The time course of perceptual choice: the leaky, competing accumulator model.

The time course of perceptual choice is discussed in a model of gradual, leaky, stochastic, and competitive information accumulation in nonlinear decision units. Special cases of the model match a classical diffusion process, but leakage and competition work together to address several challenges to existing diffusion, random walk, and accumulator models. The model accounts for data from choice tasks using both time-controlled (e.g., response signal) and standard reaction time paradigms and its adequacy compares favorably with other approaches. A new paradigm that controls the time of arrival of information supporting different choice alternatives provides further support. The model captures choice behavior regardless of the number of alternatives, accounting for the log-linear relation between reaction time and number of alternatives (Hick's law) and explains a complex pattern of visual and contextual priming in visual word identification.

No right to speak? The relationship between object naming and semantic impairment: neuropsychological evidence and a computational model.

The processes required for object naming were addressed in a study of patients with semantic dementia (a selective decline of semantic memory resulting from progressive temporal lobe atrophy) and in a computational model of single-word production. Although all patients with semantic dementia are impaired in both single-word production and comprehension, previous reports had indicated two different patterns: (a) a parallel decline in accuracy of naming and comprehension, with frequent semantic naming errors, suggesting a purely semantic basis for the anomia and (b) a dramatic progressive anomia without commensurate decline in comprehension, which might suggest a mainly postsemantic source of the anomia. Longitudinal data for 16 patients with semantic dementia reflected these two profiles, but with the following additional important specifications: (1) despite a few relatively extreme versions of one or other profile, the full set of cases formed a continuum in the extent of anomia for a given degree of degraded comprehension; (2) the degree of disparity between these two abilities was associated with relative asymmetry in laterality of atrophy: a parallel decline in the two measures characterized patients with greater right- than left-temporal atrophy, while disproportionate anomia occurred with a predominance of atrophy in the left-temporal lobe. In an implemented computational model of naming, semantic representations were distributed across simulated left- and right-temporal regions, but the semantic units on the left were more strongly connected to left-lateralized phonological representations. Asymmetric damage to semantic units reproduced the longitudinal patient profiles of naming relative to comprehension, plus additional characteristics of the patients' naming performance. On the basis of both the neuropsychological and computational evidence, we propose that semantic impairment alone can account for the full range of word production deficits described here.

Deficits in irregular past-tense verb morphology associated with degraded semantic knowledge.

Two distinct mechanisms are often considered necessary to account for generation of the past-tense of English verbs: a lexical associative process for irregular forms like speak-->spoke, and a rule-governed process ("add -ed") for regular and novel forms like talk-->talked and wug-->wugged. An alternative account based on a parallel-distributed processing approach proposes that one complex procedure processes all past-tense types. In this alternative view, neuropsychological dissociations are explained by reduced input from word meaning that plays a greater role in successful generation of the past-tense for lower frequency irregular verbs, and by phonological deficits that disproportionately affect regular and novel forms. Only limited evidence has been available concerning the relationship between knowledge of word meaning and verb-tense processing. The study reported here evaluated the past-tense verb abilities of 11 patients with semantic dementia, a neurodegenerative condition characterised by degraded semantic knowledge. We predicted and confirmed that the patients would have essentially normal ability to generate and recognise regular (and novel) past-tense forms, but a marked and frequency-modulated deficit on irregular verbs. Across the set of 11 patients, the degree of impairment for the irregular past-tense was significantly correlated with the degree of comprehension impairment as measured by verb synonym judgements. These results, plus other features of the data such as the nature of the errors to irregular verbs, are discussed in relation to currently developing theories of the language system.

The Morton-Massaro law of information integration: implications for models of perception.

Information integration may be studied by analyzing the effect of 2 or more sources (e.g., auditory and visual) on participants' responses. Experiments show that ratios of response probabilities often factorize into components selectively influenced by only 1 source (e.g., 1 component affected by the acoustic source and another 1 affected by the visual source). This is called the Morton-Massaro law (MML). This article identifies conditions in which the law is optimal and notes that it reflects an implicit assumption about the statistics of the environment. Adherence to the MML can be used to assess whether the assumption is being made, and analyses of natural stimuli can be used to determine whether the assumption is reasonable. Feed-forward and interactive models subject to a channel separability constraint are consistent with the law.

The basis of hyperspecificity in autism: a preliminary suggestion based on properties of neural nets.

This article reviews a few key ideas about the representation of information in neural networks and uses these ideas to address one aspect of autism, namely, the apparent hyperspecificity that is often seen in autistic children's application of previously acquired information. Hyperspecificity is seen as reflecting a possible feature of the neural codes used to represent concepts in the autistic brain.

Repetition priming of words, pseudowords, and nonwords.

In 5 experiments, the authors assessed repetition priming for words, pseudowords, and nonwords using a task that combines an implicit perceptual fluency measure and a recognition memory assessment for each list item. Words and pseudowords generated a consistently strong repetition effect even when there was a failure to recognize the stimulus. In 2 of the experiments, the repetition effect for nonwords was reliably above chance even when there was a failure to recognize the stimulus. The authors propose a parallel distributed processing (PDP) model based on the work of J. McClelland and D. Rumelhart (1985) as a way to understand the mechanisms potentially responsible for the pattern of findings. Although the error-driven nature of learning in the model results in a poor fit to the nonword priming data, this is not endemic to all PDP models. Using a model based on Hebbian learning, the authors instantiate a property that they believe is characteristic of implicit memory--that learning is primarily based on the strengthening of connections between units that become active during the processing of a stimulus. This model provides a far more satisfactory account of the data than does the error-driven model.

Neural models of memory.

Neural models assist in characterizing the processes carried out by cortical and hippocampal memory circuits. Recent models of memory have addressed issues including recognition and recall dynamics, sequences of activity as the unit of storage, and consolidation of intermediate-term episodic memory into long-term memory.

Familiarity breeds differentiation: a subjective-likelihood approach to the effects of experience in recognition memory.

With repeated exposure, people become better at identifying presented items and better at rejecting items that have not been presented. This differentiation effect is captured in a model consisting of item detectors that learn estimates of conditional probabilities of item features. The model is used to account for a number of findings in the recognition memory literature, including (a) the basic differentiation effect (strength-mirror effect), (b) the fact that adding items to a list reduces recognition accuracy (list-length effect) but extra study of some items does not reduce recognition accuracy for other items (null list-strength effect), (c) nonlinear effects of strengthening items on false recognition of similar distractors, (d) a number of different kinds of mirror effects, (e) appropriate z-ROC curves, and (f) one type of deviation from optimality exhibited in recognition experiments.

Complementary learning systems in the brain. A connectionist approach to explicit and implicit cognition and memory.

Freud's ideas about the role of non-conscious processes relates to contemporary thinking about explicit and implicit memory, and his early efforts to understand cognition and behavior in terms of neural mechanisms share several themes in common with contemporary connectionist models. The present paper presents a connectionist perspective of the neural basis of learning and memory and their organization in the brain. The central claim of the article is that the neocortex and many other forebrain learning systems learn slowly so as to become sensitive to the overall structure of experience. Slow learning is crucial for sensitivity to this structure and for organizing specific information with other information in a structured way. The hippocampus and related areas in the medial temporal lobes complement these slow learning systems by providing a mechanism that allows the rapid learning of arbitrary conjunctions of elements that go together to make up an episodic memory.

Rethinking infant knowledge: toward an adaptive process account of successes and failures in object permanence tasks.

Infants seem sensitive to hidden objects in habituation tasks at 3.5 months but fail to retrieve hidden objects until 8 months. The authors first consider principle-based accounts of these successes and failures, in which early successes imply knowledge of principles and failures are attributed to ancillary deficits. One account is that infants younger than 8 months have the object permanence principle but lack means-ends abilities. To test this, 7-month-olds were trained on means-ends behaviors and were tested on retrieval of visible and occluded toys. Means-ends demands were the same, yet infants made more toy-guided retrievals in the visible case. The authors offer an adaptive process account in which knowledge is graded and embedded in specific behavioral processes. Simulation models that learn gradually to represent occluded objects show how this approach can account for success and failure in object permanence tasks without assuming principles and ancillary deficits.

Considerations arising from a complementary learning systems perspective on hippocampus and neocortex.

We discuss a framework for the organization of learning systems in the mammalian brain, in which the hippocampus and related areas form a memory system complementary to learning mechanisms in neocortex and other areas. The hippocampal system stores new episodes and "replays" them to the neocortical system, interleaved with ongoing experience, allowing generalization as cortical memories form. The data to account for include: 1) neurophysiological findings concerning representations in hippocampal areas, 2) behavioral evidence demonstrating a spatial role for hippocampus, 3) and effects of surgical and pharmacological manipulations on neuronal firing in hippocampal regions in behaving animals. We hypothesize that the hippocampal memory system consists of three major modules: 1) an invertible encoder subsystem supported by the pathways between neocortex and entorhinal cortex, which provides a stable, compressed, invertible encoding in entorhinal cortex (EC) of cortical activity patterns, 2) a memory separation, storage, and retrieval subsystem, supported by pathways between EC, dentate gyrus and area CA3, including the CA3 recurrent collaterals, which facilitates encoding and storage in CA3 of individual EC patterns, and retrieval of those CA3 encodings, in a manner that minimizes interference, and 3) a memory decoding subsystem, supported by the Shaffer collaterals from area CA1 to area CA3 and the bi-directional pathways between EC and CA3, which provides the means by which a retrieved CA3 coding of an EC pattern can reinstate that pattern on EC. This model has shown that 1) there is a trade-off between the need for information-preserving, structure-extracting encoding of cortical traces and the need for effective storage and recall of arbitrary traces, 2) long-term depression of synaptic strength in the pathways subject to long-term potentiation is crucial in preserving information, 3) area CA1 must be able to exploit correlations in EC patterns in the direct perforant path synapses.

Understanding normal and impaired word reading: computational principles in quasi-regular domains.

A connectionist approach to processing in quasi-regular domains, as exemplified by English word reading, is developed. Networks using appropriately structured orthographic and phonological representations were trained to read both regular and exception words, and yet were also able to read pronounceable nonwords as well as skilled readers. A mathematical analysis of a simplified system clarifies the close relationship of word frequency and spelling-sound consistency in influencing naming latencies. These insights were verified in subsequent simulations, including an attractor network that accounted for latency data directly in its time to settle on a response. Further analyses of the ability of networks to reproduce data on acquired surface dyslexia support a view of the reading system that incorporates a graded division of labor between semantic and phonological processes, and contrasts in important ways with the standard dual-route account.

Hippocampal conjunctive encoding, storage, and recall: avoiding a trade-off.

The hippocampus and related structures are thought to be capable of 1) representing cortical activity in a way that minimizes overlap of the representations assigned to different cortical patterns (pattern separation); and 2) modifying synaptic connections so that these representations can later be reinstated from partial or noisy versions of the cortical activity pattern that was present at the time of storage (pattern completion). We point out that there is a trade-off between pattern separation and completion and propose that the unique anatomical and physiological properties of the hippocampus might serve to minimize this trade-off. We use analytical methods to determine quantitative estimates of both separation and completion for specified parameterized models of the hippocampus. These estimates are then used to evaluate the role of various properties and of the hippocampus, such as the activity levels seen in different hippocampal regions, synaptic potentiation and depression, the multi-layer connectivity of the system, and the relatively focused and strong mossy fiber projections. This analysis is focused on the feedforward pathways from the entorhinal cortex (EC) to the dentate gyrus (DG) and region CA3. Among our results are the following: 1) Hebbian synaptic modification (LTP) facilitates completion but reduces separation, unless the strengths of synapses from inactive presynaptic units to active postsynaptic units are reduced (LTD). 2) Multiple layers, as in EC to DG to CA3, allow the compounding of pattern separation, but not pattern completion. 3) The variance of the input signal carried by the mossy fibers is important for separation, not the raw strength, which may explain why the mossy fiber inputs are few and relatively strong, rather than many and relatively weak like the other hippocampal pathways. 4) The EC projects to CA3 both directly and indirectly via the DG, which suggests that the two-stage pathway may dominate during pattern separation and the one-stage pathway may dominate during completion; methods the hippocampus may use to enhance this effect are discussed.

Nonword pronunciation and models of word recognition.

Nonword pronunciation is a form of generalization behavior that has been at the center of debates about models of word recognition, the role of rules in explaining behavior, and the adequacy of the parallel distributed processing approach. An experiment yielded data concerning the pronunciation of a large corpus of nonwords. The data were then used to assess 2 models of naming: a model developed by D. C. Plaut and J. L. McClelland (1993), which is similar to the one described by M. S. Seidenberg and J. L. McClelland (1989) but uses improved orthographic and phonological representations, and the grapheme-phoneme correspondence rules of M. Coltheart, B. Curtis, P. Atkins, and M. Haller's (1993) dual-route model. Both models generate plausible nonword pronunciations and match subjects' responses accurately. The dual-route model does so by using rules that generate correct output for most words but mispronounce a significant number of exceptions. The parallel distributed processing model does so by finding a set of weights that allow it to generate correct output for both "rule-governed" items and exceptions. Some ways in which the two approaches differ and other issues facing them are also discussed.

The organization of memory. A parallel distributed processing perspective.

Parallel distributed processing (PDP) provides a contemporary framework for thinking about the nature and organization of perception, memory, language, and thought. In this talk I describe the overall framework briefly and discuss its implications of procedural, semantic, and episodic memory. According to the PDP approach, the processing of information takes place through the interaction of a large number of simple processing units organized into modules. Storage occurs through the modification of connection weights based on the system's response to its input, that provides an opportunity for incremental storage. I will describe how connection modification may give rise through the course of experience to procedural learning and to the formation of semantic memories, structured by their semantic content. I will argue that the discovery of semantic structure requires gradual learning, with repeated exposure to representative samples of the structure to be learned. I will then describe two neuropsychological implications of the PDP approach. First, I will consider the possible modular organization of semantic information in the brain. Then, I will examine the role of the hippocampus in learning and memory. In the first case, we will see how the PDP approach leads us to see how brain damage might produce apparent dissociations between categories, when in fact the underlying organization is not by category but by modality. In the second case, we will see that the PDP approach gives us a new way to understand why it is important that unique, arbitrary associations not be stored all at once in the same memory's systems used for semantic information.(ABSTRACT TRUNCATED AT 250 WORDS)

Category learning. Learning the general but not the specific.

Amnesia patients have a normal ability to learn categories from examples, even though they fail to learn the examples themselves; computational models of brain function suggest how and why.

Computational approaches to cognition: top-down approaches.

Computational models are useful tools for exploring the nature of human cognitive processes. In particular, connectionist models are providing researchers with new ways of thinking about the basic nature of cognition and its implementation in the brain. They support novel explanations of important aspects of perception, memory, language, thought and cognitive development, and allow cognitive processes to be linked with the underlying physiological mechanisms. The models also aid our understanding of how disorders of brain function lead to disorders of cognition.

A parallel distributed processing approach to automaticity.

We consider how a particular set of information processing principles, developed within the parallel distributed processing (PDP) framework, can address issues concerning automaticity. These principles include graded, activation-based processing that is subject to attentional modulation; incremental, connection-based learning; and interactivity and competition in processing. We show how simulation models, based on these principles, can account for the major phenomena associated with automaticity, as well as many of those that have been troublesome for more traditional theories. In particular, we show how the PDP framework provides an alternative to the usual dichotomy between automatic and controlled processing and can explain the relative nature of automaticity as well as the fact that seemingly automatic processes can be influenced by attention. We also discuss how this framework can provide insight into the role that bidirectional influences play in processing: that is, how attention can influence processing at the same time that processing influences attention. Simulation models of the Stroop color-word task and the Eriksen response-competition task are described that help illustrate the application of the principles to performance in specific behavioral tasks.

A computational model of semantic memory impairment: modality specificity and emergent category specificity.

It is demonstrated how a modality-specific semantic memory system can account for category-specific impairments after brain damage. In Experiment 1, the hypothesis that visual and functional knowledge play different roles in the representation of living things and nonliving things is tested and confirmed. A parallel distributed processing model of semantic memory in which knowledge is subdivided by modality into visual and functional components is described. In Experiment 2, the model is lesioned, and it is confirmed that damage to visual semantics primarily impairs knowledge of living things, and damage to functional semantics primarily impairs knowledge of nonliving things. In Experiment 3, it is demonstrated that the model accounts naturally for a finding that had appeared problematic for a modality-specific architecture, namely, impaired retrieval of functional knowledge about living things. Finally, in Experiment 4, it is shown how the model can account for a recent observation of impaired knowledge of living things only when knowledge is probed verbally.