvinick, T. S. Braver, D. M. Barch, C. S. Carter, and J. D. Cohen (2001) triggered several research programs investigating various aspects of cognitive control. One problematic aspect of the Botvinick et al. model is that there is no clear account of how the cognitive system knows where to intervene when conflict is detected. As a result, recent findings of task-specific and context-specific (e.g., item-specific) adaptation are difficult to interpret. The difficulty with item-specific adaptation was recently pointed out by C. Blais, S. Robidoux, E. F. Risko, and D. Besner (2007), who proposed an alternative model that could account for this. However, the same problem of where the cognitive system should intervene resurfaces in a different shape in this model, and it has difficulty in accounting for the Gratton effect, a hallmark item-nonspecific effect. The authors of the current article show how these problems can be solved when cognitive control is implemented as a conflict-modulated Hebbian learning rule.     

A modified model of the Hebbian synapse and its role in motor learning

The `Hebbian synapse', an old neuro-psychological concept which describes the process of associative learning at the synaptic level, is being increasingly confirmed by neuro-biological explanations. The purpose of the present paper is to show that sensorimotor learning, which is essentially non-associative, can also be explained by Hebbian learning. This requires a re-definition of the Hebbian synapse ensuring convergence of the synaptic weight. Re-defined Hebbian learning then appears as a linear adaptive filter known in the field of adaptive signal processing, which in turn is equivalent to the delta rule used to train artificial neural networks. For motor learning, the modified Hebbian synapse must be embedded into a special learning algorithm called `auto-imitation'. This is a non goal-oriented inductive learning algorithm, enabling a controller to adopt a general rule from being shown only a few examples of that rule. When applied to motor learning, the neural controller can acquire the capability to online invert the behavior of the plant to be controlled. The complete learning process then can be described as a relaxation process between the controller and the controlled system, which is governed by long-term potentiation (LTP), the neuro-physiological process underlying Hebbian synaptic shaping.PsycINFO classification: 2330     

Success and failure of new speech category learning in adulthood: Consequences of learned Hebbian attractors in topographic maps

The influence of a native language on learning new speech sounds in adulthood is addressed using a network model in which speech categories are attractors implemented through interactive activation and Hebbian learning. The network has a representation layer that receives topographic projections from an input layer and has reciprocal excitatory connections with deeper layers. When applied to an experiment in which Japanese adults were trained to distinguish the English /r/-/l/ contrast (McCandliss, Fiez, Protopapas, Conway, & McClelland, 2002), the model can account for many aspects of the experimental results, such as the time course and outcome of the learning, how it varies as a function of feedback, the relative efficacy of adaptive and initially easy training stimuli versus nonadaptive and difficult stimuli, and the development of a discrimination peak at the acquired category boundary. The model is also able to capture some aspects of the individual differences in learning. nt]mis|This research was supported by NIMH Grant MH64445 and Training Grant 5T32-MH019983-07. The authors thank Susan G. Guion for generously providing the recordings from her experiment and Mary L. Conway for her assistance with data analysis. Correspondence concerning this article should be addressed to J. L. McClelland, Department of Psychology, 450 Serra Mall, Jordan Hall, Bldg. 420, Stanford, CA 94305     

Demystifying social cognition: a Hebbian perspective

For humans and monkeys, understanding the actions of others is central to survival. Here we review the physiological properties of three cortical areas involved in this capacity: the STS, PF and F5. Based on the anatomical connections of these areas, and the Hebbian learning rule, we propose a simple but powerful account of how the monkey brain can learn to understand the actions of others by associating them with self-produced actions, at the same time discriminating its own actions from those of others. As this system appears also to exist in man, this network model can provide a framework for understanding human social perception.     

Time scales in motor learning and development

A theoretical framework based on the concepts and tools of nonlinear dynamical systems is advanced to account for both the persistent and transitory changes traditionally shown for the learning and development of motor skills. The multiple time scales of change in task outcome over time are interpreted as originating from the system's trajectory on an evolving attractor landscape. Different bifurcations between attractor organizations and transient phenomena can lead to exponential, power law, or S-shaped learning curves. This unified dynamical account of the functions and time scales in motor learning and development offers several new hypotheses for future research on the nature of change in learning theory.     

The learning and motor development transfer process

The aim of this study is to analyze the transference process in motor skill learning. For this purpose, 320 boys and girls, with ages ranging from 3 to 12 years (M= 7.61; SD= 2.61), took part in nine object movement reception drills in which the following variables were cross-examined: the presence-absence of displacement (static or in motion), the corporal segments utilized (hands or arms), the movement direction (right or left), and the moving object (volleyball or tennis ball). The results indicate that what is being transferred is the common factor among them, the ocular-kinesthetic regulating system, which is constructed according to a generalized motor program and a predictive strategy of continuous control. The way that individuals group by levels of skill that represent the developmental levels of the aforementioned regulating system can also be observed. Lastly, the results are discussed, and strategies to improve the learning process in sports and physical education are provided.     

Emergence and retention of learning in early fetal development

Prior research has demonstrated that the late-term fetus is capable of learning and then remembering a passage of speech for several days, but there are no data to describe the earliest emergence of learning a passage of speech, and thus, how long that learning could be remembered before birth. This study investigated these questions. Pregnant women began reciting or speaking a passage out loud (either Rhyme A or Rhyme B) when their fetuses were 28 weeks gestational age (GA) and continued to do so until their fetuses reached 34 weeks of age, at which time the recitations stopped. Fetuses’ learning and memory of their rhyme were assessed at 28, 32, 33, 34, 36 and 38 weeks. The criterion for learning and memory was the occurrence of a stimulus-elicited heart rate deceleration following onset of a recording of the passage spoken by a female stranger. Detection of a sustained heart rate deceleration began to emerge by 34 weeks GA and was statistically evident at 38 weeks GA. Thus, fetuses begin to show evidence of learning by 34 weeks GA and, without any further exposure to it, are capable of remembering until just prior to birth. Further study using dose–response curves is needed in order to more fully understand how ongoing experience, in the context of ongoing development in the last trimester of pregnancy, affects learning and memory.     

Vicarious learning and the development of fears in childhood

Vicarious learning has long been assumed to be an indirect pathway to fear; however, there is only retrospective evidence that children acquire fears in this way. In two experiments, children (aged 7-9 years) were exposed to pictures of novel animals paired with pictures of either scared, happy or no facial expressions to see the impact on their fear cognitions and avoidance behavior about the animals. In Experiment 1, directly (self-report) and indirectly measured (affective priming) fear attitudes towards the animals changed congruent with the facial expressions with which these were paired. The indirectly measured fear beliefs persisted up to 3 months. Experiment 2 showed that children took significantly longer to approach a box they believed to contain an animal they had previously seen paired with scared faces. These results support theories of fear acquisition that suppose that vicarious learning affects cognitive and behavioral fear emotion, and suggest possibilities for interventions to weaken fear acquired in this way.     

Strategy development and learning differences in supervised and unsupervised categorization

The processes that determine unsupervised categorization, the task of classifying stimuli without guidance or feedback, are poorly understood. Two experiments examined the emergence and plasticity of unsupervised strategies using perceptual stimuli that varied along two separable dimensions. In the first experiment, participants either classified stimuli into any two categories of their choice or learned identical classifications by supervised categorization. Irrespective of the complexity of classification, supervised and unsupervised learning rates differed little when both modes of learning were maximally comparable. The second experiment examined the plasticity of unsupervised classifications by introducing novel stimuli halfway through training. Whether or not people altered their strategies, they responded to novel stimuli in a gradual manner. The gradual and continuous evolution and adaptation of strategies suggests that unsupervised categorization involves true learning which shares many properties of supervised category learning. We also show that the choice of unsupervised strategy cannot be predicted from the properties of early learning trials, but is best understood as a function of the initial distribution of dimensional attention.     

Concepts of intellect in relation to learning and adult development

This article is a summary and integration of results from studies of the organization and development of human abilities. Several kinds of results are interpreted as converging toward support of theories stipulating a hierarchy of intellectual functions. Near the top of this hierarchy, most related to intelligence as this is usually conceived, are the two broad sets of abilities known as fluid intelligence (Gf) and crystallized intelligence (Gc). The major focus of the review is on the decline of Gf through the “vital years of adulthood,” from 20 to 60 years of age. Major conclusions are:1. Sensory detector functions, near the bottom of the hierarchy of intellectual capacities, decline with development through the vital years, but this kind of decline is not responsible for aging decline of the major features of fluid intelligence: keenness of sensory function is not of the sine qua non of mature expressions of human intelligence.2. Memory abilities are somewhat higher in the hierarchy of intellectual functions, and thus are more strongly related to intelligence, than are sensory detectors. With increase in time interval over which retention is required, there is increase in the relationship of memory to Gf and to decline of Gf over the vital years. Accompanying this increase is an increase in the extent to which memory is dependent upon adequacy of organization at the stage of encoding. This, in turn, reflects the adequacy of eduction of relations and correlates, which is an essential feature of Gf.3. To the extent that very short-term memory is implicated in fluid intelligence, and the decline of this, it is intimately related to either a capacity or inclination to maintain close attention under conditions of high demand for attention.4. The speed with which one obtains correct answers to difficult problems is not closely related to the number of correct answers obtained when all subjects attempt all items in fluid intelligence tests, but clerical/perpetual speed, as indicated in tasks of quickly finding a particular symbol among others or tasks of making same/different judgments, is related to aging decline of Gf. As with short-term memory, this form of intellectual speediness appears to be closely related to capacities (or inclinations) for maintaining attention, dividing attention, and holding things in mind while doing other things.5. Crystallized intelligence probably increases over the vital years. This reflects dynamic restructuring of knowledge systems. It appears that there is not only an aging increase in knowledge but also an increase in the accessibility of information, a result being that humans become more creative (in one sense) as age increases through the vital years.6. Results from multivariate cross-sectional studies of intelligence are seen to be consistent with comparable results obtained with other kinds of research strategies.     

Learning words and learning sounds: Advances in language development

Phonological development is sometimes seen as a process of learning sounds, or forming phonological categories, and then combining sounds to build words, with the evidence taken largely from studies demonstrating ‘perceptual narrowing’ in infant speech perception over the first year of life. In contrast, studies of early word production have long provided evidence that holistic word learning may precede the formation of phonological categories. In that account, children begin by matching their existing vocal patterns to adult words, with knowledge of the phonological system emerging from the network of related word forms. Here I review evidence from production and then consider how the implicit and explicit learning mechanisms assumed by the complementary memory systems model might be understood as reconciling the two approaches.     

Phase-specific relationships between team learning processes and transactive memory development

With this study we seek to contribute to the research on the emergence of transactive memory, that is, the knowledge of “who knows what”, as part of the structural component of transactive memory systems (TMSs). Based on our theoretical model of team learning and transactive memory development, we analyse the relationships among different team learning behaviours and transactive memory during various phases of team development. We use a longitudinal design with three points of measurement at the beginning, in the middle, and at the end of a semester, to examine real student project teams over their full life cycle. The results support our presumptions of phase-specific influences of team learning behaviours on transactive memory development. This shows that knowledge-based processes (storage and retrieval) play a more important role during early stages of project-based teamwork, followed by a shift to a higher relevance of communication-based processes (reflection and co-construction) in later stages. Implications for further research and team management phenomena are discussed.