Brain networks of explicit and implicit learning

Are explicit versus implicit learning mechanisms reflected in the brain as distinct neural structures, as previous research indicates, or are they distinguished by brain networks that involve overlapping systems with differential connectivity? In this functional MRI study we examined the neural correlates of explicit and implicit learning of artificial grammar sequences. Using effective connectivity analyses we found that brain networks of different connectivity underlie the two types of learning: while both processes involve activation in a set of cortical and subcortical structures, explicit learners engage a network that uses the insula as a key mediator whereas implicit learners evoke a direct frontal-striatal network. Individual differences in working memory also differentially impact the two types of sequence learning.     

Disjunctive models of Boolean category learning

Four connectionistic/neural models which are capable of learning arbitrary Boolean functions are presented. Three are provably convergent, but of differing generalization power. The fourth is not necessarily convergent, but its empirical behavior is quite good. The time and space characteristics of the four models are compared over a diverse range of functions and testing conditions. These include the ability to learn specific instances, to effectively generalize, and to deal with irrelevant or redundant information. Trade-offs between time and space are demonstrated by the various approaches.     

Dynamics and context dependence of visual category learning

Visual category learning by humans is observed within a paradigm of supervised learning. Mental representations for recognition are reconstructed from the observed data structures by fitting to them predicted classification data obtained from similarity-based models of recognition on the one hand and machine vision systems for image understanding on the other hand. These principles are illustrated with examples concerning the dynamics and the dependence on context of processes of category learning.     

The neural basis of perceptual learning

Perceptual learning is a lifelong process. We begin by encoding information about the basic structure of the natural world and continue to assimilate information about specific patterns with which we become familiar. The specificity of the learning suggests that all areas of the cerebral cortex are plastic and can represent various aspects of learned information. The neural substrate of perceptual learning relates to the nature of the neural code itself, including changes in cortical maps, in the temporal characteristics of neuronal responses, and in modulation of contextual influences. Top-down control of these representations suggests that learning involves an interaction between multiple cortical areas.     

A neural model for category learning

We present a general neural model for supervised learning of pattern categories which can resolve pattern classes separated by nonlinear, essentially arbitrary boundaries. The concept of a pattern class develops from storing in memory a limited number of class elements (prototypes). Associated with each prototype is a modifiable scalar weighting factor () which effectively defines the threshold for categorization of an input with the class of the given prototype. Learning involves (1) commitment of prototypes to memory and (2) adjustment of the various factors to eliminate classification errors. In tests, the model ably defined classification boundaries that largely separated complicated pattern regions. We discuss the role which divisive inhibition might play in a possible implementation of the model by a network of neurons.This work was supported in part by the Alfred P. Sloan Foundation and the Ittleson Foundation, Inc.     

Computational model of motor learning and perceptual change

Motor learning in the context of arm reaching movements has been frequently investigated using the paradigm of force-field learning. It has been recently shown that changes to somatosensory perception are likewise associated with motor learning. Changes in perceptual function may be the reason that when the perturbation is removed following motor learning, the hand trajectory does not return to a straight line path even after several dozen trials. To explain the computational mechanisms that produce these characteristics, we propose a motor control and learning scheme using a simplified two-link system in the horizontal plane: We represent learning as the adjustment of desired joint-angular trajectories so as to achieve the reference trajectory of the hand. The convergence of the actual hand movement to the reference trajectory is proved by using a Lyapunov-like lemma, and the result is confirmed using computer simulations. The model assumes that changes in the desired hand trajectory influence the perception of hand position and this in turn affects movement control. Our computer simulations support the idea that perceptual change may come as a result of adjustments to movement planning with motor learning.     

A common framework for perceptual learning

In this review, we summarize recent evidence that perceptual learning can occur not only under training conditions but also in situations of unattended and passive sensory stimulation. We suggest that the key to learning is to boost stimulus-related activity that is normally insufficient exceed a learning threshold. We discuss how factors such as attention and reinforcement have crucial, permissive roles in learning. We observe, however, that highly optimized stimulation protocols can also boost responses and promote learning. This helps to reconcile observations of how learning can occur (or fail to occur) in seemingly contradictory circumstances, and argues that different processes that affect learning operate through similar mechanisms that are probably based on, and mediated by, neuromodulatory factors.     

Modeling perceptual learning with multiple interacting elements: A neural network model describing early visual perceptual learning

We introduce a neural network model of an early visual cortical area, in order to understand better results of psychophysical experiments concerning perceptual learning during odd element (pop-out) detection tasks (Ahissar and Hochstein, 1993, 1994a).The model describes a network, composed of orientation selective units, arranged in a hypercolumn structure, with receptive field properties modeled from real monkey neurons. Odd element detection is a final pattern of activity with one (or a few) salient units active. The learning algorithm used was the Associative reward-penalty (Ar-p) algorithm of reinforcement learning (Barto and Anandan, 1985), following physiological data indicating the role of supervision in cortical plasticity.Simulations show that network performance improves dramatically as the weights of inter-unit connections reach a balance between lateral iso-orientation inhibition, and facilitation from neighboring neurons with different preferred orientations. The network is able to learn even from chance performance, and in the presence of a large amount of noise in the response function. As additional tests of the model, we conducted experiments with human subjects in order to examine learning strategy and test model predictions.     

Implicit associative learning engages the hippocampus and interacts with explicit associative learning

The hippocampus is crucial for conscious, explicit memory, but whether it is also involved in nonconscious, implicit memory is uncertain. We investigated with functional magnetic resonance imaging whether implicit learning engages the hippocampus and interacts with subsequent explicit learning. The presentation of subliminal faces-written profession pairs for implicit learning was followed by the explicit learning of supraliminal pairs composed of the same faces combined with written professions semantically incongruous to those presented subliminally (experiment 1), semantically congruous professions (experiment 2), or identical professions (experiment 3). We found that implicit face-profession learning interacted with explicit face-profession learning in all experiments, impairing the explicit retrieval of the associations. Hippocampal activity increased during the subliminal presentation of face-profession pairs versus face-nonword pairs and correlated with the later impairment of explicit retrieval. These findings suggest that implicit semantic associative learning engages the hippocampus and influences explicit memory.     

Higher olfactory processes: perceptual learning and memory.

The past year has seen several important findings emerge from studies of higher olfactory processes. The identification of synaptic long-term potentiation in the olfactory cortex, induced via repetitive burst stimulation at the theta rhythm, and physiological activity patterns associated with learning, some of which mimic long-term potentiation induction patterns, have suggested relationships between rhythmic activity, behavioral learning and synaptic plasticity. In addition, the construction of computational models of the olfactory bulb and cortex have generated testable behavioral and physiological predictions which have been supported by experimental evidence.     

Blocked and test-stimulus exposure effects in perceptual learning re-examined

Three experiments re-examined the effects of blocked or alternated exposure to the conditioning and test stimuli and the effect of simple exposure to the test stimulus, on stimulus generalization. In all experiments rats received conditioning where a compound flavor, AX, was paired with LiCl-induced illness. All rats were tested for generalization with another flavor, BX. In Experiment 1, rats that received alternating exposure to the two flavor compounds, AX and BX, prior to conditioning showed less generalization to BX than rats that received no exposure. Exposure to BX or AX alone was also somewhat effective in reducing generalization. In Experiment 2 blocked exposure to AX and BX prior to conditioning was effective in reducing generalization, as was alternated exposure, and extended exposure to BX was more effective than the other procedures. In Experiment 3, exposure to X alone prior to conditioning produced generalization equal to that produced by alternated or blocked exposure and replicated the effect of extended exposure to BX found in the previous experiment. The relevance of the results to the theories proposed by McLaren and Macintosh [Anim. Learn. Behav. 28 (2000) 211] and Hall [Q. J. Exp. Psychol. B 56 (2003) 43] is discussed.     

Bayesian natural selection and the evolution of perceptual systems

In recent years, there has been much interest in characterizing statistical properties of natural stimuli in order to better understand the design of perceptual systems. A fruitful approach has been to compare the processing of natural stimuli in real perceptual systems with that of ideal observers derived within the framework of Bayesian statistical decision theory. While this form of optimization theory has provided a deeper understanding of the information contained in natural stimuli as well as of the computational principles employed in perceptual systems, it does not directly consider the process of natural selection, which is ultimately responsible for design. Here we propose a formal framework for analysing how the statistics of natural stimuli and the process of natural selection interact to determine the design of perceptual systems. The framework consists of two complementary components. The first is a maximum fitness ideal observer, a standard Bayesian ideal observer with a utility function appropriate for natural selection. The second component is a formal version of natural selection based upon Bayesian statistical decision theory. Maximum fitness ideal observers and Bayesian natural selection are demonstrated in several examples. We suggest that the Bayesian approach is appropriate not only for the study of perceptual systems but also for the study of many other systems in biology.     

Parasite-driven extinction in spatially explicit host-parasite systems

General host-parasite theory suggests that parasites may be implicated in the extinction of their hosts by causing instability that leads to increased risk of stochastic extinction. In contrast, spatially explicit models suggest that the parasite may directly drive the host population to extinction. Here we examine the ecological characteristics of host-parasite interactions that favor parasite-driven host extinction. Pair approximations and simulations show that parasites only drive their hosts to extinction when they significantly reduce host reproduction. As a matter of interest, parasites that have a relatively small effect on host death rate are more likely to cause host extinction. Parasite-driven host extinction occurs at any population size, whereas extinction caused by stochastic effects is less likely to occur in large host populations. Populations may therefore be under threat from parasites that stop host reproduction, and this type of parasite may prove to be the most effective biological pesticide.     

Retrograde interference in perceptual learning of a peripheral hyperacuity task

Consolidation, a process that stabilizes memory trace after initial acquisition, has been studied for over a century. A number of studies have shown that a skill or memory must be consolidated after acquisition so that it becomes resistant to interference from new information. Previous research found that training on a peripheral 3-dot hyperacuity task could retrogradely interfere with earlier training on the same task but with a mirrored stimulus configuration. However, a recent study failed to replicate this finding. Here we address the controversy by replicating both patterns of results, however, under different experimental settings. We find that retrograde interference occurs when eye-movements are tightly controlled, using a gaze-contingent display, where the peripheral stimuli were only presented when subjects maintained fixation. On the other hand, no retrograde interference was found in a group of subjects who performed the task without this fixation control. Our results provide a plausible explanation of why divergent results were found for retrograde interference in perceptual learning on the 3-dot hyperacuity task and confirm that retrograde interference can occur in this type of low-level perceptual learning. Furthermore, our results demonstrate the importance of eye-movement controls in studies of perceptual learning in the peripheral visual field.     

Modeling perceptual learning: difficulties and how they can be overcome

We investigated the roles of feedback and attention in training a vernier discrimination task as an example of perceptual learning. Human learning even of simple stimuli, such as verniers, relies on more complex mechanisms than previously expected – ruling out simple neural network models. These findings are not just an empirical oddity but are evidence that present models fail to reflect some important characteristics of the learning process. We will list some of the problems of neural networks and develop a new model that solves them by incorporating top-down mechanisms. Contrary to neural networks, in our model learning is not driven by the set of stimuli only. Internal estimations of performance and knowledge about the task are also incorporated. Our model implies that under certain conditions the detectability of only some of the stimuli is enhanced while the overall improvement of performance is attributed to a change of decision criteria. An experiment confirms this prediction. Received: 23 May 1996 / Accepted in revised form: 16 October 1997